JOURNAL OF SHELLFISH RESEARCH VOLUME 20, NUMBER 1 JUNE 2001 The Journal of Shellfish Research (formerly Proceedings of the National Shellfisheries Association) is the official publication of the National Shellfisheries Association Editor Dr. Sandra E. Shumway Natural Science Division Southampton College, Long Island University Southampton, NY 11968 Dr. Standi.sh K. Allen. Jr. (2002) School of Marine Science Virginia Institute of Marine Science Gloucester Point, VA 23062-1 1346 Dr. Peter Beninger (2001) Laboratoire de Biologic Marine Faculte des Sciences Universite de Nantes BP 92208 44322 Nantes Cedex 3 France Dr. Andrew Boghen (2001) Department of Biology University of Moncton Moncton, New Brunswick Canada El A 3E9 Dr. Neil Bourne (2001) Fisheries and Oceans Pacific Biological Station Nanaimo, British Columbia Canada V9R 5K6 Dr. Andrew Brand (2001) University of Liveipool Marine Biological Station Port Erin, Isle of Man Dr. Eugene Burreson (2001) Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Dr. Peter Cook (2002) Department of Zoology University of Cape Town Rondebosch 7700 Cape Town, South Africa EDITORIAL BOARD Dr. Simon Cragg (2002) Institute of Marine Sciences University of Portsmouth Ferry Road Portsmouth P04 9LY United Kingdom Dr. Leroy Creswell (2001) Harbor Branch Oceanographic Institute US Highway 1 North Fort Pierce, Florida 34946 Dr. Lou D'Abramo (2002) Mississippi State University Dept of Wildlife and Fisheries Box 9690 Mississippi State, Mississippi 39762 Dr. Ralph Elston (2001) Battelle Northwest Marine Sciences Laboratory 439 West Sequim Bay Road Sequim, Washington 98382 Dr. Susan Ford (2002) Rutgers University Haskin Laboratory for Shellfish Research P.O. Box 687 Port Norris, New Jersey 08349 Dr. Raymond Grizzle (2001) Randall Environmental Studies Center Taylor University Upland, Indiana 46989 Dr. Mark Luckenbach (2001) Virginia Institute of Marine Science Wachapreague, Virginia 23480 Dr. Bruce MacDonald (2002) Department of Biology University of New Brunswick P.O. Box 5050 Saint John, New Brunswick Canada E2L 4L5 Dr. Roger Mann (2002) Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Dr. Isla'y D. Marsden (2002) Department of Zoology Canterbury University Christchurch, New Zealand Dr. Tom Soniat (2002) Biology Department Nicholls State University Thibodaux, Louisiana 70310 Dr. J. Evan Ward (2002) Dept. of Marine Sciences University of Connecticut Groton, CT 06340-6097 Dr. Gary Wikfors (2002) NOAA/NMFS Rogers Avenue Milford, Connecticut 06460 Journal of Shellfish Research Volume 20, Number 1 ISSN: 0730-8000 June 2001 www.shelltish.org/pubs/jsr.htm JoKi-iuil oj Shellfish Research. Vol. 20. Nci. I. 1-12. 2(101 SEP 0 & 2001 IN MEMORIAM THOMAS CLEMENT CHENG NOVEMBER 5, 1930-NOVEMBER 28. 2000 Thomas C. Cheng, an international authority in the fields of parasitology, molluscan biology, and shellfish diseases died on November 28. 2000 in Charleston, South Carolina. He was 70 years old and is survived by three children, Thomas. Jr., Brad, and Allison. Tom was born on November 5, 1930 at Nanking, China. He moved to the United States with his parents he was six years old. Tom spoke \ery little English but airived without any problems and soon attended local public schools. After graduating from Greenbrier Military School, Lewisburg, West Virginia he attended Wayne State University at Detroit. Michigan, with a major in biology. While at Wayne State, he enrolled in a parasitology course taught by Professor Dominc L. DeGiusti and became very interested in the public health importance of parasites. After graduation, he joined the U.S. Army and for four years he worked in a clinical diagnostic laboratory performing parasitology, serology, and bacteriology during the Korean conflict. These experiences triggered even a greater interest in parasites, and upon discharge from military service, he enrolled in the graduate program at the University of Virginia and received his M.S. in biology in 1956, and his Ph.D. in 1958 under the of codirection of Professor Bruce Dodson Reynolds and Professor Horton Holcome Hobbs. Jr. His first faculty position was at the University of Maryland. Baltimore. Maryland as an instructor-assistant professor. One year later, he joined the Department of BioUigy at Lafayette College. Easton. Pennsylvania. During his six years at Lafayette College, he taught various courses including parasitology, microbiology, general biology, physiology, and invertebrate zoology. During these six years, he formulated and wrote his first book. The Biology of Animal Pcinisiws. The book became a popular textbook for an introductory course in parasitology. He was awarded a National Institutes of Health Fellowship in 1964 to travel to the Pacific Biomedical Research Center at the University of Hawaii. Honolulu. Hawaii. He worked as a visiting scientist on the modes of infection of Acharina fiilica by the larvae of Anfiioslroiifiyhis cantonensis. After this one-year fellowship, he returned to the U.S. mainland and was appointed Chief of the Immunology and Parasitology section at the U.S. Public Health Service Northeast Marine Health Sciences Laboratory at Narrangansett. Rhode Island. It was here that he started one of his most significant academic accomplishments, the writing oi Marine Molluscs us Host for Symbioses with a Review of Known Parasites ofCoinmericiallv Iniportaiu Species that was published in the 5th volume of Advances in Marine Biology. He then returned to the University of Hawaii to join the Department of Zoology. After five years, he returned to Pennsylvania and created and directed the Institute for Pathobiology at Lehigh University. The next 12 years were very productive and exciting for Tom. However, 1980, Tom was approached by the Medical University of South Carolina and asked if he would be interested in creating a Marine Biomedical Research Program. He accepted their offer and the challenge and joined the Medical University at South Carolina 2 RODRICK at Charleston, South Carolina. He served as Director of this program until his retirement in 1993. While the Director of the Marine Biomedical Research Program, he accepted the position as Acting Chairman of the Department of Cell Biology and Anatomy of the Medical University for 18 months. It was during this time that he was recognized by the scientific community for his significant contributions to the advancement of moUuscan immunology and host parasite interactions. He received numerous awards including a Senior Fulbright Scholar. Directeur de Recherche Award, and the Montpellier Medal and was in very active in lecturing and chairing many national and international meetings and congresses. Tom retired in 1993. but still remained active at the laboratory bench working on molluscan diseases and imnunnilogy until the end. During his seven years of retirement, he continued to write research grants and gain funding for his ideas. When diagnosed with terminal esophageal cancer, his pace slowed but not his interest in his research. He continued to work daily for several hours until he could no longer fight the fatigue and pain. Tom Cheng was a bright and energetic teacher and scientist who helped shaped the careers of many individuals who worked in the areas of parasitology and molluscan immunology. He published over 350 scientific papers, 22 books, and served as Editor of the Journal of Invertebrate Pathology for twenty-three years and twenty years as coeditor of Experimental Parasitologv. He directed twenty-three graduate students and ten postdoctoral fellows. The tleld of molluscan shellfish pathobiology and parasitology will miss Tom Cheng. However, his contributions to these important areas will continue to live through his scholariy works, his students, and colleagues he touched during his remarkably productive life. Gary Rodrick Food Science & Human Nutrition Institute of Food and Agricultural Sciences Gainesville. Florida 3261 1-0370 PUBLICATIONS 1. Cheng. T. C. 19.'i6. Ta.xonomic and morphological studies on the genus Aiuiulnnniini (Tremutoda: Lecithodendriidae). Va. J. Sci. 7:286. 2. Cheng. T. C. 1957. Studies on the genus Acwnhatrium Faust, 1919 (Trematoda: Lecithodendriidae); with the description of two new species. / Parasilol. 43:60-65. 3. Cheng. T. C. 1957. A study of the melacercarial cyst and metacercaria of Crepiihisldiiiuiii conuttum (Trematoda: .Mlocreadiidae); with notes on the similarity of the larval forms of the genus. Bull. Assoc. SE Biol. 4( 1 1 ); 5. 4. Cheng. T. C. 1957. A study of the metacercaria of Crepidostomiim cormnum (Oshi>rn, 19031 (Trematoda: Mlocreadiidae). Proc. Heimiiith. Soc. Wash. 24: 107-109. 5. Jones, A. W., T. C. Cheng & R. F. Gillespie. 1957. A new Ophiomema from a frog. Bull, .\ssoc. SE Biol. 4:13. 6. Jones, A.W.. Cheng, T.C. & Gillespie, R. F. 1958. Ophiotaenia gracillis n. sp.. a proteocephaiid cestode from a frog. J. Tenn. Acad Sci. 53:84-88. 7. Ph. D. Dissertation University of Virginia. 1958. Systematic, morphological, and life history studies on the trematode family Brachycoeiiidae Johnston, 1913. Dissert. Abstr. 20 (4). 8. Cheng, T C. 1958. Studies on the trematode family Dicrocoeliidae. 1. the genera Brmhycoeliiim (Dujardin, I 845) and Leptopluillus Luhe. 1909. (Brachycoeiiinae). -4m. Midi. Nat. 59:67-81. 9. Cheng. T. C, & B. J. Bogitsh. 1958. The description of Hymenolepis turdi n. sp.. a parasite of Turdus migratonus L: with notes on the systematic validity of rostllar hooks in the species of passeriform birds. Trans. Am. Microsc. .Snc. 77: 295-298. 10. Lautenschiager. E. W. & T. C. Cheng, 1958. Laretminlha polyorchis (Trematoda: Heterophyidae): a new fluke from the herring gull. Trans. .Am. Microsc. Soc. 77:451-454. U. Cheng, T. C. 1959. Studies on the trematode family Brachycoeiiidae. revision of the general Glxpthelmms (Stafford, 1900). Stafford, 1905, and Margeana Con, 1910; with the description of Rcynoldstema n. gen. .Am. Midi. Nat. 61:68-88. 12. Provenza. D. V., W. E. Biddington & T. C. Cheng. 1959. Studies on the etiology of periodontosis. I. the role of vascular changes in periodontium. Oral Med.. Oral Path. Oral Surg. 12:676-684. 13. Cheng. T. C. & D. V. Provenza. 1959. Histologic observations on the morphology of the blood vessels of canine and human tooth pulp. ./. Dent. Res. 38:552-557. 14. Bogitsh. B. J. & T. C. Cheng. 1959. Pisciamphisloma reynoldsi (Paramphistomatidae), a new trematode parasite of Lepomis spp. in Virginia. J. Tenn. Acad. Sci. 34:159-161. 15. Cheng. T. C. 1959. The histology of the prostate mass in the genus Acaniharriiim (Trematoda: Lecithodendriidae). Proc. Helminth. Soc. Wash. 26:111-113. 16. Cheng. T. C. 1959. The description of Acanthatrium beuschleini n. sp., a new trematode parasite of bats; and a re-e\aluation of the reproductue system of Acanthatrium sogandaresi Coil and Kuntz, 1958, J. Parasitol. 45: 323-326. 17. Cheng. T. C. 1960. A survey of Enterobius vermicularis infestation among preschool age transient children, with the description of a modified Graham swab. / Tenn. Acad. Sci. 35:49-53. 18. Cheng. T. C. & H. A. James. 1960. The histopathology of Crepidostomum sp. infection in the second intermediate host, Sphaerium striatuuim. Proc. Helminth. Soc. Wash, ll-.bl-bl,. 19. Cheng. T. C, & H. A. James. 1960. Bothriocephalus schilbeodis n. sp. (Cestoidea: Bothriocephalidae), an intestinal parasite of Schilbeodes insignis. J. Tenn. Acad. Sci. 35: 164-168. 20. Cheng, T. C. & D, V. Provenza, 1960. Studies on the trematode family Brachycoeiiidae III. the subfamilies subordinate to the Brachycoeiiidae and the status of the genus Cymatocarpus Looss, 1899. .\m. Midi. Nat. 63:162-168. 21. Cheng. T. C. 1960. Studies on the trematode family Brachycoeiiidae. IV. a revision of the genus Mesocoelium Odhner, 1911, and the status of Pintnaria Poche, 1907. .Am. Midi Nat. 63:439-169. 22. Provenza, D. V. & T. C. Cheng. 1960. A simplitied pariodion method for sectioning teeth; with notes on the decalcification of teeth. Trans. .Am. Microsc. Soc. 79:103-104. In Memoriam: Thomas Clement Cheng 3 23. Cheng. T. C. 1960. The life history of Bnwhycoelium obesum Niciill. 1414. with a discussion of the systematic status of the trematode family Brachycoeiiidae Johnston. 1912. J. Parasitol. 46:464-474. 24. Cheng, T. C. & D. V. Provenza. 1960. The sphmcter muscle pattern as found in the microcirculation of the dog's liver. J. Teim. Acad. Sci. 35; 1281-1238. 25. Cheng. T. C. & D. V. Provenza. I960. Studies on cellular elements of the mesenchyma and of tissues of Haemaloloechus cgici 2:267-275. 60. Cheng. T. C. 1965. Histochemical observations on changes in the lipid composition of the American oyster. Crassostrea virginica (Gmelin), parasitized by the trematode Bucephalus sp. J. Invert. Pathol. 7:398— +07. 61. Cheng. T. C. 1965. Parasitological problems associated with food protection. J. Environ. Health 28:208-214. 62. Cheng. T. C. 1965. On the structure and alteration of cellular polarity of the sporocysts of Bucephalus sp. as a function of age. Proc. Penn. .Acad. Sci. 39: 180-186. 63. Cheng. T. C. 1966. Perivascular leucocylosis and other types of cellular reactions in the oyster Cra.ssoslrea virginica experimentally infected with the nematode Angiostrongylu.i cantoneiisi.'^. J. Invert. Pathol. 8:52-58. 64. Cheng, T. C. & R. W. Burton. 1966. Relationships between Bucephalus sp. and Crcis.s(Ktrca virginica: a histochemical study of some carbohydrates and carbohydrate complexes occurring in the host and parasite. PmiLsilology 56:1 1 1-122. 65. Cheng, T. C, C. N. Schuster, Jr. & A. H. Anderson. 1966. A comparative study of the susceptibility and response of eight species of marine pelecypods to the trematode Himasthia cpiissentensis. Trans. Am. Microsc. Soc. 85:284-295. 66. Cheng. T. C. 1966. The coracidium of the cestode Tylocepluihim and the migration and fate of this parasite in the American oyster Crassostrea virginica. Trans. Am. Micros. Soc. 85:246-255. 67. Cheng. T. C. R. W. Snyder, Jr., A. W. Rourke & A. B. Blumenthal. 1966. Ionic calcium concentrations in nonparasitized Nitocris Jilaiaius and those parasitized by the larvae of Prosthodendrium (Acanthatrium) anaplocami (Trematoda) /\;h. Zool. 6:225. 68. Cheng, T. C, C. N. Shuster, Jr. & A. H. Anderson. 1966. Effects of plasma and tissue extracts of marine pelecypods on the cercarai of Huuasthia quissentensis. Explt. Parasit. 19: 9-4. 69. Cheng, T. C. 1966. Perivascular leucocytosis and other types of cellular reaction in Crassostrea virginica (Gmelin) experimentally infected with the metastrongylid nematode Angiostrongylus cantonensis (Chen). Proc. Natl. Shellfish. Assoc. 56:2. 70. Cheng, T. C. 1966. On the structure, mode of infection, and fate of Tvlocephalum in the American oyster, Cras.so.^trea virginica. Proc. Natl. Shellfish. Assoc. 56:1-2. 71. Cheng, T, C, & A. S. Thakur. 1967. Thermal activation and inactivation of Philopluhabmis glairi metacercariae. J. Purasitol. 53: 212-213. 72. Cheng, T. C. 1967. Review of The Physiology ofTrematodes by J. D. Smyth. Science 1(56): 1-587. 73. Cheng. T. C. 1967. Review of The Physiologv ofTrematodes by J. D. Smyth. Quart. Rev. Biol. 42:315-316. 74. Cheng. T. C E. Rit\in & H. W. F. Yee. 1967. The role of certain parenchymal cells of Terpios zeteki (Porifera: Demospongiae) in phagocytosis and elimination of foreign particles. Am. Zool. 7:771-772. 75. Cheng, T. C. 1967. Review of Ecology of Parasites by N. A. Croll. Quart. Rev. Biol. 42:553-554. 76. Cheng, T. C. 1967. Review of Biochemisliy of Parasites by T. von Brand. Quart. Rev. Biol. 42:554. 77. Rifkin, E. & T. C. Cheng. 1968. The origin, structure, and histochemical characterization of the encapsulating cysts in the oyster Cra.K.wstrea virginica parasitized by the cestode Tylocephaimn sp. J. Invert. Pathol. 10:54—64. 78. Cheng, T. C. & E. Rifkin. 1968. The occurrence and resorption of Tylocephalum metacestodes in the clam Tapes semidccussata. J. Invert. Pathol. 10:65-69. 79. Cheng, T, C. 1968. The compatibility and incompatibility concept as related to trematodes and molluscs. Pacific Sci. 22:141-160. SO. Cheng, T. C. & H. W. F. Yee. 1968. A histochemical demonstration of aniinopeptidase activity associated with the intramolluscan stages of Philophthaimus grain (Trematoda: Philophthalmidae). Trans. Am. Microsc. Soc. 87:128-129. 81. Cheng, T. C. & H. W. F. Yee. 1968. Histochemical demonstration of aniinopeptidase activity associated with the intramolluscan stages of Philophthaimus gralli Mathis and Leger. Parasitology 58:473—180. 82. Cheng, T. C, H. W. F. Yee & E. Rifkin. 1968. Studies on the internal defense mechanisms of sponges. 1. the cell types occurring in the mesoglea of Terpios zeteki (de Laubenfels) (Porifera: Demospongiae). Pacific Sci. 22:395^01. 83. Cheng, T. C. Riflcin, E. & Yee, H. W. F. 1968. Studies on the internal defense mechanisms of sponges. 11. phagocytosis and elimination of India ink and carmine particles by certain parenchymal cells of Terpios zeteki. J. Invert. Pathol. 1 1:302-309. 84. Cheng, T. C, H. W. F. Yee. E. RiflJ.s a.s Final A/o.?/.? (if Nemaunies and Nematomorphs. An Annotated Bihlioj^raphy 1900-1972 by M. R. N. Shephard. Commonwealth Institute of Helminthology, United Kingdom. / Invert. Pathol. 26:424. 180. Cheng, T. C. 1975. Review: of 0.xyiiroidea(-ata) of Animals and Man. Parts 1 and 2, Essentials of Nematodology. vols. VIII and X, by K. I. Skrjabin, N. P.Shikhobalova, and E. A. Lagodorskaya. Translated from Russian by Israel Program of Scientific Translations. Jerusalem Quart. Rev. Biol. 50:479-480. 181. Cheng, T. C. 1976. Beta-glucuronidase from the serum and cells of Mercenaria mercenaria and Crasso.strea virginica (Mollusca: Pelecypoda). J. Invert. Pathol. 27:125-128. 182. Cheng, T. C. & B. M. Rudo. 1976. Chemotactic attraction of Crassostrea virginica hemolymph cells to Staphylococcus lactiis. J. Invert. Pathol. 27:137-139. 183. Cheng. T. C. 1976. The natural history of anisakiasis in animals. J. Milk & Food Technol. 39:32-46. 184. Cheng. T. C. & T. P. Yoshino. 1976. Lipase activity in the .serum and hemolymph cells of the soft-shelled clam. Mya urenaria. during phagocytosis. J. Invert. Pathol. 27:243-245. 185. Cheng, T. C. & B. M. Rudo. 1976. Distribution of glycogen resulting from degradation of '''C labeled bacteria in the American oyster. Crassostrea virginica. J. Invert. Pathol. 27:259-262. 186. Cheng, T. C. 1976. Review of Viruses and Invertebrates edited by .\. J. Gibbs. .Amsterdam: North-Holland and London Quart. Rev. Biol. 50:481^82. 187. Yoshino. T. P. & T. C. Cheng. 1976. Fine structural localization of acid phosphatase in granulocytes of the pelecypod Mercenaria niercenarui. Trans. Am. Microsc. Soc. 95(2):2 15-220. 188. Yoshino, T. P. & T. C. Cheng. 1976. Experimentally induced elevation of aminopeplidase activity in hemolymph cells of the American oyster. Crassostrea virginica. J. Invert. Pathol. 27:367-370. 189. Cheng, T. C. 1976. Identification of proliferative lesions in molluscs. Mar. Fisher. Rev. 38:5-6. 190. Cheng, T. C. & T. P. Yoshino. 1976. Lipase activity in the hemolymph of Biomphalaria glahrata (Mollusca) challenged with bacterial lipids. J. Invert. Pathol. 28:143-146. 191. Sullivan, J. T. & T. C. Cheng. 1976. Comparative mortality studies on Bionipluilaria glahrata (Mollusca: Pulmonatal exposed to copper internally and externally. J. Invert. Pathol. 28:255-257. 192. Cheng, T. C. 1976. Aspects of substrate utilization and energy requirement during molluscan phagocytosis. / Invert. Pathol. 27:263-268. 193. Cheng. T. C. 1976. Review of Genetics and Biochemistr\- of Pseudomonas edited by P. H. Clarke and M. H. Richmond. New York: John Wiley & Sons. Quart. Rev. Biol. 51: 314-315. 194. Cheng, T. C. 1976. The role of lysosomal enzymes in molluscan immunity. In: Programme and proceedings of the 19th annual meeting of the Canadian Federation of Biological Societies. Halifax, Nova Scotia, Dalhouise LIniversity. p. 158. 195. Cheng, T. C. 1977. Biocheinical and ultrastructural evidence for the double role of phagocytosis in molluscs: defense and nutrition. Comp. Pathohiol. 3:21-30. 196. Cheng. T. C. & J. T. Sullivan. 1977. Alterations in the osmoregulation of the pulmonale gastropod Biomphalaria glahrata due to copper. J. Invert. Pathol. 29: 101-104. 197. Renwrantz, L. R. & T. C. Cheng. 1977. Identification of agglutinin receptors on hemocytes of Heli.x pomatia. J. Invert. Pathol. 29:88-96. 198. Renwrantz. L. R. & T. C. Cheng. 1977. Agglutinin-mediated attachment of erthrocytes to hemocytes of Heli.x pomatia J. Invert. Pathol. 29:97-100. 199. Cheng, T. C, M. J. Chomey & T. P. Yoshino. 1977. Lysozyme-like activity in the hemolymph of Biomphalaria glahrata challenged with bacteria. J. Invert. Pathol. 29:170-174. 200. Cheng. T. C. 1977. The control of parasites: the role of the parasite. Uptake mechanisms and metabolic interference in parasites as related to chemotherapy. Proc. Helminth Soc. Wash. 44:2-17. 201. Foley. D. A. & T. C. Cheng. 1977. Degranulation and other changes of molluscan granulocytes associated with phagocytosis. / Invert. Pathol. 29:321-325. 202. Cheng, T. C. 1977. Scholarship in science. J. Invert. Pathol. 29: 1 15-1 17. 203. Yoshino. T. P., T. C. Cheng & L. R. Renwrantz. 1977. Lectin and blood group determinants of Schistosoma mansoni: alteration following in vitro transformation of miracidium to mother sporocyst. J. Parasitol. 63:818-824. 204. Cheng, T. C. & K. R. Auld. 1977. Hemocytes of the pulmonale gastropod Bi\- by G. D. Schmidt & L. S. Roberts. St. Louis: C. V. Mosby. Quart. Rev. Biol. 57: 211. 261. Cheng, T. C. 1982. Review of Parasitoloiiical Topics: presentation volume to P. C. C. Garnham F.R.S. on the occasion of his SOth hirlliday 1981. edited by E. U. Canning. Lawrence, KS: .Allen Press, Kansas Quart. Rev. Biol. 57:357-358. 262. Vasta. G, R„ J. T. Sullivan. T. C. Cheng, J. J. Marchalonis & G. W. Warr. 1982. A cell membrane-associated lectin of the oyster hemocyte. J. Invert. Pathol. 40:367^377. 263. Cheng, T. C. 1982. The (re-lemerging area of marine biomedicine. ./. Invert. Pathol. 40:155-158. 264. Cheng, T, C. 1983. The role of lysosomes in molluscan intlammation. Am. Zool. 23:129-144. 265. Cheng. T, C. 1983, Triggering of immunologic defense mechanisms of molluscan shellfish by biotic and ahiolic challenge and its applications. Mar. Teclwol. Soc. J. 17:18-25. 266. Cheng, T, C, K, H, Howland, H, J. Moran & J. T. Sullivan. 1983. Studies on parasitic castration: aminopeptidase activity levels and protein concentrations in lllynassa. obsoleta (Mollusca) parasitized by larval trematodes. J. Invert. Pathol. 42:42-50. 267. Cheng, T. C, K. H. Howland & J. T. Sullivan. 1983. Enhanced reduction of T4D and T7 coliphage titers from Bioniphalaria filabrata (Mollusca) hemolymph induced by previous homologous challenge. Biol. Bull. 164:418^27. 268. Cheng. T. C, J, T. Sullivan. K. H. Howland, T. F. Jones & H. J. Moran Studies on parasitic castration: soft tissue and shell weights of lllyanassa obsoleta (Mollusca) parasitized by larval trematodes. J. Invert. Pathol. 42:143-150. 269. Cheng, T. C. 1983. Review of Biocheinistiy of Parasitic Helminths by J. Barrett. Baltimore: University Park Press. Quart. Rev. Biol. 58:264-265. 270. Cheng. T, C, 1983. Internal defense mechanisms of molluscs against invading microorganisms: personal reminiscences. Trans. Am. Microsc. Soc. 102:185-193. 271. Cheng, T. C. 1983. Review of Platyhelnnnth Parasites of the Amphibia by S. Prudhoe and R. A. Bray. Oxford University Press. Quart. Rev. Biol. 58:445^46. 272. Cheng, T. C. 1983, Review of Pest Sluqs ami Snails: Biolof;y ami Control by D. Godan. Berlin; Springer. Quart. Rev. Biol. 58:574-575. 273. Cheng, T, C, 1983, Review of Molecular Biology of Parasites, vol, 2, edited by J. Guardiola. L. Luzzatto & W. Trager. Serono Symposia Publications. Quart. Rev. Biol. 59:60-61. 274. Cheng. T. C. 1984. A classification of molluscan hemocytes based on functional evidences. Coinp. Pathobiol. 6:1 1 1-146. 275. Cheng. T. C. 1984. Evolution of receptors. Coinp. Pathobiol. 5:33-50. 276. Cheng. T.C. & J, T. Sullivan 1984. Effects of heavy metals on phagocytosis hy molluscan hemocytes. Mar. Environ. Res. 14:305-3! 5. 277. Cheng, T. C, K, H. Howland & J. T. Sullivan. 1984. Comparison of the efficacy of twci maintenance iiiedui for gonads of llxanassa obsoleta (Mollusca: Gastropoda). Trans. Am. Microsc. Soc. 103:249-262. 278. Sullivan, J, T,. T. C. Cheng & C. C. Chen. 1984. Genetic selection for tolerance to niclosamide and copper in Biomphalanu alahruta (Mollusca: Pulmonata). Tropenmed. Parasit. 35:189-192. 279. Cheng. T. C, 1984, Review of Neoplasms — Comparative Patholo.i>y of Growth in .Animals. Plants, ami Man edited by H. E. Kaiser. Baltimore: Williams & Wilkins. J. Inverl Pathol. 44:120. 280. Cheng, T. C. 1984. Review of Black Apollo of Science: The Life of Ernest Everett Just by K. R. Manning. New York: Oxford University Press, 1 983. J Invert. Pathol. 43:443-444. 281. Vasta, G. R,, T, C, Cheng & J. J. Marchalonis. 1984. A lectin on the hemocyte membrane of the oyster {Crassostiea virginica). Cell Immunol. 88:475^88. 282. Sullivan. J. T.. T. C. Cheng & K. H. Howland. 1984. Mitotic responses of the anterior pericardial wall of Bioniphalaria glabrata (Mollusca) subjected to challenge. J. Invert. Pathol. 44:1 14-1 16. 283. Cheng, T, C, 1984, Several hats, J. Invert. Pathol. 44:1-2. 284. Sullivan, J. T.. T. C. Cheng & K. H. Howland. 1985. Studies on parasitic castration: castration of Llyanassa obsoleta (Mollusca: Gastropoda) by several marine trematodes. Trans. Am. Microsc. Soc. 104:154-171. 285. Cheng, T. C. 1985. Review of Diseases of Marine Animals, vols, 1 and II. edited by O. Kinne, New York: John Wiley & Sons, New York (vol, I) and Hamburg: Biologische Anstalt Helgoland, (vol. II). / Invert. Pathol. 46: 21-25. 286. Cheng, T. C. 1985. Evidence for molecular specificities involved in molluscan intlammation. Coinp. Pathobiol. 8:129-142. 287. Mohandas. A, & T, C, Cheng. 1985. An electron microscope study of the structure of lysoscmes relea.sed from Merceiiuria mercenaria granulocytes. J. Invert. Pathol. 46:332-334. 288. Cheng, T. C. & A. Mohandas Effect of high dosages of bacterial challenge on acid phosphatase release from Biomphalaria cilahrata hemocytes. J. Invert. Pathol. 45:236-241. 289. Pearson, E. J. & T. C. Cheng. 1985. Studies on parasitic castration: occurrence of a gametogenesis-inhibiting factor in extract of Zooiiomis lasiiis (Trematoda). / Invert. Pathol. 46:239-246. 290. Mohandas, A, cS: T, C. Cheng. 1985, Release pattern of aminopeptidase from Biomphalaria glahrata hemocytes subjected to high4evel bacterial challenge, / Invert. Pathol. 45:298-303. 291. Mohandas. A,. T, C, Cheng & J, B, Cheng, 1985, Mechanism of ly,sosomal enzyme release from Mercenaria mercenaria granulocytes: A scanning electron microscope study, J. Invert. Pathol. 46:189-197, 292. Cheng, T, C, & E, J, Pearson, 1986. Modification and evaluation of Burch and Cuadros' medium for the maintenance of the testes of a marine gastropod. Malacologia 27:173-183. 10 RODRICK 293. Brody. T., T. D. Matthews & T. C. Cheng. 1986. A23187 and EDTA stimulated enzyme release is associated with release of /inc from blood cells of Crassostrea virqinica. FASEB Prnc. 46(3):580. 294. Cheng, T. C. 1986. Specificity and the role of lysosomal hydrolases in moUuscan inflammation. lull. J Tiss. Reac. 8:439-445. 295. Cheng, T. C. 1986. Review oi Synopsis of Invertebrate Pathology Exclusive of Insects by A. K. Sparks. Amsterdam: Elsevier. J. Invert. Pathol. 48:131. 296. Cheng. T. C. 1986. Review oi Neurobiology of Arachnids edited by F. G. Barth. Berlin: Springer. J. Invert. Pathol. 48:388. 297. Cheng. T, C. 1986. Review of Biology of the Acanthocephala edited by D. W. T. Crompton and B. B. Nickol. Cambridge, UK: Camhndge University Press. Quart. Rev. Biol. 61:555-556. 298. Cheng. T. C. 1986. Biological control studies: bacteria associated with moribund Biomplaria glabrata (Mollusca) in the laboratory. J. Invert. Pathol. 47:219-224. 299. Jourdane, J. & T. C. Cheng. 1987. The two-phase recognition process of allografts in a Brazilian strain of Bioniphalaria glabrata. J. Invert. Pathol. 49: 145-158. 300. Combes, C. & T. C. Cheng. 1987. Control of biomedically important molluscs. Arch. Inst. Pasteur Algcrie 55:153-193. 301. Cheng, T. C. 1987. Some cellular mechanisms governing self and nonself recognition and pathogenicity in vertebrates and invertebrates relative to protistan parasites. Aquaculture 67:1-14. 302. Cheng, T. C. & J. Jourdane. 1987. Transient cellular reaction in Binmphalaria glabrata (Mollusca) to heterotopic isografts. J. Invert. Patliol. 49:273-278. 303. Cheng, T. C. 1987. Review of Heniocvtic and Humoral Imnuintx in Arthropods, edited by A. P. Gupta. New York: Wiley-Interscience. Science 236:1684-1685. 304. Cheng. T. C. 1987. Review of Z./r;>;,i; Together: The Biology of Animal Parasitism by W. Trager. New York: Plenum Press. Quart. Rev. Biol. 62:200-201. 305. Cheng. T. C, J. Jourdane & C. Combes. 1988. Les Mecanismes d'evitement dii rejet des parasites par I'hote: Revue syntheiique. I'.Annee de Biologic 27:73-92. 306. Cheng, T. C. & J. C. U. Downs. 1988. Intracellular acid phosphatase and lysosyme levels in subpopulations of oyster ^Cras.wstrea virginica) hemocytes. y./Hvcrt. Pathol. 52:163-167. 307. Cheng, T. C. 1988. In vivo effects of heavy metals on cellular defense mechanisms of Crassostrea virginica: total and differential cell counts. / Inven. Pathol. 51:207-214. 308. Cheng. T. C. 1988, In vivo effects of heavy metals on cellular defense mechanisms of Crassostrea virginica: phagocytic and endocytotic indices. J. Invert. Pathol. 51:215-220. 309. Jourdane. J. & T. C. Cheng. 1988. Approche evolutive des interactions entre schistosomes du groupe a oeuf a eperon terminal et leurs mollusques vecteurs. Bull. Soc. France Parasit. 6: 159-164. 310. Cheng. T. C. 1989. Immunodeficiency diseases in marine mollusks: measurements of some variables. J. Aqnat. Anim. Hc(dih I: 1 17-138. 311. Cheng. T. C, & W. J. Doughertv, 1989. Ultrastructural evidence for the destruction of Schistosoma mansoni sporocysts associated with elevated lysosomal enzyme levels in Biomplialaria glabrata. J. Parasitol. 75:928-941. 312. Rodrick. G. E. & T. C. Cheng. 1989. Parasites: occurrence and significance in marine animals. Food Technol. 43:98-102. 313. Cheng. T, C, 1990, Review of Parasite Lives edited by M. Cremin. C. Dodson & D. E. Moorhouse. Queensland, Australia: University of Queensland Press, Australia. Quart. Rev Biol. 64:206. 314. Cheng. T, C, 1990. Review of Free-living and Symbiotic Platheiminthes edited by P. Ax, U. Ehlers. B. & Sopott-Ehlers. Stuttgart: Gustav Fischer. Quart. Rev. Biol. 65:91-92. 315. Cheng. T. C. 1990. Review oi Principal Diseases of Marine Fish and Shellfish, vol. 2. Diseases of Marine .Shellfish. 2nd ed., by C. J. Sindermann. San Diego: Academic Press. / Invert. Pathol. 46:289. 316. Cheng. T.C. 1990. Review oi Comparative Biochemistiy of Parasitic Helminths, edited by E. M. Bennet. C. Behm & C. Bryant. London: Chapman & Hall. Quart. Rev. Biol. 65:371-372. 317. Cheng. T. C. 1990. Review of Biochemical Adaptation in Parasites by C. Bryant & C. A. Behm. London: Chapman & Hall. Quart. Rev. Biol. 65:370-371. 318. Cheng. T. C. 1990. Review of The Physiology and Biochennsliy of Cestodes by J. D. Smyth & D. P. McManus. Cambridge, UK: Cambridge University Press. Quart. Rev. Biol. 65:236-237. 319. Cheng. T. C. 1991. Review oi Myxosporidia of the USSR by S. S. Shuman. Rotterdam: A. A. Balkema. Quart. Rev Biol. 66:99. 320. Cheng. T. C. 1991. Review of Modern Parasite Biology: Cellular. Immun, 'logical, and Molecular Aspects, edited by D. J. Wyler. New York: W.H. Freeman. Quart. Rev. Biol. 66:1 10-1 1 1. 321. Cheng. T. C. 1992. Selective induction of release of hydrolases from Cras.sostrea virginica hemocytes by certain bacteria. J. Invert. Pathol. 59: 197-200. 322. Cheng. T. C. 1992. Requirement of a chelator during ionophore-stimulated release of acid phosphatase from Crassostrea \'irginica hemocytes. J. Invert. 323. Dougherty. W. J., T. C. Cheng & V. G. Burrell. Jr. 1993. Occurrence of the pathogen Haplosporidium iiclsoni m oysters, Crassostrea virginica. in South Carolina. Trans. Am. Microsc. Soc. 112:75-77. 324. Cheng. T. C. W. J. Dougherty & V. G. Burrell, Jr, 1993. Lectin-binding differences on hemocytes of two geographic strains of the American oyster. Crassostrea virginica. Trans. Am. Microsc. Soc. 112: 151-157. 325. Cheng. T. C„ W. J. Dougherty & V, G, Burrell. Jr. 1994. A possible hemocyte surface marker for resistance to Haplosporidiim nelsoni in the oyster Crassostrea virginica Res. Rev. Parasitol. 54:51-54. 326. Cheng. T, C. & W, J, Dougherty, 1994. Occurrence of "lalhyrose" on hemocytes of Haplosporidium nclsimi-free oysters from the Chesapeake Bay. Res. Rev. Parasitol. 54:117-120. 327. Cheng. T. C, & W, J. Dougherty 1994. Inhibition of hemocyte agglutination by Lathyrose odoratus lectin in the Chesapeake Bay. USA, oysters (Crassostrea virginica) infected with Perkiiisiis marinus. Res. Rev. Parasitol. 54:1 21-1 23. 328. Cheng. T, C. 1994. Molecular basis for ecological diversity as related to the American oyster and two symbionts. Agenda Abstr. 6th Intl. Colloq. Pathol Mar. Aquacult. p.l 2. In Memoriam: Thomas Clement Cheng II 329. Cheng, T. C. J. J. Manzi & V. G. Burrell. Jr. 1995, Differences in lectin binding by hemocytes of oysters (Crassoslrea vin/inica) from three regions and further evidence for the correlation between the presence of lathyrose and the absence of Haplosporicliuni iielsoiii. J. Shellfish Res. 14:477-48 1 . 330. Cheng. T. C. & W. J. Dougherty. 1995. Partial inhibition of hemocyte agglutination by Luihvni.s odorams lectin in Cras.sostrea virginicci infected with Perkinsus mariiius. Mem. Insi. Oswaldo Cruz 90: 407—110. 331. Cheng. T. C. V. G. Burrell, Jr., J. J. Manzi & F. O. Perkins. 1995. Natural and experimental mfection oi Mercenarui mercenaria by Perkiiisu.s muriniis from Crassostrea virginica. Res. Rev. Parasit. 55:31-34. 332. Cheng. T. C. & J. J. Manzi. 1996. Correlation between the presence of lathyrose with the absence of Haplosporidium nelsoni in Crassostrea virginiea from two South Carolina tributaries where Perkinsus mariniis also inhibits hemocyte agglutination by the Lathyrose odoratus lectin. ./. Shellfish Res. 15:391-394. Bobo, M. Y.. D. Richardson. T. C. Cheng, E. McGovern & L. Cohen. 1996. Seasonal cycle of Haplosporidium nelsoni (MSXl in intertidal oysters, Crassostrea virginica. in South Carolina. National Shellfish. Association Annual Meeting. Abstract, p. 59. 333. BOOKS AND EDITED VOLUMES Cheng. T. C, editor. 1963. Some biochemical and immunological aspects of host-parasite relationships. New York: Annals of the New York Academy of Science, vol, 1 13. 509 pp. Cheng, T. C. 1964. The biology of animal parasites. Philadelphia: W. B. Saunders, 727 pp. Cheng, T. C. 1967. Manne molluscs as hosts for syinbioses: with a review of known parasites of commercially important species. In: F. S. Russell editor. Advances in marine biology, vol. 5. London and New York: Academic Press. 424 pp. Cheng, T. C. 1970. Symbiosis: organisms living together, Indianapolis: Bobbs Merrill Co. 150 pp. Cheng, T. C, editor. 1971. Aspects of the biology of symbiosis. Baltimore: University Park Press. 327 pp. Cheng, T. C, editor. 1971. Current topics in comparative pathobiology. vol. 1. San Diego: Academic Press. 277 pp. Cheng. T. C. 1973. General parasitology. San Diego: Academic Press. 998 pp. (Translated into Spanish and published by Editorial AC Libros Cientificos y Technicos, Madrid, Spain, 1978.) Cheng. T. C. editor. 1973. Current topics in comparative pathobiology. vol. 2. San Diego: Academic Press. 334 pp. Cheng. T. C. 1973. Human parasites transmissible by seafood and related problems. In: C. O. Chichester & H. Graham, editors. Microbiological quality of seafoods. San Diego: Academic Press, pp. 163-189, Malek. E, A. & T. C, Cheng, 1974. Medical and economic malacology. San Diego: Academic Press. 398 pp. Cheng, T. C, editor. 1974. MolUiscicides in schistosomiasis control. San Diego: Academic Press. 266 pp. Bulla, L, A.. Jr. & T, C. Cheng, editors. 1975. Pathobiology of invertebrate vectors of disease. Neew York: Annals of the New York Academy of Sciences, vol. 266. 540 pp. Bulla. L. A., Jr. & T. C. Cheng, editors. 1976. Comparative pathobiology, vol, I. biology of the microsporidia. New York: Plenum, 371 pp. Bulla, L. A.. Jr. & T. C. Cheng, editors. 1977. Comparative pathobiology. vol. 2. systematics of the microsporidia. New York: Plenum. 510 pp. Bulla. L. A.. Jr. & T. C. Cheng, editors. 1977. Comparative pathobiology, vol. 3. invertebrate immune responses. New York: Plenum. 192 pp. Bulla, L. A., Jr. & T. C. Cheng, editors. 1978. Comparative pathobiology, vol 4. invertebrate models for biomedical research. New York: Plenum. 167 pp. Cheng, T. C. editor. 1984. Comparative pathobiology. vol. 5. structure of membranes and receptors. New York: Plenum, 297 pp. Cheng, T. C, editor. 1984. Comparative pathobiology, vol. 6. invertebrate blood: functions of serum factors and cells. New York: Plenum. 21 1 pp, Cheng, T, C. editor. 1984. Comparative pathobiology. vol. 7. pathogens of invertebrates: applications in biological control and transmission mechanisms. New York: Plenum. 278 pp. Cheng, T. C. editor. 1985. Comparative pathobiology, vol. 8. parasitic and related diseases. New York: Plenum. 168 pp. Cheng, T. C. 1986. General parasitology. 2nd edition. San Diego: Aca- demic Press. 827 pp. Bogitsh, B. J. & T. C. Cheng. 1990. Human parasitology. San Diego: Academic Press. 435 pp. (2nd edition. 1998). Perkins. F. 0. & T. C. Cheng, editors. 1990. Pathology in marine science. San Diego: Academic Press. 538 pp. Cheng, T, C. 1964. Comparative electrophoretic studies on the sera of marine and freshwater molluscs. In: C. .\. Leone, editor. Taxonomic biochemistry and serology. New York: Ronald Press, pp. 659-666. Cheng, T, C. 1971. Enhanced growth as a manifestation of parasitism and shell deposition in parasitized mollusks. In: T, C, Cheng, editor. As- pects of the biology of symbiosis. Baltimore: University Park Press, pp. 101-137. Cheng, T. C. & G. E. Rodrick. 1974. The use of respirometry as a bioassay method for pollutants in sea water. In: G. LaRoche. editor. Proceedings of the marine bioassays workshop, Washington, DC: Marine Techno- lology Society, pp. 174—181. Cheng, T. C. 1974. Future research on molluscicides: a summary. In: T C. Cheng, editor. Molluscicides in schistosomiasis control, San Diego: Academic Press, pp. 259-266. Cheng. T. C. A. Call & D. .A. Foley. 1974. Cellular reactions in marine pelecypods as a factor inlluencing endosymbioses. In: W. B. Vernberg. editor. Symbiosis in the sea. Columbia. SC: University of South Caro- lina Press, pp. 61-91. Cheng, T. C. & J, T. Sullivan. 1974. Mode of entry, action, and toxicity of copper molluscicides. In: T. C. Cheng, editor. Molluscicides in schis- tosomiasis Control. San Diego: Academic Press, pp. 89-153. BOOK CHAPTERS Cheng, T. C. 1975. Schistosomiasis: the worid's major medical problem. Proceedings of 1975 International Controlled Release Pesticide Sym- posium, Dayton. OH: Wright State University Press, pp, 133-156. Cheng, T. C. 1976. Humoral immunity in molluscs. Proceedings of the IXth International Colloquium on Invertertabrate Patholology. Kings- ton, Ontario: Queens University Press, pp. 190-194. Healy. G. R., G. J. Jackson. J. R. Lichtenfels. G. L. Hoffman & T. C. Cheng. 1976. Foodbome parasites. In: M. L. Speck, editor. American Public Health Association compendium of microbiological methods for parasites in foods. Chicago: American Public Health Association, pp, 471-483. Cheng, T. C, 1976. Liver and other digestive organs. In: C. R. Kennedy, editor. Ecological aspects of parasitology. Amsterdam: Elsevier Ex- cerpta Medica North Holland, pp. 287-302. Cheng, T. C. 1979. Use of copper as a molluscicide. In: J. O. Nriagu. editor. Biogeochemistry of copper. New York: John Wiley & Sons. pp. 401-432. Cheng. T. C. 1981. Parasitology. In: R. L. Moon & D. D, Whitt. editors. Highlights in microbiology 1979-1980.1 1 Am. Soc. Microbiol, pp. 3-1- 35. 12 RODRICK Cheng. T. C. 1981. Bivalves. In: N. A. Ratcliffe & A. F. Rowley, editors. Invertebrate blood cells. London: Academic Press, pp. 233-300. Cheng, T. C. 1982. Cutaneous lesions due to nonarthropod parasites. In: L. C. Parish. W. B. Nutting & R. M. Schwartzman, editors. Cutaneous infestations of man and animal. New York: Praeger. pp. 237-254. Cheng. T. C. 1982. Anisakia.sis. In: .J. H. Steele, editor. Handbook of zoonoses, vol. 11. sect. C. Boca Raton. FL: CRC Press, pp. 37-54. Cheng. T. C. 1983. Parasitism and symbiosis. In: M. Rechcigl. editor. Handbook of nutrition and food. Boca Raton, FL: CRC Press. Healy, G. R.. G. J. Jackson, J. R. Lichtenfels. G. L. Hoffman & T. C. Cheng. 1984. Foodborne parasites. In: M. L. Speck, editor. Compen- dium of methods for the microbiological examination of foods, 2nd ed. Washington. DC: American Public Health Association, pp. 542-556. Cheng. T. C. J. J. Marchalonis & G. R. Vasta. 1984. Role of molluscan lectins in recognition processes. In: E. Cohen, editor. Recognition pro- teins, receptors, and probes: invertebrates. New York: Alan R. Liss. pp. 1-15. Cheng. T. C. 1985. Newly recognized pathogens. In: R. R. Colwell. editor. Natural toxins and human pathogens in the marine environment. Mary- land Sea Grant Publication. College Park: University of Maryland Press, pp. 39-42. Cheng, T. C. 1986. Summary and some comments (to Proceedings of the 2nd International Colloquium on Pathology in Marine Aquaculture). Oporto. Portugal: University of Oporto Press, pp. 169-173. Combes. C. & T. C. Cheng. 1986. Les ba.ses conceptuelles du controle des Schistosomoses. In: Congres sur la Protection de la Sante Humaine et des cultures en Milieu tropical. Marseille: Press Chambre de Com- merce et d'lndustrie de Marseille, pp. 529-532. Cheng, T. C. 1987. /;; vivo effects of heavy metals on cellular immunity m molluscs. In: J. V. Dorigan & F. L. Harrison, editors. Physiological responses of marine organisms to environmental stresses. Washington, DC: Office of Energy Research, U.S. Department of Energy (DOE/ ER-0317). pp. 59-81. Cheng. T. C. 1988. Strategies employed by parasites of marine bivalves to effect successful establishment. In: W. S. Fisher, editor. Disease pro- cesses in marine bivalve molluscs. Bethesda. MD: American Fisheries Society Special Publication 18. pp. 112-129. Cheng, T. C. 1990. Comments on nutritional diseases. In: F. O. Perkins & T. C. Cheng, editors. Pathology in marine science. San Diego: Aca- demic Press, pp. 439—440. Cheng, T. C. 1990. Effects of in vivo exposure of Cnissoslrea virginica to heavy metals on hemocyte viability and activity and levels of lysosom- al enzymes. In: F. O. Perkins & T. C. Cheng, editors. Pathology in marine science. San Diego: Academic Press, pp. 513-524. Cheng. T. C. & C. Combes. 1990. Influence of environmental factors on the invasion of molluscs by parasites: with special reference to Europe. In: F. di Castri. A. J. Hansen & M. Debussche, editors. Biological invasions in Europe and the Mediterranean basin. Dordrecht. Germany: Kluwer Academic Publishers, pp. 305-330. Cheng, T. C. 1991. Is parasitism symbiosis? A definition of temis and the evolution of concepts. In: C. A. Toft. A. Aeschlimann & L. Bolis. editors. "Parasite-host associations, coexistence or conflict?" Oxford. UK: Oxford University Press, pp. 15-36. Cheng. T. C. 1993. Noninfectious diseases of marine mollusks. In: J. Couch & J. W. Fournie. editors. Pathobiology of marine and estuarine organisms. Boca Raton, FL: CRC Press, pp. 289-318. Cheng. T. C. 1996. Oyster hemocytes: forms and functions. In: V. S. Kennedy, R. I. E. Newell & A. F. Eble. editors. The eastern oyster Crassostrea virginica. College Park: University of Maryland Press, pp. 299-333. Journal of Slh'llthh Resfunh. Vol. 21). No. I. 13-14. 2001. ANOTHER INTRODUCED MARINE MOLLUSK IN THE GULF OF MEXICO: THE INDO-PACIFIC GREEN MUSSEL, PERNA VIRIDIS, IN TAMPA BAY, FLORIDA DEBRA A. INGRAO,' * PAULA M. MIKKELSEN," AND DAVID W. HICKS' Center for Coastal Ecology. Mote Marine Ljihoratory. 1600 Ken Thompson Parkway. Sarasota. Florida. 34236: 'Division of Invertebrate Zoology. American Museum of Natural History. Central Park West at 79th Street. New York. New York 10024-5192: ''Department of Biology. Lamar University. P.O. Box 10037. Beaumont. Texas. 77710 ABSTRACT In July 19yy. green niu^^els were seen for the first time at Tampa Electric Company's (TECO) Gannon Station Power Plant (Gulf of Mexico, Hillsborough Bay portion of Tampa Bay. western coast of Florida) during routine maintenance of the seawater intake system. The mussels were identified as Permi viridis (Linnaeus 1758) based on shell and anatomical characters. This identi- fication has been confirmed by cytological analysis. Periui viridis had previously not been found in the United Stales or in the Gulf of Mexico. Penui viridis is a native species in the Indian and Pacific Oceans and was discovered in the Caribbean Sea in Trinidad in 1990 and in Jamaica in 1998. The populations are widespread locally and well established; the largest specimen collected was estimated to be 15 months of age. This is the second species of Periia introduced to the Gulf of Mexico, joining Periui periui (Linnaeus 1758). which was discovered in Texas in 1990. Since the initial discovery of Penni viridis in Tampa Bay, range surveys have found it as far north as John's Pass in St. Petersburg, Pinellas County. Florida and as far south as Boca Grande. Charlotte County. Florida. KEY \yORD.S: Pirna viridis. Bi\alvia. exotic species, invading species. Mytilidae. Florida, aquaculture INTRODUCTION The introduction of non-native species through natural, inten- tional or accidental transport by human activities is a threat to natural flora and fauna and can cause severe economical impact (Meineke 1999). According to Carlton (1992), the major human activities resulting in the introduction of non-native species in North America are the following: the transponation of organisms on the outside (fouling species) or on the inside (boring species) of ships: the transportation of organisms inside vessels in solid ballast (rocks, sand and detritus) or ballast water from coastal, transoce- anic, and interoceanic vessels; the transport of oysters (including organisms, sediment and detritus on or in the oyster shells); and the intentional release of species for fisheries purposes. United States coastal and inland waters now support a large list of invasive bivalve species (e.g., Perna perna. Mytilus galloprovincialis Lamarck IS 19, Dreissemt polymorpha (Pallas 1 77 1), D. bugeusis Andrusov 1897. Musculista senhoiisia (Benson 1842), Corhicula flmninea (Miiller 1774): see Turgeon et al. (1998) for full list). In July 1999, Tampa Electric Company (TECO) requested Mote Marine Laboratory (Sarasota, Florida) to identify "green mussels" that were discovered fouling the water intake structures of TECO Gannon Station Power Plant in the Hillsborough Bay portion of Tampa Bay, Florida (27°54'25"N: SP^S'SS'^). The green mussels were immediately recognized as members of the tropical/subtropical genus Perna comprised of three extant spe- cies: P. perna (Linnaeus 1758). P. viridis (Linnaeus 1758), and P. canaliculus (Gmelin 1791) (reviewed by Siddall 1980a), The three Perna species are similar morphologically and exhibit considei- able variation in taxonomically important tnorphological charac- ters (Siddall I98()a; Sadacharan, 1982, Holland et al. 1999), mak- ing it difficult to separate the species. In the past, Perna species were easily distinguished from one another on the basis of their mutually exclusive geographical dis- *Corresponding author. E-mail: debi@mote.org; Fax: +1-941-388-4312. Editor's note: manuscript received January 26. 2001; accepted for publi- cation March 3, 2001. tributioiis. Perna perna occurs natively in warm tropical and sub- tropical coastal regions of South America and Africa (Jeffs et al. 1999) and also extends into the Mediterranean (Berry 1978). Perna viridis is a tropical euryhaline species, widely distributed natively throughout the Indo-Pacific having a western limit from the Per- sian Gulf an eastern limit at Papua New Guinea (Morton 1987), and northern limit to southern Japan (Cheung 1993), including southern China (Huang et al. 198.3). Perna canaliculus is only found in coastal waters of New Zealand (Siddall 1980a) and is found on North. South, and Stewart Islands but is most abundant in central and northern New Zealand (Powell 1979; Jeffs et al. 1999). Today, these species can no longer be distinguished from one another based only on geographical location. Segnini de Bravo et al. (1998) mention that Perna perna colonized the Caribbean Sea several decades ago. Diaz Meriano and Hegedus (1994) report a series of fresh shells found on the beach near Bocachia along the Columbian coast. The geological range of P. perna expanded again in 1990 when two small specimens (2 cm in length) were collected in the western Gulf of Mexico from the Port Aransas Jetty on the east coast of Texas in February 1990 (Hicks & Tunnell 1993). The Texas P. perna are now found from the mouth of the Colorado River, Texas south to Veracruz. Mexico (Sea Grant News Media Center 'Website 1999). The geographical range of P. viridis has also expanded. Morton (1987, p. 159) reports "Siddall (1980a) does not consider P. viridis to be naturally distributed along the coast of China or Japan and Arakawa . . . believes the species was introduced into Japan sometime around 1967." In 1990 P. viridis was first observed in the southern Caribbean Sea at Port Lisas in Trinidad and by 1992 had spread along the entire Trinidad coast of the Gulf of Paria (Agard et al. 1992). In 1993 P. viridis appeared on the 'Venezuelan coast of the Gulf of Paria and by 1995 was discovered in Isla Margarita (Segnini de Bravo et al. 1998). According to Segnini de Bravo et al. (1998), P. viridis is driving P. perna from its natural beds in La Esmeralda, Guatapa- nare and El Morro de Chacopata. Perna viridis was later discov- ered in the northern Caribbean in Kingston Harbour. Jamaica in 1998 (D. Buddo and T. Bowes, pers. comm.). Perna canaliculus is 13 14 Ingrao et al. still geographically isolated, however Carlton (1992) expresses concern that P. canaliculus will spread to California where it is now imported daily in large numbers for human consumption. The discovery of introduced populations of P. perna and P. viridis make the identification of these species based solely on geographic location unworkable. Since Perna perna and P. viridis are no longer geographically separated and are difficult to differentiate from each other because of their highly variable morphological characters, other laboratory procedures are necessary to positively identify these species. Perna viridis and P. perna have been shown to express dimor- phism in chromosome numbers and can be distinguished from each other using cytological characters. Perna viridis has 15 pairs of homologous chromosomes (Ahmed 1974) and P. perna (includ- ing introduced Texas populations) has 14 pairs of homologous chromosomes (Ahmed 1974; Jacobi et al. 1990: Holland et al. 1999). Unfortunately, no karyotype analysis is available for P. canaliculus. In this report, we confirm the identity of green mus- sels in Hillsborough Bay as P. viridis utilizing both morphometric and cytogenetic characters. We further describe its present well- established distribution throughout and beyond the Tampa Bay estuarine system. MATERIALS AND METHODS Initial specimens for identification were collected at the TECO Gannon Station Power Plant in Hillsborough Bay (Fig. 1). Speci- mens were collected from the screens of the water intake channel by a diver and from inside the intake tunnels during routine main- tenance on 15 July 1999. Preserved voucher specimens are depos- ited in the Division of Invertebrate Zoology. American Museum of Natural History. New York ( AMNH 290101 ), Mote Marine Labo- ratory (MML 8027.00. 8028.00. 8029.00) and al Lamar University (LU 100399). Specimens for cytological analyses were collected from the intake boot (flared flange at the base of the water intake pipe) and tunnels of the same power station on 13 October 1999. Chromo- some preparations were obtained from gill tissue using a modified colchicine-Giemsa technique (Yaseen 1995; Holland et al. 1999). Specimens were immersed in 0.05% colchicine in aerated 35 ppt seawater at rooiu temperature for 12 h. After this, gill tissues were isolated and placed into a hypotonic solution of 0.8% sodium citrate for 40 min. Tissues were fixed twice in freshly prepared mixtures of 3 parts methanol to 1 part glacial acetic acid for 20 min each, then left in the fixative for 12 h. Cell suspensions were prepared by gently grinding tissues with a microtube pestle in a 1 .5 ml microcentrifuge tube containing 60% acetic acid. Disaggre- gated tissues were then transferred back into fresh fixative by pipette. Cell suspensions were then dropped on cleaned glass slides held at a 45 ° angle from a height of 15 cm. Slides were air-dried then stained with a 24:1 mixture of 0.1 M phosphate buffer (pH 6.8) and Giemsa for 5 min. Photographs were taken of represen- tative mitotic spreads and karyotypes were formed. Florida Fish and Wildlife Conservation Commission (Florida Figure L ( A) Confirmed sightings of Penm viridis in the Caribbean Sea and (iulf of Mexico. (B) Selected documented sightings of Perna viridis along the central west coast of Florida. Perna viRiDis Discovered in Tampa Ba^, Florida 15 Marine Research Institute. St. Petersburg, Florida), Mote Marine Laboratory (Sarasota. Florida), U. S. Coast Guard Cutter Josluui Appleby (based in St. Petersburg, Florida), Florida Caribbean Sci- ence Center (U. S. Geological Survey. Gainesville, Florida), and Great Lakes Science Center (U. S. Geological Survey. Ann Arbor. Michigan) conducted several surface and diving surveys to deter- mine the extent of the introduction of P. riridis along the western coast of Florida. RESULTS Morphological characters of 26 mature specimens ranging from 24-87 mm in shell length were examined. All specimens have a two-part retractor muscle scar and no anterior adductor muscle scar that is consistent with mussels belonging to the genus Perna (mussels belonging to the genus Mytilus have a single retractor muscle and an anterior retractor muscle) (Siddall 1980a; Vakily 1989). The specimens were predominantly brilliant green and blue-green sometimes having concentric blue-green bands and the central portion of the shell light brown in some specimens, match- ing descriptions for P. viridis according to Siddall ( 1980a,b), Va- kily ( 1989), and Dance ( 1974). Adult shell shape and muscle scars (Fig. 2) corresponded closely to P. viridis. as depicted in Siddall's (1980a) comparative review. The mantle margin of preserved and living specimens showed no enlarged sensory papillae, which are characteristic of P. perna (Siddall 1980a). Morphological obser- vations of the 26 specimens suggested identification as P. viridis but did not conclusively exclude P. perna. known to occur along the nearby Gulf coast of Texas. A diploid number of 2n = .30, characteristic of Perna viridis. was scored in 18 mitotic metaphases from two individuals. Quali- tative morphological groupings of chromosomes indicated 10 meta-submetacentric and .'i subtelocentric pairs (Fig. 3), consistent with the chromosome morphology for P. viridis presented by Ahmed ( 1974). These results confirmed the tentative morphologi- cal identification as P. viridis. and indicate the introduction of a second alien pernid mussel to the Gulf of Mexico. Since its initial discovery at TECO Gannon Station Power Plant in Hillsborough Bay, TECO has reported the presence of P. viridis at their Big Bend Station Power Plant in Lower Hillsborough Bay since July 1999 (E. Bietia. pers. comm.); and Florida Power Cor- poration, Bartow Power Plant on Weedon Island near the entrance to Old Tampa Bay, Florida (27°5r40"N; 82°36'09"W) first no- ticed these mussels in July 1999 (D. Bruzek, pers. comm.). Range surveys have found living P. viridis specimens as far north as John's Pass in St. Petersburg. Pinellas County, Florida, as far south as the Venice Fishing Pier (Venice, Sarasota County. Florida) in the Gulf of Mexico (M.A. Blouin, A. J. Benson, pers. comm.). The largest specimens (many exceeding 100 mm) have been found in Tampa Bay. Specimens found outside of Tampa Bay have been relatively small (1— M) mm). Well-established communities of P. viridis have been found in Hillsborough Bay. Old Tampa Bay, Middle Tampa Bay, Lower Tampa Bay, Boca Ciega Bay, and Lower Manatee River (D.C. Marelli, A.J. Benson, M.A. Blouin. C. O'Neil, pers. comm.). The first specimen (10 mm) discovered in Sarasota Bay was found on an invertebrate collecting box in New Pass in January 2000. Four out of eight specimens collected from the Mote Marine Laboratory's New Pass intake pump (19 April 2000) were completely brown (including the inside lip). Over 800 specimens of P. viridis, sizes ranging from 0.6-10 mm, were in- advertently collected on a PVC tidal gauge that had been attached to the Apollo Beach Pier for 41 days (29 April to 8 June 2000). Juvenile specimens collected off this tidal gauge which ranged in size from 0.6 to approximately 2.5 mm were golden colored with concentric reddish-brown zigzag markings; speciinens greater than 2.5 mm started to develop a green lip; specimens greater than 3 mm showed alternating concentric green banding with the reddish- brown zigzag markings; specimens greater than 4.5 mm started developing their green lip with no banding. Several completely brown juveniles were found. DISCUSSION Carlton's 1992 review of introduced marine and estuarine mol- lusks of the Atlantic, Gulf and Pacific coasts of North America revealed that 36 introduced species have become established. Perna perna was the only clearly introduced marine mollusk in the Gulf of Mexico (Cariton 1992). Turgeon et al. (1998) states there are now 47 introduced species of marine and estuarine inollusks in North America; 25 bivalves and 22 gastropods. Perna viridis now Figure 2. Shell morphology i>( Perna riridis from Tampa Bay, Florida. (\( Kxternal view, shell length S7 nmi. |I5) Internal view of right valve, shell length 104 mm: lower view with enhanced muscle scars and pallial line. APR, anterior pedal retractor muscle; .MPR, middle pedal retractor muscle; PPR, posterior pedal retractor muscle; PA, posterior adductor muscle. 16 INGRAO ET AL. B ..^ •« <0 f 9 •% 0 mm shell length — approximately 2-3 months after fertili/ation (references in Siddall 1980b; Vakily 1989). In stable tropical environments, spawning can occur year round exhibiting two more-or-less promi- nent peaks (Vakily 1989: Siddall 1980b; Huang 1985). Spawning can also be easily induced by rapid changes in water temperature (Vakily 1989; Siddall 1980b). Larvae settle within l?-20days if a suitable substratum (i.e. some type of hard surface — natural or man-made) is present. II II M «l *i • t li II It •• 9 10 11 12 13 Ai 14 15 Figure 3. .Mitotic chromosomes (A and B) and standard karyotype (Cl (from B) prepared from gill tissue from I'enia viridis specimens col- lected October 13. 1999 from TECO's Gannon Power Station. Hills- borough Bay (in Tampa Bay), Florida. The 15 pairs of chromosomes are diagnostic for P. viridis. joins these lists as the second introduced mollusk in the Gulf of Mexico and warrants concern as a potential ecological threat to coastal waters of North America (Morton 1997). Penui vindis shares many characters common to other invasive bivalve species (Morton 1997). For example: P. vindis has an extensive capacity for larval and adult dispersal, early maturity, rapid growth rate, and high productivity. Periia viridis has also been very successful at finding suitable habitat along the south- eastern coast of the Gulf of Mexico. Extensive Capacity for Dispersal The planktotrophic larval stages of Penui vindis remain free- swimming for 10-20 days (Siddall 1980b. Vakily 1989). during which they can be dispersed over great distances by ocean currents Rapid (ho will The combined effects of a number of environmental factors such as water temperature, salinity, currents, food availability, densities and pollution have a considerable effect on the growth rates of P. viridis (see Vakily 1989). Morton ( 1987. p. 163) stated, ■■Throughout its broad range. P. viridis has been reported to have a phenomenal growth rate of some 10 mm per month . ."' Vakily ( 1989) examined previously published data and compiled growth rates for wild cultures of P. viridis ranging from 2.2-10.6 mm per month. Huang et al. (1983) projected an annual growth rate of 60 mm in the relatively polluted waters of Tolo Harbour. Hong Kong. A rough estimate of growth rate of Florida specimens was calcu- lated using the largest specimen (10 mm) collected from the PVC tide gauge at Apollo Beach Pier using shell length/number of days substratum was submerged. This calculation assumes P. vindis larvae settled almost immediately, the largest specimen collected was the first to settle, no predation. and a constant growth rate. The estimated growth rate for this specimen was 7.2 mm per month. The following age approximations are based on this growth rate. The specimens collected initially from the TECO Gannon Station Power Plant (15 July 1999) ranged from 24.4-86.6 mm in length, and are estimated to be 3-12 months old. suggesting that initial settlement occurred in July 1998. Some of the largest specimens (over 100 mm) were collected from shipping channel marker buoys in Middle Tampa Bay (14 December 1999): the largest specimen (109.8 mm) is estimated to be over 15 months old sug- gesting initial settlement occurred in September 1998. The speci- mens collected on the PVC tide gauge at Apollo Beach (8 June 2000) range in size from 0.6-10 mm and are estimated to be 2-41 days old. There is considerable fluctuation in growth rates and popula- tion densities of all species of Perna caused by the combined effects of several environmental factors such as water temperature, food availability (especially chl-a and particulate organic carbon (Rivonkar et al. 1993)], setting densities, currents exposure, and Perna viRiDis Discovered in Tampa Bay, Florida 17 pollution (Vakily 1989). The highest densities of P. viriilis in Hong Kong are consistently recorded from Tolo Harbour (4000 adult individuals m~-) and Victoria Harbour (200-246 adult individuals m"") which are polluted by domestic, agricultural and industrial effluents (Huang et al. 1983). The lowest densities in Hong Kong waters (<100 m"") are recorded from eastern (exposure to strong wave action), southern (exposure to strong wave action) and west- em waters (salinities consistently <5 ppt) (Morton 1987). Huang et al. (1983) reported the highest density recorded from Tolo Harbour was 35,000 m"". When living in a favorable area, P. viridi.s has been found to spawn year round (reference in Huang et al. 1983). Cheung ( 1993) studied P. viridis population dynamics in the polluted Tolo Har- bour. Hong Kong and states (p. 20) "With high densities and a fast growth rate, the productivity of the mussel bed is considerable and is higher than many bivalve populations." Hicks and Tunnell (1995) reported that the closely related P. perna could be found forming a distinct mussel belt on the lower mid-intertidal zone of granite jetties in densities of up to 27.200 m"" small individuals (mean = 16 mm). These jetties commonly support 10,000-12,000 individuals m"" (Hicks and Tunnell 1995). Hicks and Tunnell have found on average 25.000-30,000 individuals m"" of P. penia spat; the maximum density found was 127,000 m^* (Sea Grant Media News Center website 1999). Habitat Tampa Bay provides hard substrata and temperature/salinity regimes siinilar to those where P. viridis naturally occurs in the Indian and southwestern Pacific Oceans as well as areas such as Japan and Venezuela where P. viridis has been introduced. It is a subtropical estuary located on the west central coast of peninsular Florida and is the largest open water estuary in the state (Estevez 1989; Lewis and Estevez 1988). It is a Y-shaped embayment, 56 km in length and 16 km wide encompassing over 1000 km" (Es- tevez 1989; Pribble 1999). The average depth of the bay is ap- proximately 4 m. The deepest natural area of Tampa Bay (27 ni) is near the mouth of the bay (Pribble 1999). Tampa Bay is also a highly urbanized area and supports many shipping and recreational boating activities. The Port of Tampa is Florida's largest port and consistently ranks among the top 10 ports nationwide in shipping tonnage (Tampa Bay Estuary Program 2000; Meineke 1999). Over 25 million tons of phosphate and related products move through the Port of Tampa annually — more than any other port in the world (Tampa Bay Port Authority 2000). Ships dumped an esti- mated 543 million gallons of ballast water into Tampa Bay in 1996 (Pittman 1999) from around the world, thus providing an excellent transport mechanism for P. viridi.s. Perna viridis has been able to maintain dense populations in Tampa Bay and is spreading along the southeastern Gulf of Mexico shore and estuaries despite the lack of continuous, natural, rocky intertidal habitats just as P. perna has done on the south- western Gulf of Mexico shore since its introduction to Texas in 1990. Perna viridis can be found attached to hard substrata such as rock faces, wharf piles, among algal holdfasts, and mangrove roots. In deeper waters, where hard substrata are not available, initial attachment may be to sediment grains, shell fragments or other mussels resulting in mussel beds on sandy or muddy bottoms (Agard et al. 1992; Hickman 1991; Jeffs et al. 1999; Berry 1978). Disjunct man-made structures including pier pilings, channel markers, buoys, jetties and bulkheads provide excellent surfaces for Perna larvae to settle and grow on. Structures like these are common along the southeastern Gulf of Mexico. Perna viridis prefer temperatures from 3 1-32 °C in the summer and 26-28 °C in the winter (Vakily 1989; Agard et al. 1992). Nearly linear growth was evident when temperatures were from 23.6-30.8 °C and cessation of growth occurred when winter tem- peratures were from 17-18 C in Tolo Harbour, Hong Kong (Che- ung 1993). The low (6 °C) and high (37.5 °C) lethal temperatures for P. viridis were recently determined through laboratory experi- ments (Segnini de Bravo et al. 1998). Water temperatures range from 11-32 "C in subtidal areas of Tampa Bay and vary more widely in intertidal areas (Lewis & Estevez 1988). Maximum wa- ter temperatures of 28-30 °C are found from June through August; minimum temperatures of 15-18 °C occur from December through February (Pribble 1999). Perna viridis prefers salinities 27-33 ppt although laboratory tests showed 50% survival rate after two weeks of exposure to 24 and 80 ppt (reference in Vakily 1989; Agard et al 1992). The low (0 ppt) and high (64 ppt) lethal salinities for P. viridis were re- cently determined through laboratory experiments (Segnini de Bravo et al. 1998). Perna viridis can also survive extended expo- sures to salinities as low as 20 ppt that can occur after a heavy rain (Agard et al. 1992). As would be expected, salinity varies through- out Tampa Bay — higher salinities are found near the mouth of the bay in aieas that interact strongly with the Gulf of Mexico; lower salinities are found in areas affected by the inflow of fresh water from rivers and occur in regions farthest away from the Gulf of Mexico (Pribble 1999). Maximum salinities, up to 38 ppt usually occur in June (Pribble 1999). Lowest salinities, down to 15 ppt usually occur in September (Pribble 1999). The rapid dispersal of P. viridis. its high thermohaline tolerance and establishment of successful colonies in the southern Caribbean Sea and Tampa Bay suggest that it will continue to spread in the Gulf of Mexico. SUMMARY Rayl ( 1999, p. 6) states: "Experts estimate that aquatic invasive species cost the United States at least $2 billion every year, in- cluding fisheries losses, removal and recovery efforts, and sport fishing losses — and the figure is projected to rise." Considering the reproductive rate, growth rate of Perna viridis and the simi- larity in environmental conditions between the southwestern coast of Florida and this mussel's natural habitats, area scientists, re- source managers, and others are concerned about its potential im- pact to native communities as well as coastal economics. Intro- duced aquatic species frequently cause massive transformation of the structure and function of native eco.systems. Like other intro- duced species, P. viridis has the biotic potential to flourish and to out-compete native species. Economically, the presence of dense mussel colonies on water intake systems, channel markers and buoys, threatens to increase the frequency of maintenance activi- ties and their associated costs. The capacity of P. viridis to attach byssally to hard surfaces makes it a major macrofouling threat to industrial facilities drawing raw seawater for operations (Rajago- pal et al. 1995). Perna viridis is already creating serious opera- tional problems for many power plants and other water intake systems around the worid such as Madras Atomic Power Station in southern India (Rajagopal et al. 1991) and Jamaica Private Power Company (T. Bowes, pers. comm.). Byssal mats that are formed by dense population of P. viridis are already a problem for local power plants. These byssal mats (and all that grows in them) are 18 Ingrao et al. ripped off the hard substratum by the water current flowing to the condensers. The inflowing water pushes the mats against the small pipes in the end of the condensers and the mats act like a clogged air-conditioning filter. This causes pumps to work harder, overheat and potentially cause plant shut down (E. Beitia and D. Bruzek. pers. comm.). In April 2000. P. viridis was found growing inside one of the water intake pumps at Mote Marine Laboratory and caused the pump to overheat requiring the pump to be shut down decreasing the amount of water available for daily operations (C. Tra.xler. pers. comm.). Heavy colonies could also potentially sink channel marker buoys, creating a hazard to navigation as was seen in Texas with P. penia (Hicks pers. comm.). All of the news may not be bad. All three Pevna species are a valuable food source high in protein and are harvested from wild population or aquaculture facilities for human consumption (Va- kily 1989). In 1955 the first commercial mussel farm started in the Philippines (Vakily 1989). Thailand (P. viridis) and New Zealand (P. canaliculus) are the only two other countries where Perna is being cultured on an intensise commercial scale (references in Vakily 1989). India (P. viridis) is the only country in Asia with a substantial mussel industry not based on culture; over 3.000 tonnes are harvested annually (Vakily 1989). Rafts, racks, bamboo poles, date-palm stakes, iron bars, ropes, rope-webs and longlines are used to culture Perna (Vakily 1989). Mussel seed are rarely pro- duced artificially but are collected on spat collectors during peak spawning seasons (Vakily 1989). Aquaculture of P. canaliculus started about 20 years ago and has continued to expand, with production almost doubling in the last 4 years to reach 70.000 tonnes y"' (Jeffs et al. 1999). Perna canaliculus, marketed as Green.shell® mussel, is internationally recognized as a premium- eating mussel and is thought to possess unique remedial properties in rheumatoid and osteoarthritis (references in Jeffs et al. 1999). The success of P. viridis in the southeastern Gulf of Mexico, the demand for these mussels in the food industry and potential bio- medical importance could spawn new aquaculture opportunities. In rapid response to the discovery of Perna viridis in Tampa Bay. staff at the Florida Fish and Wildlife Conser\'ation Commis- sion (at the Florida Marine Research Institute. St. Petersburg) and Mote Marine Laboratory (Sarasota) have issued public announce- ments in the printed and audio media, as well as on the World Wide Web (http://www.mote.org~debi/perna.phtml. http:// www.fmri.usf.edu/invert/pviridis.htm) to infonn the public and so- licit information about the spread of the mussel (Associated Press. 1999; Pittman. 1999; Roberts. 1999; Florida Fish and Wildlife Conservation Commission. 2000). Meeting presentations and post- ers have been given (Ingrao et al. 1999; Benson et al. 1999; Ingrao et al. 2000) and live-animal displays and requests for further in- formation were put in place at the publicly available Mote Marine Aquarium. In spite of the activity surrounding this newly recorded introduction, the mussel is already well established in its new environment and its eradication is not expected. Clearly, priorities must be directed at monitoring its ecological and economic im- pacts, and at developing methods for its control and use. ACKNOWLEDGMENTS Special thanks are due to C. Barringer and E. Beitia of Tampa Electric Company for reporting the initial sighting and sending specimens to Mote Marine Laboratory. We also thank E. Estevez and J. Culter (MML) for reviewing the manuscript. The authors would like to thank the following for sharing Perna viridis sighting and size information; A. Benson (Florida Caribbean Science Cen- ter, U. S. Geological Survey, Gainesville, Florida), M. Blouin (Great Lakes Science Center, U. S. Geological Survey, Ann Arbor, Michigan), T. Bowes (Jamaica Private Power Plant), D. Bruzek (Florida Power Corporation, Paul L. Bartow Power Plant), D. Buddo (University of the West Indies, Jamaica), D, Marelli [Florida Fish and Wildlife Conservation Commission (at the Florida Marine Research Institute, St. Petersburg, Florida)], and C, O'Neil (U. S. Coast Guard Cutter Jusluia Appleby. St. Petersburg, Florida). Special thanks to the following MML staff, volunteers and interns for their assistance with data entry and image preparation, literature searches, laboratory and field work: S. Arbuckle, Andrea Baird. K, Churchill, J. Goldstein, D. Harayda, W. Hoebel, S. How- ard, K. Ingrao, N. Jacobs, N. Levin, R. Marzek, J. Massoud, M. Oliver, J Osterhoudt, T. Pancake, J. Perry, J. Sadler, L. Stevenson, S. Stover. C. Traxler. J. Van Horn, and V. Zoll. Figure 2 was prepared by S. Thurston (AMNH). David Hicks wishes to thank M. Haiduk for use of his laboratory at Lamar University in pre- paring karyotypes. LITERATURE CITED Agard. J.. R. Kishore & B. Bayne. 1992. Perna viridis (Linnaeus, 1758): first record of the Indo-Pacific green mussel ( Mollusca: BivaK ia) in the Caribbean. Cunbb. Mar. Snid. 3: 59-60. Ahmed, M. 1974. Chromosomes of two species of the marine mussel Perna (Mytilidae: Pelecypoda). Bol. Inst. Oceanogr. Venez. Orienle. 13: 17-22. Associated Press (AP). 1999. Exotic mussels invade Tampa Bay. Sarasota Herald-Tribune 29 October 1999 p 7B. Benson, A. J.. D. A. Ingrao & M. A. Blouin. 1999. Range expansion and colonization of the green mussel. Perna viriiii.s. along Florida's west coast [abstract]. Program and abstract from Florida Association of Benthologists 14th Annual Meeting, Mote Marine Laboratory. Sara- sota. Florida. 07-08 December 2000. Berry, P. F. 1978. Reproduction, growth and production in the mussel. Perna perna (Linnaeus), on the east coast of South Africa. Investiga- tional Report no. 48, Oceanographic Research Institute, South African Association for Marine Biological Research, Durban. 28 pp. Carlton, J. T. 1987. Patterns of transoceanic marine biological invasion in the Pacific Ocean. Bull. Mar Sei. 41: 452-465. Carlton, J. T. IW2. Introduced marine and estuarine mollusks of North America: an end-of-the-20th century perspective. / Shellfrsh Res. 11(2): 489-505. Cheung, S. G. 1993. Population dynamics and energy budgets of green- lipped mussel, Perna viridis (Linnaeus), in a polluted harbor. J. Exp. Mar Biol. Ecol. 168: 1-24. Claudi. R. & G. L. Mackie. 1994. Practical manual for Zebra Mussel monitoring and control. Boca Raton. FL: CRC Press. 227 pp. Dance. S. P. (ed). 1974. The Collector's Encyclopedia of Shells. New York: McGraw-Hill Book Company. 288 pp. Diaz Merlano. J. M. & M. P. Hegedus. 1994. Moluscos del Carihe Co- lombiano: un catalogo ilustrado. Colciencias y Fundacion Natura Co- lumbia. Santafe de Bogota. 291 pp. Estevez. E. D. (ed). 1989. Tampa and Sarasota Bays: Issues, Resources, Status, and Management. Proceedings of NCAA Estuary-of-the-Month Seminar Series No. 1 1. NOAA Estuarine Program Office, U. S. De- partment of Commerce, Washington, DC. 215 pp. Florida Fish and Wildlife Conservation Commission [FWC]. 2000. Asian Perna viridis Discovered in Tampa Bay, Florida 19 Green Mussel. World Wide Web page http://www.fmri.usf.edu/ invert.pviridis.htm. Last accessed 8 August 200(1. Hicks, D. W. & J. W. Tunnell, Jr. 1993. Invasion of the south Texas coast by the edible brown mussel Pema perna (Linnaeus. 1758). Veliger 36(1): 92-94. Hicks, D. W. & J. W. Tunnell, Jr. 1995. Ecological notes and patterns of dispersal in the recently introduced mussel, Pema perna (Linnaeus, 1758), in the Gulf of Mexico. Am. Malacol. Bull. 1 1(2): 203-206. Hickman, R. W. 1991. Pema canaliculus (Gmelin) in New Zealand. In: W. Menzel editor. Estuarine and Marine Bivalve Mollusk Culture. Boca Raton, FL: CRC Press, pp. 326-33. Holland, B. S., D. S. Gallagher, D. W. Hicks & S. K. Davis. 1999. Cyto- taxonomic verification of a non-indigenous marine mussel in the Gulf of Mexico. Veliger 42{y): 280-282. Huang, Z. G.. S. Y. Lee & P. M. S. Mak. 1983. The distribution and population stfucture of Pema viridis (Bivalvia: Mytilacea) in Hong Kong waters. In: B. Morton & D. Dudgeon, editors. Proceedings of the Second International Workshop on the Malacofauna of Hong Kong, and Southern China. Hong Kong: Hong Kong University Press, pp. 465-17 1 . Ingrao, D. A., P. M. Mikkelsen & D. W. Hicks. 1999. The green mussel. Pema viridis. invades Tampa Bay. Florida [abstract]. Program and abstract from Florida Association of Benthologists 13th Annual Meet- ing, Harbor Branch Oceanographic Institution, Ft. Pierce. Florida. 09- 10 December 1999, 4 p. Ingrao, D. A., A. J. Benson & M. .A. Bloiun. 2000. Size, color and mor- phological variation of Pema viridis along the west coast of Florida [poster]. Program and abstract from Florida Association of Bentholo- gists 14th Annual Meeting, Mote Marine Laboratory, Sarasota, Florida, 07-08 December 2000. Jacobi, C. M., C. Rosenberg & A. M. Vianna-Morgente. 1990. The karyo- type of the brown mussel. Perna perna (L.) (Bivalvia: Mytilidae). Rev. eras. Ge\)et: 13:699-673. Jeffs. A. G., ft. C. Holland, S. H. Hooker & B. J. Hayden. 1999. Overview and bibliography of research on the greenshell mussel, Perna canali- culus, from New Zealand waters. J. Shellfish Res. 18(2):347-360. Keevin, T. M., R. E., Yarbrough & A. C. Miller. 1992. Long-distance dispersal of zebra mussels [Dreissena polymorpha) attached to hulls of commercial vessels. / Freshwater Ecol. 7(4):437. Lewis, R. R. Ill & E. D. Estevez. 1988. The ecology of Tampa Bay, Florida: an estuarine profile, U.S. Fish and Wildlife Service, Biological Repon no. 85(7.18), 132 pp. Meineke, S. 1999. Controlling invasive species: a billion dollar business. Gulfwalch 10(3):3. Morton, B. 1987. The functional morphology of the organs of the mantle cavity of Pema viridis (Linnaeus, 1758) (Bivalvia: Mytilacea). Am. Malacol. B«// 5(2):I59-164. Morton, B. 1997. The aquatic nuisance species problem: a global perspec- tive and review. In: F. M. D'ltri. editor. Zebra mussels and aquatic nuisance species. Chelsea. MI: Ann Arbor Press, pp. 1-54. Pittman. C. 1999. Asian green mussel grabs hold in bay. St. Petersburg [Florida] Times. 28 October 1999. Powell, A. W. B. 1979. New Zealand mollusca: Marine, land and fresh- water shells. Auckland. New Zealand: William Collins Publishers, Ltd. 500 pp. Pribble, R. 1999. Bay characteristics, p. 3.4. In: J. R. Pribble. A. J. Janicki, & H. Greening, editors. Baywide Environmental Monitoring Report 1993-1998. Tampa Bay Estuary Program, Technical Publication 07- 99. Rajagopal, S., N. Sasikumar, J. Azariah & K. V. K. Nair. 1991. Some observations on biofouling in the cooling water conduits of a coastal power plant. Biofiiuling 3:311-324. Rajagopal, S., V. P. Venugopalan, J. Azariah & K. V. K. Nair. 1995. Response of green mussel Pema viridis (L.) to heat treatment in rela- tion to power plant biofouling control. Biofouling 8:313-330. Ravi. A. J. S. 1999. Bioinvaders proliferate in U.S. waterways. The Sci- einist 13(16):6. Rios, E. C. 1985. Coastal Brazilian seashells. Empresas Ipiranga. Rio Grande, RS, XI, Brazil. 255 pp. Rivonkar, C. U., R. A. Sreepada & A. H. Parulekar. 1993. Growth param- eters in the cultured green mussel, Perna viridis L., from the Zuan estuary, Goa. Indian J. Mar. Sci 22:72-74. Roberts, D., Jr. 1999. New mussel arrives in Florida: state highly suscep- tible to invasions. Gulfwalch 10(3): 15. Sadacharan, D. H. 1982. Country repon: Sri Lanka. In: F. B. Davy & M. Graham, (eds.). Bivalvia culture in Asia and the Pacific: Proceedings of a workshop held in Singapore, 16-19 February. 1982. International Research Development Centre, Ottawa, Canada. Sea Grant News Media Center-Texas. 1999. Mussels: http:// www. seagrantnews.org/news/txmussels. html. Last accessed 10 No- vember 1999. Segnini de Bravo, M. I.. K. S. Chung & J. E. Perez. 1998. Salinity and temperature tolerances of the green and brown mussels, Pema viridis and Perna perna (Bivalvia: Mytilidae). Rev. Biol. Trop. 5:121-125. Siddall, S. E. 198()a. A clarification of the genus Perna (Mytilidae). Bull. Mar. Sci. 30(4):858-870. Siddall, S. E. 1980b. Effects of temperature and salinity on growth, sur- vival, metamorphosis and shell deposition of tropical mussel larvae (Perna perna and P. viridis]. PhD Thesis, University of Miami, 206 pp. Stace, G. 1997. What's on the beach? A guide to local marine life. Auck- land, New Zealand: Viking. Penguin Books (NZ) Ltd. 63 pp. Tampa Bay Estuary Program. 2000. Fast facts about Tampa Bay: http:// www.tbep.org/portrail/fastfacts.html. Last accessed 9 May 2000. Tampa Bay Port Authority. 2000. Bulk Cargo: http://tampaport. com\bulkcargo.htm. Last accessed 13 January 2000. Turgeon, D. D., J. F. Quinn, Jr., A. E. Bogan, E. V. Coan, F. G. Hochberg, W. G. Lyons, P. M. Mikkelsen, R. J. Neves, C. F. E. Roper, G. Rosen- berg. B. Roth, A. Scheltema. F. G. Thompson. M. Vecchione & J. D. Williams. 1998. Common and scientific names of aquatic invertebrates from the United States and Canada: Mollusks, 2nd ed. American Fish- eries Society, Special Publication 26, Bethesda, Maryland. 526 pp. + CD-ROM. Vakily, J. M. 1989. The biology and culture of mussels of the genus Pema. ICLARM Studies and Reviews 17. International Center for Living Aquatic Resources Management. Manila. Philippines, and Deutsche Gesellschaft fiir Technische Zusammenarbeit (GTZ) GmbH. Eschborn, Germany. 63 pp. Yaseen, A. E. 1995. The chromosomes of the Egyptian freshwater snail Melanoides luherculala (Gastropoda: Prosobranchia). J. Molluscan Stud 61:499-500. Journal of Shellfish Research. Vol. 20, No. 1, 21-29, 2001. ESTABLISHMENT OF THE GREEN MUSSEL, PERNA VIRIDIS (LINNAEUS 1758) (MOLLUSCA: MYTILIDAE) ON THE WEST COAST OF FLORIDA^ AMY J. BENSON,' * DAN C. MARELLI," MARC E. FRISCHER,' JEAN M. DANFORTH,' AND JAMES D. WILLIAMS' 't/.5. Geological Survey. Florida Caribbean Science Center, 7920 NW 71st Street. Gainesville. Florida 32653: 'Academic Diving Program. Florida State Universit}', 036 Montgomery: Tallahassee, Florida 32306-2310: ^Skidway Institute of Oceanography, 10 Ocean Science Circle. Savannah. Georgia 31411 ABSTRACT In 1999, the green mus.sel, Perna virUlis, was first observed in Tampa Bay, Florida. This was the first reported occurrence of this Indo-Pacific marine bivalve in North America. The mussels found in Tampa Bay were confirmed to be P. viridis based on both morphological and genetic characteristics. Since the initial discovery, surveys in Tampa Bay and on the west coast of Florida have documented the growth, recruitment, and range expansion of P. viridis. From November 1999 to July 2000, the mean shell length of a Tampa Bay population mcreased from 49.0 mm to 94. 1 mm. an increase of 9T7c . Populations of P. viridis are successfully reproducing in Tampa Bay. Recruitment was observed on sampling plates in May and continued through July 2000. The full extent of mussel colonization is not clear, but mussels were found outside Tampa Bay in St. Petersburg, Florida, south to Venice. Based on these studies it is evident that P. viridis has successfully invaded Tampa Bay and the west coast of Florida. The long-term impact of P. viridis on native communities off the west coast of Flonda cannot be predicted at this time. KEY WORDS: Perna viridis. introductions, green mussel. Florida, Gulf of Mexico INTRODUCTION Florida's ecosystems have been and continue to be the recipi- ents of many new nonindigenous species introductions and inva- sions (U,S. Congress 1993. Carlton & Ruckelshaus 1997). Tainpa Bay, located on the west coast of central Florida, has been par- ticularly open to exotic species introductions because of its tem- perate to subtropical climate and active industrial, commercial, and recreational use. In August of 1999, the Tampa Electric Coinpany reported finding an unknown bivalve fouling the surfaces in their cooling water intake tunnels at both their Gannon and Big Bend power stations in and near the city of Tampa, Florida (G. Shofner, personal communication). They were identified as Permi viridis (Linnaeus 1 758 ) based on their karyotype ( Ingrao et al. 200 1 ). This was the first report of this species in North America. Nonindig- enous species are potentially invasive when they display some of the following characteristics: high fecundity, short generation time, long-lived, high dispersal rate, phenotypic plasticity, broad native range, abundant in native range, and tolerant of a wide range of conditions (Williams & Meffe 1998), Penui viridis is native to the coastal marine waters of the Indo- Pacific region including the Persian Gulf to the South China Sea (Siddall 1980, Vakily 1989). It is a byssally attached, sessile bi- valve naturally inhabiting estuarine waters where the salinity ranges from 27-33 PSU, although it tolerates salinities as low as 16 PSU (Sundaram & Shafee 1989). The optimal annual temperature range for P. viridis is 26-32°C: survival rate was 50% at 10° and 35°C after two weeks of exposure (Sivalingam 1977). Perna viri- dis exhibits a wide tolerance for sub-optimal environmental con- ditions, tolerating large fluctuations in sediment flux and organic enrichment (Morton 1985. Lee 1986). Like other marine mussels. P. viridis attaches to hard substrates, but is also capable of colo- nizing soft substrates (Agard et al. 1992). Perna viridis generally occur at depths from the surface to 10 m. although settlement of juvenile mussels has been reported greatest at depths of between 2 *Corresponding author. E-mail: amy_benson@usgs.gov 'Editor's note: manuscript received November 21, 2000: accepted for pub- lication December 16, 2000. and 3 meters (Tan 1975, Cheong and Chen 1980). This species is a suspension feeder selectively feeding primarily on algae (Sival- ingam 1977, Yap et al. 1979). It has been intensively cultured within its native range as a protein source for human consumption and been shipped to many islands in the South Pacific over the last several decades for this purpose (Chalermwat et al. 1988, Eldredge 1994). Perna viridis is a dioecious broadcast spawner (Sivalingam 1977) displaying a wide variation in spawning pattern within its native range, encompassing unimodal, bimodal. and asynchronous modes (Walter 1982, Lee 1986, Tuaychareon 1991, Rajagopal et al. 1998b). Following fertilization, zygotes develop into free- swimming veligers within 16 hours, and umbonate larvae develop at about 20 hours post fertilization (Sivalingam 1977). Larvae remain free-swimming for 2-3 weeks, then settle onto hard sub- strata and byssally attach (Vakily 1989, Tuaychareon 1991). There are currently three species of Perna {P. perna (Linnaeus 1758), P. canaliculus (Gmelin 1791), and P. viridis). They inhabit tropical and subtropical coastal marine waters worldwide; how- ever, none occur naturally in North America (Siddall 1980). Within the past decade, both Penw perna (Hicks & Tunnel 1993) and Perna viridis (Ingrao et al. 2001 ) have been introduced into the Gulf of Mexico and southern Caribbean (Agard et al. 1992, Ry- lander et al. 1996, D. Buddo, personal communication). The known distribution of Perna viridis in the western hemisphere includes the following countries: Venezuela, Trinidad, Jamaica, and peninsular Florida in the United States (Fig. I). While P. perna. the edible brown mussel, is native to the southern Carib- bean. P. viridis is not, and until the recent introductions of P. perna and P. viridis into the Gulf of Mexico and southern Caribbean the species did not occur sympatrically (Siddall 1980). Perna canali- culus, known as the greenshell mussel, is native to waters sur- rounding New Zealand. Shell color, morphology, and especially geographic location have been used in the past to identify and distinguish among the species of Perna (Siddall 1980). Although color has been used as a distinguishing characteristic, it is not a definitive feature. Shells of young P. viridis mussels are a brilliant green while the adults are dark green to brown, and young P. perna 21 22 Benson et al. 80° United e States / Atlantic O c c a n N s ■y-^ Jamaica Caribbean 200 0 200 400 Kilometers 25° 15° Hgure 1. Shaded circles represent /'. viridis populations along coastlines. observed in Texas can be green but are brown as adults (D. Hicks, personal communication I. Pema viridis can be distinguished from P. pema by examining the karyotype of somatic cells: P. penui has 28 chromosomes while P. canaliculus and P. viridis have 30 (Ahmed 1974. Holland et al. 1999, Ingrao et al. 2001 ). Therefore, because color and morphological characters are unreliable for dis- tinguishing between the species in this genus (Siddall 1980). more defining characteristics are needed to determine the identity among species of the genus Pema. METHODS Study Sites From November 17-19, 1999, biologists from the U.S. Geo- logical Survey (USGS) and the Florida Marine Research Institute (FMRl) sampled bridge pilings, buoys, channel markers and shore- lines of Tampa Bay. The northern section of the Bay consists of two smaller bays. Hillsborough and Old Tampa bays. Because the two power plants that had reported finding Pema viridis were located in Hillsborough Bay. we decided to sample Old Tampa Bay to determine the species" range within the entire Tampa Bay system. Green mussels were collected by hand off bridge pilings from just above the water surface to approximately one-half meter below the surface. In addition, we qualitatively surveyed two bridges in Old Tampa Bay using scuba divers in late February 2000. Divers also attempted to locate additional populations of mussels near the mouth of Tampa Bay from March through July 2000. In the latter part of July, we expanded our survey along the west coast of Florida to further document the extent of their range. We began surveying coastal waters south of Tampa Bay at Venice and proceeded in a northerly direction to Johns Pass at the north end of Treasure Island. In addition to direct observation, we re- ceived reports of P. viridis observations from private citizens, go\emmental agencies and other organizations, which were help- ful in locating other populations. Approximately thirty adult P. viridis were collected from the wooden fenders beneath the Howard Frankland Bridge (Interstate 275) over Old Tampa Bay during late February 2000 and monthly thereafter. Mussels were fixed in 5% seawater formalin and pre- served in 70% ethanol for future analysis of reproductive condi- tion. These mussels were examined morphologically and used to determine the mean length of adults in the Old Tampa Bay popu- lation. In addition, we established recruitment collectors in the same vicinity. Recruitment collectors consisted of three 15 cm x 1 5 cm unglazed quarry tiles attached to a stainless steel frame which were protected from predation by a section of I cm x 1 cm polypropylene mesh. Tiles were replaced monthly and exposed tiles were taken to the laboratory for examination using a magni- fier lamp as well as a dissecting microscope. All P. viridis recruits settling on the tile surface were counted in the laboratory. Scrap- ings from pier and bridge pilings were also examined in the labo- ratory to determine the identity of any commensal organisms. Non-Native Green Mussel in Florida 23 Identification hy Genetic Characterization Isolation of Total DNA from Bivalves DNA was purified from specimens of P. penui (Port Mansfield, Texas), P. viridis (Pelican Island. Trinidad). P. canaliculus (Marl- borough Sound and Castlepoint. Nev\ Zealand), and Pcnui species collected from Tampa Bay, Florida. Whole animals were collected, preserved in 95% ethanol or isopropyl alcohol, and shipped by express mail to the laboratory. Total DNA was purified from ad- ductor muscle tissue {ca. 0.5-0.14 g) as previously described (Frischer et al. 2000b). This protocol routinely yielded from .^0 to 800 p,g DNA/a tissue of high molecular weight genomic DNA suitable for PCR amplification depending on the source of tissue. PCR Amplification of the Mitochondrial Cytochrome C Oxidase Subunit I Gene A 713-bp fragment of the mitochondrial cytochrome c oxidase subunit I gene (COl) was amplified using primers LCOI490 (5'- ggt caa caa ate ata aag ata ttg t) and HC02198 (5'- taa act tea ggg tga cca aaa aat ca) from each of the mussel species (Baldwin et al. 1996). Primers were synthesized using an ABI DNA/RNA syn- thesizer (model 394) by the Molecular Genetics Facility at the University of Georgia. Amplification was accomplished using the Qiagen Tai/ PCR Master Mix System following the standard pro- tocol recommended by the vendor (Qiagen, Valencia, Calif) ex- cept that to each 25 (jil reaction, 0.05 (xl of T4 Gene Protein (Roche Molecular Biochemicals, Indianapolis, IN) was added to enhance amplification. Amplification conditions were as described by Baldwin et al. ( 1996); 35 amplification cycles (95°C, I min; 54°C. I min; 72°C, 1 min) initiated by a 3-min denaturation step at 95°C and followed by a 10-min extension step at 72°C. Sequencing To facilitate sequencing of the PCR amplified COI gene frag- ment from the Tampa Bay Pcnia specimen, P. pcnui. and P. viriclis. and P. canaliciili(s specimens, the COI PCR product was cloned into the bacterial plasmid sequencing vector PCR 2.1 TOPO cloning vector using the TOPO® TA cloning system fol- lowing the instructions provided by the manufacturer (Invitrogen, Carlsbad. Calif). Sequences were determined by automated se- quencing at the Molecular Genetics Facility (University of Geor- gia) using an ABI automated sequencer (models 373 and 377) as previously described (Frischer et al. 2000b). The complete COI amplicon sequence was obtained using the plasmid vector targeted (Ml 3) sequencing primers MI3-20-F (5'- tgt aaa acg acg gcc agt) and M13-48R (5'- age gca taa caa ttt cac aca gga). Complete sequences were assembled using the assembly and editing features of the DNAsis software package version 7.00 (Hitachi Software Engineering Co.). Phylogenetic Analysis Sequences were aligned relative to each other and several other COI fragments from other molliisks available in GenBank using the CLUSTAL W version 1.7 multiple sequence alignment algo- rithm (Thompson et al. 1994). Alignments were viewed and edited using the Genetic Database Editor (GDE) (Smith et al. 1992). Percent sequence similarity between organisms was determined using the sequence alignment procedure available in the DNAsis software package version 7.00 (Hitachi Software Engineer- ing Co.). Phylogenetic trees were inferred and drawn using the TREECON for Windows software package version 1.3b (Van de Peer & De Wachter 1994, 1997) using the Kimura two-parameter model for inferring evolutionary distance. Bootstrap estimates ( 100 replicates) of confidence intervals were also made using the algorithms in TREECON. RESULTS The identity of the Perna species collected in Tampa Bay was confirmed to be the green mussel. Perna viriclis. based on both morphological and genetic characteristics. Morphological Characteristics The presence and location of muscles and muscle scars have previously been reported to be useful identifying characteristics for the genus Perna (Siddall 1980). The lack of an anterior adductor muscle confirms that the mussel in Tampa Bay is a species of Perna. Furthermore, the position of the anterior component of the posterior retractor muscles appears most similar to that of P. viri- clis. Externally, the most obvious characteristic of the mussels collected in Tampa Bay is that both juveniles and adults are a bright green to dark green color (Fig. 2). Although shell color is a poor identifying characteristic for these bivalves, to our knowledge there have been no reports of Perna perna that were green as adults. Due to the ambiguities associated with the identification of Penia species based solely on morphological and coloration char- acteristics, specimens of three Perna species were genetically compared with the mussels found in Tampa Bay. Genetic characteristics An identically sized 713 bp fragment of the mitochondrial en- coded cytochrome c oxidase subunit gene was amplified, cloned, and sequenced from each Perna species studied. These sequences have been deposited in GenBank. Accession numbers are AF298850, AF298851, AF30873I and AF298852 for P. perna (Texas), P. viriclis (Trinidad), P. canaliculus (New Zealand) and P. viridis (Tampa Bay), respectively. Phylogenetic comparison be- tween Perna species and Perna species collected from Tampa Bay indicated that the Tampa Bay specimens were most closely related to P. viridis (Fig. 3). The sequence similarity between P. perna. P. viridis. P. canaliculus, and the Tampa Bay species are shown in Table 1 . Sequence similarity between Perna species ranged from 78.2% to 99.5% with the highest similarity between the Tampa 1 cm Figure 2. The green mussel. Perna viridis. collected from Tampa Bay, Florida, November 1999. 24 Benson et al. ... r Pernaviriiiis i^TimpiBay) 100 1 Perna vindis Clhradati) 100 Per/a ctmalkiiltts (New Zealand) Perna perna (Texas) Figure 3. Inferred taxonomic relationship between P. viridis in Tampa Ba) and other Perna species. The phylogenetic tree was derived from nucleotide sequence comparison of a 713 bp fragment of the mitochon- drial cytochrome c oxidase subunit I gene (mtCOI). The tree was artificially rooted with the mtCOI sequence from the ribbed mussel {Geukensia demissa) retrieved from GenBank (LI56844). The scale bar indicates 0.1 Tixed nucleotide substitutions per site. Numbers refer to bootstrap values (from 100) for each node. Bay specimens and the P. viridis specimens collected from Trin- idad. These results confirm that the Perna species currently in Tampa Bay. Florida, is Perna viridis. Distribution, Growth, and Reproduction in Tampa Bay Region Since the initial discovery of P. viridis at two electric power plants in Tampa Bay. Florida, populations have been documented throughout the bay and extending into the Gulf of Mexico. Through July 2000, green mussels have been documented from numerou,s locations in Tampa Bay (including Hillsborough and Old Tampa bays), Sarasota Bay, and in the Gulf of Mexico from Venice, north to Treasure Island at Johns Pass (Fig. 4). One week after our final survey, mussels were discovered south of Venice at Boca Grande, a barrier island in the Gulf of Mexico just outside of Charlotte Harbor (M. Blouin, personal communication). Mussels were collected at the northernmost and southernmost sites in the final survey; therefore, the full extent of their geographic range along the west coast of Florida is still to be determined. Mussels collected from the Gannon Power Station (Hillsbor- ough Bay) in October 1999 ranged in shell length from 17.9 to 30.1 mm (mean = 23.6 mm). These animals were reproductively mature and even the smallest individuals began to spawn when they were taken into the laboratory. Perua viridis has been re- ported to achieve sexual maturity in two to three months at a shell length of approximately 20 mm (Vakily 1989, Tuaychareon 1991). The mussels collected in November 1999 from the population of P. TABLE 1. Percent sequence similarity (above diagonal) and the number of nucleotide differences between the Tampa Bay Perna species, other Perna type species, and the ribbed mussel Geukensia demissa mitochondrial cytochrome c oxidase subunit 1 gene. G. demi P. perna P. viridis P. can TBsp. G. demi 100 60.5 60.9 65.2 60.7 P. perna 243/614 100 79.1 99.5 78.2 P. viridis 240/614 129/615 100 79.3 99.2 P. can 402/617 615/618 489/617 100 78.4 TB sp. 241/614 134/615 5/617 484/617 100 Genbank accession numbers and abbreviations: G. demi (U56844. Geu- kensia demissa): P. perna (AF298850. Perna perna): P. viridis (AF298851, Perna viridis): P. can (AF308731. Pcnra canaliciili(s): TB sp. (AF298852, Tampa Bay Perna species). viridis on the Howard Frankland Bridge (Old Tampa Bay) had a mean shell length of 49.0 mm. Subsequent collections from this population demonstrate rapid growth as the mean length of mus- sels increased nearly 92% to 94.1 mm in July 2000 (Fig. 5). Growth rates have been reported from a number of areas within the native range. Perna viridis in the nutrient rich Hong Kong Harbour grew to a length of 50 mm in the first year, to 75 mm in the second year, and to 85 mm in the third year (Cheung 1993). In a power plant intake structure in India, mussels grew to a length of 27 mm in 49 days and 1 19 mm in 375 days (Rajagopal et al. 1998a). The highest growth rate recorded is from Singapore at 10.6 mm per month during the first year of culture (Cheong and Chen 1980). According to Vakily (1989), it is difficult to define maximum size and longevity of members of this genus. Perna viridis recruits were first detected in May 2000 on re- cruitment collectors placed near the Howard Frankland Bridge in Old Tampa Bay. Small mussels (<10 mm shell length) were promi- nent on both the collectors and bridge structures during June and July (Fig. 6). Cheung ( 1993) documented two recruitment periods per year in a Hong Kong harbor, the first from July to September, and the second from November to March. Tampa Bay community Tampa Bay is a large estuary approximately 57 km in length and 20 km in width. The entire Tampa Bay watershed encom- passes 6,739 km~ with 967 km" of open water. Carlton (1996) views much of the cosmopolitan species that make up the Florida marine biota as cryptogenic. Because it is probable that introduc- tions to Florida by European settlers have occurred for five cen- turies, it is difficult to differentiate native from exotic species in Florida (Carlton and Ruckelshaus 1997). Commensal organisms, endemic or cryptogenic, found during our surveys to occur with P. viridis include other species of bivalves, crustaceans, barnacles, bryozoans, cnidarians, and annelids (Table 2). Though our samples do not represent the entire biota in the bay, we consider them representative of the community in which Perna viridis is now established in Tampa Bay. TABLE 2. .\ list of commensal invertebrates present from samples collected off bridge pilings of the Courtney Campbell Parkway over Old Tampa Bay. Identincation confirmed by Jon Fajans, University of Florida. Phvlum Species Mollusca Crassoslrea virginica (Linneaus, 1758) Iscliadiiim reciiiyum (Ratlnesque, 1820) Brachiodonles e.xiisnis (Linneaus. 1758) Crepidula foniicala (Linneaus, 1758) Crepidula plana Say, 1822 Arthropoda Tanystytum orbiculare Wilson. 1878 Caprella penanlis heach, 1814 Jassa falcata (Montagu. 1808) Pelrolislhes galulliiniis (Bosc. 1802) Balanus ampliitrite Darwin, 1854 Balanus improvisiis Darwin. 1854 Balanus ehumeus Gould. 1841 Bryozoa Bowerbankia gracilis (O'Donoghue. 1926) Membranipora arborescens (Canu & Bassler. 1928) CniiJaria Bouganvillia sp. Campan\ilaria sp. Annelida NereidiJae Non-Native Green Mussel in Florida 25 83° 82° 28' 27 ^ Area Enlarged f | Old Tampa Bay j TAMPA Hillsborough Bay | Johns Pass QliLJ Of MTXICO • USGS / FMRI collection sites ▲ non-USGS / non-FMRI collection sites ■ power plant collection sites 10 0 A 10 20 Kilometers 28° 27° 83° 82° Figure 4. Locations of Penia viridh collected or observed, August 1999 to August 2000. DISCUSSION The results of this study indicate that the mitochondrial COI gene can be used to genetically identify Perna individuals to the species level. Sequence comparison of amplified mitochondrial COI gene products indicated that the Perna species discovered in Tampa Bay, Florida is Perna viridis. This confirms the reported presence of Perna viridis in the Gulf of Mexico that was based on cytogenetic karotyping (Ingrao et al. 2001). Furthermore, these studies suggest that molecular comparison of the mitochondrial COI gene can provide a diagnostic molecular marker for identify- ing Perna at the species level. However, despite the high similarity between the Tampa Bay and Caribbean Perna viridis specimens, five nucleotide differences between the two mussels were ob- served. The presence of these sequence differences suggests that COI sequences may contain sufficient variation for detecting intra- population variation within Perna viridis populations. Additional sequencing studies will be needed to test this hypothesis. Further studies will also be required to evaluate whether the COI locus will be useful for larval identification techniques (Frischer et al. 2000b. Bell & Grassle 1998, Baldwin et al. 1996). Interestingly, the simi- larity between P. perna and P. canaliculus was extremely high (95.5%) suggesting that these two species may be ecomorphs rather than true species. However, since these results were derived from a single specimen of P. perna and seven P. canaliculus. additional systematic analysis will be required to confirm this con- clusion. Green mussels were initially identified from Hillsborough Bay in August 1999 at electric power generating stations. However, green mussels were not limited to the vicinity of the power stations and furthermore, larger mussels were found in Old Tampa Bay suggesting that it was the initial colonization site for green mussels in Tampa Bay. Thousands of barnacle-encrusted green mus.sels were discovered on the bridge pilings there indicating that the mussels had been present for an extended period of time. Although we did not quantify the populations on each bridge piling, it was visually apparent that the number of green mussels on each piling decreased as we sampled from the center of the bridge to the shore. We observed no mussels along the shoreline, which consisted of suitable substrate such as rip-rap. concrete retaining walls, and mangrove prop roots. Most of the mussels were observed subtid- ally during our surveys of the bridge pilings. However, a smaller 26 Benson et al. 11/17/99 2/29/00 4/7/00 5/2/00 6/8/00 7/12/00 Date Figure 5. Mean shell length of Penia viridis individuals collected from Old Tampa Bay, 1999-2000. Data include means plus and minus one standard error. portion were observed intertidally where they were exposed to atmospheric conditions for short periods. At the sample sites along the Gulf coast from Venice to Treas- ure Island (Johns Pass), mussels were smaller and fewer when compared to the sites sampled in 1999 from Old Tampa Bay (A. Benson, personal observation). However, mussels were common enough to be detected on bridge and dock pilings and channel markers after just several minutes of diving. Having obser\ed a visible decrease in the number and size of mussels in our surveys outside of Tampa Bay reaffirmed that the Port of Tampa was the initial site of this introduction from which they subsequently were transported by currents out of the Bay along the coast. We did not observe green mussels in slow-moving backwater areas in the Tampa Bay estuary or along the Gulf coast and Intracoastal Wa- terway. Pema viridis is the most recent nonindigenous mollusk to enter Florida, probably introduced into Tampa Bay as larvae entrained in ship ballast water. Ballast water transport has become a major source for the introduction of marine organisms (Carlton and Geller 1993, Carlton and Ruckelshaus 1997). Other modes of in- troduction such as accidental releases by the seafood or aquacul- ture industries, or purposeful release of mussels seem far less likely. Although originally from the Indo-Pacific region, we sus- pect that the source of introduction of green mussels into Tampa Bay was the southern Caribbean region. Just over4(K?' of the ships arriving in the port of Tampa "in ballast" (carrying significant quantities of water ballast) originate in the Gulf of Mexico or Caribbean, and less than \% originate in areas even remotely close to the native range of P. viridis (Carlton et al. 1995). The spread of P. viridis to other areas in Florida and the southeast United States can be expected and will be limited only by the availability of transport vectors and the physiological tolerance of the mussels. Users of industrial water such as power plants may use mechanical or chemical control methods to reduce or eliminate mussels. Ef- forts to limit future introductions of nonindigenous species must include assessments of risk to native habitats, risk from species or communities in ports of origin, and the risks of potential vectors of introduction (Cariton et al. 1995, Carlton and Ruckelshaus 1997. Simberloff et al. 1997). Concurrently, measures must be developed to prevent such introductions. o o o o 0) Q. .Q E 2 1.5 1 T ..,..£. 0.5 0 4/7/00 5/2/00 6/8/00 Date Figure 6. Mean recruitment of Periia viridis to collector plates in Old Tampa Bay, Florida, 2000. Data include number of recruits per col- lector per day plus and minus one standard error. Our data and observations demonstrate that Pema viridis is able to grow, over-winter, and successfully reproduce in Tampa Bay. In the period since its introduction, estimated to be about two years, the green mussel has colonized mostly man-made habitats throughout the estuary and has entered the Gulf of Mexico where currents have begun transporting larvae south along the Gulf Coast. The rapid growth and maturity, high fecundity, high dis- persal rate, and obvious ability for P. viridis to thrive in the Tampa Bay estuary indicate that this species is already becoming invasive as defined by Williams and Meffe ( 1998). As mentioned earlier. P. viridis is a tropical species where normal water temperatures range from 26°C to 32°C. Based on water temperatures in Tampa Bay since 1998. green mussels have not yet experienced water tem- peratures below the LC50 of 10°C. but water temperatures have been as low as 12.3°C. much colder than waters in their native range. Historic data as far back as 1974 revealed winter water temperatures were below 10°C only in 1977. If green mussels persist in the Bay. cool temperatures may be selecting for a more cold-tolerant population, increasing the potential for northward expansion. Certainly other water quality parameters may be factors in the establishment of the green mussel elsewhere and deserve further investigation. We expect P. viridis to continue to spread south along the Gulf Coast of Florida, possibly to include the Florida Keys and other Caribbean islands (Fig. 7). Based on surface water temperature data, it is likely that green mussels could survive almost anywhere along the Gulf of Mexico coastline, from Florida to Texas and Mexico. Pema viridis may survive in less than optimal water temperatures as far north in the Atlantic Ocean as Charleston, South Carolina. If green mussels were able to survive in sub- optimal water temperatures in the United States, occasional below- normal winter air temperatures may lower water temperatures enough to kill or limit populations. We might also expect them to survive along the Hawaiian Islands coasts in the Pacific Ocean and along the coast of Mexico but probably not as far north as Cali- fornia in the United States. We must also be aware of the fact that thermal pollution can provide overwintering refugia in temperate climates such as in Japan, resulting in an even greater distribution in North America. The biological and economic impacts of Pema viridis on the Tampa Bay region are difficult to predict. Areas currently colo- nized by green mussels are human created substrates and not of Non-Native Green Mussel in Florida 27 SC LA TX MS -~;yi. AL jfflliilkiiiif.,,,, Q u f f of Mexico MEXICO 100 0 100 200 Kilometers Figure 7. Predicted range of Perna viridis in the Gulf of Mexico and western Atlantic waters based on near-surface water temperatures. The lower limit of the sub-optimal range based on LC,^ of 10°C with two weeks exposure. particular concern as far as issues of biodiversity are concerned. We are concerned that P. viriclis will eventually colonize prop roots of the red mangrove, Rhizophoni mangle, and affect the diversity of the native communities associated with this habitat. Green mussels have already become established on red mangrove prop roots in Trinidad within the past decade (Agard et al. 1992) and more recently in Jamaica (D. Buddo. personal communica- tion). They may also affect the dynamics of the phytoplankton assemblage and increase water clarity in Tampa Bay through ac- tive suspension feeding, much as Potamocorbula amurensis has in San Francisco Bay and Dreissena polymorpha has in numerous freshwater bodies (Alpine and Cloem 1992; Stoermer et al. 1996). However, it is expected that many of the nutrients will be cycled into the benthic community as feces and pseudofeces, possibly causing local shifts in benthic community dynamics (Tsuchiya 1980; Dobson and Mackie 1998. Frischer et al. 2000a). Impacts are expected to include fouling of unprotected boats and in-water structures such as docks, seawalls, and aids to navi- gation and interference with cooling water intakes in electric power generating stations. Oddly, green mussels are biofoulers of power plants in their native India (Nair & Murugan 1991; Raja- gopal et al. 1995). They are also a major part of the biofouling community in China where they are notorious for fouling naviga- tion buoys with biomasses of as much as 72 kg/m" (Yan et al. 1994). Perna viridis is also a new member of the biofouling com- munity in Japan. Winter water temperatures normally decimate populations there; however, overwintering of mussels does occur in Tokyo near warmwater discharges (Umemori and Horikoshi 1991 ). Tampa Bay already has an active fouling community con- sisting mainly of barnacles, tunicates, and native bivalve mollusks. but none of these animals approaches either the size or the poten- tial accumulated mass of Perna viridis. Local electric utilities are already experiencing problems associated with green mussel settlement in intake structures (D. Marelli. personal observation) and it is anticipated that the cost of maintenance in these structures, as well as on navigational buoys and vessels will increase sub- stantially as the P. viridis population continues to increase. ACKNOWLEDGMENTS We thank Daniel O'Connell of Newcastle University, New- castle, England and Gary Hill, Robert Lewis, Marc Blouin. and Don Hickey of USGS for collecting samples of green mussels in Florida; Suzanne Dilworth of the Center for Coastal Studies. Cor- pus Christi. Texas for supplying Perna perna samples; Dr. Peter Smith of the National Institute of Water & Atmospheric Research Ltd (NIWA) Wellington. New Zealand for Perna canaliculus samples; and Paul Gabbadon and Addison Titus of the Institute of Marine Affairs. Trinidad for Perna viridis samples. We also thank Jon Fajans of the University of Florida for identifying the com- mensal invertebrates. LITERATURE CITED Agard. J.. R. Kishore c& B. Bayne. 1992. Perna viridis (Linnaeus 1758): tlrst record of the Indo-Pacific green mussel (Mollusca: Bivalvia) in the Caribbean. Caribbean Marine Studies 3:59-60. Ahmed. M. 1974. Chromosomes of two species of the marine mussel Perna (Mytilidae: Pelecypoda). Bio. Inst. Oceanogr. Univ. Oriente 13:17-22. Alpine. A. & J. E. Cloem. 1992. Trophic interactions and direct physical 28 Benson et al. effects control phytoplankton biomass and production in an estuary. Limnol. Oceanogr. 37:946-955. Baldwin. B. S.. M. Black. O. Sanjur. R. Gustafson. R. A. Lutz & R. C. Vrijenhoek. 1996. A diagnostic molecular marker for zebra mussels {Dreissena polymorpha) and potentially co-occurring bivalves: mito- chondrial COI. Moi Mar. Biol. Biotecbnol. 5:9-14. Bell. J. L. & J. P. Grassle. 1998. A DNA probe for identification of larvae of the commercial surfclam iSpisula solidissima). Mol. Mar. Bio. Bio- tech. 7:127-137. Carlton. J. T. 1996. Biological invasions and cryptogenic species. Ecology 77:1653-1655. Carlton, J. T. & J. Geller. 1993. Ecological roulette: The global transport and invasion of nonindigenous marine organisms. Science 261:78-82. Carlton, J. T.. D. Reid & H. van Leeuwen. 1995. Shipping study: The role of shipping in the introduction of nonindigenous aquatic organisms to the coastal waters of the United States (other than the Great Lakes) and analysis of control options. Connecticut Sea Grant Project R/ES-6. Prepared for the U. S. Coast Guard. U. S. Department of Transporta- tion. Report no. CD-D-1 1-95. Carlton, J. T. & M. H. Ruckelshaus. 1997. Nonindigenous inanne inver- tebrates and algae. In: D. Simberloff, D. C. Schmitz & T. C. Brown, editors. Strangers in Paradise: Impact and Management of Non- indigenous Species in Florida. Washington, DC: Island Press, pp. 187- 201. Chalermwat. K.. R. A. Lutz & P. Menasveta. 1988. Mussel culture in Thailand. Journal o) Shellfish Researcli l\i(ft (Summary only). Cheong. L. & F. Y. Chen. 1980. Preliminary studies on raft method of culturing green mussels. Pcnui \iriili.s (L.) in Singapore. Singapore J. Prim. bid. 8:119-133. Cheung. S. G. 1993. Population dynamics and energy budgets of green- lipped mussel Pema viridis (Linnaeus) in a polluted harbor. / E.\p. Mar. Biol. Ecol. 168:1-24. Dobson, E. P. & G. L. Mackie. 1998. Increased deposition of organic matter, polychlorinated biphenyls, and cadmium by zebra mussels (Dreissena polymorpha) in western Lake Erie. Can. J. Fish. Aqual. Sci. 55:1131-1139. Eldredge, L. G. 1994. Perspectives in aquatic e.xotic species management in the Pacific Islands. Volume I: Introductions of commercially sig- nificant aquatic organisms to the Pacific Islands. Noumea, New Cale- donia: South Pacific Commission. 127 pp. Frischer, M. E., S. A. Nierzwicki-Bauer, R. H. Parsons. K. Vathanodom & K. R. Waitkus. 2000a. Interactions between zebra mussels {Dreissena polymorpha) and microbial communities. Can. J. Fish. Aqual. Sci. 57:591-599. Frischer. M. E., J. M. Danforth, L. C. Tyner. J. R. Leverone, D. C. Marelli, W. S. Arnold & N. J. Blake. 2000b. Development of an .\rgopec- ten-specific 18S rRNA targeted genetic probe. Mar. Biolechnol. 2:11- 20. Hicks, D. W. & J. W. Tunnel, Jr. 1993. Invasion of the south Texas coast by the edible brown mussel Pema pema (Linnaeus 1758). The Veliger 36:92-94. Holland. B. S., D. S. Gallagher. D. W. Hicks & S. K. Davis. 1999. Cyto- taxonomic verification of a non-indigenous marine mussel in the Gulf of Mexico. The Veliger 42:280-282. Ingrao, D. A.. P. M. Mikkelsen & D. W. Hicks. 2001. Another introduced marine mussel in the Gulf of Mexico: the Indo-Pacific green mussel, Pema viridis (Linnaeus 1758), in Tampa Bay, Florida. J. Shellfish Res. 20:13-19. Lee. S. Y. 1986. Growth and reproduction of the green mussel Perna viridis (L.) (Bivalvia: Mytilacea) in contrasting environments In Hong Kong. Asian Marine Biology 3: 1 1 1-1 27. Morton. B. 1985. Marine pollution induced environmental changes in Hong Kong— the Tolo Harbour case study. In: M. H. W. Chan. R. W. M. Hoare, P. R. Holmes. R. J. S. Law and S. B. Reed, editors. Pollution in the Urban Environment. Plomel '85. Hong Kong. London: Elsevier Applied Science Publishers, pp. 548-558. Nair. K. V. K. & P. Murugan 1991. Biofouling by Perna viridis in a deep submarine tunnel system at Kalpakkam. Journal of the Marine Bio- logical Association of India 33:336-373. Rajagopal. S., V. P. Venugopalan, J. Azariah & K. V. K. Nair. 1995. Response of the green mussel Pema viridis (L.) to heat treatment in relation to power plant biofouling control. Biofouling 8:313-330. Rajagopal. S., V. P. Venugopalan. K. V. K. Nair, G. Van der Velde & H. A. Jenner. 1998a. Settlement and growth of the green mussel Pema viridis {L. I in coastal waters: mfiuence of water velocity. Aciiiatic Ecol- ogy 32:313-322. Rajagopal, S.. V. P. Venugopalan, K. V. K. Nair, G. Van der Velde, H. A. Jenner & C. den Hartog. 1998b. Reproduction, growth rate and culture potential of the green mussel, Perna viridis (L.) in Edaiyur backwaters, east coast of India. Aquaculture 162:187-202. Rylander, K., J. Perez & J. A. Gomez. 1996. Status of the green mussel, Perna viridis (Linnaeus 1758) (Mollusca: Mytilidae), in north-eastern Venezuela. Caribbean Marine Studies 5:86-87. Siddall. S. E. 1980. A clarification of the genus Pema (Mytilidae). Bulletin of Marine Science. 30:858-870. Simberloff, D., D. C. Schmitz & T. C. Brown. 1997. Why we should care and what we should do. In: D. Simberioff. D. C. Schmitz & T. C. Brown, editors. Strangers in paradise: impact and management of non- indigenous species In Florida. Washington. DC: Island Press, pp. 359- 367 Sivalingam, P. M. 1977. Aquaculture of the green mussel. Mytilus viridis Linnaeus, in Malaysia. Aquaculture 1 1:297-312. Smith, S. W., R. Overbeek, C. R. Woese, W. Gilben, P. M. Gillevet. 1994. The genetic data environment: an expandable GUI for multiple se- quence-analysis. Computer Applications in the Biosciences 10(6):671- 675. Stoermer. E. P.. G. Emmert, M. L. Julius & C. L. Schelske. 1996. Paleo- limnologic evidence of rapid recent change in Lake Erie's trophic status. Can. J. Fish. Aquat. Sci. 53:1451-1458. Sundaram. K. S. & M. S. Shafee 1989. Salinity tolerance of some bivalves of Ennore Estuary. Joumal of the Marine Biological Association of India 31:299-302. Tan. W. H. 1975. The effects of exposure and crawling behaviour on the survival of recently settled green mussels (Mytilus viridis L.). Aqua- culture 6:357-368. Thompson, J. D.. D. G. Higgins & T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive a multiple .sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673—4680. Tsuchiya, M. 1980. Biodeposit production by the mussel Mytilus edulis L. on rocky shores. J. E.xp. Mar Biol. Ecol. 47:203-222. Tuaychareon, S. 1991. Reproductive biology of the green mussel in Thai- land. Special Publication of the Phuket Marine Biological Center No. 9. pp. 75-79. Umemori, T. & M. Horikoshi. 1991. Death and survival during winter season in different populations of the green mussel, Pema viridis (Lin- naeus), living in different sites within a cove on the western coast of Tokyo Bay. Umi/la mer. 29:103-107. U.S. Congress. Office of Technology Assessment. 1993. Harmful nonin- digenous species in the United States. OTA-F-565. Washington. DC: U.S. Government Printing Office. 391 pp. Vakily, J. M. 1989. The biology and culture of mussels of the genus Pema. ICLARM Studies and Reviews 17. International Center for Living Aquatic Resources Management, Manila, Philippines and Duetsche Gesellshaft fiir Technische Zusammenarbeit (GTZ), GmbH, Eschbom, Federal Republic of Germany. 63 pp. Van de Peer, Y. & R. De Wachter. 1997. Construction of evolutionary distance trees with TREECON for Windows: accounting for variation in nucleotide substitution rate among sites. Comput. Applic. Biosci. l3:227-2.m Van de Peer, Y. & R. De Wachter. 1994. TREECON for Windows: a Non-Nativh Green Mussel in Florida 29 software package for the construction and drawing of evolutionary Vol. 1. Reston. VA: U.S. Department of the Interior. U.S Geological trees for the Microsoft Windows environment. Comintl. Applic. Biosci. Survey, pp. 1 17-129. 10:569-570. Van, Y.. Y. Dong & W. Yan. 1994. An ecological study on the hiofouling Walter, C. 1982. Reproduction and growth ui the tropical mussel Pfnui on buoys in Zhanjiang Port. Tioi>ic Oceimology 13(2);68-74. viridis (Bivalvia: Mytilidae). Kalikasan. Philippine Joiinuil of Biolof>y Yap. W. G., A. L. Young, C. E. F. Orano & M. T. de Castro. 1979. Manual 11:83-97. on mussel fanning. Aquaculture Extension Manual No. 6. Uoilo, Phil- Williams, J. D. & G. K. Meffe. 1998. Nonindigenous species. In: M.J. Mac ippines: Southeast Asian Development Center, Aquaculture Depart- et al., editors. Status and Trends of the Nation's Biological Resources. nicnt. 17 pp. Joiirmil of Shellfish Research, Vol, 20, No, I. 31-38. 2001, SPECIES COMPOSITION OF BLUE MUSSEL POPULATIONS IN THE NORTHEASTERN GULF OF MAINE PAUL D. RAWSON, SUSAN HAYHURST, AND BROOK VANSCOYOC School of Marine Sciences, 5751 Murray Hall, University of Maine, Orono, Maine 04469-5751 ABSTR.ACT We e.xammed the species composilion of eleven blue mussel populations in eastern and central Maine. USA using a set of PCR-hased genetic markers. Previous reports suggested that mussel populations m the Gulf of Maine were composed of only a single species. Mviilus edulis. In contrast, our results clearly indicate that the range of a congener. M. irossulus, extends well into the Gulf. The two blue mussels are sympatric in eastern Maine populations, including all of those we sampled within Cobscook Bay. ME, The frequency of A/, trossulus. however, declines dramatically in the vicinity of Little Machias Bay, ME, so that populations along the coast of central Maine are composed predominantly of M. edulis mussels. Among populations containing a mi.xture of M. edulis and M. rrossidus-specifk alleles we ob.served a low but significant frequency of mussels with hybrid genotypes including putative backcross genotypes, indicating the potential for introgression between these two species. We suggest that larval supply and recruitment, thermal tolerance and perhaps the interplay of these factors likely delimit the southern extent of the range of A/, trossulus and influence the .species composition of blue mussel populations in the northwest .Atlantic. KEY WORDS: blue mussels. Mytilus spp,. hybridization, hiogeography. molecular markers INTRODUCTION Mussels in the genus M\tihis are widely distributed in the tem- perate and sub-polar zones of both the northern and southern hemi- spheres (Gosling 1992a) and are often dominant members of the rocky intertidal community. This genus includes a group of mor- phologically similar smooth shelled or blue mussels belonging to the Mytilus edulis species complex. Blue mussels have been the focus of a wide variety of studies on topics ranging from physi- ology, biochemistry, ecology, biogeography and pollution ecol- ogy, and are economically important as a source of food in many parts of the world (Gosling 1992a). Due to the lack of rehable morphological characters, the taxonomy of blue mussels has been highly confused historically. Based on the work of McDonald et al. ( 1991 ), however, who employed a combination of allozyme elec- trophoresis and multivariate analysis of morphological characters there are currently three recognized species of blue mussel world- wide, M. edulis (Linnaeus 1758). M. galh^pnivincialis (Lamark 1819), and M. trossulus (Gould 1850). In the northwest Atlantic. allozyme surveys have fuinher indicated that M. edulis occurs from Cape Hatteras. NC to the Canadian Maritimes while M. trossulus has been observed at multiple locations in southeastern Nova Scotia and eastern Newfoundland, often in mixed populations with M.edulis (Koehn et ak 1976, 1984, Varvio et al. 1988, Bates and Innes 1995). Based on these surveys, it has been generally as- sumed that mussel populations within the Gulf are composed en- tirely of M. edulis. The species composition of blue mussel populations in the northeast Gulf of Maine, however, had not been examined until Hennigar et al. (1996) used allozyme electrophoresis to assay variation at the mannose phosphate isomerase (Mpi) locus for mus- sels satnpled from several sites in the Gulf, including the Bay of Fundy. The Mpi allozyme locus is considered to be the most di- agnostic enzyme marker for identifying M. edulis and M. trossulus (e.g. Mallet & Carver 1 995, Comesaiia et al. 1 999 ). Hennigar et al. (1996) observed a high frequency of M. /)Y),v.v((/».v-specific Mpi alleles al Digby, Nova Scotia and Hospital Island, New Brunswick, providing the first indication that the range of A-/, irossulus extends into the Gulf of Maine. The presence oi M. trossulus in this region is a serious concern to the Maine mussel culture industry. Although the mussel fishery in Maine has traditionally relied on harvesting from both natural beds and bottom culture leases seeded with juvenile mussels, the industry has recently begun to employ suspended culture systems. Reports from mussel farms in the Maritime Provinces of Canada suggest that M. trossulus tnussels contain less meat than M. edulis mussels and their shells are prone to fracturing during processing (Freeman et al. 1994, Mallet and Carver 1995). Given species- specific growth rates, survival, and production losses. Mallet and Carver (1995) have estimated that, under raft culture conditions, the economic value of M. edulis is 1 .7 times that of M. trossulus. Blue mussels also figure prominently in environmental moni- toring efforts within the Gulf of Maine. The Gulfwatch program, initiated in 1991 by the Gulf of Maine Council on the Marine Environment, assesses coastal habitat exposure to inorganic and organic contaminants by examining the accumulation of heavy metals, PCBs. and other toxins in mussels collected from 60 sites located between southern Massachusetts and Nova Scotia (Sowles et al. 1997). Significant differences have been observed in the seasonal patterns of lead, chromium, zinc, and mercury accumu- lation between M. edulis and M. trossulus (Mucklow 1996) em- phasizing the need for accurate species identification and consid- eration of taxonomic affinity when interpreting data from the Gulf- watch program (Lobel et al. 1990, Sowles et al. 1997). In this study, we set out to determine the southern-most extent of M. trossuhi.t' range in the Gulf of Maine. Using a set of diag- nostic molecular markers, we have assessed the species composi- tion of eleven blue mussel populations sampled froin central to eastern Maine. Recently, Bates and Innes (1995) and Comesaiia et al. (1999) have used allozyme electrophoresis as well as DNA- based genetic markers to investigate the degree of hybridization between these two species and test hypotheses regarding the eco- logical factors which affect their distribution on the coast of New- foundland. Although our study was not designed to specifically test for associations between species abundance and ecological factors, we compare our results to their findings to assess whether similar ecological and genetic mechanisms may be shaping the local distribution of and interaction between M. edulis and M. tro.ssulus throughout the region where they are sympatric. 31 32 Rawson et al. MATERIALS AND METHODS Approximately 30 to SO adult blue mussels (10.9-61.8 mm shell length) were collected from eleven locations along the coast of central and eastern Maine, including two separate sites near the Darling Marine Center (see Fig. 1). Sampling was conducted be- tween February and July of 1999 at the lowest tide level during spring tides. Mussels from all twelve locations were transported live to the University of Maine were they were dissected. DNA was isolated from a tissue biopsy using the protocol of Schizas et al. (1997). wherein a small piece of mantle tissue (-10 mg) from each mussel was digested with 40 ng of proteinase K in 20 p-l of a buffer containing 50 mM KCl. 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-100 at 55°C for 3 hours. After the digestion process, the samples were heated to 95°C for 15 minutes and 20 p,l of Gene Releaser (BioVentures Corp.) was added. The resulting slurry was incubated at 65''C for 30 s, 8°C for 30 s, 65°C for 90 s. 97°C for 180 s. 8°C for 60 s, 65°C for 180 s, 97°C for 60 s, 65°C for 60 s and centrifuged at 13.000 rpm for 2 min. The supernatant contain- ing total cellular DNA was then removed to a new microcentrifuge tube. Isolated DNA was used as template in each of four polymerase chain reaction (PCR)-based genetic assays. The application of the mtl6S-F. Glu-5'. and ITS markers followed the protocols provided by Rawson and Hilbish (1995), Rawson et al. (1996a), and Heath et al. (1995), respectively. In this study, the use of the Mytiliis anonymous locus I (MAL-I) marker essentially followed the pro- tocol in Rawson et al. (1996b) except that the oligonucleotide primers were redesigned to amplify a smaller PCR product. Three of these markers, ITS, MAL-I, and Glu-5', target the nuclear ge- nome while the mtl6S-F marker targets the mussel female mito- chondrial DNA lineage. For each assay, 50 ng of template DNA was added to a 12.5 |xl reaction (total volume) containing 50 mM KCl. 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-IOO, 1.5 mM MgCU 2.5 nmol dNTPs, I U Tacj DNA polymerase (Promega) and 50 pniol of assay-specific forward and reverse oligonucleotide primers (Table I ). The reactions were initially denatured for three minutes at 94''C and then incubated for 30 cycles at 94°C for 30 s, followed by an assay-specific annealing temperature (see Table I) for 30 s. and 72°C for 1 minute. The ITS, MAL-I and mtl6s-F markers relied on PCR to generate equal length PCR products for both M. ediilis and M trossiiliis. When these products were then digested with the appropriate restriction enzymes (see Table 1) species-specific restriction fragment length patterns were gener- ated. For the Glu-5' marker, which is similar to that designed by Inoue et al. ( 1995), PCR amplification produced M. cdiilis and M. trossiiliis-ipec\{ic products that differ significantly in size (see Rawson et al. 1996a for details). For each marker the species- specific PCR products or PCR/RFLP fragments were resolved on 1 .59c to 2.5% agarose gels. Individual genotypes and haplotypes were scored as outlined for ITS in Heath et al. (1995), Glu-5' in Rawson et al. (1996a). and mtl6S-F in Rawson and Hilbish (1995, 1998). Although Rawson f Quoddy Head Grand Manaan Island So. Trescott Little Machias Bay Machjasport Jones port M. trossulus M. edulis Darling Marine Center Glu-5' marker Figure i. Map of the blue mussel populations in central and eastern Maine sampled in this study indicating the frequency of Mytilus edulis and M. frossw/Hs-specific Glu-5' alleles observed at each location. The frequency of (;iu-5' alleles at the two Darling Marine Center sites were combined for clarity. Gulf of Mainf Blue Mussel Populations 33 TABLE 1. Olisonuclcotide primer sequences, annealin;; temperature, and restriction enzymes used in each of the four DNA-based markers employed in this study. Marker Primer Sequence Temperature Enzyme ITS Glu-S' MAL-1 nit 1 6S-F Forward: GTTTCCGTAGGTGAACCTG Reverse: CTCGTCTGATCTGAGGTCG Forward: GTAGGAACAAAGCATGAACCA Reverse: GGGGGGATAAGTTTTCTTAGG Forward: GAAGCGTATTTGGTCACTGGCAC Reverse: GTCATAAAATGGAACATCTGAGTC Forward: CCGGTCTGAACTCAGATCACGT Reverse: CTGCCCAGTCGAACTAGAGTAAT 50°C 57X 50°C 47°C Hhal NA Spel Spel + EcoRV et al. ( 1996b) described the applieatioii of the MAL-I marker tor distinguishing M. ii-dsskIks and M. galloproviiuialis. their study did not inckide M. ediilis. As they reported, when MAL-1 PCR products generated from M. trossuliis DNA are digested with the restriction enzyme Spel two different restriction profiles are ob- served. In contrast, when MAL-I PCR products generated from M. cdulis DNA are digested with the same restriction en/yme. a single profile distinct from those observed for M. trossulus results (Fig. 2). Allopatric populations of M. eilulis (Lewes. DE and Cape Anne. MA) and M. trossulus (Port Orford and Newport. OR) are fixed for alternate PCR/RFLP profiles at all four of these DNA- hascd markers (Table 1). M 1 2 3 4 Figure 2. Restriction fragment length polymorphism for .1/. cdulis and M. tmssiiliis at the NL\L-I locus analyzed on a 2.5% agarose gel. PCR amplitlcation produces a single -650 base pair band for both species as shown in the lane 1. Digestion of the PCR product produces a single banding pattern for M. edulis (lane 2) and two separate patterns for M. Irossidiii. Lane 3 shows the RFLP pattern for an individual homozy- gous for the Tl allele while lane 4 was produced by an individual heterozygous for the TI and T2 alleles. Allele and haplotype frequencies were estimated for each of the twelve sampling locations. Post hoc tests of heterogeneity were conducted using the allele frequencies at each locus to determine whether there were significant changes in the frequency of M. trossulus-specific alleles across all eleven study sites as well as across subsets of sites in eastern and central Maine. For each test, an R X C contingency table was constructed and an e.xact test for heterogeneity in allele frequency performed using the STRUC pro- gram distributed with the GENEPOP .software package (Raymond and Rousset 1995). To estimate the frequency of mussels with multi-locus M. irossidiis. M. edulis and hybrid genotypes, a hybrid index was constructed based on the number of M. trossulus- specific alleles a given individual carried at the ITS. MAL-I. and Glu-5' markers. Thus, mussels homozygous for M. trossulus al- leles at each nuclear marker received a score of 6. those homozy- gous for M. edulis alleles received a score of 0. while mussels with mixed multi-locus genotypes received a score between 1 and 5. For example, mussels which were heterozygous at two loci and ho- mozygous for M. /roMif/wi-specific alleles at the third locus re- ceived a score of 4. The frequency of mussels falling into each hybrid index score was estimated for the Whiting. East Bay. Gove Point. Lubec, Quoddy Head. South Trescott and Little Machias Bay samples combined. Each of the.se samples contained M. tros- sulus alleles at a frequency in excess of 5%. RESULTS We successfully genotyped 5\S individual adult mussels iMyli- lus spp.) for three, and in most cases, four PCR-based genetic markers. There was highly significant spatial heterogeneity in the frequency of M. trossulus and M. edulis-specific alleles and hap- lotypes across the 1 1 locations sampled in this study (Table 2). This was illustrated by the variation in the frequency of M. tros- sulus and M. edulis Glu-5' alleles shown in Figure 1. In mussel populations east of the Little Machias Bay site, including all of the sites sampled in Cobscook Bay. M. trossulus alleles and haplo- types were quite common, with frequencies ranging from a low of 34.39f at South Trescott to a maximum of 97.4<7r at Gove Point. West of Little Machias Bay. however, the frequency of M. tros- sul US-specific alleles declined precipitously. There was no evi- dence of spatial heterogeneity in allele frequency for any of the four markers we employed among the Machiasport. Jonesport. Seal Cove, and both of the Darling Marine Center populations (Table 2) indicating that populations along the central coast of Maine consistently contained predominantly M. edulis alleles. The observed spatial variation for each of the four markers was 34 Rawson et al. TABLE 2. Allele frequencies and haplotype frequencies in reference mussel populations and at tv>elve sites sampled in Gulf of Maine. ITS" Glu-S'" MAL-r mtl6S-F'' Location n E T n E T n E Tl T2 n A B Allopatric Populations Newport. OR 52 0.0 100.0 52 0.0 100.0 52 0.0 71.0 29.0 52 0.0 100.0 Port Orford, OR 27 0.0 100.0 27 0.0 100.0 27 0.0 74.1 25.9 27 0.0 100.0 Lewes. DE 40 100.0 0.0 40 100.0 0.(1 40 100.0 0.0 0.0 40 100.0 0.0 Cape Anne. MA 36 100.0 0.0 36 100.0 0.0 36 100.0 0.0 0.0 36 100.0 0.0 Eastern Maine Whiting 82 25.6 74.4 82 28.0 72.0 76 21.1 55.3 23.6 41 22.0 78.0 East Bay 74 54.0 46.0 82 57.3 42.7 80 55.0 36.3 8.7 42 .54.7 45.3 Gove Point 74 1.4 98.6 76 2.6 97.4 72 1.4 65.3 33.3 37 0.0 100.0 Lubec 40 16.7 83.3 94 21.3 78.7 86 16.3 53.5 .30.2 47 14.9 85.1 Quoddy Head 114 44.1 50.9 114 ,54.4 45.6 112 50.0 29.5 20.5 57 49.1 50.9 South Trescott 68 64.7 .V5.3 70 65.7 34.3 68 6L8 27.9 10.3 35 62.9 37.1 Little Machias Bay 118 .55.9 44.1 120 56.7 43.3 118 54.2 33.1 12.7 60 55.0 45.0 Central Maine Machiasport 74 100.0 0.0 94 100.0 0 70 100.0 0.0 0.0 46 100.0 0.0 Jonesport 94 95.7 4.3 96 95.8 4.2 84 95.2 4.8 0.0 47 95.7 4.3 Seal Cove 58 97.-3 2.7 62 100.0 0 60 100.0 0.0 0.0 30 100.0 0.0 Darling Center A 48 91.6 8.4 42 90.5 9.5 34 94.1 5.9 0.0 26 92.3 7.7 Darling Center B 76 97.3 2.7 88 97.7 2.3 82 97.6 2.4 (1,0 46 97.8 2.2 Tests for Heterogeneity All Sites *** *** *** *** Eastern Maine'' ^:** :h** *** *** Central Maine' NS NS NS NS " Genotypes at the ITS and Glu-5' markers were scored as in Heath et al. (1995) and Rawson et al. ( 1996al and lor MAL-I marker as described in the text. E = M. edulis-ipec\{\c alleles and T = M f;'o.MH/»i'-specit~ic alleles. '' Haplotypes at mtl6S-F were scored as in Raw.son and Hilbish (1995. 1998). A = Af . (>rfi(//i-specific haplotype and B = M. trossulus-^peafk haplotype. ■■' Tests for heterogeneity in allele/halpotype frequency were conducted across all sites as well as across sites within two regions. Eastern Maine (Whiting. East Bay, Gove Point. Lubec. Quoddy Head. South Trescott, and Little Machias Bay) and Central Maine (Machiasport, Jonesport. Seal Cove. Darling Center A and B). ***p< 0.001. NS not siizniticant. highly concordant (Table 2). This was further reflected by the predominance of mussels homozygous for M. edulis or M. trossii- liis alleles at ail three nuclear loci. For each of the mixed popula- tions containing appreciable frequencies of M. trossiihis alleles (i.e. Whiting. East Bay. Gove Point. Lubec. Quoddy Head. South Trescott and Little Machias Bay), a hybrid index was calculated for each mussel based on the number of M. trossiiliis-specific alleles at each nuclear marker. Among the 320 mussels sampled from these seven sites, the vast majority (>87'7f ) had index scores of 0 or 6, i.e. nearly all of the mussels sampled were homozygous for all three nuclear makers (Fig. 3). Less than 1% of the mussels we sampled had index scores of 3 and no mussel heterozygous at all three nuclear markers was observed. Of the remaining mussels, most (97i ) had index scores of 4 or 5 indicating that these mussels carried more M. irossuliis-specific than M. ccliilis-specii'ic alleles. In addition, there was a high concordance between multi-locus nuclear genotypes and mitochondrial haplotypes. Of the 117 mus- sels with an index score of 0, all but one carried the A mtl6s-F haplotype fixed in allopatric populations of M. edidis. while 161 of 163 mussels with an index score of 6 had the B mtl6S-F haplotype found in allopatric populations of M. trossidiis. DISCUSSION Our results indicate that a marked transition in the species composition of blue mussel populations occurs in the northeastern 06 05 O 0 4 Z m 3 0.3 o UJ q: 02 0 1 0 12 3 4 5 6 HYBRID INDEX Figure 3. Frequency of mussels hIiIi nuiltl-locus hybrid index scores based on the number of A/. rro.svH/Hv-specific alleles at each of three nuclear markers. \ score of II indicates mussels homozygous for M. edulis-syu'vW'K alleles at all three nuclear loci, a score of 6 indicates mussels homozygous for M. (TOvvH/Hv-speeinc alleles and scores of 1-5 indicate mussels containing a mixture of M. edulis and A/, tnissiilus alleles. The analysis includes only mussels from the seven populations which contained >5% M. (tossh/hj -speeitic alleles. Gulf of Maine Blue Mussel Populations 35 Gulf of Maine. This contrasts witln the conclusions of Koehn et al. (1976. 1984), who suggested that Gulf of Maine populations were composed solely of M. edulis mussels. Their studies, however, did not sample mussel populations between Bar Harbor in central Maine and Halifax on the eastern side of the Nova Scotian pen- insula. Together with the results of Hennigar et al. (1996). who found high frequencies of M. r(OM»/i«-specific Mpi alleles in neighboring Passamaquoddy Bay. New Brunswick and Digby, Nova Scotia, our observation of fairly high frequencies of M. tmssKliis at several sites in eastern Maine, especially in the Cobs- cook Bay region, clearly indicates that mussels with M. trossulKs alleles occur throughout the Bay of Fundy region and well into the Gulf of Maine. Our results also suggest that there has been a limited amount of introgression between M. edulis and M. rmssiiliis. Within popula- tions containing appreciable frequencies of both M. edulis and M. trossulus alleles in eastern Maine, the vast majority of mussels had multi-locus genotypes consistent with their being pure M. edulis or M. edulis mussels. Among these multi-locus homozygotes (n = 280). all but three had species-specific mitochondrial haplotypes that were in agreement with their nuclear genotypes. For mussels with multi-locus hybrid genotypes, we found most to have hybrid index scores of 4 or 5. genotypes which are indicative of back- crosses formed by the mating of Fl hybrids with M. trossulus individuals. Thus, hybridization between M. edulis and M. trossu- lus to produce Fl mussels must occur, at least occasionally, even though we observed no tri-locus heterozygous mussels in our samples and the observation of putative backcross genotypes sug- gests that a low level of introgression occurs between these spe- cies. A thorough analysis of introgression would benefit from addi- tional nuclear markers (e.g. Boecklen and Howard 1997) and a more rigorous statistical framework (e.g. Nason & Elstrand 199.^). From the reports presented to date, however, hybridization be- tween M. edulis and M. trossulus appears to be relatively limited. Estimates of the frequency of hybrids ranges from a low of <2% (Mallet & Carver 19951 to as much as 23'7f in No\a Scotia (Saave- dra et al. 1996) and 269c in Newfoundland (Comesaiia et al. 1999). This level of hybridization is comparable to that reported by Raw- son et al. (1999). Sarver and Foltz (1993) and Suchanek et al. (1997) for M. trossulus and M. galloproviiicialis on the Pacific coast of North America. In contrast, several studies have reported a frequency of hybrids in excess of 50 to 80% at locations in Western Europe where M. galloprovincialis hybridizes with M. edulis (Sanjuan et al. 1994. Comesaiia & Sanjuan 1997, Hilbish et al. 1994. Gardner 1996). Although many of these studies have employed different marker systems with varying ability to dis- criminate hybrid classes, taken together these observations are consistent with phylogenetic studies which suggest that M. edulis and M. galloproviiicialis are sister taxa and M. trossidus is more distantly related to these other two taxa (Rawson & Hilbish 1998). Gardner ( 1996) has proposed that differential adaptation to en- vironmental conditions is the primary force determining the dis- tribution of all three species of blue mussel, M. edulis. M. gallo- provincialis, and M. trossulus, in the Northern Hemisphere. For example, within the most extensively studied blue mussel hybrid zone, several studies have shown that habitat-specific selection based on differential growth or physiology, strength of attachment to the substrate and other factors, plays a major role in the distri- bution of mussel genotypes across the region of sympatry between M. edulis and M. galloprovincialis in western Europe (e.g. Hilbish et al. 1994, Comesaiia and Sanjuan 1997. see review by Gosling 1992b). Among the other blue mussel hybrid zones studied to date, the importance of differential adaptation is less well documented. Sarver and Foltz ( 1993) observed a correlation between local tem- perature and salinity and the distribution of M. galloprovincialis and M. trossulus on the Pacific coast of North America. Rawson et al. ( 1999), however, pointed out that hybridization between these two species occurs between Monterey and Cape Mendocino. Cali- fornia, a region that is characterized by extremely complex hydro- dynamic conditions, so that limited dispersal may also be involved in determining the distribution of parental and hybrid mussels. Likewise, in the northwest Atlantic it is presently unclear what mechanisms operate to structure the species composition of blue mussel populations. It has been suggested that M. trossulus may be better adapted to estuarine conditions and therefore is more toler- ant of low salinity than is M. edulis (e.g. Gosling 1992b). Nearly all of the sites we sampled, including those with high frequencies of M trossulus alleles, however, experience oceanic salinities (35 ppt) year round. The only exception was the Machias Bay popu- lation which was far enough upriver to occasionally receive pulses of fresh water. Yet, mussels at this site were fixed for M. edulis alleles and haplotypes. These results suggest that salinity cannot be the sole selective agent regulating the local distribution of M. trossulus and M. edulis. Bates and Innes ( 1995) and Comesai'ia et al. (1999) have also investigated the role of wave exposure in structuring mussel populations off the coast of Newfoundland. Their studies failed to find a clear association between the fre- quency of M. trossulus and degree of wave exposure in the popu- lations they sampled. Similarly, we observed high frequencies of both M. edulis and M. trossulus alleles in all populations sampled in eastern Maine, including relatively sheltered sites within Cobs- cook Bay as well as the exposed sites on the outer coast such as Quoddy Head and South Trescott. Thus, neither salinity nor wave exposure is likely to explain the precipitous decline in the fre- quency of M. trossulus that we have observed in eastern Maine. The spatial variation in the frequency of M. trossulus mussels that we observed among eastern Maine populations may reflect variation in larval supply and recruitment. Johnson and Black (1982) and Watts et al. (1990) suggested that genetic patchiness among populations of Siphonaria sp. and Echinoinetra mathaei were likely due to stochastic variation in larval recruitment. On a broader scale, regional current patterns can potentially limit larval dispersal and proinote genetic divergence among conspecific populations. Although many population genetic surveys of marine species with highly dispersive stages have found little evidence to support the hypothesis that present day current patterns influence genetic divergence among contemporary marine populations (e.g. Benzie and Williams 1995, Shulman and Bermingham 1995), other studies such as those of Tracey et al. (1975) and Rocha- Olivares and Vetter ( 1999) observed that genetic structure for lob- sters and rockfish. respectively, coincided with regional oceanic circulation patterns. In the Gulf of Maine, the Eastern Maine Coastal Current (EMCC) is a prominent oceanographic feature and is known to transport herring larvae from spawning grounds near Grand Manan Island westward along the coast of Maine (Townsend 1992). The EMCC also appears to have a significant influence on the distribution of Ale.xandrium sp., a dinotlagellate responsible for paralytic shellfish poisoning. The lowest incidence of paralytic shellfish poisoning along the coast of Maine occurs between the western edge of Penobscot Bay and Jonesport. ME (Shumway et al. 1988). In this region, Ale.xandrium is found pre- 36 Rawson et al. dominantly offshore and its distribution coincides with an off- shore-directed tongue of cold water near Jonesport (Dave Townsend, pers. eomm.). The dispersive larval stage in blue mus- sels typically spans upwards of at least 3 to 4 weeks (Bayne 1976). Given a mean current velocity in the EMCC approaching 13-25 cm s~' (Townsend 1992) there is the potential for any M. trossiilus larvae entrained in the EMCC to be transported great distances to the west. Moreover, the offshore flow of the EMCC near Jonesport. ME is particularly evident in satellite imagery from the summer months (Fig. 4; see also Pettigrew et al. 1998) when the bulk of mussel larvae are in the water column along the coast of Maine. Thus, it is possible that any M. trossiilus larvae transported west further into the Gulf of Maine may not be retained in near- shore waters at a time when they are competent to settle, and the relative strength and direction of the EMCC then, determines the extent to which the range of M. trossulus extends into the Gulf. Temperature variation has also been investigated as a potential mechanism structuring the species composition of blue mussel populations in the northwest Atlantic (Mallet and Car\er 199?). Given the broad-ranging distribution of M. edidis. the more re- stricted sub-polar distribution of M. irossiihis. and the absence of M. trossulus from the Northumberland Straits between Prince Ed- ward Island and New Brunswick. Canada, it has been suggested that the latter species is less tolerant of warmer waters (e.g. Gos- ling 1992b, Mallet & Carver 1995). Mallet and Carver (1995) examined the relative survival and growth of M. trossulus and M. edulis at two locations differing in mean temperature in Nova Scotia. Their findings, that M. trossulus demonstrated higher sur- vival rates at the warmer site, were the reverse of what would be predicted based on the large scale distribution of M. edulis and M. trossulus. Their study, however, was limited to a comparison of the survival of adult mussels in two size classes (20-30 and 40-55 mm shell length). In contrast, Bayne (1976) has argued that the environmental requirements for normal embryonic development are likely to be a major factor limiting the distribution of blue mussel populations. If so, then attempts to correlate adult distributions with spatial varia- tion in environmental parameters are likely to be of limited value in determining the mechanisms regulating the species composition of mussel populations in the Gulf of Maine. More attention should be focused on processes affecting the early life history stages of mussels. In this regard, satellite imagery suggests that there may be a significant change in coastal water temperature along the coast of Maine, coincident with the offshore mo\ emenl of the EMCC. East of Jonesport. where the EMCC remains close to shore, the near- shore waters remain relatively cool throughout the summer com- pared to those west of Jonesport. Thus, the EMCC may impact blue mussel larvae in two ways, in a direct manner wherein M. trossulus larvae are advected away from potential settlement sites and an indirect manner wherein the EMCC creates a temperature differential between eastern and western Maine that affects the survival of M. trossulus larvae. We have recently initiated a series of experiments designed to assess the relative importance of these two mechanisms in limiting the range of M. trossulus in Maine. In conclusion, our results clearly show that M. trossulus occurs, and in some cases is the predominant species in mussel popula- tions in eastern Maine. This study sampled two locations, Machi- asport and Go\e Point, frequently sampled by the Gulfwatch pro- gram (Hennigar et al. 1996). The marked change in species com- position between these two sites illustrates the importance of monitoring the frequency of M. trossulus at Gulfwatch sites in the northeastern Gulf of Maine to facilitate the interpretation of Gulf- watch data. The findings of Mallet and Carver ( 1995) regarding the economic impact of A-/, trossulus on mussel aquaculture also high- light the importance of monitoring species composition at current and future mussel culture lease sites to enable the industry to select sites that limit the presence of this species. Additional work is now underway in our laboratory. We are attempting to better define the factors that control the distribution of M. trossulus throughout the Gulf of Maine, to further understand how the species composition of Gulf of Maine populations may change in the future. ACKNOWLEDGMENTS This research was supported by the University of New Hamp- shire/L'ni\ersity of Maine Sea Grant program under contract 99- 343 and Hatch funds provided by the U.S. Department of Agri- culture and the Maine Agriculture and Forestry Experiment Station 70° 68°44° _l I NEW BRUNSWICK ■46° -42° Figuri' 4. Ihe map on the left provides a diaHriiiiiniatic representation of the general circulation patterns (arrows! in the (iulf of Maine during stratified conditions typical in this region from May to September (after Pettigrew, unpublished). The figure on the right is an AVHRR satellite image of sea surface temperature in the Gulf of Maine on July 19, 1998. Cooler water appears as darker shades and warmer water as lighter shades in this image. The eastern .Maine coastal current and an offshore-directed plume are indicated by the arrows. Image courtesy of the University of Maine Satellite Oceanography Data Laboratory, Gulf of Maine Blue Mussel Populations 37 (Project No. ME08510). The authors would like to thank Dr. P. Yund and D. DenDanto for graciously providing some of the samples used in this study. Thanks also to Carter Newell of Great Eastern Mussel Farms Inc. for an introduction to the mussel culture industry in Maine and to John Sowles of the Maine Department of Environmental Protection for providing background information on the Gulfwatch program. In addition. R. Luerssen and A. Thom- as of the Satellite Oceanography Data Laboratory at the University of Maine provided the satellite imagery and thanks to D. Townsend for several enlightening discussions on the current pat- terns in the Gulf of Maine. Constructive comments on an earlier draft of this manuscript were graciously provided by M Gordon, S. Lindsay and an anonymous reviewer. LITERATURE CITED Bayne. B. L. ( 1976). The biology of mussel larvae. In: B. L. Bayne, editor. Marine mussels: their ecology and physiology. Cambridge, UK: Cam- bridge University Press, pp. 81-120. Bates, J. A. & D. J. Innes. 1995. Genetic variation among populations of Myliliis spp. in eastern Newfoundland. Men: Biol. 124:417^24. Benzie, J. A. H. & S. T. Williams. 1995. Gene flow among giant clam (Triilacna gigas) populations in the Pacific does not parallel ocean circulation. Men: Biol. 123:781-787. Boecklen, W. J. & D. J. Howard. 1997. Genetic analysis of hybrid zones: numbers of markers and power of resolution. Ecology 78:261 1-2616. Comesaiia, A. S. & A. Sanjuan. 1997. Microgeographic allozyme differ- entiation in the hybrid zone oi Mytilus galloproviiicialis Lmk. and M. edulis L. on the continental European coast. Hclgol Mcercswucrs 51: 107-124. Coniesana, A. S., J. E. Toro, D. J. Innes & R. J. Thompson. 1999. A molecular approach to the ecology of a mussel iMytilus ctliilis-M. rro.s.siiliii) hybrid zone on the east coast of Newfoundland. Canada. Mai: Biol. 1.^3:213-221. Freeman, K. R., K. L. Perry, T. G. DiBacco & D. J. Scarratt. (1994). Observations on two Mytilid species from a Nova Scotian mussel farm. Canadian Technical Report of Fisheries and Aquatic Sciences, No. 1969. Halifax, NS, Canada: Dept. Fisheries and Oceans. Gardner, J. P. A. 1996. The Mytilus ediilis species complex in southwest England: effects of hybridization and introgression upon interlocus associations and morphometric variation. Mai: Biol. 125:385-399. Gosling, E. M. {1992a). Systematics and geographic distribution of Myli- liis. In: E. M. Gosling, editor. The mussel Mytilus: ecology, physiology, genetics and culture. Amsterdam: Elsevier, pp. 1-20. Gosling. E. M. (1992b). Genetics of Mytilus. In: E. M. Gosling, editor. The mussel Mytilus: ecology, physiology, genetics and culture. Amsterdam: Elsevier, pp. 309-382. Heath, D. D., P. D. Raw.son & T. J. Hilbish. 1995. PCR-based nuclear markers identify alien blue mussel {Mytilus spp.) genotypes on the west coast of Canada. Can. J. Fish. Aquat. Sci. 52:2621-2627. Hennigar, P., J. Sowles, R. Crawford & G. Harding. ( 1996). Evaluation of Gulfwatch 1993. Augusta, ME: The Gulf of Maine Council on the Marine Environment. State Planning Office. Hilbish, T. J., B. L. Bayne & A. Day. 1994. Genetics of physiological differentiation within the marine mussel genus Mytilus. Evolution 48: 267-286. Inoue, K., J. H. Waite, M. Malsuoka. S. Odo & S. Harayama. 1995. Interspecific variation in adhesive protein sequences of Mytilus edulis. M. galloprovincialis. and M. Irossulus. Biol. Bull. Mai: Biol. Lab. Woods Hole 1 89:370-375. Johnson, M. S. & R. Black. 1982. Chaotic genetic palchiness in an inter- lidal limpet. Siphonaria sp. Mar. Biol. 70:157-164. Koehn, R. K.. R. Milkman & J. B. Muton. 1976. Population genetics of marine pelecypods. IV. Selection, migration and genetic differentiation in the blue mussel Mxtilus edulis. Evolution 30:2-32. Koehn, R. K.. J. G. Hall. D. J. Innes & A. Z. Zera. 1984. Genetic differ- entiation of Mvtilus edulis in eastern North America. Mar. Biol. 70: 117-126. Lobel, P. B., S. P. Belkhode. S. E. Jackson & H. P. Longerich. 1990. Recent taxonomic discoveries concerning the mussel Mxtilus: implica- tions for biomonitoring. .4/r/). Environ. Contamination & Toxicology 19:508-512. Mallet, A. L. & C E. Carver. 1995. Comparative growth and survival patterns of Mytilus trossulus and Mytilus edulis in Atlantic Canada. Can. ,/. Fish. Acpiat. Sci. 52:1873-1880. McDonald. J. H., R. Seed & R. K. Koehn. 1991. Allozymes and morpho- metric characters of three species of Mytilu.s in the Northern and South- ern Hemispheres. Mar. Biol, 111:323-333. Mucklow. L. C. (1997). Effects of season and species on physiological conditions and containment burdens of mussels (Mytilus edulis L. and Mytilus trossulus G.): implications for the Gulfwatch Program. Au- gusta, ME: The Gulf of Maine Council on the Marine Environment, State Planning Office. Nason, J. D. & N. C. Ellstrand. 1993. Estimating the frequencies of ge- netically distinct classes of individuals in hybridizing populations. J. Heredity 84:\-\2. Pettigrew, N. R.. D. W. Townsend. H. Xue. J. P. Wallinga & P. Brickley (1998). Observations of the Eastern Maine Coastal Current and its offshore extensions in 1994. J. Geophy.s. Res. l03(CI3)30:623-639. Rawson, P. D. & T. J. Hilbish. 1995. Evolutionary relationships among the male and female mitochondrial DNA lineages in the Mytilus edulis species complex. Mol. Biol. Evol. 12:893-901. Rawson, P. D., K. Joyner & T J. Hilbish. 1996a. Evidence for intragenic recombination within a novel genetic marker that distinguishes mussels in the Mytilus edulis species complex. Heredity 77:599-607. Rawson, P. D., C. L. Secor & T. J. Hilbish. 1996b. The effects of natural hybridization on the regulation of doubly uniparental mtDNA inheri- tance in marine mussels {Mytilus spp.). Genetics 144:241-248. Rawson, P. D. & T. J. Hilbish. 1998. .Aisyinmetric introgression of female and male lineage mitochondrial DNA (mtDNA) haplotypes within the European hybrid zone between Mylilus edulis and M. galloprovincialis. Evolution 52:100-108. Rawson, P. D.. V. Agrawal & T. J. Hilbish. 1999. Hybridization between the blue mussels Mytilus galloprovincialis and M. Irossulus along the Pacific coast of North America: evidence for limited introgression. Mar. Biol. 134:201-211. Raymond. M. & F. Rousset. 1995. GENEPOP. Vers. 1.2. A population genetics software for exact tests and eccumenicism. J. Hered. 86:248- 249. Rocha-Olivares, A. & R. D. Vetter. 1999. Effects of oceanographic circu- lation on the gene flow, genetic structure, and phylogeography of the rosethom rockfish (Sebastes helvoiiiaculatiis). Can. J. Fish. .Acpiai. Sci. 56:803-813. Saavedra, C, D. T Stewart, R. R. Slanwood & E. Zouros. 1996. Species- specific segregation of gender-associated mitochondrial DNA types in an area where two mussel species {Mytilus edulis and M. irossulus) hybridize. Genetics 143:1359-1367. Sanjuan, A., C. Zapata & G. Alvarez. 1994. .Mytilus galloprovincialis and M. edulis on the coast of the Iberian peninsula. Mar. Ecol. Prog. Set: 113:131-146. Sarver, S. K. & D. W. Fohz, 1993. Genetic population structure of a species' complex of blue mussels {Mytilus spp.). Mar. Biol. 117:105- 112. Schizas, N. V., G. T. Street. B. C. Coull, G. T. Chandler & J. M. Quauro. 1997. An efficient DNA extraction method for small metazoans. Mol. Mar. Biol. Biotech. 6:381-383. Shulman. M. J. & E. Bermingham. 1995. Early life histories, ocean cur- rents, and the population genetics of Caribbean reef fishes. Evolution 49:897-910. Shumway, S. E., S. Sherman-Caswell & J. W. Hurst. 1988. Paralytic 38 Rawson et al. shellfish poisoning in Maine: monitoring a monster. J. Shcllfi.sli Res. 7:643-652. Sowles, J., R. Crawford, P. Hennigar, G. Harding, S. Jones. M. Chase. W. Robinson, J. Pederson, K. Coombs, D. Taylor & K. Freeman. ( 1997). Gulfwatch Project Standard Procedures: Field and Laboratory. Au- gusta, ME: The Gulf of Maine Council on the Marine Environment, State Planning Office. Suchanek, T. H.. J. B. Geller, B. R. Kresier & J. B. Mitton. 1997. Zoo- geographic distributions of the sibling species Mytiliis gallopmyiiuialis and M. trossulus (Bivalvia: Mytilidae) and their hybrids in the North Pacific. Biol. Bull. Mar. Biol. Uih. Woods Hole 193:187-194. Tracey, M. L., K. Nelson, D. Hedgecock, R. A. Shleser & M. L. Pressick. 1975. Biochemical genetics of lobsters: genetic variation and the struc- ture of American lobster (Homanis americaiuLS) populations. / Fish. Res. Bd. Can. 32:2091-2101. Townsend, D. W. 1992. Ecology of larval herring in relation to the ocean- ography of the Gulf of Maine. J. PUmkhm Res. 14:467-+93. Varvio, S.-L., R. K. Koehn & R. Vainola. 1988. Evolutionary genetics of the Myriliis ediilis complex in the North Atlantic region. Mar. Biol. 98:51-60. Watts, R. J.. M. S. Johnson & R. Black. 1990. Effects of recruitment on genetic patchiness in the urchin Echinometra malhaei in Western Aus- tralia. Mar. Biol. 105:145-151. i i I Juimml oj Shcllfiili Research. Vol. 20. No. I, 39-47, 2001. THE EFFECT OF TEMPERATURE ON THE REPRODUCTIVE MATURITY OF THE PENSHELL ATRINA MAURA (SOWERBY, 1835) (BIVALVIA: PINNIDAE) C. RODRIGUEZ-JARAMILLO,' A. N. MAEDA-MARTINEZ,' ^ * M. R. VALDEZ,' T. REYNOSO-GRANADOS.' P. MONSALVO-SFENCER,' D. PRADO-ANCONA.' F. CARDOZA-VELASCO,' M. ROBLES-MUNGARAY,' AND M. T. SICARD' ^Centra de Investigaciones Biologicas del Noroeste, S.C. (CIBNOR), P.O. Bo.x 128, La Paz. California Sur. 23000 Mexico: 'Centro de Investigacion en Alimeniacion y Desarrollo. A.C. la Victoria. Hennosillo. Son. 83000 Mexico Baja Km. 0.6 a ABSTRACT The effect of temperature on reproductive maturity of recently .spawned penshell broodstock Alriiui inaiiia (Sowerby 1835) was studied. Penshells were maintained in an open-flow system and were fed to satiation with a mixture of microalgae. Condition index (tissue wet weight/total wet weight x 100). the rate of maturation, and oocyte quality were followed histologically until spawning. A scheme of six stages, based on histological preparations, was found best to describe the maturation process in the penshell. Results indicate that penshell maturation must be at low temperatures over two months to obtain high quality oocytes in the hatchery, even though the maturation period can be shortened to half that tiine at higher temperatures. A reduction of the quality of oocytes is obtained at higher temperatures. KEY WORDS: reproductive maturity, temperature, condition mdc.\. penshell. atresia. Atriiia mauni INTRODUCTION Along the Pacific CDast of Mexico, five species of penshell bivalves occur (Keen Wl\ ). Atrina maura (Sowerby 1835) is the species with the highest valtie adductor muscle. Over the past ten years, exploitation of penshell natural stocks has led to a decline in production (Velez-Barajas & Fajardo-Leon 1996. Cardoza- Velasco & Maeda-Martinez 1997). Penshells are vulnerable to overexploitation because they have a long life, limited reproduc- tive investment, and sporadic recruitment (Butler et al. 199.3). Because of the decline in production, an increased interest in the cultivation of A. maura was stimulated (Baqueiro & Castagna 1988, Arizpe-Covarrubias 1995. Reynoso-Granados et al. 1996). Information on the biology and aquaculture of A. maura is still scarce. In this species, tolerance to temperature and salinity and the optimum temperature for growth (Leyva-Valencia 1999). as well as the gonadic cycle in Bahia Magdalena (Maeda-Martinez unpub- lished data) and the growout method in the field (Cardoza-Velasco & Maeda-Martinez 1997) was determined. Two Mexican institutes (Centro de Reproduccion de Especies Marinas in Sonora (CREMES) and al CIBNOR) produced a liinited number of spat used mainly for scientific purposes. However, there are no studies on hatchery reproductive maturation of this species. In other members of the family Pinnidae. such as Pinna nigosa. reproductive biology (Noguera & Gomez- Aguirre 1972, Coronel 1981) and recruitment (Arizpe & Felix 1986, Arizpe-Covarrubias 1987, Arizpe-Covarrubias 1995) was studied. Oogenesis and sper- matogenesis have been described at an ultrastructural level in Pinna nohilis (de Gaulejac et al. 1995a. b). as having the ultra- structural changes of maturing oocytes in Atrina pcctinata ( Yong- qiang & Xiang 1988). Endogenous and exogenous factors control both gametogenesis and spawning (Barber & Blake 1991 ). Among the exogenous fac- tors, the inost important are temperature and food (Sastry 1963, Loosanoff & Davis 1963, Sastry 1966. 1968, Giese & Pearse 1974. Lowe et al. 1982. Taylor & Capuzzo 1983, Barber & Blake 1991 ). *Corresponding author. E-mail: amaedaCs'cibnor.mx Mollusks can be caused to mature artificially out of season by increasing temperature and food (Loosanoff & Davis 1963). A threshold tetnperature is needed for vitellogenesis to continue and effects the transfer of nutrients necessary for the growth of oocytes (Sastry 1968, 1970, Sastry & Blake 1971. Blake 1972. Sastry 1979). The quality of oocytes can be assessed by the size of the oocytes at the postvitellogenic stage, and by examining oocyte morphology. The mean oocyte dimensions observed in histological sections is a reflection of the ganietogenic cycle, because oocytes gradually increase in size as they develop, reaching maximum size prior to spawning (Barber & Blake 1991 ). However maximum size can vary depending on oocyte nutrition during vitellogenesis and therefore this size has to be determined experimentally under dif- ferent exogenous conditions. Oocyte morphology is another crite- ria used to determine its quality. A common phenomenon usually observed in mollusks is oocyte degeneration (atresia) at the end of vitellogenesis because of the presence of lytic enzymes that spread throughout the acinus (Beninger & Le Pennec 1991 ). To find the optimum temperature for reproductive maturation of recently spawned A. maura. we determined both the condition index and the rate of maturation and oocyte quality at different temperatures. As a basic reference, a histological study was done to describe the phases of reproductive maturation in this species. MATERIALS AND METHODS About 100 penshell (A. maura) specimens (98 ± 0.1 min shell length) were brought to the laboratory at CIBNOR from the aqua- culture farm Rancho Bueno, located at the southernmost end of Bahia Magdalena on the Pacific coast of Baja. 137 km northwest of La Paz. These animals were produced originally at CREMES and were then sent to Rancho Bueno for growout in Nestier trays suspended from a lOO-m longline. By the time of collection on July 15. 1995. the animals were six months old and had fully developed gonads. Specimen spawning was easily induced in the laboratory by using a thermal shock. After spawning, the penshells were washed with seawater and divided randomly into three groups. They were then placed in three 1 ,500-L fiberglass tanks at 20, 25, and 30°C. vertically in plastic racks to simulate their natu- 39 40 Rodriguez-Jaramillo et al. ral position in the field. To eliminate an effect of food shortage on gonad maturation, mieroalaae was supplied continuously during the experiments. Each tank received a constant tlow of 83-L h" of seawater at 37%c salinity, containing a mixture of niicroalgae (Iso- chiysis galhana, Monocluysis liitheri. and Chaetoceros gracilis) at a concentration of 130-155 cells |a.L"'. At this flow rate, the tanks received a daily water exchange of 133<7r. Counts of niicroalgae at the outlet ports of the tanks varied between 15 and 60 cells (xL"' depending on the number of animals held in the tanks throughout the experiments. The quality of these niicroalgae species on bi- valve nutrition is well documented (Walne 1970). The water used in the experiments was drained from a 200-L tank where the cul- tured algae were mixed with flowing filtered seawater. The algae were kept well mixed in suspension by using an air bleed at the bottom of the tank. The mixture was monitored automatically with a Hach turbidimeter Model 1 720C to make certain the algal con- centration was constant. The turbidimeter operated a 1/32-hp sub- mersible pump that transferred the niicroalgae to the mixing tank. The turbidimeter was set to turn i.in the pump when turbidity dropped below a preset value equivalent to the desired algal sus- pension of a standard calibration curve. The temperature in the 20°C tank was maintained with a 1/8 hp Acrytech water chiller. The temperature in the 25°C tank was held by setting the air conditioning system of the laboratory to that temperature, and the temperature in the 30 C tank was achieved by using six 200-W thermostatically controlled immersion heaters. The tanks were cleaned every day. removing feces with a plastic siphon hose. To determine the influence of temperature on gonad matura- tion, six penshells were analyzed histologically at the start of the experiment and an equal number of animals from each tank was taken, and examined every 15 days (15. 30, 45, and 60 days). Tissues were fixed in Davidson's solution, dehydrated with an increasing concentration of ethanol (IQ^c to 100%). and embedded in Paraplast at 56°C. Thin sections (6 |jim) from the gonad area were cut with a rotary microtome and stained with hematoxylin and eosin (Howard & Smith 1983). From these sections, a sched- ule of the phases of gonad maturation in the penshell was termed. The quality of postvitellogenic oocytes at the different tempera- tures was evaluated by determining the nucleus-cytoplasm ratio and the presence or absence of atresia. The nucleus and cytoplasm of oocytes were measured with the Sigma Scan image analyzer software on images taken with a Zeiss light microscope Model 16. Only oocytes sectioned through the nucleolus were measured. Atresic oocytes were considered those in which the basophilic properties of the nucleus was lost, and the oocytes took a "jigsaw- puzzle" appearance (Barber and Blake 1991). Care was taken to record any spontaneous spawning in the tanks and also to fix tor histological analyses the tissues of any specimen at the moment of spawning. The histological sections were used to define the dif- ferent stages of gonad development in this species, taking into consideration morphology, presence of particular structures, and degree of vitellogenesis. Before cutting and embedding the tissues, a condition index was estimated by dividing the tissue wet weight of recently dissected animals by total wet weight, multiplied by 100. RESULTS Microscopic observations of gonadal sections from the differ- ent temperature treatments allowed us to determine that A. imiiini is a dioecious species with two previtellogenic (Stages 1 and VI). two vitellogenic (Stages II and HI), one postvitellogenic (Stage IV), and one spent stage (Stage V) in the female gonad. Six stages were found that clearly describe speniiatogenesis in the male go- nad. We observed no hermaphroditism. Female Gonad Maturation Stages Stage I. Early Active The acini are irregular in shape and are supported by interfol- licular connective tissue (Fig. la). The acini stem cells give rise to primary oogonia 3-5 \x.m in diameter. The nucleus is round and contains a round nucleolus. In this stage, previtellogenic oocytes are round with a maximum diameter of 25 ptm. and aie attached to the acini. Stage II. Developing Acini walls are well defined and the interfollicular connective tissue has decreased (Fig. lb). Acini acquired a circular shape with oocytes attached to the periphery. Oocytes enter the vitellogenic phase, which is evidenced by a rapid growth of the cytoplasm (lengthwise), reaching up to 60 |xni in length. The nucleus also attains an elongated shape. It has a single nucleolus (5 |xm) located centrally or marginally within the nucleoplasm. Oocytes are pe- dunculated and show dense aggregates in the stalk region, arranged in narrow rods perpendicular to the acini walls (Fig. Ic). These aggregates may correspond to the clusters of mitochondria ar- ranged in line alongside cytoplasmic microtubules described by de Gaulejac et al. (1995a) in Pimm luMlis. Some elongated cells appear in the stalk region and are in close contact with the oocytes (Fig. Id). These may play a role as auxiliary cells in the transfer of nutrients to the oocytes (de Gaulejac et al. 1995a). Stage III. Late Active Oocytes continue to grow and attain a polyhedral shape (Fig. le) ranging from 45-57 |j.m in diameter. Some oocytes appear free in the lumen and others remain attached to the acini walls. The nucleus is round and occupies a large area in the middle of the oocyte. A dense amorphous mass (probably chromatin) is fre- quently observed in the nucleoplasm. The nucleolus decreases in size to 3.3 |xm diameter, and appears surrounded by a nucleolar ring (Fig. le). From this stage on. a large and conspicuous dense aggregate (Fig. le) appears outside the nucleus, which may cor- respond to the 7-8 layered Golgi complex observed by Yongqiang and Xiang (1988) in Atrina pecrinata. Auxiliary cells are no longer observed. Stage IV. Mature This stage marks the end of the vitellogenic process. Postvitel- logenic oocytes measure a maximum of 56 p.m along the major axis, and maintain their polyhedral shape because of oocyte crowd- ing (Fig. 2a). Interfollicular and interoocyte spaces are minimal. Oocytes appear free in the lumen of the acinus. The dense aggre- gate still appears outside the nucleolar envelope. Stage V. Spawning Acini appear full of mature postvitellogenic oocytes with a broken germinal vesicle (Fig. 2 b). The dense aggregate and the nucleoplasm with the nucleolus appear spread along the ooplasm. The polyhedral shape is maintained because of oocyte crowding in the acini. Temperature Effect on Penshell Reproductive Maturity 41 ^^^^^p^;^^ ^;s^?^?^ ^^m^i^^^ m '^■/^;;\ V.:^:^% ^<^ M. -,:^ Figure I. Light niicrographsof stages of reproductive maturity in i'ema\e Alrina maiira, at 20 C. (a) K;ii l> iiilJM' (stage 11. Arrow shows an oocyte under t'ormution within the acinus, bar = 5() pni. (hi Developing (stage II). Tho acini with oocytes growing toward the lumen, bar = 50 pni. (c) Oocytes form stage II. showing clusters of mitochondria at the stalk, in line with microtubules (arrows!, bar = 25 pm. (dl Oocyte from stage II with auxiliary cells at the stalk region, bar = 25 pm. (e) Oocytes at late active stage (stage III), bar = 5(1 pm. (f) Dense aggregate (arrow) on the external wall of nuclear envelope of a stage III oocyte, bar = 25 pm. Stage VI. Spent The acini are collapsed as a result of oocyte evacuation (Fig. 2c). A few unspent oocytes with the germinal vesicle intact remain in the acini in the process of degradation. Hemocytes proliferate during this stage, contributing to gonad repair. Male Gonad Maturalum Stages Stage I. Early active Acini look elongated (Fig. 3a). The germinal epithelium pro- duces spherical spermatogonia 3-|j.m diameter after a centripetal 42 Rodriguhz-Jaramillo et al. V'. ^ i)-:s r:.T\ ■^¥^ -a. »VX'''- •^i??.^'*^^^ Figure 2. Light niicriigriiphs of stages olreproduclive nialiirit> in (vina\K Aliiiia tiiaiira. (a I Polyhedral oocytes at mature stage (stage l\ )at 20°C, bar = 5() pni. (b) Oocytes al spawning stage (stage \ ) uith brolven germinal vesicles at 20 C, bar = 50 pm. (c) Spent gonad (stage VI) with residual non-spawned oocytes at 20C bar = 50 pni. (d) Overmaturated (atresic) oocytes (stage IV) showing vacuoles probably resulting from distension of endoplasmic reticulum at 30C. bar = 25 fim. (e) (ionad with "jigsaw-puzzle" shaped overmaturated (atresic) oocytes (stage IV) at 30C. Nuclei of some oocytes have lost their basophilic properties (arrows), bar = 20 pm. evolution troni the internal wall to the lumen. Spermatogonia are located along the internal wall of the acini in bands of several cells. Some spermatocytes are present in the lumen. Stage 11. Developing Acini begin to show stratification (Fig. 3b). and all develop- mental stages are present including spermatogonia, abundant sper- matocytes in the following layer, spermatids, and scarce sperma- tozoa located toward the acinus lumen. Cell diameter decreases from 3 to 0.5 \xm. Stage III. Late Active Spermatogenesis takes place in the whole area of acini (Fig. 3c). The spermatogonia layer becomes thinner. The number of spermatozoa increase and their tails are directed toward the lumen. Temperature Effect on Penshell Reproductive Maturity 43 Figure 3. I-ii;ht niicroKraphs of stages of reproducti\c maturity in malt' Atrina maura. from 20 C treatment, (al Early active (stage I) showing spermatoogonia at the periphery of acini, bar = 50 ^m. (b) Developing (stage 11). bar = 50 pm. (c) Late active (stage llll, bar = 50 fim. (d) Mature (stage l\ ), bar = 50 uni. (e) Spawning (stage V). bar = 50 pni. (fl Spent gonad (stage \'1| partially spawned, bar = 50 pni. Stage IV. Mature Acini look full ol spennato/.oa, with their tails toward the acini lumen (Fig. 3d). Merging of some acini is evident. Stage V. Spawning Acini become elongated and the boundaries between acini are not easily distinguished (Fig. 3e). Evacuating ducts are evident from w hich a great number of spermatozoa are spent. Stage VI. Spent Numerous empty spaces toward acini lumen are evident as a result of the release of spermatozoa during spawning (Fig. 3f ). .As in the female gonad, spawning is not complete. Some acini are empty but others remain full of ripe sperm. The empty acini show a great number of hemocytes. No signs of active spermatogenesis are observed in anv region of the gonad. 44 Rodriguez-Jaramillo et al. Effect of Temperature on Maturation The results of the histological analyses at different tempera- tures and sampling dates indicate synchronous gonad maturation in male and female gonads. In addition, a faster gonad maturation rate was observed at 2.S and 30X than at 20°C (Fig. 4). At the beginning of the experiments, more than 80*7^ of the organisms showed immature gonads (stages I and VI) as a product of spawn- ing induced by thermal shock. Only 20% were in stage II. indi- cating that aametogenesis can occur rapidly just after spawning. At day 15, samples from 20°C were only at stages I and II. At 25°C, 70% of the organisms were also in stage II. but the rest had reached developing stage III. At 30°C. the majority of the organisms (70%) were in stage III and the remaining .■?0% had reached maturity (stage IV). By day 30. a small proportion of animals ( 18%) at 20°C was still in stage II. but the rest had advanced to stage III. All specimens analyzed at 25°C had advanced to stage IV. but matu- ration rate in 30°C organisms did not show any progress. After this sampling date, spontaneous spawning was recorded at 25°C on day 34 and at 30*^0 on day 41. The results from spawning were only confirmed histologically on day 45 at 25°C. where 60% showed spent gonads. The remaining 40% had not spawned, showing go- nads still in stage IV. Histological confirmation was not obtained at 30°C by day 45. In this treatment and date, relative frequencies remained the same as in days 15 and 30. with 70% and 30% in stages III and IV. Relative frequencies of penshells at 20°C on day 45 were 50% and 50% in stages II and IV. At 20°C. nearly 80% of the organisms were in stage III and the rest in stage IV on day 60. and spawning of these organisms was also recorded on day 60. ss 100 an >. o HO c 0) 40 -> o- 20 Q n 3? 100 > 80 o 60 « 40 §■ 20 £ 0 u c 0) 3 0) 100 80 60 40 20 0 n n Day 1 a III IV VI Day 15 1. Ill IV VI Day 30 IV VI Oocyte Size and Nucleus-Cytoplasm Ratio Results from postvitellogenic oocyte measurements at different temperatures are shown in Table 1 . From this, oocytes at 20"C show on average a larger area ( 1660 |xm") and major a.\is (56 |j,m) than those at 25* C (1416 \x.m- and 53 |a.m) and 30°C (1290 |j.nr and 49 jj-m). ANOVA indicates a significant difference between oocyte sizes at different temperatures at a = 0.05. A Tukey's multiple range test showed that oocyte areas at 25 and 30"C were similar at a = 0.05. but were significantly different from those at 20"C. The areas of the nucleus varied between 578 and 564 p.m" at different temperatures, but the ANOVA identified no significant differences between temperatures. With these data, we calculated a nucleus-cytoplasm ratio that was lower (0.34) at 20*'C than at 30°C (0.43). This indicates a reduction in cytoplasm area at higher maturation temperatures. Atresia A second way to demonstrate oocyte quality related to matu- ration temperature was the presence of atresic oocytes. Results from histological observations indicate that atresic oocytes were not present in animals held at 20°C but were ob.served in stage III and IV female gonads at 25°C and 30°C (Fig. 2e). In this Figure, one can see the nucleus of atresic oocytes has lost its basophilic properties, and the oocytes have a "jigsaw-puzzle" appearance. Atresic oocytes were smaller than normal oocytes. The cytoplasm of nonatresic oocytes at 25 and 30^'C showed large vacuoles (Fig. 2f) at the dense aggregate outside the nucleus. Vacuolation of this kind was reported in atresic oocytes of Pinna nobilis (de Gaulejac et al. 1995a). as a result of distension of endoplasmic reticulum. ss 100 80 >. u HO r o 40 , 3 a- 20 - £ 0 u c 3 o 100 80 60 40 20 0 Day 45 Lii_ III IV VI Day 60 Jl IV VI Gonad maturity stages Figure 4. Relative Irequentj of stages of reproductive maturity in Atrina inaiira, at different maturation times and temperatures (white columns = 20C, gray columns = 25C, black columns = 30'C|. Condition Index Condition index (CI) at 20'C increased steadily from 32% at the beginning of the experiment to 52% at day 60 when spawning occurred (Fig. 5). However, at 25°C and 30°C. only small in- creases were measured during the maturation period. At these temperatures, CI increased rapidly from 32% to 38% in the first 15 I Temperaturh Effect on Penshell Reproductive Maturity 45 TABLE 1. Dimensions of postvitelogenic oocytes and ntuiciis of pcnsht'li (Alrina mania) and the nucleus-cytoplasm ratio (N:C), at different temperatures. Oocyte Nucleus Temperature Area Major Axis Minor Axis Area Major Axis M inor Axis (°C) Ipm") Ipni) (pni) (pm") Ipml (pm) N:C Ratio 20 I66()± 30(1 56 + 6 42 ±6 578 ± 101 31 ±3 25 ± 3 0.34 25 1413 ± 1S2 5? ± 4 39 + 5 564 ± 83 3 1 ± 3 24 ±3 0.39 30 1 2yO ± 234 49 ±6 37 ± 5 564 ± 110 30 ± 3 24 ±3 0.43 Figures are the mean + standard deviation (;; 90). days, but no liullier increases were observed from day 15 to day M) at both temperatures. From days 30 to 45. minor increases of \% and 8% were recorded at 25°C and 30°C. Spawning at these tem- peratures occurred at days 34 and 41. No effect of spawning on CI was detected at these temperatures. DISCUSSION hi the literature, many schemes for classifying gonad condition in moUusks can be found (Chipperfield 1953. Lubet 1957. Sastry 1963, Wilson & Hodgkin 1967. Lunetta 1969, Seed 1969, Villa- lejo-Fuerte & Ochoa-Bae/ 1993. de Gaulejac et al. 1995a, de Gaulejac et al. 1995b, Viilalejo-Fuerte et al. 1995). We developed a specific scheme for Atrina mawa based on morphological ob- servations of gametes and specific structures. The scheme devel- oped included six stages that clearly describe the gonad maturation process in A. matira. Although our work was conducted using light microscopy, interpretation of several structures was possible using the ultrastructural descriptions of gonad maturation made by Yongqiang and Xiang (1988) and de Gaulejac et al. (1995a,b), for the pinnids Alrina pectinata and Pinna hicolor. This study re- vealed maturation of male and female gonads occurs synchronous- ly, and progressed evenly among the animals in the same treat- ment. Morphological observations allowed us to compare the oocytes of Atrina maiira with those described in A. pectinata and Pinna nohilis. Results indicate that the oocytes of Atrina species are morphologically similar, but three interesting differences exist with the oocytes of P. nohilis. The margin of the nucleus of P. nobilis is polylobed and twisted, whereas in Atrina the nuclear envelope is even and lacks indentations throughout development. The nucleolus in P. nchilis is located at the edsie of the nucleus. 20 °C 30 45 60 Time (days) Figure 5. Changes in condition index of Atrina maura at different temperatures. Arrows indicate the day of spontaneous spaHuing. whereas in Alrina it can either be at the edge or in the center. Finally, the dense aggregate (presumably a Golgi complex) found outside the nucleus of Atrina oocytes is not conspicuous in P. nchilis. Previous studies on the reproducti\e biology of bivalve mol- lusks have demonstrated that temperature and food are the inost important factors influencing the reproductive cycle of marine in- vertebrates (Loosanoff & Davis 1963. Sastry 1963. Giese & Pearse 1974. Lowe et al. 1982). We found an effect of temperature on the rate of gonad maturation and the quality of oocytes in A. maura. The effect of food on these factors was eliminated by feeding the organisms to satiation with high quality species of microalgae in all treatments. Differences in the rates of gonad maturation and quality of oocytes are therefore attributed exclusively to tempera- ture. The gonad maturation period was shorter at 25°C and 30°C than at 20°C. possibly because of the higher metabolic rates found at higher temperatures in this species. The scope for activity and shell growth rate measurements indicate the range of optimum temperature for growth in juvenile A. maura was 25-29"'C (Ley\ a- Valencia 1999. Leyva- Valencia et al. 2001). With this knowledge, and desiring the shortest conditioning period possible, we suggest that penshell gonad maturation should be done at a temperature above 25"C. However, the results from histological observations on the quality of oocytes may lead to a different conclusion. Larger and nonatresic oocytes were only found at 20°C, probably because of a longer vitellogenic phase that allowed the incorporation of larger amounts of nutritive materials. Spawning from unfed or stressed broodstock results in weak batches of larvae that fail to undergo normal pelagic development (Loosanoff & Davis 1963, Bayne et al. 1984). The correlation between both factors on the fecundity and larval viability in A. maura needs to be determined. Coincidentally. spawning of A. maura in Bahia Magdalena takes place during the coldest season of the year (from January to March) when water temperature varies between 19.5 and 20°C ( Maeda-Martinez unpublished data). We measured condition index (CI) based on tissue and total wet weights of the animals. This is not a refined method but it proved to be a good alternative for measuring CI in A. maura. One of the problems that limited the use of a more precise method was the gonad in this species is included in a visceral mass together with other tissues, and it is not possible to obtain its weight excluding other tissues. Clear differences were observed between CI from the different temperature treatments and these correlated positively with oocyte size. Therefore, CI reflected variation in gonad size assuming that fecundity was the same among treatments. This means that smaller ("immature") oocytes were spontaneously spawned at 25°C and 30''C. The lower cytoplasmic content in 46 Rodriguez-Jaramillo et al. these oocytes was confirmed with the results from the nucleus- cytoplasm ratio. The relation between oocyte size and viability of resulting penshell larvae and spat remains to be tested. ,\nother factor that may have contributed to the depletion of CI at higher temperatures (25-30°C) was the formation of atresia. CI depletion was observed after the first 15 days from the start of the experiments (Fig. 5) at the beginning of vitellogenesis when atresia normally occurs (Barber & Blake 1991 ). The acceleration of matu- ration at high temperatures (23 and 3()°C) resulted in necrotic oocytes, probably by the secretion of lytic enzymes (Beninger & Le Pennec 1991). Prolonged exposure of Argopecren irradians to subthreshold temperature, with oocytes undergoing vitellogenesis. results in vacuolization of the cytoplasm as we observed in A. maiira (Fig. 2d) and lysis of oocytes (Saslry 1966. 1968). Although oocyte degeneration is a coinmonly observed phenomenon in spe- cies whose gametogenesis is under the control of natural environ- mental conditions (Pipe 1987). further studies are needed to find the correlation between temperature and secretion of lytic en- zymes. Temperature did not affect the morphology of male go- nads. Again, the viability of sperm released from animals matured at different temperatures remains to be tested. In conclusion, results from the present work indicate that to obtain high-quality oocytes in the hatchery, penshell maturation must be done at low temperatures (20°C) over two months. Even though the maturation period could be shortened to half this time at higher temperatures, the quality of oocytes is greatly reduced. ACKNOWLEDGMENTS This wxirk was supported by Consejo Nacional de Ciencia y Tecnologi'a. Mexico, project 1775P-B. We are grateful to Dr. Reyna Castro from DICTUS for her support with the photomicro- graphic equipment. Thanks to Dr. Ellis Glazier for editing the Enalish-laniiuasze text. LITERATURE CITED Arizpe. O. & R. Felix. I9S6. Crecimiento de Pinna lugnsa (Sowerby. 1835) en la Baliiade La Paz. Mexico. /4/j. Inst. Cwn. Mar. Lininnl. Me.\. 13:167-172. Arizpe-Covarrubias. O. 1987. Reclutamiento y mortalidad de Pnnui riigosa (Sowerby, 1835) en condiciones semi-conlroladas en la Bahia de La Paz. Mexico. An. Inst. Cien. Mar. Limnoi Mex. 14:249-254. Arizpe-Covarrubias. O. 1995. Mortality, growth and secondary somatic production of the bivalve Pinna rugosa (Sowerby), in suspended and bottom culture in Bahfa de La Paz. Mexico. Aqiiacultiire Research 26:843-853. Baqueiro. E. & M. Castagna. 1988. Fishery and culture of selected bivalves in Mexico: Present, past, and future. J. Slicllfish Res. 7:433—143. Barber, B. J. & N. J. Blake. 1991. Reproductive physiology. In: S. E. Shumway, editor. Scallops: biology, ecology and aquaculture. New York: Elsevier, pp. 377^28. Bayne. B. L.. D. A. Brown, K. Bums, D. R. Dixon. A. Ivanovici. D. R. Livingstone. D. M. Lowe. M. N. More. A. R. D. Stebbing & J. Wid- dows. 1984. The effects of stress and pollution on marine animals. Weslport: Greenwood Publishing. 384 pp. Beninger. P. G. & M. Le Pennec. 1991. Functional anatomy of scallops. In: S.E. Shumway. editor. Scallops: biology, ecology and aquaculture. New York: Elsevier, pp. 1 33-223. Blake, N. J. 1972. Environmental regulation of neurosecretion and repro- ductive activity in the bay scallop. Aeqiiipecten irradians (Lamarck). PhD. Thesis. Kingston: University of Rhode Island. 161 pp. Blake, N. J. & A. N. Sastry. 1979. Neurosecretory regulation of oogenesis in the bay scallop Argopeeten irradians irradians (Lamarck). In: E. Neylor & R.G. Hartnoll. editors. Cyclic Phenomena in Marine Plants and Animals. New York: Pergamon Press, pp. 181-190. Butler, A.. N. Vicente & B. de Gaulejac. 1993. Ecology of the pterioid bi\ alve^ Pinna hicolor Gmelin and Pinna nobilis L. Marine Life 3:37- 45. Cardo/a-Velasco, F.. & ,\. N. Maeda-Martinez. 1997. An approach to aquacultural production of the penshell .Atrina mania Sowerby. 1835 (Bivalvia: Pinnidae) in northwest Mexico. J. Shellfisli Res. 16:311. Chippertleld. P. N. I. 1953. Observations on the breeding and setdement of Mytilus edulis (L.) in British waters. J. Mar. Biol. Ass. U.K. 32:449- 476. Coronel. J. S. 1981. Cicio sexual de Pinna riigosu. B. Sc. Thesis. Univer- sidad Aulonoma de Baja California Sur. Mexico. 36 pp. de Gaulejac, B.. M. Henry & N. Vicente. 1995a. An ultrastructural study of gametogenesis of the marine bivalve Pinna nahilis (Linnaeus. 1758) I. Oogenesis. J. Moll. Stud. 61:375-392. de Gaulejac, B.. M. Henry & N. Vicente. 1995b. An ultrastructural study of gametogenesis of the marine bivalve Pinna nobilis (L. 1758). II Spermatogenesis. J. Moll. Stud. 61:393-403. Giese. A. C. & J. S. Pearse. 1974. Introduction: general principles. In: A.C. Giese & J. S. Pearse, editors. Reproduction of marine invertebrates. New York: Academic Press, pp. 1^9. Howard, D. W. & C. S. Smith. 1983. Histological techniques for marine bivalve mollusks. National Oceanic and Atmospheric Administration. Technical Memorandum NMFS-F/NEC-25. U.S. Department of Com- merce. 95pp. Keen. M. A. 1971. Sea shells of Tropical West American Marine Molluscs from Baja California to Peru. Stanford. California: Stanford University Press. 1025 pp. Leyva-Valencia, I. 1999. Temiotolerancia. halotolerancia y teniperatura optima en juveniles de callo de hacha Atrina maiira (Sowerby, 1 835). B.Sc. Thesis. ENEP-Iztalaca UNAM.Mexico. 54pp. Leyva-Valencia, I.. A. N. Maeda-Martinez, M. T. Sicard. L. Roldan & M. Robles-Mungaray. 2001. Halotolerance. upper Ihermotolerance. and optimum temperature for growth of the penshell .Atrina maura (Sow- erby. 1835) (B1VALVIA:P1NN1DAE). J. Sltellfish ««. 20:49-54. Loosanoff. V. L. & H. C. Davis. 1963. Rearing of bivalve molluscs. .4(A'. Mar. Biol. 1:1-36. Lowe. D. M.. M. N. Moore & B. L. Bayne. 1982. Aspects of gametoge- nesis in the marine mussel Mxliliis ediilis L. J. Mar. Biol. Ass. U. K. 62:133-145. Lubet. P. 1957. Cycle sexuel de Mytilus edulis L. et de Mytilus gallopro- vincialis Lmk. dans le Bassin d'.^rcachon (Gironde). Aiwe biologique 33:19-29. LuneUa. I. E. 1969. Reproductive physiology of the mussel Mytilus perna. Boletin da Facultade de Filosofia, Ciencias e Letras. Universidade de Sao Paulo. Zoologia e Biologia Marhina 26:33-1 1 1. Noguera, O. M. & S. Gomez- Aguirre. 1972. Ciclo sexual de Piivia rugosa Sowerby. 1835 (Lamellibranchia: Pinnidae) de La Paz, B.C.S. In: J. Carranza. editor. Memorias del IV Congreso Nacional de Oceanogra- fia. Mexico, pp. 273-283. Pipe. R. K. 1987. Oogenesis in the marine mussel Mytilus edulis: an ul- trastructural study. Mar. Biol. 95:405^14. Reynoso-Granados. T.. A. N. Maeda-Martinez. F. Cardoza-Velasco c& P. Monsalvo-Spencer. 1996. Cultivo de hacha. In: M. Casas-Valdez & G. Ponce-Diaz, editors. 2. Estudio del Potencial Pesquero y Acuicola de Baja California Sur. La Paz: CIBNOR. pp. 545-550. Sastry. A. N. 1963. Reproduction of the bay xMop. Aec/iiipeeten irradians Lamarck. Influence of temperature on maturation and spawning. Biol. Bull. 125:146-153. Saslry, A. N. 1966. Temperature effects in reproduction of the hay scallops. Temperature Effect on Penshell Reproductive Mati'rit^' 47 Aequipecten inaJiuns Lamarck. J. Mar. Biol. Assuc. U.K. I3();1IX- 134. Sastry. A. N. I%8. Relationships among food, temperature and gonad development of the bay scallop, Aequipecten irnuliims cimcenlricus Say, reared in the laboratory. Bull. Mar. Sci. 15:417-^35. Sastry, A. N. 1970. Environmental regulation of oocyte growth in the bay ■icMop Aeqiiipeden irriuliuii.s Lamarck. Experientki 26:1371-1372. Sastry. A. N. 1979. Pelecypoda (excluding Ostreidae). In: A. C. Giese and J. S. Pearse. editors. Reproduction of marine invertebrates. New York: Academic Press, pp. 113-292. Sastry. A. N. & N. J. Blake. 1971. Regulation of gonad development ui the bay scallop, Aequipecten irraiiians Lamarck. Biol. Bull. 140:274-282. Seed, R. 1969. The ecology oi Mytilus eiiulis L. (Lamellibranchiata) on exposed rocky shores. 1. Breeding and settlement. Oecaliiiiia 3:277- 316. Taylor. R. & J. Capuzzo. I9S3. The reproductive cycle of the Bay scallop Aequipecten irnulians irraciicms (Lamarck), in a small coastal embay- nient on Cape Cod Massachusetts. Estuaries 6:431-435. Velez-Barajas, J. A. & M. C. Fajardo-Leon. 1996. Pesqueri'a de hacha. In: M, Casas-Valdez & G. Ponce-Diaz, editors. I. Estudio del Potencial Pesquero y Acuicola de Baja California Sur. La Paz: CIBNOR. pp. 101- 111. Villalejo-Fuerte. M. & R. 1. Ochoa-Baez. 1993. The reproductive cycle of the scallop Arifopecten circuluris (Sowerby, 1835) in relation to tem- perature and photo period, in Bahia Concepcion, B.C.S., Mexico. Cien- cias Marinas. Me.x. 19:181-202. Villalejo-Fuerte. M., F. Garcia-Domi'nguez & R. 1. Ochoa-Baez. 1995. Reproductive cycle of Glycyineris gigantea (Reeve. 1834)(Bivalvia: Glycymerididae) in Bahi'a Concepcion. Baja California Sur. Mexico. The VW/ger 38:126-132. Walne. P.R. 1970. Studies on the food value of nineteen genera of algae to juvenile bivalves of the genera Ostrea. Crassostrea, Mercenaria and Mytilus. Fishery Investigations 26:1-62. Wilson, B. R. & E. P. Hodgkin. 1967. A comparative account of the reproductive cycles of five species of marine mussels (Bivalvia; Mytil- idae) in the vicinity of Freemantle, W. Australia. .Australian J. Mar. Freshwater Res. 18:175-203. Yongqiang, F. & Q. Xiang. 1988. Studies on ultrastructure of oocyte in process of maturing in pen shell. Acta Oceunologica Sinica 7:459— f72. Jourmil ,>l Shellfish Research. VoL 20. No. 1, 49-.'i4, 2001. HALOTOLERANCE, UPPER THERMOTOLERANCE, AND OPTIMUM TEMPERATURE FOR GROWTH OF THE PENSHELL ATRINA MAVRA (SOWERBY, 1835) (BIVALVIA: PINNIDAE) I. LEVVA-VALENCIA,' A. N. MAEDA-MARTINEZ,' " * M T. SICARI),' L. ROLDAN,' AND M. ROBLES-MUNGARAY' ^Ceiiiro cle Investigaciones Biologkcis del Noroeste S.C. (CIBNOR). P.O. Box 128, La Paz. B.C.S.. 23000 Me.xico; 'Centra de Investigacion en Alimeutacidn y Desanollo A.C. (ClAD). Km. 0.6 Canetera a la Victoria. Hermosillo. Son.. 83000 Me.xico ABSTRACT Halotolerance. upper tlierniotoleiance. and optimum temperature for growth wa.s investigated in juvenile penshell (Airina mauni). Halotolerance and therniotolerance were determined measuring LD,o-96h within the salinity and temperature ranges of 15-50%f and 19-35°C. Optimum temperature for growth was measured estimating scope for activity, ingestion, and clearance rates at 16. 19, 25, and 29°C. Growth rates were also measured over 42 days at the same temperatures. Results indicate that salinity LD5o-96h ranged from I6%<' to more than JC/ft. and the upper temperature LD5„-96h was 33.2''C. Scope for activity, ingestion, clearance, and growth rates were highest at 29°C indicating that optimum temperature in penshell juveniles is 29 °C or higher. Because of the wide halotolerance range and the high therniotolerance and optimum temperature, the penshell is a good candidate for aquaculture in subtropical and tropical areas. KEY WORDS: .\iima muuni. penshell. salinity tolerance, temperature tolerance, temperature optimum INTRODUCTION The penshell Atriiia iiiaiira Sowerby. 18.35, is a bivalve niol- lusk thai lives in shallow coastal lagoons and estuaries along the Pacific coast from Baja California to Pei-u (Keen 1971). h has a large adductor muscle of great value in Mexican and international markets. Natural populations in Mexico have decreased because of overfishing (Reynoso-Granados et al. 1996). Because of this, there is an increasing interest to develop aquacultuial techniques for this species. Aquacultural methods are under development (Maeda- Marti'nez et al. 1996). It appears that a combination of suspension culture for the nursery stage and bottom culture for growout will be adequate for the species (Reynoso-Granados et al. 1996, Maeda-Martinez, A.N. unpublished results). Growout in bottom culture lasts two years (Cardoza-Velasco 1998), allowing then harvesting of meats of 14-26 g fresh weight (Cardoza-Velasco unpublished results) at 9-llUSD/kg in the Mexican market (Velez-Barajas and Fajardo-Leon 1996, Cardoza-Velasco 1998). The tolerance limits to salinity and temperature and the opti- mum temperature for growth have not yet been determined. For A. nuiiini. knowledge of the limits of these variables are important to optimize hatchery production and to select sites for aquaculture in the field. Tolerance limits have been determined in other mollusks (Al-Habbib & Graingner 1977, Wilson 1978, Ansell et al. 1980, Ansell et al. 1981, Poza-Boveda & Rodriguez 1987, Quinn et al. 1994, Sicard-Gonzalez 1999). Physiological methods such as scope for activity and clearance and ingestion rates are good tools to determine optimum values, as these correlate well with actual growth measurements (Sicard et al. 1999). Scope for activity has been used as a good indicator of the energy available to the organism to do other functions, apart from basic maintenance, such as growth, reproduction, movement, and regulation against variations in external factors (Fry 1947). Scope for activity is the arithmetic difference between active and stan- dard respiration rates. The former is obtained from an individual fed to satiation and the latter from an organism kept under long- term starvation. Under starvation, respiration rate gradually drops to a steady-state in several days. Scope for activity is then deter- *Corresponding author. E-mail: amaeda@cibnor.mx mined over a wide range of temperatures, and the optimum value would be where the difference between active and standard respi- ration rates is a maximum. Other physiological factors such as clearance and ingestion rates have been found in the catarina scallop to correlate with scope for activity over a wide range of temperatures (Kirby-Smith 1970, Lu & Blake 1997, Sicard et al. 1999). This needs to be confirmed because optimum values for growth could be deter- mined in the future by making only clearance rate experiments. In the present work, tolerance limits to salinity, the upper tol- erance limit of temperature, and the optimum temperature for growth of penshell juveniles have been studied. MATERIALS AND METHODS Experimental Animals Penshell (A. Hiaior;) juveniles were produced in the hatchery at CIBNOR following a technique developed by Robles-Mungaray (unpublished results). These were approximately 5-months old when tised. and measured 23 ± 10 mm shell height. To standardize the results, the relationship between shell height and dry tissue weight of 24 juveniles of the size range used in the experiments was calculated. Shell height was found to be a better characteristic than shell length, because the shell margin could be easily dam- aged during tiieasurenients. To obtain the dry weight of the tissues, the whole animals were dried in an oven at 65°C for 36 h and then the tissues removed from the valves with a dissection needle. The tissues were weighed on an electronic OHAUS microbalance, model Galaxy 1 10 with 0.1 mg resolution. Halotolerance and Upper Thermololerance Halotolerance of penshell juveniles was determined following the standard procedure of LD5„-96 described by Rand and Petro- celli ( 1985). This consists of recording survival at predetermined times up to 96 h in animals exposed to a range of a given variable, i.e. temperature, salinity, toxic substance. Before the experiments, the juveniles were maintained for 14 days in 70-L plastic tanks containing 40 L of seawater at 22°C and 38%c salinity. During this acclimation period, the tanks received constant aeration and the animals were fed 4 x 10'' cells/penshell/day of a 1:1 mixture of the 49 50 Leyva-Valencia et al. micoalgae, Isochrysis galbana and Chaetoceros gracilis. Fifty per- cent of the water was changed every day. At day 14. groups of seven juveniles, in triplicate, were trans- ferred from the acclimation tanks to 3.5-L containers with seawater at 22°C and at 13. 20. 25. 30. 38, 42. 45. and SO'/rr salinity. The water for salinity treatments below 38 %( was prepared by diluting seawater with distilled water, and those at higher salinity with sea salts added to give the desired salinity. During the experiments. 50% of the water was changed daily and the animals were fed with the same mixture of microalgae. Survival was recorded at 4. 8. 72. and 96 h. Upper thermotolerance was determined in penshell juveniles following the same method for halotoleraiice. The range investi- gated was 19°C to 35°C. Groups of seven juveniles in triplicate were transfen'ed to the 3.5-L containers with water at 38'7f(. and at 19. 22. 24. 28. 30, 32. and 35°C. The 19T treatment, which was colder than the temperature in the laboratory (22°Cl. was achieved by placing the containers in a refrigerated waterbath. and in the rest of the treatments, a 200-Watt thermostatically-controlled heater was introduced to each container. The water was constantly aer- ated to assure good mixing. Fifty percent of the water was changed daily and the animals were fed with the same mixture of microal- gae used in the previous experiments. Survival was recorded at 4. 8. 72. and 96 hours. Optimum Tcmperuliire Optimum temperature of penshell juveniles was determined by measuring the scope for activity, and the rates of clearance, inges- tion, and growth. Groups of five bivalves in triplicate were trans- ferred to 19-L cylindrical tanks containing sea water at 38'?r and ambient temperature (22°C). The water of these tanks was gradu- ally adjusted to the experimental temperatures (16. 19. 25. and 29°C) changing the temperature l°C/day. Temperatures of 16 and 19°C were produced by dipping the tanks into a 1.100-L fiberglass tank, connected to a 1-hp titanium water chiller recirculating the water. In those treatments warmer than 22°C (25 and 29^C). a 200-Watt thermostatically controlled heater was introduced into each tank. Once all treatments were at the correct temperatures, a 10-day acclimation period was given, feeding the penshells a mix- ture of 3.6 X 10^ cells/individual/day of/, galbana and C. gracilis in a 1:1 proportion. Half of the water volume was changed every day. Scope for Aclivily Scope for activity is the arithmetical difference between active and standard O, uptake rates. In this work, active respiration rates were measured by incubating, for 30 min. groups of five accli- mated animals in 30()-mL BOD bottles filled with O^-saturated seawater at the experimental temperatures and 389Jf salinity, in triplicate. The volume of the bottles was determined at each ex- perimental temperature by calculating the difference between the weight of the bottle filled with distilled water, minus the weight of the empty bottle. The seawater was previously filtered through a 0.75-|jLm GF/F membrane to reduce respiration from other organ- isms. Three bottles treated in the same manner without animals served as controls. After incubation, the oxygen content (PO,) of the bottles was accurately estimated by determining the PO, of four subsamples of water siphoned into ground-neck borosilicate glass tubes of approximately 7 niL. Subsamples were fixed for further 0-, determinations following the method of Maeda- Martinez (1985). This method uses a miniaturized version of a micro-Winkler titrator developed by Bryan et al. (1976). This ti- irator has been used by Sicard et al. ( 1999) with good results. Acti\e O, uptake rate ( AVO,) was calculated with the equation AVO, : PO,, - PO,, SDTW X 0.5 h where PO,, is the oxygen concentration in the controls. PO21 is the oxygen concentration in the incubated bottles, i DTW is the dry tissue weight of juveniles incubated, and h is the incubation time in hours. Once the active rates were determined, the animals were transferred to their original acclimation tanks, and were kept at the same conditions as before, but without the addition of food. The VO, of the organisms from all treatments was frequently mea- sured, observing a gradual decline in VO,. until a minimum stable level was reached. This level was equal to the standard respiration rate (SVO,). The scope for activity of the penshell juveniles was calculated by AVO, - SVO, at the different temperatures tested. A one-way analysis of variance (ANOVA) was done to find sig- nificant differences of AVO-, and SVO, at different temperatures. A Tukey's test was applied to find similarities among temperature treatments. Optimum temperature was where the difference was a maximum. Ingestion and Clearance Rates Ingestion rates (IR) and clearance rates (CRI were measured following the principle of the method of Winter (1973). This es- timates IR and CR from the amount of algae removed by the animals incubated at a constant algal concentration. In this way variations on feeding rates because of a decrease in food concen- tration are eliminated. The amount of algae cleared by the animals is calculated from the algae replenished into the incubating cham- ber as removal occurs. Replenished algae will be equivalent to the algae consumed. In the present work we used a system that oper- ates under this principle (Sicard et al. 1999) (Fig. 1), based on the design of Gallager and Mann (1980). Five juveniles of A. maura were incubated in a 1-L cylindrical chamber containing seawater filtered at 7 |xm at the same experimental temperatures of the scope for activity experiments. 38%r salinity, and 8x10"* cells/mL of /. galbana. The water was pumped to a flow-through cell of a Turner 1 12 tluorometer with a peristaltic pump at 87 mL/h. The tluorometer was optimized for chlorophyl a readings by using a 478 filter and 420-500 nm wavelength. The relationship between microalgae concentration (MC) (cells/(j.L) and the tluorometer output voltage (Volts) was described by the linear regression equa- tion (/?' = 0.98;)! = 10) MC = 226 X Volts - 22.3 The water from the cell was returned to the incubation cham- ber. When the algal concentration dropped, the system operated a second peristaltic pump which transferred concentrated microalgae to the chamber at 1 .0 mL/min until the concentration was reestab- lished. The algae in the chamber were kept homogeneous in sus- pension with an air bleed. The system recorded the volume of algae pumped (V) during a 2-h incubation time (t). IR (cells in- gested/g DTW/h) was then calculated with the equation IR = VxS SDTW X t Salinity and Temperature Tolerances for Penshell Growth 51 Solid state relay on/off ''AfiCjy^^yAfl?? Computer Penstaltic pump 2 Digital-to-analog converter Fluorometer Continuous flow cell Thermoregulated water batti Figure I. Apparatus employed for measuring clearance and ingestion rates in juvenile penshells {Atrina inaiira). Modified from Sicard et al. (1999). where S was the concentration of microalgae in the replenishing stock. From this. CR (Liters of water/g DTW/h) was calctilateci with the equation CR = IR X Incubation chamber volume (L) Alaal concentration in incubation chamber The experiments were run in triplicate at each experimental temperature. .An ANOVA was ijone to find significant differences of IR and CR at different temperatures. A Tulcey's test was apphed to find similarities among temperature treatments. Growth and Siinhal Growth and survival of penshell juveniles were recorded for 42 days. Groups of eight animals (18.7-21.0 mm shell height) were placed in tanks containing 19-L aerated seawater. 38'/ff salinity and at the same experimental temperatures as in the previous experi- ments. The animals were fed 1.8 ± 0.6 x 10** cells/individual/day of a mixture of/, galbana and C. gracilis at equal proportions. Half of the water was changed every day. Shell height was measured with plastic calipers every seven days over 42 days. Survival was measured daily, removing the dead animals from the tanks to avoid contamination of the water. RESULTS Dry Tissue Weight The relation between dry tissue weight (DTW) and shell height (SH) of penshell juveniles is shown in Figure 2 and is described by the equation (/■ = 0.95; /; = 24l DTW = 2.3 X 10 "xSH'' Halololerance LD^(j-96 of penshell juveniles to salinity is shown in Figure 3. Results show that the penshell withstands a range of salinity from 16.5%c to more than the highest salinity tested (50S?f ) in the present study. As 859f mortality was only obtained at 509cc salinity at 96 h. the median lethal time was calculated instead, extending the experiment until half of the population died at this salinity. Results indicate that the median lethal time for penshell juveniles was at 168 h (7 days) at this salinity (Fig. 3). TlieniiDtolerance The upper temperature LD^,|-96 of penshell juveniles is shown in Figure 4. Results indicate that A. matira tolerates high tempera- 450 400 F 350 "r 300 O) n) ■/M 5 0) ?no -) 150 *- ;=- 100 a 50 0 • •• 10 15 20 Shell height (mm) 25 30 Figure 2. Relation between shell height and dry tissue weight in ju- venile penshells {Atrina maura). 52 Leyva-Valencia et al. t i 00 - 90 - 80 - 70 - ( / ■ N ^ g Q. ^^ 1^- 60 - 50 - 16.25%. 1 40 - 30 - 20 - 10 - 0 - |168h j go 15 1 1 20 25 30 38 Salinity (%o) 42 45 50 o Figure 3. Median lethal dose at 96 h (LD511-96) of salinity, on penshell (Atrina mania) juveniles. Because LDj,, Has not obtained at % h and at the highest salinity tested (50'7fl, the median lethal time was deter- mined (broken linel. Bars are the standard deviation of the means. n = 3. tures. with the LD,||-96 of 33.2°C. Survival was not affected be- tween 19 and 30°C. and at 32°C only 10% mortality was recorded. Scope for Aclivity Active (AR) and standard (SR) respiration rates at 16. 19. 25. and 29°C are shown in Figure 5. Results indicate that AR rates varied directly with temperature from 0.65 mLO,/g/h at 16°C. to 1.15 niLO,/g/h at 29"C. However SR rates remained stable be- tween 0.2 and 0.4 mLO^/g/h and independent of temperature. SR was reached within 15 and 20 days from starvation in all treat- ments. With these results, the scope for activity (AR to SR) (Fig. 6) increased with temperature from 0.4 niLO,/g/h at 16"C to 0.9 mLOi/g/h at 29°C. The ANOVA indicated significant differences between temperature treatments at P > 0.05. However the Tukey's test revealed significant differences between all treatments but not between 16 and 19°C. No inflection of the curve was observed even at higher temperatures tested, which did not allow the deter- mination of the temperature optimum in this species. Ingestion and Clearance Rates IR and CR of penshell juveniles are shown in Figures 7a and 7b. Results indicate that both rates increased with temperature, in the same manner as the scope for activity. Mean IR increased from 2x10^ cells/g/li at 16°C. to 6.5 x 10^ cells/g/h at 29°C. CR varied from a mean of 0.6 L/g/h at 16°C. to 2.25 L/g/h at 29°C. Large variations of IR and CR were observed between replicates of the all treatments, as indicated by the large standard deviation bars in > 'Eg. to 100 -1 90. 80 - 70 - 60 - 50 - ____ \ 40 - 33.2 'C \ 30 - \ 20 - \ 10 - 0 J ' 1 1 1 ■ r I 1 — f — 19 22 25 28 30 32 35 Temperature (°C) Figure 4. I'pper median lethal dose at % h of temperature on penshell [Atrina Hifli/ra), juveniles. Bars are the standard deviation of the means. n = 3. 19 _ .25 Temperature rc) Figure 5. Active and standard respiration rates of penshell {Atrina inaura) at different temperatures. Bars are the standard deviation of the means, n = 3. Figures 7a and 7b. ANOVA indicated no significant differences of IR and CR at different temperature treatments (P > 0.05). Growth and Survival The results of growth experiments at 16. 19. 25. and 29°C are shown in Figure 8. Restilis indicate that growth in penshell juve- niles correlated with temperature. Linear shell height increments of 0.02 1 . 0.029 and 0.058 mm/day were obtained at 1 6, 1 9. and 25°C. However at 29°C, the height increment was low (0.5 mm) during the first 14 days, and then increased rapidly to 2.9 mm in 42 days. No explanation for this finding was found. Mortality was not re- corded during the 42 days of experimentation in all treatments, with the exception of a slight decrease in survival ( 159H on day 21 at 29°C. DISCUSSION The results indicate that A. nuiiini is a tropical species that withstands a large range of salinity and high upper temperatures. Because of this, A. maitra appears to be suitable for aquaculture in a wide variety of environments from hypersaline lagoons like those in Baja California, estuaries located along the Pacific coast of Mexico. Central, and South America, and probably in earthen ponds similar to those used in shrimp aquaculture. With our find- ings, temperature does not seem to impose a limit in the selection of sites, because the upper thermotolerance of penshells is greater than the maximum temperatures recorded in most bays of the Pacific coast of Mexico and of the Gulf of California (Lluch-Cota et al. 2000). In the present work, optimum temperature for growth 1.2 >^ ~> 0 iZ ro O) 1— CN !^ 0 0) Q. _l E o o CO Temperature (°C) Figure 6. Scope for activity of Juvenile penshells {Atrina maitra) at different temperatures. Bars are the standard deviation of the means. n = 3. Salinity and Ti;mpkrature Tolerances for Penshell Growth 53 0) TO 0) o c TO i_ TO _QJ O Temperature (°C) 12 -| 0 10- 8 ■ c o o 6- X (1) (« 4 ■ D) (D c O 2, n ■ Temperature (°C) Figure 7. Clearanct (al and ingestion (b) rates of penshell {Alriiui maiira) juveniles at different temperatures. Bars are the standard de- viation of the means. ;/ = 3. in penshell juveniles was not found within the range of tempera- tures studied. This could probably have been found if a higher temperature was tested. However, optimum temperature in A. maura is close to 29°C, because apart from the higher growth rate measured, reproductive maturation in the laboratory is faster at 25 and 30°C than at lower temperatures (Rodriguez-Jaramillo et al. 2001). Scope for activity and feeding rates gave identical results and coincided with growth experiments. This confirms the findings of previous work in catarina scallop {Argopecten venlricosus) (Si- card et al. 1999) where scope for activity, IR, CR, and growth gave similar results indicating an optimum temperature for growth of 19-22°C. in the range of temperatures tested (12-2S"C). There- fore, any of the methods studied here appears to be adequate for the determination of optimum temperatures in molluskan species. However precision of feeding rates is lower than scope for activity and growth rates. Our ingestion and clearance rates experiments showed large variations between replicates of the same treatment. Variations do not seem to result from technical problems, but from 26 25 24 - 23 f I" w 21 20 19 18 II (1 II i^-l —] — 14 21 28 35 1 1 42 49 1 56 1 63 1 70 Time (days) Figure 8. Growth of penshell [Atrina maura) juveniles, cultured at different temperatures. JS'Jr, and fed 1.8 ± (1.6 x 10* cells/individual/ day of a mixture n( Isochrysis galhana and Chaetoceros gracilis at equal proportions. \ alues are the mean of 8 individuals ± standard devia- tion. a discontinuous feeding behavior of the animals. Large dispersion of IR and CR data have also been reported in Argopecten irradians (Kirby-Smith 1970). Mytihis edidis (Widdows and Bayne 1971) and in A. ventricosus (Sicard et al. 1999). In contrast. Thompson and Bayne ( 1972). Bayne ( 1973). Bayne et al. ( 1973). Widdows (1973). and Sicard et al. ( 1999) have con- sidered the scope for activity as a practical way to find the envi- ronmental conditions at which the energy available to the organism for growth and for other \ ital functions is a maximum. Both active and standard respiration rates are normally dependent on tempera- ture as seen in M. edulis (Bayne 1976) and A. ventricosus (Sicard et al. 1999). However in our work, only the active rate varied with temperature. Standard respiration rate remained fully independent of temperature, which is not common. This is only found when in.sensitivity to temperature change may help to conserve meta- bolic energy reserves (Bayne 1976), as in situations of environ- mental stress (extremes of temperature, during starvation, or when exposed to air at low tide). Further work is needed to clarify these findings. In the present work, salinity and upper temperature tolerance limits of penshell juveniles have been determined. We have dem- onstrated that the optimum temperature for growth in this species is 29 C or higher, making A. maura a good candidate for aqua- culture in subtropical and tropical waters. ACKNOWLEDGMENTS The authors thank Consejo Nacional de Ciencia y Tecnologia CONACyT (Mexico) for finanical support (Project 1775P-B). and to Dr. Ellis Glazier for editing this English-language manuscript. LITERATURE CITED Al-Habhib. O. A. M. & J. N. R. Graingner. 1977. The effect of constant and changing temperatures of the thermal resistance of Lyinnuea per- es^rci (Muller). / Thenn. Biol. 2:191-195. Ansell. A. D.. P. R. O. Barnelt & A. Bodoy. 19S0. Upper temperature tolerances of some European mollusks. 2. Ddiuix vimitiis. D. semiiria- liis and D. Irunculus. Mar. Biol. 58:41-46. 54 Leyva-Valencia et al. Ansell. A. D.. P. R. O. Barnett. A. Bodoy & H. Masse. 1981. Upper tem- perature tolerances of some European mollusks Canlium glaiiciim, C. hibeniihiumi and C. edute. Mar. Biol. 65:177-183. Bayne. B. L. 1973. Aspects of the metabolism of Myiihis ediiUs during starvation. Nether. J. Sea. Res. 7:399-410. Bayne. B. L.. R. J. Thompson & J. Widdows. 1973. Some effects of tem- perature and food on the rate of oxygen consumption by Myiiliis ediilis L. In: W. Wieser, editor. Effects of temperature on ectothermic organ- isms. Berlin: Springer- Verlag. pp. 181-193. Bayne. B. L. 1976. Marine mussels: their ecology and physiology. Cam- bridge: Cambridge University Press. Bryan. J. R.. J. P. Riley, P. J. Le. B. Williams. 1976. A Wmkler procedure for making precise measurements of oxygen concentration for produc- tivity and related studies. J. Exp. Mar Biol. Ecol. 21:191-197. Cardoza-Velasco. F. 1998. Penshell aquaculture in northwestern Mexico: The effect of tidal height and stocking density on growth of .\trina maiira Sowerby. 1835. cultured in bottom plots. Bull. Malcol. Soc. Ltmd. 31:11. Fry. F. E. J. 1947. Effects of the environment on animal activity. Univer- sity of Toronto Studies. Biological Series 55:5-52. Gallager, S. M. & R. Mann. 1980. An apparatus for the measurement of grazing activity of filter feeders at constant food concentrations. Mar Biol. Lett. 1:341-349. Keen, A. M. 1971. Sea shells of tropical west America. Marine mollusks from Baja California to Peru. 2nd. Ed. Stanford: Stanford Univ. Press, Kirby-Smith. W. W. 1970. Growth of the scallops .Argopecten irradians concentricus (Say) and Argopecten gibbiis (Linne), as influenced by food and temperature. PhD. thesis. Durham: Duke University. 126 pp. Lu. Y. T. & N. J. Blake. 1997. Clearance and ingestion rates of Isochiysis galbana by larval and juvenile bay scallops. .4rgo;)e(7c/i irradiam con- centricus (Say). J. Shellfish Res. 16:47-54. Lluch-Cota, S. E.. C. A. Pacheco-Ayub. J. J. Bautista-Romero. S. Hernan- dez-Vazquez & D. B. Lluch-Cota. 2000. Coleccion de informacion ambiental para el Pacffico Mexicano. Maeda-Martinez, A. N. 1985. Studies on the physiology of shell formation in molluscan larvae with especial reference to Crepidula fornicata Ph.D. thesis. U.K. University of Southampton. 155 pp. Maeda-Martinez. A. N.. P. Monsalvo-Spencer & T. Reynoso-Granados 1996. Acondicionamiento. induccion al desove y obtencibn de embri- ones de callo de hacha. iAtrina inaiira). Oceanologia (Me.\.) 4:167- 170. Poza-Boveda. J. V. & R. J, Rodriguez. 1987. Survival of mangrove oyster Crassostreii rhizopliorae (Gilding. 1928) in respect to temperature. salinity and pH variations. Mcni. Soc. Cienc. Nat. La Salle. 47:127- 128. Quinn, J. M., G. L. Stelle, C. W. Hickey & M. L. Vickers. 1994. Upper thermal tolerances of twelve New Zealand stream inxertebrate species. N. Z. Mar Freshwat. Res. 28:391-397. Rand, G. M. & S. R. Petrocelli. 1985. Fundaments of aquatic toxicology: methods and applications.. In: A. E. Greenberg. S. Lenore. Clesceri & A. D. Eaton, editors. Standard Methods 1992. 18th ed.. Section 8-2. Washington: Hemisphere Pub. Corp. Reynoso-Granados, T., A. N. Maeda-Martinez, F. Cardoza-Velasco & P. Monsalvo-Spencer. 1996. Cultivo de hacha.. In: M. Casas-Valdes & G. Ponce-Diaz, editors. Estudio del Potencial Pesquero y Acuicola de Baja California Sur. II. La Paz: Centro de Investigaciones Biologicas del Noroeste Press, pp. 545-546. Rodriguez-Jaramillo. C, A. N. Maeda-Martinez. M. E. Valdez, T. Rey- noso-Granados, P. Monsalvo-Spencer. D. Prado-Ancona, F. Cardoza- Velasco, M. Robles-Mungaray & M. T. Sicard. 2001. The effect of temperature on the reproductive maturity of the penshell .Atrina maiira (Sowerby. 1835) (B1VALV1A:PINNIDAE). / Shellfish Res. 20( 1 ):39- 47. Sicard-Gonzalez. M. T. 1999. Temperalura letal superior y temperatura optima en una poblacion de almeja catarina ^Argopecten ventricosus. Sowerby II, 1842). M.Sc. thesis. CICIMAR, IPN. Mexico. 96 pp. Sicard, M. T., A. N. Maeda-Martinez, P. Orniart, T. Reynoso-Granados & L. Carvalho. 1999. Optimum temperature for growth in the catarina sc^Wop {Argopecten ventricosus-circularis. Sowerby U. 1842). 7. Shell- fish «e,s-. 18:385-392. Thompson. R. J. & B. L. Bayne. 1972. Active metabolism associated with feeding in the mussel Mytilus ediilis L. J. E.xp. Mar. Biol. Ecol. 8:191- 212. Velez-Barajas. J. A. & M. C. Fajardo-Leon, 1996, Pesqueria de hacha. In: M. Casas-Valdes & G. Ponce-Diaz, editors, Estudio del Potencial Pesquero y Acuicola de Baja California Sur. I. La Paz: Centro de Investigaciones Biologicas del Noroeste Press, pp. 101-1 1 1. Widdows. J. 1973. The effects of temperature on the metabolism and activity of Mytilus edulis. Neth. J. Sea Res. 7:387-398. Widdows. J. & B. L. Bayne. 1971. Temperature acclimation of Mytilus edulis with reference to its energy budget. / Mar. Biol. Assoc. U.K. 51:827-843. Wilson, J, G. 1978. Upper temperature tolerances of Tellina tenuis and T. fahula. Mar Biol. 45:123-128. Winter. J. E. 1973. The filtration rate ot Mytilus edulis and its dependence on algal concentrations, measured by a continuous automatic apparatus. Mar Biol. 22:317-328. Jdunnil 11] Shellfish Research. Vol. 20. No. 1. 53-61, 21)01. AN IN VITRO ASSAY TO DETECT PARALYTIC SHELLFISH POISON JOSE LUIS CORDOVA,' * ADOLFO JAMETTr JUAN AGUAYO.^ MARIA TERESA FAURE,' ORIALIS VILLARROEL/ AND LEONIDAS CARDENAS' ' Fiinckicidii Ciencia para la Vida. and Millcuniiim Institute for Fiiiuhinwntal and Applied Biology. Av. Marathon 1943. Nimoa. Santiago. Chile: 'BiosChile I.G.S.A.. Av. Marathon 1943. Nitiloa. Santiago, Chile: ' Serxicio de Salud de Magullanes. Jose Menendez 589. Punta Arenas. Chile: *lxiboratorio del Amhiente. Institiito de Salad Piibliea de Chile. Av Marathon 1000, Santiago. Chile ABSTRACT We have developed an in viiro assay named Paralytic Shellfish Poison-lest (PSP-test) that is capable ot detecting low quantities of saxitoxin (STX), the first toxin studied from PSP. The PSP-test is an immunoassay based on the development of specific monoclonal antibodies (MAbs) capable of recognizing soluble saxitoxin (STX) and able to agglutinate in minutes, a latex containing STX adsorbed to its surface. The PSP-test is a two-step indirect immunoassay. During the first step, the MAbs are incubated with an extract suspected of containing STX. If the solution contains STX. this molecule w ill occupy the MAbs binding sites and block the latex agglutination. In the second step. MAbs-ST.X are incubated with an STX-latex. An inhibition indicates the presence of STX (PSP) in the sample. The PSP-test can detect within the range of 1.2.'i-2.3 p.g/ml of STX standard. It is also capable of recognizing STX in whole shellfish extract, acidic shellfish extract, and supernatants from aged toxic cultures of bacteria or dinotlagellates. The assay is simple, economic, and quick. Training of in its use is fast and harmless to the user. It can be performed in the field as a first screening procedure because an instrument to interpret the results is not required. However, the PSP-test was not developed to replace the mouse bioassay but rather to he complementary, and the strategy used for its development can be easily adapted to other toxins. KEY WORDS: Paralytic shellfish poison, saxitoxin. monoclonal antibodies, dinofiagellates. assay INTRODUCTION The world oceans are affected by temporary massive prolifera- tion of microalgae called Harmful Algal Blooms (HAB). com- monly known a.s "Red Tides." Among the toxins produced by these HABs, the most feared is the Paralytic Shellfish Poison (PSP). a neurotoxic toxin produced by dinoflagellates like Alex- andriiim. Gymnodinium and Pyrodinium. There have been reports that some bacteria like Anabaena. Aphanizomenos. Moraxella and PseiidoiiiDiuis are capable of producing PSP (Rausch and Lassu 1991). Monitoring PSP toxicity potential in shellfish is carried out using the mouse bioassay. the only international assay accepted for certification of shellfish safety (AOAC 1990). The mouse bioassay has a detection limit of about 40 |j.g of ST.X for 100 g of total shellfish tissue. However, reports indicate that this assay has high variability (Park et al. 1986) and the maximum concentration of STX accepted for shellfish commercialization (80 (a,g/100 g) is too close to the limit of the mouse detection assay. This situation has stimulated the de\ elopment of new methods for detection of toxin concentrations lower than those detected by the mouse bioassay. Among the new assays is high pressure liquid chromatography (HPLC) which is one of the most versatile and sensitive (Sullivan et al. 1983.. Oshima et al. 1984). The fact that STX is able to block the activity of the voltage-dependent sodium channel and displace radioactive STX bound to brain membranes by cold STX present in the shellfish extract, allows quantification of the STX (Davio and Fontelo 1984). Other methods also include the cytopathic effect of STX on neuroblastoma cells cultures (Kogure et al. 1988) and the use of monoclonal and polyclonal antibodies in several immunoassays formats (Chu and Fan 1985; Carlson et al. 1984; Davio. et al. 1985 and Usleber et al. 1991 ). The most accurate and reliable assay to detect PSP is liquid chromatography followed by mass spectrometric analysis. LC-MS (Quilliam et al. 1993). The 'Corresponding author. E-mail: jcordova&bionova.cl; Fax: -1-56-2-239-425(1. main advantage of the assays listed above is. that they allow the evaluation of several shellfish samples simultaneously within a relatively short period of time when compared to the mouse bio- assay. However, these assays can be performed only under labo- ratory conditions requiring expensive infrastructure and most of them cannot be easily adapted to field conditions as required to peiform daily monitoring. Furthermore, the mouse bioassay is a time consuming and an expensive procedure. The main reason to maintain the mouse bioassay is that it detects the whole shellfish toxicity regardless of which toxin or toxin-complexes are present in the shellfish sample. On the other hand, animal rights defen- dants groups are encouraging and pressing the scientific commu- nity to look for alternative assays that will allow the elimination of the mouse bioassay or at least reduce the number of animals used. Here, we describe an indirect immunoassay that detects low quantities of STX from a single shellfish sample. PSP-test is a qualitative assay, economical, easily adaptable to other toxins or toxic profile, and does not require instruments to interpret the results. Thus, the PSP-test can be adapted to field conditions al- lowing to rapidly alert shellfish farmers of the presence of con- taminated shellfish and can be used as primary screening of samples before analysis of mouse bioassay. MATERIALS AND METHODS Generation of Monoclonal Antibodies (MAb) STX was purchased from Sigma Chemical Company (St, Louis, MO). It was bound to hemocyanin ("blue carrier" protein), Biosonda (Santiago. Chile) to be used as immunogen and to bovine serum albumin (BSA) to be used as antigen for screening as de- scribed elsewhere by Levine et al. (1988) ). Briefly. STX's hy- droxyl group was modified using CMA (carboxymethylamine). It was then, incubated with a water-soluble carbodiimide EDC. fol- lowed by incubation for 1 h at room temperature with blue carrier 55 56 Cordova et al. protein. The ration STX-Protein was 100:1. Free STX was sepa- rated from the STX-protein conjugate by centrifugation through a niicroconcentrator Centricon® 30 tube (Amicon. Danvers, MA). Similarly- -STX was adsorbed to 1 |j,m latex following manufac- tures indications. Bangs Labs. (Fishers. IN). To immunize the animals, 50 iJig of STX-bhie canier conjugate was mixed with Complete Freund Adjuvant. Gibco (Grand Island. NY) just before immunization. Female lO-week-old BALB/c mice were immu- nized subcutaneously. After the first immunization, four additional boosts were done at 20-day intervals. Five days after the last boost, serum samples were drawn from each mouse and their titers were determined by ELISA using polystyrene plates DK 4000, Nunc (Roskilde, Denmark) coated with the STX-BSA as antigen as de- scribed elsewhere by Chu ( 1985). The mouse with the highest titer for STX-BSA was used to generate the monoclonal antibodies. After immunization, spleen cells from the selected mouse were fused with 2.5 x 10^ NSO/2 mouse myeloma cells using 50% polyethylene glycol 4000 Merck (Rahway, NJ) as described by Kohler and Milstein (1975). Then, the cells were suspended in Dulbecco supplemented with 1% nonessential amino acids. 10% fetal bovine serum (FBS), HyClone Labs. (Logan, UT), penicillin (100 lU/ml), streptomycin (100 p-g/ml), 2 mM L-glutamine. 10"' M hypoxanthine. 4 x 10"^ M aminopterin, and 1.6 |ji 10"' M thy- midine, HAT selecting medium, Sigma, (St. Louis, MO). Aliquots of 100 |xl were plated on 96-well tissue culture plates. The plates were incubated at 37°C in a moisturized atmosphere with 10% CO, in the air. ELISA, as described previously, screened 10 days after fusion, supernatants from cultures with hybridoma growth for specific monoclonal antibody secretion. All positive cultures were expanded onto 24-well plates, and cell samples were frozen in liquid nitrogen in a medium containing 90% FBS and 10% di- inethylsulfoxide. Positive clones were re-cloned by limited dilu- tion. Three hybridomas (8C3/A7, 6A4/H7 and 3F7) were selected for the test. Shellfish Sample Extracts Two types of shellfish samples were utilized for evaluation using the PSP-test. One, freshly frozen untreated shellfish from Aysen, XI region of Chile (43°45' to 48°45', Fig. 2 and Table 2). These same shellfish samples were subjected to acid extraction in performing the mouse bioassay (AOAC 1990). The other type was acidic shellfish extracts previously used to perform the mouse bioassay, obtained from different Chilean geographic areas af- fected by PS P. Dimiflagellate Cultures The OF875-I7 clone of /I. tanutrense from Ofunato, Japan and clone GTCA28 of A. fimdyense from New Hampshire, USA were Assay Components MAb anti-STX STX STX-Latex Extract (-) Extract(+) Q □ -jO. A.- Positive agglutination means "Absence STX" Phase I (90 seconds) + Phase II (90 seconds) Positive agglutination (ISO seconds) B.- Negative Agglutination means "Presence of STX" Phase I (90 seconds) IIWWilBwf + — Phase II (90 seconds) -c Negative agglutination (180 seconds) Figure 1. PSP-test assay detection principle. The latex agglutination indicates the absence of STX in the sample (A), while the inhibition indicates STX in the sample. A.SSAY TO DETECT PaR.ALYTIC SHELLFISH POI.SON 57 TABLE 1. Assay detection limit and recognition of STX-isomers. A Toxin C oncentrations Saxitoxin standard (|x Agglutination i/ml) 10 5 No No 2.5 No 1 .25 0.625 Yes Yes B Toxin Concentration Detection limit Neosaxitoxin GTX,„ - 1 GTX,„ - 1 GTX, or B 1 * 28 p,g/ampule 14.8/6.4 |jLg/ampule 24.8/5.8 |xg/ampule -100 (jLg/ampule 25 (j.g/1 mL 39 (ig/1 mL 36 [xg/] mL 200 |jLg/l mL Oligoclonal anti-STX antibodies cross-reactivity was assessed using com- mercially available STX and its isoforms. The lower STX concentration that induces inhibition of the agglutination was 1.7 |j,g/ml (A) and for the isoforms are as indicated (B). * Indicates that this standard is not lOO'r pure. kindly donated by Dr. Donald M. Anderson. Biology Dept.. Woods Hole Oceanographic Institution. MA. U.SA. The ACC07 clone oi' A. cateiwlla was isolated on April 1994 from Costa Chan- nel, XI region. Chile (45°32'S, 73°34'W). and was kindly donated by MSc. Miriam Seguel IFOP. Puerto Monlt. Chile. Dinotlagei- late cultures were maintained in glass tubes at 10^'C with constant light until they reached stationary phase. Culture media consisted of sea water sterilized by passage through 0.2 |j,in filtration units Nalgene (Rochester. NY), supplemented with Guillard f/2 marine enrichment basal salts and vitamins (Sigma), as described by Guil- lard. (197.5). Bacterial Culliires PseiidoiiKiiias diminuta was isolated from the ACC()7 clone of A. catenella. The Momxella sp. strain was isolated froin A. uiina- rense and kindly provided by Dr. Masaki Kodama, School of Fisheries Sciences. Kitasato LIniversity Sanriku. Iwate, Japan. Free-living bacteria identified as Pseudonumas vcrsiciilari.s and Proteus vulgaris were isolated from the Magellan Strait. Chile. All bacterial strains were grown in Zobell marine media and kept at 10°C (Kodama et al. 1988). PSP-Test Assay Procedures Determination of the .\ssay Detection Limit Ten microliters of a 1;30 dilution of the oligoclonal anti-STX (a mixture of three MAbs that recognize STX) was mixed with 10 (j.1 of? a serial dilution of STX standard in distilled water (Sigma) and placed over a spot in the agglutination plate, mixing for 90 seconds at room temperature. Then. 10 |jl1 of STX-latex was added and mixed by gentle swirling of the agglutination plate until aggluti- nation was visible in the control (without free STX). Also, the anti-STX cross-reactivity of the oligoclonal antibody mixture to STX-isoforms, neosaxitoxina (Neo-1), gonyautoxins-1 (GTXl/4- 1), gonyautoxins-2 (GTX 2/3-1). and gonyautoxin-5 acetate (GTX-5 or Bl). all obtained from the Certified Reference Mate- rials Program. Institute for Marine Biosciences. National Research Council of Canada, were evaluated. Evaliialitiii of the Presence of STX in Freshly Frozen Shellfish Samples were allowed to thaw at room temperature. Approxi- mately 0.1 g of digestive individual gland was mixed in an Ep- pendorf tube with 100 (xl of distilled water and minced with a pestle until the suspension was homogeneous. The tubes were then centrifuged at 14.000 rpm for 3 min (Eppendorf. Nathalar. Ger- many) and 10 |jl1 of the supernatant was placed on the agglutina- tion slide and evaluated, as described previously. Agglutination occurs in the control spot about five minutes after the initiated reaction. The reaction for each sample was evaluated and scored as positive or negative for agglutination. Evaluation of Whole Acidic Shellfish Extracts Extracts from toxic and non toxic shellfish were obtained ac- cording to the AOAC procedure (AOAC 1990). Ten microliters of each sample was placed on the agglutination slide without neu- tralization and evaluated, as described previously. Evaluation of Supernatants from Toxic Organisms One milliliter of culture from stationary phase (10-12 days) either bacteria or dinofTagellate. was individually placed in an Eppendorf tube, after a quick spin to pellet the cells. 10 (xl of supernatant was placed on the agglutination slide and evaluated, as described previously. RESULTS Assay Rationale A mixture of MAbs (oligoclonal) generated against STX rec- ognizes STX present in the latex surface (STX-latex) and agglu- TABLE 2. Evaluation of freshly frozen shellfish toxlcit) from Aysen, XI region of Chile by the PSP-test and the mouse bioassav. Shellfish Mouse Bloassa.> Sample Death time or PSP- # Place Species Result concentration test 1470 Melimoyu chorito N/D + 1473 Melimoyu almeja N/D + 1475 Seno Gala chorito + r:08" r:26" Sve. + 1479 S. Gala almeja + 31 (xg/lOOg + 1480 Seno Miller culengue + 0':59" r:07" Sve. + 1483 S.Miller cholga + r:04" l':09" r:08" + 1484 S. Miller chorito + r:35" r:51" ':02" + 1485 S. Miller almeja + 42 jjLg/100 g + 1486 Isla. Toto chorito + 2':16"2':00" r:4.5" - 1489 I. Toto culeniiuc + r:25" r:35" r:2.5" + 1490 I. Toto almeja + 55 jjig/lOO g -1- 1491 Isla Manuel chorito + 2':24" 2':07" Sve. + 1492 1. Manuel cholga + l':50" r:54" 2':09" - 1493 1. Manuel almeja + 59 |xg/100 g + 1496 Isla Gamal almeja N/D + 1498 S. Magdalena almeja + 41 (ig/lOOg + 1499 S. Magdalena cholga + T 07" 2': 17" r:55" - 1502 Pta. Calqueman culenaue + 1' 38" l':37" l':38" - 1504 Pta. Calqueman almeja + 3' 11"3':25" 2':50" + 1 505 Pta. Calqueman chorito -t- r 08" l':15" r:3l" - If two out of three mice died, the sample was considered positive. N/D: Non-detectable or negative using mouse bioassay. Sve: Survival. 58 Cordova et al. > 4200 km REGION DET.3 (Cip IQUIC ^^ACA XI REGION DE ANTOFAGASTA ■(.tip AHTUTAGASTA) REGION DE ATACARU (Cip COPUPO) XII REGION DE COOUIMBO (Cap LA SEMSa;! REGION DE VALPARAISO (Cap VALPARAISO) REGION METROPOLIT,\NA (Cip SANTIAGO) REGION DEL MAIILE (C»p TXLCA) REGION DELA AR.A.13CANL^ (C»pT-EMUC01 REGION DE AISEN (C»p COIHAIQUE) REGION DE MAGALL.'VNES Y ANTARTICA CHILENA tC4p, PUNT A ARENAS) Figure 2. Geographical origin of Chilean shellflsh samples used in this paper (Regions I, II. XI, and XII). tinates it. The MAbs should not agglutinate the STX-latex it' their binding sites are oeeupied by STX present in the sample (Fig. 1 ). Thus, if a sample, that does not contain ST.X. is mi.xed with the anti-STX MAbs, their binding sites will remain free and able to agglutinate the STX-latex. This positive agglutination indicates the absence of STX (PSP) in the sample. However, if the sample does contains STX. then the MAbs binding sites will be occupied and would not be able to agglutinate the STX-latex. Thus, the lack of agglutination indicates the presence of STX (PSP) in the sample. Assay Deteclioii Limit and Recognition of ST\-lsomers The detection limit was established as the minimal concentra- tion of standard STX that inhibits agglutination after five minutes of reaction at room temperature. For STX this limit was in the range of 1.2,'i-2..'^ (j.g/ml (Table lA). However, additional experi- ments with different STX sources indicate that this minimal in- hibitory concentration may vary according to the sample condition. i.e. acid shellfish extract, culture, fresh, or canned (data not shown). The PSP-test was evaluated to determine whether the MAbs recognize STX-isoforms. The MAbs anti-STX recogni/e different STX-isoforms evaluated as indicated in Table IB. The STX- isoforms concentration that inhibits the agglutination is greater, as expected, compared to that of STX from monoclonal antibodies raised and screened for that purpose. Analysis of Fresh Frozen Samples Once the PSP-test assay conditions were established, we were interested in determining whether the assay can detect PSP in individual non-chemical extracts from freshly frozen samples. The results are presented in Table 2. Also, the same samples were used to detect PSP using the mouse bioassay with ?i mice of the F-l strain. Death times and PSP concentration calculations are in- cluded in Table 2. Table 2 shows the results of the analysis done on 20 samples of 4 different shellfish species obtained from eight different places of XI region (Fig. 2). These frozen samples were all from one indi- vidual organism and stored under non-standardized conditions to ASSA'i' TO DETECT PARALYTIC SHELLFISH POISON 59 reflect field conditions more closely. Results obtained with the mouse bioassay and with the PSP-test are compared in Table 2. The analysis shows that the mouse bioassay detected 17 (85%) positives out of 20 samples, and when the same shellfish samples were evaluated, the PSP-test detected 15 positives out of 20 (75%). Analysis of Whole Acidic Shelljish Extracts Whole acidic shellfish extracts from four different Chilean re- gions (Fig. 2) including eight different shellfish species from dif- ferent years, were evaluated for PSP with the mouse bioassay. the data obtained are presented in Table 3. The same samples were evaluated with the PSP-test. giving a 100% of positive correlation with the mouse bioassay. Twenty-five samples of whole acidic shellfish extracts from non toxic shellfish (as determined by the mouse bioassay) gave negative results with the PSP-test assay (data not shown). Analysis of Dinoflagcllalcs and Bacterial Cnllnres To determine whether the PSP-test assay could be used lo detect toxic dinoflagellates reported as primary sources of PSP, we carried out the PSP-test in culture supernatants as well as in their supematants obtained after cell homogenization. The results are shown in Table 4A. Thev indicate that the PSP-test detects PSP in both samples, regardless of which Alcxandriimi species we tested or where they came from. In addition, the PSP test was shown as being capable of de- lecting toxins on culture supernatants from P. diminiita. a bacte- rium isolated from a Chilean strain of ,4. cateiiella. (Cordova el al.. 2000). As positive control, we used a supernatant from cultures of Moraxella sp. a bacterium isolated from A. tamarense (Kodama et al. 1988). The confirmation that the bacteria isolated produced STX was done by HPLC (data not shown). Finally, several free- living marine bacteria isolated from XII region. Chile, were also evaluated for PSP production using the PSP-test. The results in- dicate that only two i.solates, P. vcrsicukiris and P. vulgaris, were capable of producing PSP. DISCUSSION The mouse bioassay is used to monitor shellfish toxicity for legislative purposes (open or closing areas for shellfish harvest- ing), as well as to evaluate shellfish wholeness, fresh or canned, before commercialization and human consumption. Thus, evalua- tion of shellfish toxicity has been the best alternative to mitigate the negative impact of PSP poi.son in the community. Despite the effectiveness over the years of the mouse bioassay in detecting shellfish toxicity, there are problems with accuracy TABLE 3. Toxicity detection in acid extracts of shellfish species harvested from different geographical area.s. Samples evaluated by the mouse bioassay and the PSP-lest as indicated under Materials and Methods. Geographic location ShellHsh region Sample # Species Place Mouse (Mg/1()0 gl PSP-test I and 11 03-12-1997 524 ostion Caleta Las Verdes de Iquique -1-65 + 05-02-1998 118 locos Literal Iquique -1-37* + 09-02-1998 123 locos Punta Atala. Tocopilla +49 + 15-05-1998 293 machas Sector Plava Machas +34* + 294 ostion Sector Lagunilla -^35* + XI 13-05-1998 1817 cholga Seno Gala +402 + 1821 culengue Seno Miller +2875 + 1831 culengue Isla Toto +3875 + 1845 chorito Caleta Puvuhuapi +471 + 1847 picoroco Islas Orestes +157 + i 850 almeia Islas Orestes +316 + 1867 picoroco Islas Huichas +108 + 1870 cholga Islas Huichas +2297 + 1879 cholga Puerto Amparo +651 + 1881 chorito Puerto .Amparo +445 + XII 02-03-1994 18 chorito Bahi'a Porvenir. Tierra del Fuego + 1224 + 04-04-1994 601 cholga Isla Isabel. Magallanes +443 + 655 chorito Bahia Buena, Magallanes +6516 + 669 chorito Cabo San Isidro, Magallanes +3528 + 30-04-1994 878 cholga Estero Nufiez, Magallanes +601 + 07-07-1994 1.^04 cholga Bahia Afio Nuevo, U. Esperanza +145 + 01-12-1995 996 almeia Isla Magadalenu, Magdalena +137 + 19-04-1996 452 cholga Bahfa Stokes +270 + 09-01-1997 17 chorito Los Nires, Magallanes + 1021 + 19-01-1998 115 chorito Cabo San Isidro. Magallanes +2011 + 117 chorito Cabo San Isidro. Magallanes +2041 + * Indicates that these values are slightly lower than the lower limit of the mouse bioassay test (Fernandez and Cembella 1995) 60 Cordova et al. TABLE 4. Evaluation of toxic cultures using the PSP-test. PSP-Test Sup. Cell A Dinoflagellates species Medium Homo. Alexciiulriiim catenella clone ACC()7. Chile + + Alexandrium fundyense clone GTCA28 .USA + + Alexandnwn taniarcnse clone F875-17. Japan + + PSP-test B Bacteria Species Culture 1 Pseudomonas diminuta Intracellular + 2 Moraxetia sp.-like Intracellular + 3 Flavobacleriuin breve Intracellular - 4 Aeromona salmonicida Intracellular - 5 Pseudomonas versicidares Intracellular - 6 Pasleurella haemolyticus Intracellular - 1 Moraxella sp.-like* Intracellular + 8 Pseudomonas versicularis Free Living + 9 Proteus vulgaris Free Living + 10 Escherichia coli Free Living - ! 1 Pseudomonas mallei Free Living - 1 2 Vibrio parahaemolyticus Free Living - Bacterial and dinotlagellate cell cultures were evaluated detecting toxic dinoflagellate in their culture supernatant from stationary phase, as well as in their supernatants from cell homogenization (A). Isolated bacteria from toxic dinoflagellate ( 1 and 2) as well as free living bacteria from the XII region of Chile (8 and 9) were detected positive for PSP production (Bl. Other intracellular as well as extracellular bacteria tested negative for PSP production. * PSP positive Moraxella sp. from Dr. Kodaina. and poor reproducibility when shellfish is contaminated with low concentrations of PSP. This indicates that there is a need for quick, sensitive, and economic assay, complementary to the mouse bio- assay. The PSP-test could also be used for fast screening of to.xic shellfish, allowing better management, hopefully, just before shell- fish harvesting or to monitor toxic shellfish that undergo detoxi- fication procedures, or shellfish acidic extracts before mouse bio- assay, thereby reducing the number of mice used. The PSP-test assay reported here has proven to be effective in detecting PSP in freshly frozen shellfish species, 15/20 (75%) when compared with positive samples in the mouse bioassay 17/ 20, (85%). The negative PSP-test results could indicate the pres- ence of new STX isomers in the sample, different from those evaluated in Table IB, which are not recognized by the MAbs anti-STX. However, this hypothesis, although possible, is unlikely because the death-times of the mice suggest high toxin concentra- tion. Another explanation for this result could be that the shellfish among those pooled were diverse and possibly, not all contami- nated, thereby resulting in different amounts of toxins in an indi- vidual assay. The mouse bioassay uses extracts from 100 g of shellfish; approximately 12 to 15 types of shellfish, increasing the chances of detecting toxicity. However. PSP-test uses an indi- vidual sample and a low amount of tissue. Finally, the possibility exists, although minimal, that some deaths might be due to a toxin that resembles PSP and consequently there would be no detection with PSP-lests. Unfortunately, samples to perform such, as HPLC or liquid chromatography-mass spectrometry, were not considered because we have no access to these technologies. When the PSP-test was used in evaluating positive acidic shell- fish extracts that were positive to the mouse bioassay. the results indicated 100% positive coiTelation. Therefore, based on this re- sult, the PSP-test detects PSP under acidic conditions, suggesting that there is a potential for it to be used as a primary screening assay for these types of samples. Perhaps the number of mice used to evaluate PSP in shellfish acid extracts could be reduced. We believe that 100% correlation was obtained because toxins are concentrated from 100 g of shellfish. Regardless of what type of STX isoforms were present in the samples, under this acidic con- dition, they tend to transform to STX, isoform that is detected by PSP-test. The versatility of the PSP-test was also demonstiated when toxic cultures of dinoflagellates and bacteria were detected. However, a significantly higher number of samples will be neces- sary to assess the full potential of the PSP-test to detect PSP from different sources. It is important to note that the PSP-test assay was not developed with the intent to replace the mouse bioassay. because it is inca- pable of recognizing "whole shellfish toxicity" as detected by the mouse (Table 2). which can be due to the highest standardization and reproducibility of sample preparation. However, the PSP-test has proven useful for detection of lower concentrations of PSP in fresh frozen shellfish, which indicates its potential use in testing field samples and for different shellfish species. The most impor- tant Chilean shellfish for local human consumption and for export were evaluated using the PSP-test. The following: "almeja del none". Priilhuca thaca; "almeja del sur", Venus antiqua: "cho- rito". Mytilus chilensis; "cholga". Aulacomya alter, "culengue", Gari solida: "loco", Concholepas concholepas: "macha". Me- soileswa doiuishwr. "ostion del norte", Chlamys (Argopeclen) piir- piiratus: "picoroco", Megabalamis psittaciis: indicates 9 out of these 19 species (47%) used. Moreover, the PSP-test can be im- proved by developing new MAbs characteristics to specific STX isomers present in the PSP. adapting it to different conditions and toxin profiles. Finally, the PSP-test uses a new technology that can be easily adapted for other shellfish toxins, and becau.se it is eco- notnical and quick has the potential for uncomplicated shellfish monitorinl phosphate (Sigma Chemical Co.. St. Louis. MO). 200 mM Tris/HCl. 20 mM EDTA. 20 mM DTT. and 2 mM MgCK (pH 8.31 ) was mixed with enzyme solution, and the initial rate of liberation of /)-nitrophenol was measured by recording the change in absorbance at 420 nm in a Beckman spectrophotometer model 25 with a pen recorder (Beck- man Instruments, Fullerton. CA). The samples or PP2A inhibitors as positive control were mixed with the enzyme solution 5 min prior to initiating the reaction. Okadaic acid (OA) and Microcys- tin-LR (both from Sigma Chemical Co) were used as the specific inhibitors for PP2A activities. Complete inhibition was observed with 5 ng of Microcystin-LR or OA. All assays were carried out at 22-24°C. Each determination was done in duplicate. The reaction was stopped with 25 |jil of methanol (Rivas et al. 2000). RESULTS Phytoplankton Distribution Sampling on March 6 was made under high cloudiness, which did not permit us to distinguish any water discoloration. A total of 2 1 diatoms and 23 dinotlagellates. including a cystic form, were observed in the net samples collected (Table I ). The inner station showed the lower species richness ( 1 2 spp. ). and the others ranged from 23 to 35 spp.. most of them dinotlagellates. D. acuminata and P. ronindatuin were detected in these samples. Cell contour variation of D. acuminata seen in the samples showed some deviation from the characteristic oval or elliptical form typically found in Magellanic waters (Fig. 2). Antapex shape varies from rounded (Fig. 2a) to slightly acute (Fig. 2d), having in all cases two or three small protuberances. The light photomicro- graphs clearly illustrate that the Dinophysis sp.. collected in the four stations of monitoring, correspond to D. acuminata. Quantitative samples showed an absolute predominance of Me- Miclinium rubriim Lohmann (Table I ). Dinoflagellate was the sec- ond group best represented, and diatoms were present in scarce number (Table 2). From the dinotlagellates, GyroJinium lachiyma (Meunier) (Kof. & Swezy) was the most abundant, followed by Polykrikos schwarlzii (Biitschli) and D. acuminata. The concen- tration of D. acuminata ranged from 600-960 cells/1 in the four sampled sites. P. rotundatum was also present in these samples, although in lower concentrations (Table 1 ). Idi'iiliflcatiiin and (Jiianlitalive .Analysis of DSP Toxins HPLC with fluorescence detection is the most suited approach for the quantitative analysis of DSP toxins. In our study, we used a detection system involving a precolumn derivatization of DSP toxins with a fluorescent chromophore 9-anthryldiazomethane, ADAM (Lee et al. 1987), which allowed us to detect nanogram quantities of DSP toxins. Figure 3A shows a typical chromatogram of DSP reference toxins, such as OA, dinophysistoxin-1 (DTX-1 ), and the internal standard deoxycholic acid (DEO), showing reten- tion times (Rt) of 9:60, 13:20. and 18:40 min. respectively, for the 9-Anthrylmethyl esters. Figure 3B shows the chromatographic properties of the extract obtained from the mussel sample collected in site A (see Fig. lb and Table 2). The elution profile showed a single peak indistin- guishable from 9 MA-DTX-I (Fig. 3A). with a Rt = 13:20 min. All samples extracted and analyzed showed the same elution pro- file, displaying DTX-1 as the only DSP toxin present in the samples. The amounts of DTX-1 measured by the HPLC analysis of each sample collected in the four sites of Estero Nunez are showed in Table 2. These values were confirmed bv inhibition of TABLE L Phytoplankton concentration (cells/1) in Estero Nufiez (53 19'(I()"S, 72 30'00"W). Sites Species Diatoms BaciUaria paiado.m Coscinodiscus sp. Fragilaria sp. Nitzschio longissima Pseudo-nitzschia cf. australis Thalassiosira sp. Dinotlagellates Ceratium furca Ceratiwn peniagomiw Diiuiitli\.\i.s Licumhuila Diplopsalis sp. Gyinnodinium sp. Gyrodinium lacryma Phalacroma rotimdutum Polykrikos sclnvarzii Prohiperidinium conicwn Protoperidinium sp. Silicotlagellates Dyctiocha speculum Subtotal Mesodiniiim riihruni 400 0 160 0 800 40 20 200 960 200 400 14.070 400 10.553 0 0 0 28.20? 1 54.795 200 0 0 0 0 0 0 0 680 0 18.704 7.484 340 3,208 200 200 0 31.016 .16.353 200 0 0 200 0 200 20 0 800 0 1,500 4,500 240 500 0 0 200 8.360 133.000 400 200 0 0 0 0 200 200 600 0 0 0 0 2.000 0 0 0 3.600 135,200 72 A B Uribe et al. A ■■■'#•**' # Figuru 2. Contour variation of Dinopliysis acuminata (Clarapete & Lachmann) Ibund in Estero Nunez. PP2A assays. The extracts of mussel samples collected from sites A. B, and C showed inhibition over PP2A, the amounts of DSP toxin calculated from the inhibition curve using the mu.ssel extracts were the same measured by HPLC (Table 2). DISCUSSION From the low phytoplankton concentrations and the absence of water discoloration it is possible to infer that the sampling done in March 6 was made in a period after a bloom. Samples came from the head of a small, enclosed fjord that receives fresh water dis- charge from two creeks, so the water column stability is a typical situation in this area. During the monitoring visit, done on January 26 and March 23. the temperature and salinity were recorded from a monitoring station located approximately 2 miles from the sites where samples were collected. These data showed a low density in surface water with an average temperature of 1 1.1 8°C and a mean salinity of 28.78 psu in January. During the monitoring visit of March, these parameters were 9. U'^C and 29.49 psu, respectively. At both sampling visits a pycnocline at approximately 5 m depth was observed. According to the literature, such stability is a con- dition that permits planktonic dinotlagellates and ciliates to reach high densities (Margalef et al. 1979, Holligan 1985). D. acuminata is one of the seven species of Dinophysis sp. that 10000 5000 9 MA - DEO 10 15 20 Minutes 30 > E JbUU 9 MA - DTX1 3000 - \\ 2500 - 2000 - 1500 : 1 1000 - j l^^ 1 500 1 . 1 . 1 , 1 . 1 , r . f , 10 20 25 30 15 Minutes Figure 3. A. Fluoronietric high-performance liquid chromatography (HPLC) chronialogram of tiiarrhetic slielitlsh toxin standards deriva- tlzed with ADAM: yMA-OA (y-anthryhiiethyl Okadalcol. 9MA-DTX1 (9-anthr>lmethyl Dinophysistoxin 1), and 9MA-DFO (y-anthrylniethyl Deoxycholic). .All esters are from constituents that have carboxylic acid. B. Fluoronietric HPLC chromatograni of diarrhetic shelltlsh toxin present in mussel extracts derivatized with .ADAM: (9- anthrylniethyl Dinophysistoxin I). have been confirmed to produce OA and/or DTX-1 (the other six species are D. fi>rtii Pavillard. D. acuta. D. norvegica clarapede & Luclunann. D. mi Ira (Schutt) Abe, D. tripos Gourret, and Ph. rotimdatum: Lee et al. 1989). The quantitative analysis of phy- toplankton samples collected in the Estero Nufiez showed two possible sources of DSP toxins: D. acuminata and P. rolundatum. P. rotundatuin is the most likelv source because it showed the TABLE 2. Amount of DTX-1 detected in mussel extracts collected in the four sites of Estero Nunez. Samples Location Date OA (ng/g) DTX-1 (ng/g) Mytilus chitensis Hupe Mytilus cliilensis Hupe Mytilus cinlensis Hupe Mytilus cliilensis Hupe Estero Nuiiez (site A; see map. Fig. IB) Estero Nunez {site B; see map. Fig. IB) Estero Nuiiez (site C, see map. Fig. IB) Estero Nuiiez (site D. see map. Fig. IB) March 6 March 6 March 6 March 6 ND ND ND ND \\1.5 65. (J ND DTX-1 = dinophysistoxin- 1; OA = okadaic acid; ND = none detected. DlARRHETlC SHHLLMSH TOXlNS IN MAGELLANIC FjOKDS 73 higher cell density. D. nciinunata is a common species in the phytoplankton of the Magellanic fjord waters (Uribe et al. 1998). The variation of the contour found this time in the D. cinmiiiniid cells is not common in the Magellanic fjords, but it adjusts quite well in (he range presented by Balech (19761 in a study of Nor- wegian Dinophysis. Eslero Nunez has the typical geomorphology of small Magel- lanic fjords, with the water column stability as the most common hydrographic feature. Consequently, small toxic blooms could be frequent along the region and probably are being overlooked by monitoring programs because sampling stations are located far from the tjord heads. In fact, mussels gathered in the head of Estero Nuriez caused eight human cases of PSP in April 1989 in this region (Uribe 1989. Lagos 1998). A similar situation occurred with mussels collected in the head of Bahfa Nash, in March 1991, where there were 9.^ human cases of PSP and two deaths (LIribe 1992. Lagos 1998). Historically in our country. HABs associated with DSP have been confined to Patagonic fjords north of 46 S. Now. this finding extends the known DSP toxins distribution approximately ."iOO miles further to the South of Chile in the Austral Magellanic fjords. The frequent presence of Dinophysis sp. in Magellanic tjord waters, now associated with DSP toxins, represents a latent threat to the public health, which demonstrates the need to develop a more thorough evaluation for the presence of DSP toxins and its primary source in many small fjords along the Magellanic region. This is especially important after the demon.stration that DTX-1 causes severe injuries on the intestinal mucosa (Terao et al. 1986). Furthermore, another important aspect of OA and DTX-I is their potent tumor-promoting activity (Suganuma et al. 1988). so atten- tion should be paid to the continuous uptake of subacute levels of DSP toxins through seafood. ACKNOWLEDGMENTS This study was supported by FONDECYT 1961 122 and Fun- dacion Andes. Also, we want to thank the Chilean Navy for al- lowing us to cruise in the Coast Guard vessel LPC Villarrica to collect the samples on March 6. Miss Marcia Vera drew the map. LITERATURE CITED Anderson. D. M. 1989. Toxic algal blooms and red tides: a global per- spective. In: T. Okaichi. D. M. Anderson & T. Nemoto. editors. Red tides: Biology, environmmenlal science and toxicology. New York: Elsevier, pp. 1 1-16. Balech. E. 1976. Some norwegian Dinophysis species (dmotlagellata), Sar- sici 6l:7.';-y4. Compagnon. D.. G. Lembeye. N. Marcos. N. Ruiz-Tagle & N. Lagos. 1998. Bioaccumulation of PSP toxins in bivalve Aiilacomya ater and two carnivorous gastropods Concholepas concholepas and Ari>(>hucci- ntim ranellifimnes during Alexandrium catenella bloom in the southern Chile. / Shellfish Res. 17:67-74. Falconer. I. R. 1996. Potential impact on human health of toxic cyanobac- teria. Phycologia. 36:6-1 1. Guzman. L.. I. Campodonico & S. Hermosilla. 1975. Estudios sobre un tloreciento toxico causado por Gonyaulax catenella en Magallanes. I.- Distribucion espacial y temporal de G. catenella. Anales. Inst. Pa!.. Puma Arenas (Chile). 4:173-183. Hallegraeff. G. M. 1993. A review of harmful algal blooms and their apparent global increase. Phycologia 32:79-99. Hasle. G. R. 1978. ."1.2.1. The inverted-microscope method. In: A. Sournia. editor. Phytoplankton manual, monographs on oceanographic method- ology. Paris: UNESCO, pp. 88-96. Holligan. P. 1985. Marine dinoflagellate blooms - growth strategies and environmental exploitation. In: D. M. Anderson. A. W. White & D. G. Baden, editors. Toxic dinoflagellates. Oxford: Elsevier, pp. 133-140. Lagos. N.. D. Compagnon. M. Seguel & Y. Oshima. 1996. Paralytic shell- fish toxin composition: A quantitative analysis in Chilean mussels and dinoflagellate. In: T. Yasumoto. Y'. Oshima & Y. Fukuyoeditors. Harmful and toxic algal blooms. Paris: Intergovernmental Oceano- graphic Commission of UNESCO, pp. 121-124. Lagos. N.. D. Compagnon. D. Andrinolo & K. Salas. 1998. Quantitative analysis of PSP toxicity in the fjord system of the southern regions of Chile: HPLC technology. Proc. II Reimiao Ibero-Americana de Flco- logia Vol. 1:51-62. Lagos. N. 1998. Microalgal bloom: a global issue with negative impact in Chile. Biol. Res. 31:375-386. Lee. J. S.. T. Yanagi. R. Kenma. & T. Yasumoto. 1987. Fluorometric determination ofdiarrhetic shellfish toxins by high-performance liquid chromatography. Agile. Biol. Cheni. 51:877-881. Lee. J. S.. T. Igarashi. S. Fraga. E. Dahl. P. Hovgaard & T. Yasumoto. 1989. Determination ofdiarrhetic shellfish toxins in various dinoflagel- late species. J. .\ppl. Phycol. 1:147-152. Lembeye. G.. L. Guzman & 1. Campodonico. 1975. Estudios sobre un florecimiento toxico causado por Gonyaula catenella en Magallanes. III.- Fitoplancton asociado. Anales. Inst. Pat. Piinta Arenas 4: 197-208. Lembeye. G.. T. Yasumoto. J. Zhao & R. Fernandez. 1993. DSP outbreaks in Chilean fiords. In: T. J. Smayda & Y. Shimizu. editors. Toxic phy- toplankton blooms in the sea. New York: Elsevier, pp. 525-529. Margalef. R.. M. Estrada & D. Blasco. 1979. Functional morphology of organisms involved in red tide, as adapted to decaying turbulence. In: D. L. Taylor & H. H. Seliger. editors. Toxic dinoflagellate blooms in the sea. New York: Elsevier, pp. 89-94. Muiioz. P.. S. Avaria. H. Sievers & R. Prado. 1992. Presencia de dinoflage- lados toxicos del genero Dinophysis en el seno de Aysen. Chile. Re\: Biol. Mar. Valparaiso 27:187-212. Proenca. L. A. O.. L. Rorig. M. Silva. S. Guimaraes & N. Lagos. 1997. PSP outbreak in Brazil. Harmful algae news. Paris: Intergovernmental Oceanographic Commission (IOC). UNESCO. No. 16. pp. 1-3. Proenca. L. A., L. Rorig. M. A. Barreiro & N. Lagos. 1998. A possible case of diarrhetic shellfish poisoning in Santa Catarina. Southern Brazil. Proc. II Reiiniao Ibero-Americana de Flcologia 11:257-263. Rivas, M.. C. Garcia, J. L. Liberona & N. Lagos. 2000. Biochemical characterization and inhibitory effects of dinophysistoxin-1. okadaic acid and microcystine L-R on protein phosphatase 2A purified from the mussel Mxtitns chilensis. Biol. Res. 33:197-206. Smayda. T 1992. Global epidemic of noxious phytoplankton blooms and food chain consequences in large ecosystems. In: S. Klma & B. D. Gold, editors. Food chains yields, models, and management of large ecosystems. Boulder: Westview. pp. 275-307. Steindinger. K. 1993. Some taxonomic and biologic aspects of toxic di- noflagellates. In: I. Falconer, editor. Algal toxins in seafood and drink- ing water. New York: Academic Press, pp. 1-28. Suganuma. M., H. Fujiki. H. Suguri. S. Yoshizawa. M. Hirota. M. Naka- yasu. M. Ojika, K. Wakamat.su, K. Yamada & T. Sugimura. 1988. Okadaic acid: An additional non-phorbol-12-tetradecanoate-13- acetate-type tumor promoter. Proc. Natl. Acad. Sci. USA. 85:1768- 1771. Terao. K.. E. Ito. T Yanagi & T. Yasumoto. 1986. Histopathological studies on experimental marine toxin poisoning. I. Ultrastructural changes in the small intestine and liver of sucking mice induced by dinophysistoxin-1 and pectenotoxin-l. Toxicon 24:1141-1151. Uribe. J. C. 1989. Antecedents about a third PSP outbreaks in the Magellan Region. Ans Inst. Pat. Ser Cs. Nts 18:97-101. 74 Uribe et al. Uribe. J. C. 1992. Toxic shellfish kill twelve people in Chile. Red Tide Newsletter 5( 1 ):3. Uiihe, J. C. 1993. PSP outbreaks in Chile. In: Sixth international confer- ence on toxic marine phytoplankton. Nantes. France. 18-22 October. Abstracts, pp. 210. Uribe. J. C. L. Guzman. H. Pacheco. R. Igor, G. Pizarro. E. Barticevic. M. Frangopulos. C. Alarcon. T. Hromic. C. Labbe. A. Atalah & P. Chang. 1998. Red tide monitoring program in the Magellan region, first stage. Final report presented to Regional Economy Secretary. XIl Region. Tomos 1(111 pp.), Tomo II (257 pp.), Tomo 111 (231 pp.). White, A. W. 1988. Blooms of toxic algae worlwide: Their effects on fish farming and shellfish resources. Proc. Intl. Conf Impact of Toxic Al- gae on Mariculture. Aqua-Nor 187 Exibition. Trondhein. Norway, 13- 18 August 1987, pp, 9-14. Yasumoto, T. & M. Murata. 1993. Marine toxins. Chem. Rev 93:1897- 1909. Yasumoto, T,, M. Fukui, K, Sasaki & K. Sugiyama. 1995. Determination of marine toxins in foods. J. of AOAC International 78:574-582. Zhao, J.. G. Lembeye, G. Cenci. B. Wall & T. Yasumoto. 1993. Determi- nation of okadaic acid and dinophysi.stoxin -I in mussels from Chile, Italy and Ireland. In: T. J. Smayda & Y. Shimizu, editors. Toxic phy- toplankton blooms in the sea. New York: Elsevier, pp. 587-592. I JiHinwl oj Shellfish Research, Vol. 20, No. I, 75-S4. 2001, AMNESIC SHELLFISH POISONING IN THE KING SCALLOP, PECTEN MAXIMUS, FROM THE WEST COAST OF SCOTLAND D. A. CAMPBELL.' M. S. KELLY.' M. BUSMAN," C. J. BOLCH,' E. WIGGINS," F. D. R. MOELLER.- S. L. MORTON." P. HESS,' AND S. E. SHUMWAY^ 'Scotlish Association for Marine Science. Olnm. Argyll. Scotland. UK, PA34 4AD: 'National Ocean Sen'ice, South Carolina. 29412: ^FRS Marine Laboratory, PO Box 101. Victoria Rd.. Aberdeen, Scotland. UK. AB31 6BF: Natural Science Division. Southaiiipton College. Long Island University. Southatnpton, New York, 1 19(iH ABSTRACT The king scallop. Pecten maximus, is a valuable economic resource in (he UK. The industry relies on supplying premium "roe-on" processed scallops to the continental market. In July 1999. king scallops harboring the amnesic shellfish poisoning (ASPl toxin, domoic acid (DA), in gonadal tissue at levels above the regulatory limit (20 |a.g DA g"') were detected across a wide area ot northern and western Scotland. In response, a survey of the southern extent of the closed harvest areas was initiated to describe variability of ASP toxin levels over varying spatial scales (<5 m to >5 km): determine the anatomical distribution of the toxin, and identify, Lsolate, and culture causative Pseudo-niizschia species. Toxin analysis was conducted using a liquid chromatography-tandem mass spectroscopy (LC-MS/MS) procedure. The DA content of tissues followed the predictable rank order: all other tissue — > gonad — > adductor. The toxin levels within all other tissue (959i- CI = 580-760 |j,g DA g"'. /; = 170) consistently accounted for 99'7r of the total individual toxin burden. DA levels in the gonad (95% CI = 8.2-1 1.0 fig DA g~'. /; = 170) were an order of magnitude below levels in all other tissue and contributed to less than 0.5''f of the total individual toxm burden, although levels above the regulatory limit were detected in individual gonad samples. Adductor muscle tissue contained the lowest concentrations of DA (95'?f CI = 0.38-0.82 |jLg DA g"', ;; = 170). and was typically within two to three orders of magnitude below levels in all other tissue. None of the scallops examined had DA toxicities in adductor muscle tissue exceeding the regulatory limit. Toxin variability among individuals and sites was high (range of coefficients of variation (CV) in all other tissue = 29%-120% and gonadal = 45'7r-85%). The results do give an indication of the scale on which microhabitat differences may influence ASP toxicity in P. maximus populations, because significant differences were found in all other and gonadal tissue toxin levels between groups of individuals only 25-m apart. In total, seven species of PseKdo-nilzschia were identified from west coast waters. A suspected causative species, P. australis, was found to produce high levels of DA. in culture. The high individual variation in toxicities and the occurrence of DA in the gonad at levels above the regulatory limit clearly demonstrate the complexity of managing the king scallop fishery during ASP events. KEY WORDS: amnesic shellfish poisoning, domoic acid, Pseiuid-nitzschia. Pecten maximus. scallop fishery INTRODUCTION Marine algal toxins ctimprise a diverse group of biologically active compounds with high acute toxicities in humans (Shtimway & Cembella 1993). Scallops, opportunistic filter feeders exploiting both pelagic and benthic microorganisms as food sources, are li- able to the accumulation and concentration of phycotoxins from toxic algal species present in the water column (.Shumway et al. 1987, Bricelj & Shumway 1991). The risk of human illness as a result of toxic scallop consumption poses a significant threat to both public health and shellfish industries (Shumway & Cembella 1993). Amnesic shellfish poisoning (ASP), a relatively new type of seafood toxicity, was first described from Prince Edward Island, Canada, in 1987 (Bates et al. 1989). Over 100 people who con- sumed mussels contaminated with a naturally occurring neuro- excitatory toxin, domoic acid (DA), experienced gastroenteritis and neurological symptoms (Wright et al. 1989, Todd 1993). In this first episode, the source of the domoic acid was identified as the pennate diatom, Pscudo-nitzschia pugens f. multiseries. which was ingested and accumulated by the mussels during normal filter feeding (Bates et al. 1989). Global awareness of ASP has since been raised, and, to date, ASP toxin-producing species of Pseiido- nitzsehia have now been reported from the gulf of Mexico region. North America, Canada. Europe. Australia, Japan, and New Zealand (Hallegraeff 1995). From laboratory studies, it is now clear that several species of Pseiido-nitzschia. and two species from separate genera {.Amphora and Nitz.schia). are capable of producing DA, but the levels of production are highly variable (Kotaki et al. 2000, Bates 2000, Lundholm & Moestrup 2000). The king scallop, Pecten nuixiiuiis. is a valuable economic re- source in the UK. The UK scallop industry is principally a wild fishery exploited by scallop dredgers, which account for an esti- mated 97% of UK landings. However, small quantities are landed by divers, and in Scotland, there is an emergent aquaculture in- dustry. The industry is largely reliant on supplying premium "roe- on" processed scallops to the continental market. In 1998, 9.700 tons of P. maximus were landed in Scotland, with a first sale value of £13.3 million, equating to approximately 25% of all EU scallop landings (Denton 1999). An estimated 957^ of the king scallops are processed as meat and roe product, of which 60% is distributed as premium chilled product and 40% frozen. Dive collected and farmed scallops are sold whole to a smaller, yet higher \alue market for live shellfish. The incorporation of systematic ASP/domoic acid testing of shellfish into the Food Standards Agency (FSA) Scottish waters surveillance program was initialed early in 1999. By July 1999, P. maximus harboring DA in the gonad at levels above the interna- tionally accepted closure limit (20 |ji.g DA g"') were detected across a wide area of northern and western Scotland. This prompted a widespread closure of the king scallop fishery, which persisted in excess of 9 months, resulting in financial hardship for scallop dredging, diving, and cultivation industries. The nature of 75 76 Campbell et al. the ubiquitous and prolonged high toxicity seemed to be confined to the king scallop, because only sporadic, short-term toxicities were noted in the queen scallop. Aquipecten opevciilaiis. and neg- ligible levels detected in other shellfish (FSA pers. comm.. 1999). The direct cost to the industry to date has been estimated at £10 million, and the loss of skilled processing staff and disruption of established supply routes to continental markets led to serious concern for its survival (Denton 1999). The restriction on all scal- lop landings provoked controversy, stimulating much media inter- est. To date, there has been no documented history of human illness caused by ASP in the UK. In response, an opportunistic survey of the southern region of the closed harvest areas was initiated to provide fundamental in- formation on the ASP incident. The objectives were to: describe ASP toxin variability among individual and neighboring scallop populations over varying spatial scales (<5 m to >5 km); deter- mine the anatomical distribution of ASP toxin within scallop body parts: adductor muscle and gonad, and all other tissue (digestive gland, mantle, gill); assess any influence of size. age. and depth on scallop toxicity levels; and isolate, culture, and identify causative Pseiulo-nhzschia species. The data collected provide basic infor- mation to assist with the development of rational management strategies to continue to protect public health while minimize the economic constraints of future ASP events. Area J5 «■ 2.5km THE WEST COAST OF SCOTLAND t i N I y-^ V 3 ^ ^gaehan r/ ^^^ .20 no. >10()0 Tissue II Mean SE Minimum Maximum (^H CI Mgg"' Mgg"' Adductor 170 0.60 0.114 0.011 15.42 247 0.38-0.82 0 Gonad 170 9.58 0.722 0.131 75.5 98 8.16-11.01 19 All other 170 669 45.7 0.2 3689 89 580-760 169 10 Total 170 295 10.1 0.9 1569 85 256-333 168 8 Nliicl\ -five percent confidence intervals (CI ). mminunii and nia.\imuni levels of DA obtained, coefficient of variation (CV %) and number of individuals with toxin burdens >20 and >1000 \xg DA g"' are given. MATERIALS AND METHODS Sampling In December 1999. 10 specimens of aiJult P. imuimKs (shell height >90 mm) were collected by SCUBA at three subsites (total of 30 inciivWuals) within Loch Sligachan (.57'L5'5' N 06°15'5' W) and at four subsites (total of 40 individuals) within Tobermory Bay (56°35'5' N 06°05'5' W) (Fig. 1). To provide an assessment of ASP toxin levels among scallop populations on a larger spatial scale, in context with the current monitoring program; a third sampling location, the scallop fishing box-Area J5, was included in the sampling regime (Fig. 1). Twenty adult scallops (shell height >90 nim) were selected at random from the landings dredged from each of five subsites (total of 100 individuals). These locations were routinely used for the monitoring of ASP toxin under the Food Standards Agency Program and were chosen as a result of previously consistent high DA levels (above the 20 p-g DA g"' statutory level) within scallop gonad. Upon collection, all scallops were individually sealed in zip-lock polythene bags, placed in cool boxes, and transported to the Scottish Association for Marine Sci- ence (SAMS) Laboratory within 6-12 h. for immediate dissection. The age of each scallop was estimated by enumerating shell growth bands and shell length and breadth (Mason 1983! measured to the nearest 0.1 mm. The scallops were dissected into the body components: adductor muscle, gonad, and all other tissue (diges- tive gland, mantle, and gills!. Special care was taken to avoid artifactual contamination fiom adjacent tissues, by careful dissec- tion, washing, and drying of individual body components. The digestive material of the intestinal loop within the gonad was physically removed. All body components were weighed to the nearest 0.001 g. sealed separately in zip-lock polythene bags, and frozen at -20"C before DA extraction (Quilliam et al. 1989). DA Extraction and Quantificalinn in Scallop Tissue Tissues were homogenized in a blender (3 min). which was cleaned and rinsed with methanol and then distilled water, between each sample. Four grams of tissue homogenate was rehomog- enized (4 min) with 10 mL of 100% methanol, centrifuged (10 min at 5.000 rpm). and a 5-mL subsample of the resulting supernatant filtered through a disposable 45-|j.m filter membrane. The extiact was stored at -20°C before DA detection and quantification. The extracts were evaporated to dryness using vacuum cen- trifugation and resolubilized in 50/50 methanol and water before triplicate analysis. The samples were analyzed on a liquid chro- matography-tandem mass spectrometry (LC-MS/MS) system con- sisting of an Agilent Model 1 100 high-pei-formance liquid chro- matography (HPLC) system, coupled to either a SCIEX API-Ill triple quadrupole or a Finnigan LCQ ion trap mass spectrometer. The chromatography was performed on a CI 8 reversed phase col- umn with a 0.2 niL/min flow of a 1-95% gradient of methanol: 3000 — r -a 5 2000 . 00 < 1000 . Q 00 Loch Sligachan I I I I II I I T 10 30 -a C 2000 . 5 00 < Q 00 Tobermory Bay lllllll 20 3000 _ < Q 0 . Area J5 0 1 5 too 2 3 4 Subsites and Individuals Figure 2. DA toxicity levels (fig \)\ g ' wet weight! in all other tissue of P. maximus collected from subsites within locations. 78 Campbell et al. water. All solvents had 0.\% trinuoroacetic acid added. A portion of the effluent from the HPLC system was directed into the elec- trospray ionization source of the mass spectrometer via a flow splitter. The mass spectrometer was operated in positive mode with |M+H|+ ions 1-312 m/z) being isolated in a first stage of MS analysis. The isolated ions were subjected to "collision- induced dissociation" reaction conditions, which are expected to stimulate the fragmentation of the |M+H]+ ions into characteristic product ions. All quantitations were based on the integrated chromato- graphic intensity areas of one of the fragment ions (at 267 m/z), and the appearance of other characteristic ions was used as con- firmatory e\'idence for the DA ion's identity (Scholin et al. 2000). Phytoplankton Sampling and Isolation During August and October, plankton tows were taken from Orkney. Dunstaffnage. and Jura, to record P\ei(do-nirzschiii spp. presence. In December, at each subsite within Loch Sligachan and Tobermory Bay, surface plankton trawls, quantitative water samples (NIO bottle, at 5 m), cores (water/sediment interface samples), and benthic sediment samples were taken to examine the vertical distribution of Psciido-nitzschiu spp. present. At the sub- sites within Area J5. plankton trawls and quantitative water samples with (NIO bottle, at 5 m) were performed. Four replicates of each sample were obtained from each site; of which two were preserved with LugoFs iodine for cell counts, and two were en- riched by addition of f/2 -i- Si growth medium. Samples were examined for actively growing chains of PseiiJo-nitzschiu cells and candidate chains of cells isolated by micro-pipette and re- peated washing in sterile f/2 growth medium. Individual chains of cells were incubated with f/2 -i- Si growth medium (Guillard & Ryther 1962) at 15°C under an approximate light intensity of SO- SO ixmol PAR m"- s"' (12:12 light/dark cycle). Stock cultures were grown for three weeks (stationary phase) before cell harvest by gentle centrifugation (1.500 rpm). followed by removal of ex- cess growth medium and resuspension in sterile growth medium. Samples of cell pellets and supernatant were immediately placed on ice before DA analysis. DA Extraction and Quantification in Pseudo-nitzschia spp. Cell pellets were subjected to ultrasonication ( 10 min) in 30/50 methanol and water, before filtration (Whatman filter. 0.2 p,ni). The supernatant samples were directly filtered using the same filter size. The crude extract or fdtrate from the supernatant was ana- lyzed using HPLC coupled to a diode-array detector (10 p.L in- jected). The system (Themioquest) comprised a solvent reservoir and degasser, P4000 pump. AS3000 autosampler. and UV 6000 diode-array detector. The HPLC column was a VYDAC 20ITP54 (250 X 4.6 mm, 5 |xm) with a VYDAC guard column 20IGK54T (10 X 4 mm. 5 (xni). The mobile phase was 0. 1'7f tritluoracelic acid in 10% aqueous acetonitrile. at 1.5 niL/niin tlow rate. The column temperature was kept at 40°C. Wavelengths monitored ranged from 200-360 nm, and spectral confirmation was obtained by comparison of sample spectra to those from the certified reference standard DACS-IC. Quantitation was canied out al a wavelength of 242 nm. RESULTS Anatomical Distribution Despite the considerable variation in toxin levels within each body compartment. DA loading of the tissues followed a predict- able rank order: all other — > gonad — > adductor. The toxin levels within all other tissue consistently accounted for 99% of the total DA burden. A small proportion of individuals had DA levels in all other tissue (>l,000-3.690 p,g DA g"') an order of magnitude greater than mean levels (669 ± 45.7 pig g"' ). Mean DA levels (p.g TABLE 2. Mean and standard error (SE) of DA levels (njj D.\ g"' wet weight) in all other tissue of P. mtucimus at each sample location and subsite. Location and CV 95% no. >1000 subsite /; Mean SE Minimum Maximum (%) CI Mgg'' Loch Sligat han 30 iiO-" 44.2 58.6 820 73.4 240-f20 Tobermory Bay 40 SiT 93.5 45.2 3023 IIU 348-726 2 Area J5 100 824" 62.0 0.2 3689 75.3 701-947 8 Loch Sligac han 1 10 183" 54.7 62.5 549 94.6 59-306 2 10 352-"' 73.9 58.6 814 66.2 185-520 3 10 455" 79.3 123 820 55.1 275-634 Tobermory Bay 1 10 895' 341 307 3023 120 123-1670 2 2 10 228" 29.6 45.2 331 41.5 296-1601 3 10 407"" 57.8 137 663 45.0 276-538 4 10 61 8-' 61.1 318 984 31.2 479-756 Area J5 1 20 1076'" 224 379 3689 93.0 607-1545 3 2 20 58 1 ■' 33.5 409 998 25.8 510.6-651 3 20 586' 37.7 275 915 28.8 507-665 4 20 861" 104 0.2 2270 53.8 645-1078 2 5 20 1016" 159 525 2928 70.1 68.3-1.350 3 Means from the same location with different superscripts are significanlly different iP > 0.05. Kniskall-Wallis and Dunn's method). Ninety-tlve percent confidence intervals (CI), minimum and maxnniim levels, coefficient of variation (CV %} and number of individuals with toxinburdens >1000 |xg DA 2*' are given. ASP IN P. MAXIMUS 79 g" ) in gonad tissue were an order of magnitude below levels in all other tissue, and. on average, contributed to less than 0.5% of the total individual toxin burden. DA levels above the statutory 20 |j.g DA g"' safety level in gonads were detected in 22% of the scallops examined, although these values were not encompassed within the 95% confidence limits (S. 16-1 1.01 (xg DA g^'. ;; = 170). Ad- ductor muscle contained the lowest concentrations of DA and was typically two to three orders of magnitude below levels in all other tissue while accounting for only 0.17% (mean) of the total indi- vidual toxin burden. Although the CV (247%) of individual ad- ductor muscle DA levels was observed to be considerably greater than all other and gonad tissues (98 and 89%. respectively), none of the scallops examined, had adductor muscle toxicities that ex- ceeded the .statutory limit, and 95% of the samples had levels below 1.9 (xg g"', A weak, positive correlation was observed between logm all other tissue toxicity and log,,, gonadal toxicity (r = 0.303. P < 0.001. df = 169); whereas, no correlation could be found between DA concentrations in all other and adductor muscle tissue. No significant correlation could be made between DA toxicity in the three body compartments and scallop age and size paraineters. Spatial Dislrihiilion At all sites, a large variation in all other tissue toxin levels between individuals was observed (Fig. 2). indicated by the CV values (Table 2). However, significant differences in all other tis- sue toxin levels among locations were distinguished. Scallops from Area J5 had significantly greater levels of DA in all other tissue than individuals from Tobermory Bay and Loch Sligachan; whereas, no significant differences in toxin levels were found be- tween scallops from Tobermory Bay and Loch Sligachan, Significant differences in toxicities between individuals from different subsites within the same location were observed at all the three locations. Within Tobermory Bay. all other tissue DA toxin levels differed significantly between neighboring scallop popula- tions 25 and 1 .200 m apart (subsite 2 DA levels were significantly lower than I and 4). In scallops from Loch Sligachan, all other tissue DA toxin levels differed between neighboring scallop popu- lations 300 m apart (subsite 3 DA levels were significantly higher than 1 ). Within the scallop fishing box. Area J5, all other tissue DA toxin levels were significantly higher in scallops from subsites 4 and 5 than 2 and 3, collected 8-12 km apart. Scallops with high all other tissue toxin burdens (>1,000 |a.g g"') were not evenly dis- tributed among locations or subsites. The significance of the re- sults remained unchanged, regardless of the removal of individuals with high toxicities from the dataset. At all sites, a large variation in gonadal toxin levels between individuals was observed (Fig. 2). indicated by CV (Table 3). Despite the wide variation, gonadal toxicity among locations fol- lowed the same pattern of the toxicity as all other tissue, because scallops from Area J5 had significantly greater levels of DA in the gonad than individuals from Tobermory Bay and Loch Sligachan. Similarly, no significant differences in gonad toxin levels were found between scallops from Tobermory Bay and Loch Sligachan. Significant differences in DA levels between subsites of the same site were observed at all threesites sampled. In Tobermory Bay. gonad DA toxin levels differed between neighboring scallop populations 25-and 1.200-m apart (subsite 2 and 3 DA levels were significantly lower than 1 ). In Loch Sligachan. gonadal DA toxin levels differed between neighboring scallop populations 3(J0 m 70 . 60 . 50 40 M 30 < 20 00 a. 10 Loch Sligachan I I I I I lllllllll 10 20 30 70 . 60 ■ ^ 50 . fl) 40 ^ _oo 30 < Q ""O ou 3. 10 _ Tobermory Bay 1 1 III! 1 1 1 1 1 1 III 10 20 30 r 40 100 I 2 3 4 Subsites and Individuals Figure 3. DA toxicity levels {fig g ' wcl »eighll in the gonad of P. maximiis collected from subsites within locations. Line limit (20 ng DA g~' of tissue). statutor> apart (subsite 1 DA levels were significantly higher than 3 ). Within Area J5, gonadal DA toxin levels were significantly higher at subsite 5 than at 2. 3. and 4; whereas, gonad toxicity in subsite 1 was greater than that in 3. At each location, scallops with gonadal toxicities exceeding the 20 jjtg DA g"' liinit were encountered. However, the frequency of these individuals was not homogenous among the subsites, ranging from 0 out of 20 individuals (Area J5, subsites 2 and 3) to 7 out of 20 (Area J5, subsite 5). Scallops from Toberinory Bay had significantly lower levels of DA in the adductor inuscle than individuals from Loch Sligachan and Area J5. Levels of toxicity among locations did not correspond to the pattern of toxicity seen in all other and gonad tissue. No significant differences in adductor muscle toxin levels were found 80 Campbell et al. TABLE 3. Mean and standard error (SE) of DA levels (pg DA g~' wet weight) in tlie gonad of P. maximum at each sample location and subsite. Location and CV 95% no. >20 subsite /; Mean SE Minimum .Maximum (CH CI Mgg"' Loch Sligachan M) 6.82-' 1.35 0.77 28.62 108 4.06-9.58 3 Tobermory Bay 40 8.18^ 1.06 0.13 29.49 81.9 6.04-10.32 2 Area J5 100 10.97" 1.06 0.36 75.47 96.8 8.87-13.08 15 Loch SUgachan 1 10 3.52" 0.847 0.775 7.517 76,1 1.6-5.4 2 10 3.93° 0.622 1.513 7.939 50.1 2.5-5.3 3 10 13.02" 3.17 1.72 28.62 77.0 59-20.2 3 Tobermory Bay 1 10 14.14'^ 2.54 5.48 29.49 56.8 8.4-19.9 2 2 10 .\21-' 1.38 0.13 15.55 83.5 2.1-8.3 3 10 4.41" 1.66 0.32 14.18 119 0.7-8.2 4 10 8.95'" 1.32 3.33 14.28 46.7 6.0-11.9 Area J5 1 20 U.l?-'" 2.05 1.80 38.84 64.9 9.9-18.5 6 2 20 6.77"" 0.792 1 .593 13.28 52.3 5.1-8.4 3 20 4.86" 0.823 0.594 14.59 75.7 3.1-6.6 4 20 9.57"" 1.71 0.36 25.15 79.7 6.0-13.1 2 5 20 19.50' 3.68 5.21 75.47 84.5 11.8-27.2 7 Means from the same location with different superscripts are significantly different iP > 0.05. ICruskall-Wallis and Dunn's method). Ninety-five percent confidence intervals (CI), minimum and ma.ximum levels of DA obtained, coefficienl of variation (CV "Tr ) and number of individuals with gonadal toxin burdens >20 (jLg DA g'' statutory limit are given. between scallops from Loch Sligachan and Area J5. At all sites, an exceptionally large individtial variation in adductor muscle toxin levels was observed (Fig. 4). indicated by the CV (Table 4). Within Tobermory Bay. adductor muscle DA toxin levels dif- fered between neighboring scallop populations 25- and 1 ,20()-m apart (subsite 3 DA levels were significantly lower than I). In Loch Sligachan. no significant differences in adductor muscle toxin levels were observed between subsites. In Area J5. adductor muscle toxin levels were significantly higher in scallops in subsite 5 than in 2 and, 3, and toxicity in subsite 1 was greater than in 2. Again, the significance of the results remained unchanged, regard- less of the removal of individuals with a comparatively high toxin loading. Pseudo-nitzschia spp. Ahiindance and DA Production The August to October plankton tows samples showed several potentially toxic Pseuclo-niizschia species were present. At the peak of the blooms, P. aitstnilis was the dominant species fol- lowed by P. pungens. Several other species were present as minor components: P. mulliseries. P. seriata, P. delicatissima, P. frciudti- lenta, and P. pseudodelicatissima. The blooms were observed to subside during October 1999, with low levels of P. delicatissima persisting through to December 1999. At the time of scallop col- lection (December 1999) Pseiido-tiitzscliia spp. cell numbers were exceptionally low throughout the water column (<1 cell/niLl at the sites sampled (water temperature range 7-9. 5°C). These cell con- centrations are well below that usually associated with reported ASP events. Examination of surface sediments at Loch Sligachan and Tobermory Bay also failed to detect significant quantities of living or dormant cells Pseitda-uitzschia spp. Three Pseudo-nilzscliia cultures were established from samples collected in the August 1999 blooms, two strains of P. aiislialis. and one of P. pungens. Stationary growth-phase cultures of both P. austmlis strains produced detectable levels of DA in intracellular and extracellular fractions (Table .5). However, the presence of DA could not be detected in the P. pungens cultures (detection limit = 0.1 |j,g mL~' in cell/supernatant extracts). In both P. australis cultures, total DA was partitioned with approximately one- third being intracellular and two-thirds present in the growth medium. DISCUSSION The trend in body component toxicity of P. ma.ximus as a proportion of total scallop toxin burden (all other tissue -^ gonad — > adductor), is in agreement with previous studies of DA in P. ina.xiinus (Arevalo et al. 1998), DA in Placopecten niageUanicus (Douglas et al. 1996), and paralytic shellfish poisoning (PSP) in P. inagellanicus (Cembella et al. 1994). In the current study, 99.4% of individuals had levels of DA in all other visceral tissue over the statutory 20 [jig DA g~' limit. The maximum DA concentration in all other tissue (3,689 |j.g DA g"'). recorded in this study was approximately 180 times the regulatory limit (20 |jLg DA g"') and is among the highest levels recorded in bivalves. Arevalo et al. (1998) found the highest levels of DA in the hepatopancreas in Pecten nia.xinms (maximum = 2,083 |jL,g DA g"'). Similar high levels (approximately 3,000—4,000 |jLg DA g"') were found in the digestive gland oi Placopecten niagellaniciis (cited in Douglas et al. 1996). Thus, the levels of DA found in all other tissue in the present study are consistent with previous findings, confirming that DA is predominantly sequestered within the digestive gland in P. ina.xiiniis. Toxin levels of gonadal tissue were generally lower than the statutory 20 jjtg DA g"' limit: however, toxicities above this level were encountered. Adductor muscle toxicity contributed negligible amounts to the total body burden, and levels never exceeded the statutory limit, even when toxin levels were extremely high in all other tissue. The occurrence of PSP toxins in adductor muscle is ASP IN P. MAXIMUS 81 10.00— -a « 1.00- 5 < Q 0.10^ 01) 10.00— 1.00- 5 00 < Q 00 0.10. 0.01. 10.00 — 1.00. < Q 00 0.10. Loch Sligachan 1 li 0.01 — Subsites and Individuals Figure 4. DA t(i\ki(y levels (pg DA );"' wet weight) In the adductor muscle of /'. maximiis collected from subsites within locations, plotted against a !og,o scale. rare; and in the scallops P. mai;ellciinci(\. Paiinopeclen xessoeniss. and Crassodenna gii>anteci. adductor muscle PSP toxicity is al- ways at least 10-fold less than in corresponding digestive tissue (cited in Shumway & Cembella 1993). Bricelj and Shumway (199H1 report that, during PSP events, such tissues involved in locomotion as the muscular foot, adductor muscle, and pallial muscles, invariably attain toxicities two to three orders of magni- tude below those in the viscera, and contain a minimal proportion, typically < 19^ of the total toxin loading, despite their relatively large mass in some species. However. Pacific razor clams. Siliqua pauila. are reported to sequester DA principally within muscular tissue (Drum et al. 199.^). Confirmation of DA production by Pseudo-nitzschia australis and its dominance during the blooms indicate this species was one of the primary sources of scallop ASP contamination in 1999. Although a co-dominant species. Psciido-nitzscliia pitngens. did \w\ produce detectable concentrations of DA in culture, it is pos- sible that other Pseudo-nitzschia species, with known DA produc- tion capabilities and present as minor components, may have con- tributed to the ASP event (e.g.. P. seriata. P. fnuuhdenta. or P. pseiidiidi'licalissiiiia). The dominance of P. australis observed in the west coast waters, was not confirmed for other affected areas. Lundholm et al. (1994) .showed P. seriata produced DA at low temperatures, thus it could potentially represent a source of DA in colder, northern Scottish waters. The DA concentration of the 1999 Scottish isolates of P. australis (3-4 pg total DA celP') compare closely with previous studies of Pseudo-nitzschia, as cel- lular DA levels are reported to range from 0.1-10 pg DA cell"' for most species studied to date (cited in Bates 1998). Previous studies of western North American P. australis have indicated compara- tively high DA production capabilities (12-37 pg DA celP', Gar- rison et al. 1992). However, our data are more consistent with estimates of 2.0 pg DA cell"' for New Zealand strains (Rhodes et al. 1996), and with unpublished data from Spanish and Irish P. australis strains (S. Bates. Fisheries and Oceans. Canada, pers. comm. 2000). It is likely that the 1499 DA toxification. measured in Decem- ber 1999. in Scottish king scallops occurred as a result of Pseudo- nitzschia blooms during the May to August 1990 period and not the result of continuous intake from toxic benthic .sources (Bourne 1963). because Pseudo-nitzschia spp. concentrations were very low during October and December, and no significant quantities of living or dormant Pseudo-nitzschia cells were detected within the locations at time of sampling. Therefore, the current results sup- port the hypothesis that high DA levels in scallops are a conse- quence of low rates of toxin catabolism as a result of low winter basal metabolic rates and reduced filtration activity, further influ- enced by colder waters and reduced food supply (Shumway & Cembella 1993). The considerable degree of toxin variability observed among individual P. ina.\inius and their body components was not unex- pected and has been described for other shellfish species contami- nated with DA and PSP toxins (White et al. 1993. Arevalo et al. 1998). Characterizing variation in toxin levels among individual species of the same area is necessary both for ecological consid- erations and for development of sound management protocols (White et al. 1993). The results do give an indication of the scale on which microhabitat differences influence ASP toxicity in Pecten inaxiwus populations, because, despite wide individual variation, significant differences were found in all other tissue and gonadal toxin levels between groups of individuals only 25 m apart. Variation in bivalve toxicity is reported to result froin an interaction of such factors as timing, persistence, and magnitude of toxic blooms, microgeographic variation in exposure to toxic cells because of bloom patchiness. the specific toxicity per cell, and toxin composition of the contaminating organism, environmental effects on scallop metabolism, and. perhaps, genotypic differences among scallop populations (Bricelj & Shumway 1998). However, the reasons for the few individual scallops retaining exceptionally large toxin burdens in all other tissue 01.000-3,689 |jig DA g"') are not known. The ability to detect influences of scallop size parameters on DA accumulation may have been restricted by the limited size class (90-120 mm shell length; i.e., legal landing size) selected for use in the current study. Expanding the range of sizes used to 82 Campbell et al. TABLE 4. Mean and standard error (SE) of DA levels (jig DA g"' wet weight) in the adductor muscle of P. inaxiiniis at each sample location and subsite. Location and CV 95% no. >20 subsite 11 Mean SE Minimum Maximum <%) CI Mgr' Loch Sligachan 30 i.:?4" 0.573 0.025 15.415 250 O.OS3-2.425 Tobermory Bay 40 0.156' 0.037 0.020 1.395 151 0.081-0.231 AreaJ5 100 0.581'' 0.081 0.011 5.487 139 0.420-0.741 Loch Sligachan 1 10 2.890-' 1.630 0.030 15.42 179 -0.80-6.60 2 10 0.55y 0.219 0.045 2.310 124 0.06-1.05 3 10 0.313-' 0.040 0.168 0.501 40.5 0.22-0.40 Tohermory 1 10 0.192" 0.036 0.537 0.425 60.3 0.11-0.27 2 10 0.258* 0.132 0.021 1 .395 162 -0.04-0.56 3 10 0.(146" 0.005 0.020 0.067 38 0.03-0.06 4 10 0.129-''' 0.051 0.021 0.466 124 0.01-0.24 Area J5 1 20 0.600"'^^ 0.111 0.086 1.962 82.6 0.37-0.83 2 20 0.238" 0.0345 0.063 0.592 64.7 0.16-0.31 3 20 0.526'" 0.273 0.085 5.487 231 -0.05-1.10 4 20 0.484-"'' 0.091 0.112 1.652 83.6 0.30-0.67 5 20 1 .056' 0.235 0.225 3.962 99.7 0.60-1.50 Means from the same location with different superscripts are significantly different (P > 0.05. Kruskall-Wallis and Dunn's method). Ninety-five percent confidence intervals (CD, minimum and maximum levels of DA obtained, coefficeint of variation (CV %) and number of individuals with adductor muscle toxin burdens >20 (jLg DA g*' statutory limit are given. include juveniles may indicate any allometric influences on DA toxin accumulation in P. maximus. Under controlled conditions, weight-specific DA toxicity has been demonstrated to be inversely proportionate to body size in mussels Mylihis ecltilis (Novaczek et al. 1992). However, faster detoxification rates per unit body mass in actively growing, smaller, or younger individuals, because of toxin dilution through growth, may mask any allometric relation- ships present (Bricelj & Shumway 1998). A significant positive con-elation was observed between toxic- ity of the gonad and that of all other visceral tissue. Although little is known about transfer of DA among tissues, its likely that go- nadal toxicity is influenced by the level of digestive gland toxicity, via the intestinal loop, which passes through the gonad and may contain toxic feces. Cembella et al. ( 1993) demonstrated that PSP toxins are accumulated within gonadal follicles of P. luai^eUcmi- ciis. even after the exclusion of the intestinal loop. However, the inherent wide individual variation precludes the ability to predict gonad toxicities reliably from routine ASP toxin monitoring of the viscera. Compared with other body components, the variation in adductor muscle toxicity was proportionately larger, and no cor- relation could be found between toxicity of the adductor muscle and that of all other tissue. The variance in toxicity values in adductor tissue may be attributed to one of. or a combination of. three sources: (1) natural variation in adductor muscle toxicity; (2) variable contamination of the tissue from digestive fluid, during dissection; and (3) analytical error close to the limits of detection. The mean CV accounted for by the detection method for all other tissue, gonad, and adductor muscle was ±11.8, 4.6, and 18.69f, respectively, indicating that the variability observed between in- dividual scallops was not a result of analytical error. The extent to which toxic digestive fluid and exudates contaminates edible tis- sues should be established to ascertain the potential to reduce the ASP toxin burden by appropriate preparation of adductor muscle and gonad tissue and realize the necessity to standardize prepara- tion of these tissues before testing. During ASP events, the marketing of P. iiui.\iiiiu.s digestive gland, mantles, and gills, poses a high risk to public health, which has an impact primarily on diver-based and cultivation industries supplying markets for whole scallops. However, to allow the mar- keting of the nontoxic edible component, scallop preparation tech- niques should be promoted, such as the immediate removal of toxic tissues and thorough washing of the edible component (hav- ing ascertained the gonad is safe to consume), and this practice should be regulated and conducted by skilled processing staff be- fore the product reaches the consumer (Shumway & Cembella 1993; Curtis et al. 2000). Our results verify that strict regulatory and n-)onitoring regimes should remain compulsory for the safe marketing of "roe-on" scallops. However, when gonad toxicities are greater than the regulatory limit, discarding of tissues that selectively sequester the DA toxin may provide an effective strat- egy to enable the marketing of adductor muscle, in conformity with the domestic "roe off" market of the United States and Canada (Bricelj & Shumway 1998). The concentration of DA in gonad tissue varied by an order of magnitude (range 0.13-75.5 p-g DA g"'). Thus, if gonads with high toxicities were to be included in pooled samples, they could potentially elevate toxin levels significantly. This may explain why monitored toxicity at certain sites seemed to oscillate throughout the winter period (FSA pers. comm. 1999). Consequently, a large number of indiv iduals should be included in composite samples to reflect n-iean population toxicity accurately. However, in species where toxicities are extremely variable, it is the consensus that monitoring tissues on an individual basis proves more informative ill developing mitigating strategies for harmful algal bloom man- agement. Curtis et al. (2000) were able to propose site-specific recomiTiendations for management, on the basis of large differ- I ASP IN P. MAXIMUS 83 TABLE 5. Concentration of DA in three Pseiido-iiilzschia cultures established from the l'W» ASF e\ent (pK DA cell' of intracellular and extracellular fractions and conihined total). Cultures harvested at three «eeks. Domolc acid content (pg ccir'i Species Intracellular Extracellular (supernatant! Total (combined) P. uustralis (isolate 1) P. australis (isolate 2) P. pungens 1.32 1.20 nd 2.95 2.14 nd 4.27 3.39 ().()() nd = not detected. ences in PSP toxicity among geoduck clams. Panope abrupla. of different depths and harvest tracts. Data describing individual vari- ability of gonad toxicity within localities allow subpopulations with a low frequency of individuals of elevated gonadal toxicity to be distinguished (as seen in the current study); therefore, they permit evaluation of the level of risk, gonad tissue from specific locations with respect to its rate of consumption, poses to human health. The use of risk assessment models should be considered to assess scallop toxicity with respect to rate of consumption by hiunans. to continue to maintain public safety standards while at the same time ensuring optimum utilization of the high-quality king scallop resource. ACKNOWLEDGMENTS We thank Ms N. Lundholm (University of Copenhagen. Den- mark) for the initial identification of Pseudo-nazschia species and the members of the Scallop Association for their invaluable sup- port and assistance. This study was funded by The Scallop Asso- ciation. The Highlands and Islands Enterprise. PESCA. and The Highland Council. Arevalo. F. F.. M. Bermudez de la Puente. & C. Salgado.l99S. ASP tox- icity in scallops: individual variability and tissue distribution. In: B. Reguera. J. Blanco. M.L. Fernandez. & T. Wyalt. editors. Harmful algae. Xunta de Galicia and Intergovernmental Oceanographic Com- mission of UNESCO, pp. 499-502. Bates. S. S. 2000. Domolc acid producing diatoms: another genus added' J. Phycol 36:978-85. Bates. S. S., A. S. W. De Freitas R. A. Foxall.M. Gilgan.L. A. Hanic.G. R. Johnson.A. W. McCuUock.P. Odense.R. Pocklington.M. A. Quilli- am.P. G. Sim.J. C. Smith.D. V. Subha Rao.E. C. D. Todd. J. A. Walter. & J. L. C. Wright. 1989. The pennate diatom Niizschia pungens as the primary source of domoic acid, a toxin in shellfish from eastern Prince Edward Island. Can. J. Fish. Ac/ual. Sci. 46:1203-1215. Bates, S. S., D. L. Garrison & R. A. Homer. 1998. Bloom dynamics and physiology of domoic acid producing Pseudo-nitzschia species. In: D. Anderson. A.Cembella. & G. Hallegraeff. editors. Physiological ecol- ogy of harmful algal blooms. NATO-ASI series G: Ecological sciences, vol. 41. Berlin: Springer Bourne, N. 1965. Paralytic shellfish poisoning in sea scallops iPlacopecWn mugelUmicus. Gmelin). J. Fish. Res. Board Can. 22:1 137-1 149. Bricelj, V. M. & S. E. Shumway. 1991. Physiology: energy acquisition and utilization. In: S. E. Shumway. editor. Scallops: biology, ecology, and aquaculture. New York: Elsevier/North Holland, pp. 305-346. Bricelj, V. M. & S. E. Shumway. 1998. Paralytic toxins in bivalve mol- luscs: occurrence, transfer kinetics, and biotranformation. Rev. Fish. Sci. 6:315-383. Cembella. A. D., S. E. Shumway & R. Larocque. 1994. Sequestering and putative biotransformation of Paralytic shellfish toxins by the scallop Ptacopeclen mugellanicus: seasonal and spatial scales in natural popu- lations. J. Exp. Mar. Biol. Ecol. 180:1-22. Cembella. A. D.. S. E. Shumway & N. I. Lewis. 1993. Anatomic and spatio-temporal variation in paralytic shellfish toxin in two bivalve species from the Gulf of Maine. / Shellfish Res. 12:389-403. Curtis. K. M., V. L. Trainer & S. E. Shumway. 2000. Paralytic shellfish toxins in geoduck clams {Panope ahrupui): variability, anatomical dis- tribution, and comparison of two toxin detection methods. J. Shellfish Res. 19:313-319. Denton. W. 1999. Proposals for dealing with the algal toxin problem that has hit the scallop industry. Sea Fish Executive Report. Hull: Sea Fish Industry Authority. Douglas. D. J.. E. R. Kenchington. C. J. Bird. R. Pocklington. B. Bradford & W. Silvert. 1996. Accumulation of domoic acid by the sea scallop LITERATURE CITED [Placopeeten magellanicus) fed cultured cells of toxic Pseiido- ninschia multiseries. Can. J. Fish. Aqual. Sci. 54:907-913. Drum. A. S., T. L. Siebens. E. A. Crecelius & R. A. Elston. 1993. Domoic acid in the Pacific razor clam. Siliqua patitla (Dixon. 1789). / Shellfish Res 12:443-430. Garrison. D. L., S. M. Conrad. P. P. Eilers & E. M. Waldron. 1992. Confirmation of domoic acid production by Pseudonilzschia australis (Bacillariophyceae) cultures. J. Phycol. 29:418-^26. GuiUard. R. R. L. & J. H. Ryther. 1962. Studies of marine planktonic diatoms. 1. Cyclotella nana Hustedt. and Deionida confenalea (Cleve) Gran. Can. J. Microbiol. 8:229-239. Hallegraeff, G. M. 1995. Harmful algal blooms: a global overview. Manual on harmful marine microalgae. Intergov. Oceanogr. Comm. New York: UNESCO. Kotaki, Y., K. Koike, M. Yoshida, C. V. Thuoc. N. T. Huyden. N. C. Hoi. Y. Fukuyo & M. Kodama. 2000. Domoic acid production in Nitzschia isolated from a shrimp-culture pond in Do Son Vietnam. ./. Phycol. 36:1057-1060. Lundholm. N. & 0. Moestrup. 2000. Morphology of the marine diatom Nitzschia navis-varingica sp. nov., another producer of the neurotoxin domic acid. J. Phycol. 36:1 162-1 174. Lundholm. N.. J. Skov. R. Pocklington & 0. Moestrup. 1994. Domoic acid, the toxic amino acid responsible for Amnesic shellfish poisoning, now in Pseudo-nilzschia seriata (Bacillariophyceae) in Europe. Phyc(dogia 33:457-478. Mason. J. 1983. Scallop and queen fisheries in the British Isles. Buckland Foundation. Fishing News Books. Novaczek. I.. M. S. Madhyastha. R. F. Ablett. A. Donald. G. Johnson. M. S. Nijjar & D. E. Sims. 1992. Depuration of domoic acid from live blue mussels (Myrihis ednlis). Can. J. Fish. Ac/ual. Sci. 49:312-318. Quilham. M. A.. P. G. Sim. A. W. McCulloch & A. G. Mclnnes. 1989. High performance liquid chromatography of domoic acid, a marine neurotoxin, with application to shell-fish and plankton. Int. J. Environ. Anal. Chem. 36:139-154. Rhodes. L., D. White. M. Syhre & M. Atkinson. 1996. Pseudo-nitzschia species isolated from New Zealand coastal waters: domoic acid pro- duction in vivo and links with shellfish toxicity. In: Yasumoto.T.. Oshima, Y.. & Fukuyo. Y.. editors. Harmful and toxic algal blooms. Intergov. Oceanogr. Comm. New York: UNESCO, pp. 631-636. Scholin, C. A.. F. Gulland. G. J. Doucette, S. Benson. M. Busman. F. P. Chavez. J. Cordaro. R. Delong. A. De Vogelaere. J. Harvey. M. Hau- lena. K. Lefebvre. T. Lipscomb, S. Loscutoff. L. J. Lowenstine. R. 84 Campbell et al. Marin III. P. E. Miller. W. A. McLellan. P. D. R. Moeller. C. L. Powell. T. Rowles. P. Silvagni. M. Silver. T. Spraker. V. Trainer & F. M. Van Dolah. 2000. Mortality of sea lions along the central California coast linked to a toxic diatom bloom. Nature 403:80-84. Shumway. S. E.. R. Selvin & D. F. Schick. 1987. Food resources related to habitat in the scallop PUicopeclen magellinicus (Gmelin. 1791): a qiialilative study. J. Shellfish Res. 6:89-95. Shumway. S. E. & A. D. Cembella. 1993. The impact of toxic algae on scallop culture and fisheries. Rev. Fish. Sci. 1(2):121-150. Todd. E. C. D. 1993. Domoic acid and amnesic shellfish poisoning: a review. / Food Pwt. 56:574-576. White. A. W., S. E. Shumway. J. Nassif & D. K. Whittaker. 1993. Varia- tion in levels of paralytic shellfish toxins among individual shellfish. In: T. J. Smayda & Y. Shimizu. editors. Toxic phytopUmkton blooms in the sea. New York: Elsevier/North Holland, pp. 353-358. Wnght. J. C. L.. R. K. Boyd. A. S. W. De Freitas. M. Falk. R. A. Foxall. W. D. Jamieson. M. V. Laycock. A. W. McCullock. A. G. Mclnnes, P. Odense. V. P. Pathak. M. A. QuiUiam. A. Ragan. P. G. Sim. P. Thiba- ult. J. A. Walter. M. Gilgan. D. J. A. Richard & D. Dewar. 1989. Identification of domoic acid, a neuroexcitatory amnio acid, in toxic mussels from eastern Prince Edward Island. Can. J. Chem. 67:481- 490. .hmniiil of Shelljhh Research. Vol. 20. No. 1. X5-XX, 2001. STRANDING OF SCALLOPS RELATED TO EPIPHYTIC SEAWEEDS ON THE COAST OF NORTHERN CHILE SERGIO A. GONZALEZ, WOLFGANG B. STOTZ, AND MARCELO AGUILAR DeparUimenlo de Biologia Marina. Faciihad tie Ciencias del Mar, Lhuvcrsidad Caldlica del Norte, Ca.'iilla 1 17, Coquimbo, Chile ABSTRACT More than two million scallops lAri;oi)ecieii piirpunmis Lamarck) were stranded on the beach at Tongoy Bay, northern Chile (30°20'S) in March 1999 during a period of unusually strong wave action. The scallops, which were stranded in living condition, had frondose algae attached, principally Ulva lacnica Linneaus, to the upper valve. Settlement and growth of the algae on the scallop shells facilitated the stranding due to the large surface area thus exposed to water movement. A preponderance of larger size classes of scallops fouled with algae were stranded, suggesting thai the epibionts decreased the chance of "escape" from the effects of water movement in relation to the size of the individuals. The stranding represented about 20"^^ of the population of the bed. It is suggested that nearshore areas were mainly affected, where light conditions promoted the rapid grov\ th of the epiphytic seaweeds. KEY WORDS: scallop stranding, algal epiphytism, mass mortality. Chile INTRODUCTION Strandini; and mass mortalities on the shore are frequent oc- currences and affect distinct groups of marine organisms. There is a great deal of information on the stranding of marine mammals and turtles (Raga et al. 1Q91. Blochet al. 1997, Berrow & Rogan 1997). However, in the case of marine invertebrates, the literature is limited, particularly as it relates to Chile. For invertebrate strandings, many different causes are suggested. Natural mortality is cited for the strandings of the cnidarian Velella velella in Oregon (Kemp 1986). Some strandings are selective and affect a fraction of the population, as reported for the horseshoe crab (.Limuhis in>l\ptun\iis). in which stranding is associated with senescence and parasitism in some individuals (Penn & Brockman 1995). In other cases it is related to changes in the environment, such as the mortalities produced by the El Nino phenomenon. This phenom- enon has been implicated in strandings of the pelagic crab {Pleu- roncodes planipes) (Aurioles et al. 1994), as well as mollusks (Mesode.sma donacium) and brown algae of the genus Macrocystis (Amtz & Valdivia 1985. Amtz et al. 1985). Major storms and high tides may cause stranding, as in the case of the lobster iHiiinaru.': aincriccimis) on the coast of Prince Edward Island, Canada (May- nard & Chiasson 1988) and the yellow clam (M. mactroides). of which millions were stranded in Argentina (Oliver et al. 1971). Other strandings include those related to pollution, such as oil spills (Dyrynda et al. 1997). Strandings of scallops, resulting from strong winds and storms or due to consequences of disease, parasitism, or senescent states, have been reported (Tettelbach 1985, Orensanz et al. 1991 ). Tettel- bach ( 1991 ) reported strandings of the scallop Ar.i;opecteii irradi- ans (Lamarck, 1819) in the Poquonock River in Connecticut, which was as.sociated with strong wave and tidal currents. If the scallops are stranded live, they die due to desiccation, bird preda- tion, and human harvesting. Tettelbach (1991) included the anecdotal case of a stranded scallop having an abundant covering of the alga (Codium) on its shell. However, this is not an isolated phenomenon, as shown by the report of Kelley and Kirby (1981). who related an extensive stranding of the bay scallop on Nantucket Island. Massachusetts due to Codium fragile. Galtsoff (1964) called the epiphytic Co- dium "oyster thieves." Ansell et al. (1988) related a major strand- ing of the sandy bottom clam Donax vinatus (da Costa) in Dor- nach, Scotland, to the growth of algae on their shells. Similarly, Orensanz et al. (1991) mentioned that the risk of stranding in Chlamys telntelcha (Patagonia, Argentina) was increased by the presence of epibionts. The present report describes a stranding of the scallop A. piir- piinitiis (Lamarck) on the coast of northern Chile, produced by swell-induced waves, and the facilitation of the stranding by the presence of epibiotic algae on the scallops. MATERIALS AND METHODS Slud\ Area Observations were made at Puerto Aldea (30°17'S. 71°36'W), a small locality at the extreme southern end of Tongoy Bay, which is protected from the predominant southwest winds by Lengua de Vaca Point (Fig. 1). However, with noilh winds it experiences a major swell. Off Puerto Aldea, the bottom is composed of sand and rocky areas and a sector covered by the seagrass (Heterozostem tasinaiuca). which forms a bed of approximately 1 .5 hectares (Phillips et al. 1983, Gonzalez 1992). The scallop {A. piirpitratiis) is distributed over the entire Puerto Aldea area (Stotz & Gonzalez 1997), which is protected and managed by local fishermen. Methods During a period of unusually high swells on the northern coast of Chile in March 1999. 221 stranded scallops (-4. piirpiiratiis) were collected from a strip of beach at Puerto Aldea. The shell height (dorsoventral length) of the scallops was measured to ±1 mm. The total fresh weight of the scallops was obtained using the relationship of size to weight determined for this region by Stotz and Gonzalez (1997): Weight (g) = 2.897 x 10-" x Heiaht The occurrence of algae attached to the shells of each indi- vidual was recorded, both in cases where the entire thallus was present or where there was evidence of algae (e.g., holdfast). Given the characteristics of the algae encountered, our work cen- 85 86 Gonzalez et al. Puerto Aldea Figure I. The location of the scallop stranding: at Tonsov Hay. northern Chile. tered on the most abundant aiga. L'lvn hutma Linneaus (1753). Algae were measured for maximum length from adhesive disc to the end of the longest frond (±1 mm), and wet weights were measured to ±0.1 g on a portable balance. Algal surface area was determined using the equation 1 g wet weight of U. lactuca = 89 cnr, obtained by weighing 30 algal pieces cut to 16 cm" (0.18 g). An estimate of the magnitude of the stranding was obtained by personal interviews with fishermen at Puerto Aldea. Their deter- mination was made on the basis of the number of scallops that they collected and from their estimates of the numbers of scallops re- moved from the beach by people from outside the Puerto Aldea community. RESULTS Stranded Scallops Sixty-five percent of the stranded scallops bore an algal thallus on the shell, with a further 20% showing the presence of a holdfast or a partial algal thallus. The remainder of the scallops had no raacroalgae attached to their shells. Although a wide size range of scallops were stranded (Fig. 2). most of the scallops collected ranged from 40-80 mm in shell height. Scallops with algal epibionts were somewhat larger than those without epibionts (Table I ). A median / test showed signifi- cant differences in the sizes of the scallops in these groups {t value,7,, = 5.1338, P < 0.001). Characteristics of Algal Epibionts Algae on the scallops included principally U. lactuca. with the occurrence of only seven individuals of Ciyptoncmia ohovata (J. Agardh. 1876) and one of the brown alga Desmarestia ligulata (Lightfoot) (Lamouroux. 1813). Fronds of Ulva on the scallop shells had a broad range of sizes. 25 ?20 ri 5 15- " z " ■ lU §10 1" lU K PI >*• 5 " 0 , , rTTi rrl 1 1 1 1 1 1 1 1 1 Mlnlll 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 SHELL HEIGHT (mm) Figure 2. Size-frequency distrihution of Argopecten piirpuratus stranded at Puerto Aldea, Chile in March 1999. Algal-Induced Scallop Stranding 87 TABLE 1. Percentage of different sUe groups of scallops (Argopccten purpuratus) stranded at Puerto Aldea, Tongoy Bay, Chile, in March 1999. Range of sizes (shell height. ninil Total 0-20 21-40 41-60 61 -80 81-100 Kpibiont status n % n % // '7, // % n % II % With algae Without algae 190 33 80.2 19.8 0 — 0 — 3 2 60 40 89 23 79,5 20.5 81 8 91 9 1 100 0 — ranging from 34-971 mm in length, with an average length of 378 mm (Fig. 3). Fronds frequently exeeedeiJ 170 mm in length. The surface area of the fronds ranged from 98-6,803 cnr. and the weight ranged from 1.1-76.1 g, Scallop-Algae Relationship A tendency was observed for the algae on the scallop shells to be larger on the larger sizes of scallop. Although there was a broad dispersion of values, this relationship was statistically significant via linear regression (Fonsn.i.v,) = 46,5, P < 0,001) (Fig, 4), Magnitude of the Stranding Because of their commercial value, the scallops did not accu- niLilate on the beach; they were har\ested by a great number of people and taken away for sale. A survey of 43 persons at Puerto Aldea who harvested scallops produced a total of 2 1 7.550 scallops. This harvesting extended over the 4 days during which the strand- ing occurred. Estimates by the fishermen's association of Puerto Aldea indicated that about 1.700 other people visited the area of the stranding and removed a total of about 2 million scallops. On the basis of the sizes of scallops stranded, we estimated the total fresh weight of scallops removed to be about 101) metric tons. DISCUSSION The higher incidence of stranded scallops bearing algal epi- bionts than those without epibionts demonstrated the importance of algae to the stranding of scallops during the period of heavy swell. Observations made by diving after the passage of the strand- ing event (M. Aguilar, pers, observ.) showed an almost complete absence of algae-bearing scallops remaining in the bed. suggesting that essentially all epibiont-affected scallops had been beached by the waves. It was probable, due to the light requirements of the green algae, that the beached scallops came mainly from shallow areas near the shore, which would be more subject to water move- 14 12 b 10 >- O 8 ui 2 6 Icllhn Elm m 10 90 170 250 330 410 490 570 650 730 810 890 970 MAXIMUM LENGTH OF FROND (mm) Figure 3. Size-frequency distribution of algal thalli il'lva lactuca) at- tached to Argopeeten purpuratus stranded at Puerto Aldea, Chile. ment. However, it is also possible that the swell ripped off the epiphytic Ulva fronds from the shells of the scallops. Although the stranding was associated with heavy swells that affected Puerto Aldea when winds of about 10 m/s and waves of about 2 m in height occurred for several hours (Chilean Navy data), the stranding was facilitated by the presence of algae on the scallop shells. For example, an alga 50 cm in length with a surface area of 1.700 cm" on a scallop 6 cm in height represents an in- crement of 40x the surface area of this scallop normally exposed to water movement. The fact that large scallops not bearing epibionts were gener- ally absent froin the stranding suggests that these scallops were able to escape the effects of water movement by virtue of their size (weight), whereas scallops with algal growth on the shells were transported to the shore. This concurs with the observations of Witman and Suchanek (1984) that the presence of algae such as Laminariii saccharina and other epibiota on the shells of Mytihis edulis and M. califoniicinus significantly increased their risk of being dislodged by currents. The stranding may be related to climatic alterations related to the El Nifio phenomenon, both in reference to the heavy swells and the abundance of green algae occurring in this period. Also, changes in sea state that produce unusual wave action may be accompanied by rises in sea surface temperature associated with El Nino currents and provide improved conditions for algal growth. For example. Arntz et al. (1985) reported blooms of frondose green algae, which grew rapidly during the El Nifio event of 1982- 83 and dominated rocky shallows on the coast of Peru. An evaluation of the scallop bed at Puerto Aldea eairied out 2 mo after the stranding (Stotz & Gonzalez, unpub. technical report) .showed the presence of about 8 million scallops. The stranding ouuu y=24.389x + 238.05 R^ = 0.2787 ••• « E 6000 o < ^ Lit f^ Q ? 4000 * * ♦ |3 2000 0 * 0 20 40 60 80 100 120 140 TOTAL FRESH WElGHT(g) Argopeeten purpuratus Figure 4. Relationship between size of stranded Argopeeten purpura- tus and the surface area of algae {I'lva lactuca) attached to their shells. ***Linear regression was signiilcant at /' < 0.001. 88 Gonzalez et al. therefore represented a loss about 209c of the population (previous to the swell period) from this bed. However, in addition to the observed effeets of the stranding of scallops on the natural popu- lation, this event could have a negative effect on the perception of the local fishermen about the benefits of protecting the resource. Nature took away almost 20% of their possible income. Although the stranded scallops were harvested, it was mainly by foreign people, not by members of the Puerto Aldea community. This needs to be esaluated. ACKNOWLEDGMENTS The authors are grateful for collaboration rendered by the fish- ing community of Puerto Aldea, who aided in collecting informa- tion on the numbers of scallops stranded and harvested. We also thank M. Valdebenito and D. Lancellotti of the Ecology and Re- source Management Group (UCN) for their helpful comments and Louis DiSalvo for translation and editorial comments. An anony- mous referee is thanked for providing \aluable suggestions. LITERATURE CITED Ansell. A. D.. L. Robh & H. T. Powell. 198S. Algal-induced dislodgement as a cause of bivalve mortality on some Scottish beaches. J. Mm: Biol. Assoc. U.K. 68:219-233. Arntz. W. A. Landa & J. Tarazona. 1985. "El Niino": Su impacto en la fauna marina. Boletin Instituto del Mar. Peni - Callao. Volumen e\- traodinario. pp. 1-222. Arntz. W. E & E. Valdivia. 1985. Incidencia del femimeno "El Niiio" sobre los mariscos en el literal peruano. In: W. Arntz. A. Landa & J. Tara- zona. editors. "El Nifio"; Su impacto en la fauna manna. Callao. Peru: Boletin del Instituto del Mar. Volumen e.xtraordinario. pp. 91-101. Auricles. D., M. I. Castro & R. Perez. 1994. Annual mass strandings of pelagic red crabs. Pleuroncodes planipes (Crustacea: Anomura: Galat- heidae). in Bahia Magdalena. Baja California Sur. Me.xico. Fhliery Bull. 92(2):464-470. Berrow. S. D. & E. Rogan. 1997. Review of cetaceans stranded on the Irish coast. 1901-95. Mammal. Rev. 27(l):5l-76. Bloch. D.. G. Desportes. A. Petersen & J. Sigurjonsson. 1996. Strandings of striped dolphins iSlenella coeruleoalba) in Iceland and the Faroe Islands and sightings in the Northeast Atlantic, north of 50 degree N latitude. Mar. Mammal. Sci. I2(I):I25-132. Dyrynda. E. A.. R. J. Law. P. E. J. Dyrynda. C. A. Kelly. R. K. Pipe. K. L. Graham & N. A. Ratcliffe. 1997. Modulations in cell-mediated immunity of Myriliis ediilis following the Sea Empress oil spill. J. Mar. Biol. Assoc. U.K. 77(l):28l-284. Galtsoff. P. S. 1964. The American oyster Crassostrea virginica (Gemelin). Fishery Bulletin of the U.S. Fish and Wildlife Service. 64:l-f80. Gonzalez, S. A. 1992. Heterozoztera tasmanica (Martens ex Aschers.) den Hartog y comunidad asociada en el norte de Chile. Tesis de licencia- tura. Universidad Catolica del Norte. Facultad de Ciencias del Mar. 120 PP- Kelley, K. M. & M. Kirby. 1981. Codimn fragile and its effect on the Nantucket bay scallop industry. In: K. M. Kelley. editor. The Nantucket Bay scallop fishery: The resource and its management. Nantucket: Shellfish Marine Department, pp. 35-39. Kemp. P. F. 1986. Deposition of organic matter on a high-energy sand beach by a mass stranding of the cnidarian Vclella vclella (L.). Estua- niie. Coastal Shelf Sci. 23(4):575-579. Maynard. D. R. & Y. Chiasson. 1988. Storm related mortality of lobsters. Himiarus americamis, on the northern shore of Prince Edward Island. Canada. J. Shellfish Res. 7(1): 169. Olivier. S. R.. D. A. A. Capezzani. J. I. Carreto. H. E. Christiansen. V. J. Moreno. J. E. .'Mspun de Moreno & P. Penchaszadeh. 1971. Estructura de la comunidad. dinamica de la poblacion y biologia de la almeja aniarilla {Mesodesma mactriodes Desh. 1854) en Mar Azul. (Pdo. Graf Madariaga. Bs. As.. Argentina). Proyecto Desarrollo Pesquero FAO Serie. Informe Tecnico. 27:1-90. Orensanz. J. M.. A. M. Parma & O. Iribame. 1991. Population dynamics and management of natural stocks. Dev. Aquacuh. Fish. 21:625-713. Penn. D. & H. J. Brockmann. 1995. Age-biased stranding and righting in male horseshoe crabs. Limiihis polyphemiis. Animal Behaviour 49(6): 1531-15.^9. Phillips. R. C. B. Santehces. R, Bravo & C. P. McRoy. 1983. Hererozo- slera tasmanica (Martens ex Asch.) den Hartog in Chile. .Aquatic Botany 15:192-200. Raga. J. A.. A. Raduan. J. A. Balbuena. A. Aguilar, E. Grau & A. Borrell. 1991. Varamientos de cetaceos en las costas espafiolas del Mediterra- neo durante de periodo 1982-1988. Miscelania Zoologica 15:215-226. Stotz. W. B. & S. A. Gonzalez. 1997. Abundance, growth, and production of the sea scallop Argopecten piirpuratus (Lamarck 1819): ba.ses for sustainable exploitation of natural scallops beds in north-central Chile. Fish. Res. 32:173-183. Tettelbach. S. T. 1985. Seasonality of factors responsible for mortalitv of the northern bay •icaWop Argopecten irradians irradians (Lamarck). J. Shellfish Res. 5(I):43. Tettelbach. S. T. 1991. Seasonal changes in a population of northern bay scallops. .Argopecten irradians (Lamarck. 1819). In: S. Shumway & P.A. Sandifer. editors. Worid aquaculture workshops. I. (eds). Baton Rouge: The World Aquaculture Society, pp. 164-175. Witman. J. D. & T. H. Suchanek. 1984. Mus.sels in flow: drag and dis- lodgement by epizoans. Mar. Ecol. Prog. Ser. 16:259-268. Jdiinuil of Shellfish Research. Vol. 20, No. 1. S9-96. 20U1. REPRODUCTION, POPULATION STRUCTURE, AND RECRUITMENT LIMITATION IN A BAY SCALLOP {ARGOPECTEN IRRADIANS LAMARCK) POPULATION FROM NEW JERSEY, USA PAUL A. X. BOLOGNA,' * AMI E. WILBUR," ** KENNETH W. ABLE' ^Institute of Marine and Coaslal Sciences. Rutgers Universin Marine Field Station. 132 Great Bay Blvd.. Tuckerton. New Jersey 08087-2004: 'Florida Marine Research Insiitulc. 100 8"' Ave S.E.. St. Feterslnirii. Florida 33716 ABSTRACT A bay scallop. Argopeclen irrudians irradians (Lamarck 1819). population was studied in Little Egg Harbor. New Jersey. USA to investigate the distribution, reproduction and genetic structure of the population because virtually nothing is known about this species in this region. Argopeclen irradians irradians densities varied among eelgrass (Zoslera marina Lamarck) beds in 1998 (range 0.12-1.0 individuals per 10m transect), but were virtually absent from the region in 1999. The absence of adults in 1999 may have been due to limited spat recruitment and survival in the fall of 1998. which may be attributed to a reduction in above ground biomass of Z marina habitat. Genetic analysis (using mtDNA) indicated that this New Jersey population was intermediate between New York and North Carolina populations. Based on these data and the historical landings of bay scallops in New Jersey, it is probable that a small self-seeding population exists. Assessment of reproductive cycles dunng 1998 showed two potential peaks in reproductive condition: one occurred during June when a ma.ximuni Gonadal-Somatic Index was recorded (24.7"*) and the second was during October when a visual gonadal condition index indicated a majority of scallops in post-spawn condition. Few adult bay scallops were encountered in 1999 (;i = 8) to assess reproductive cycles, but two large individuals collected dunng July showed a post-spawn gonadal condition. However, during both 1998 and 1999 settling juveniles « 15 mm shell height) were only recorded in October, suggesting that recruitment to the population during these years resulted from the late summer-early fall spawn. Given the variability observed over the two years, future studies should concentrate on factors influencing inter-annual variation in abundance of this New Jersey population. KEY WORDS: bay scallop. Argopeclen irradians. reproduction, genetic structure, recruitment INTRODUCTION The buy scallop [Argopeclen irradians) was once a common and often abundant member of shallow marine communities along the Atlantic and Gulf coasts of the United Stales and a prized commercial and recreational shellfish. Clarke (1965) identified three distinct sub-species of A. irradians based on morphological characteristics, but recent morphological and genetic studies have suggested a fourth (Blake and Graves 1995. Marelli et al. 1997. Wilbur and Gaffney 1997). The northern sub-species. ,4. /. irradi- ans ranges from Massachusetts to New Jersey where it is thought to intergrade with the southern subspecies A. i. concentriciis. Ar- gopeclen irradians ccmcentricus was originally described to range along the mid-Atlantic coast and into the Gulf of Mexico (Clarke 1965. Waller 1969). Recent studies, however, have shown that bay scallops from the Florida gulf coast and from North Carolina are as different from one another as are the northern and southern sub- species on the Atlantic coast, indicating that these populations should be considered distinct subspecies (Blake and Graves 1995, Marelli et al. 1997. Wilbur and Gaffney 1997). The fourth sub- species. A. i. amplicostaws. is occasionally reported from the gulf coast of Texas and extends to an undefined southern limit south of the Yucatan Peninsula. This sub-species has been described solely on the basis of morphological analysis and the scarcity of animals in recent years has constrained the incorporation of representatives in the aforementioned genetic surveys. A preliminary analysis of *Corresponding author. Current address: Department of Biological and Allied Health Sciences. Fairleigh Dickinson University, M-EC-01. 285 Madison Ave.. Madison. NJ 07940. Tel.: -hi -973-443-8758; Fax: -1-1-973- 443-8766; E-mail: bologna@fdu.edu **Current address: Department of Biological Sciences and the Center for Marine Science. LIniversity of North Carolina-Wilmington. 601 South College Rd. Wilmington. NC 28403 the mitochondrial DNA variation does not support the separation of A. /. aniplicostaliis from other Gulf piipulations despite the clear demonstration of genetically determined differences in morphol- ogy (Wilbur 1995, Wilbur and Gaffney, 1997). As such, the taxo- nomic status of the western gulf form of A. irradians is unclear. Bay scallops are intimately tied to seagrass beds, which they use as a primary settlement site (Gutsell 1930. Eckman 1987). Specifically, scallops settle and cling to blades via byssal threads until they are too large to remain suspended (Thayer and Stuart 1974). During this life stage, the seagrass canopy provides protec- tion from benthic predators iPohle et al. 1991 ). However, reduced growth rates of juvenile scallops climbing higher on blades (Am- brose and Irlandi 1992), suggests that this behavior may represent a trade-off between growth and mortality. Ultimately, recruitment to the adult population may be determined by predation at this juvenile stage (Strieb et al. 1995). Because bay scallops recruit to seagrasses and use them as attachment sites (Thayer and Stuart 1974), the loss of habitat during the eelgrass (Zostera marina) wasting disease in the 1930s (see den Hartog 1987) is thought to have severely limited populations in many regions, while elimi- nating them from others. Although eelgrass has made recoveries in the subsequent decades (den Hartog 1987), bay scallops have not returned in significant numbers to many areas where they once were abundant. In coastal New Jersey, eelgrass has returned and is relatively abundant in shallow water, yet in the last several decades bay scallop densities remain below a fishable population size (Ford 1997). Historically, bay scallops were abundant and commercially fished in New Jersey, USA. The first available landing records were collected in 1956. when 52.300 bushels were harvested with an estimated value of $287,000. Continued success of scallop populations for the next 12 years yielded 317.000 bushels valued at over $1 million (Ford 1997). Subsequently, commercial bay scallop harvests were only recorded for 1973 and 1974. Despite its 89 ■ 90 Bologna et al. local importance, little information exists on the ecology and popu- lation strticture of New Jersey bay scallops, particularly since the collapse of the commercial fishery. In recent years it was generally felt that bay scallops no longer occurred in New Jersey waters. The observation of numerous scallops in 1998 in Little Egg Harbor. New Jersey prompted questions regarding the density and repro- ductive periodicity of this population, as well as its origin. New Jersey is thought to be the point of contact between the A. irradians and A. i. concentriciis sub-species, and the return of scallops suggests recruitment of larvae from extant populations north or south of New Jersey, or from a small, remnant local population. It was the goal of this research to investigate the dis- tribution, reproduction and genetic structure of New Jersey bay scallops. STUDY SITE Investigations were conducted during 1998 and 1999 in Little Egg Harbor, New Jersey. U.S.A. (39\VS'N. 74°14'W; Fig. I). which is located in the central portion of the Mid-Atlantic Bight. Little Egg Harbor is a relatively unimpacted region of coastal New Jersey and is part of the Jacques Cousteau National Estuarine Research Reserve (Psuty et al. 1993). It is a polyhaline estuary protected by a barrier island. It is relatively shallow (average depth at MLW = 1.7 m, Durand 1984) and submerged aquatic vegeta- tion covers approximately 1,305 hectares of the bottom (Bologna et al. 2000). Seasonal water temperatures range from -2"C to 28°C (Able et al. 1992) with an average tidal range of about 0.7 m (Chizmadia et al. 1984). MATERIALS AND METHODS Population Assessment Argopecten irradians densities were assessed at four sites in Little Egg Harbor, New Jersey during May 1998 and 1999. Density was determined using lO-ni x l-m transects haphazardly lain out in shallow (< 1..5m depth) Zostera marina beds (Table 1, Figure 1. Little Egg Harbor. New .Jersey (39 35'N. 74 I4'\V) with locations of Marsh Elder and Hani Island and dominant vegetation types. Abbreviations associated with Ham Island refer to sites used for population density estimates: \VH = West Ham Island, NH = North Ham Island, NEH = Northeast Ham Island. New Jersey Bay Scallops Fig. 1): (I) West Ham Island (n = 16 transects. 1998; n = 10. 1999). (2) North Ham Island (/; = 17: /; = 10). (3) Northeast Ham Island (« = 14. ;( = 10). and (4) Marsh Elder Island (/i = 8, h = 10). They were sampled by snorkeling the length of the transect and collecting all scallops within 50 cm of the transect line (1-m width). Scallop abundance was compared among sites and between years using a two-way ANOVA with site and year as independent variables and scallop abundance as the dependent variable. The size structure of the population was assessed monthly from visu- ally located field-collected individuals during 1998. Size fre- quency distributions were generated for months in which at least 16 individuals were collected to assess the potential age structure of the population. Reproditctiiin Visual inspection of the gonad condition and calculation of a Gonadal-Somatic Inde.x (GSI) assessed scallop reproduction from collections on 10 dates from April 1998 to August 1999. Scallops were frozen and returned to the laboratory where shell height was measured to 0.05 mm and gonadal and somatic tissues were then dissected out. Visual condition of gonad material for scallops was determined for each individual and assessed as undeveloped, rip- ening. \ery ripe, or post-spawn following the protocol ot Bologna (1998). Reproductive and somatic tissues were then dried at 60°C for 72 hours and weighed (g dry weight). The GSI was calculated for each scallop using the following equation: GSI = (gonad dry weight/total dry weight) *100. Evidence of reproductive success was assessed by collection of small juveniles (<15 mm shell height). RecriiilimnI Habitat Assessment Characterization of bay scallop habitat was assessed by collec- tions of benthic cores in Zostera marina beds during March. April. May. June. July, and October 1998 to determine shoot density and plant biomass (/? = 6 cores/month). The coring device (15.24 cm diameter (0.01824 m")) was pushed into the substrate to a depth of 25 cm, capped, and removed from the sediment. Samples were frozen and returned to the laboratory. Shoot abundance was deter- mined and samples were separated into Z marina above ground (shoots) and below ground (rhizomes and roots) portions, and an algal-detrital fraction. Above ground Z. marina and algae-detritus were dried to constant weight at 80°C, then ashed at 500°C for eight hours to determine ash free dry weight (AFDW). Shoot abun- TABLE 1. Argopecten irradians density comparisons amon!> sites in Little K^;; Harbor. New Jersey (see Fig. 1) for 1998 and 1999. Density Density Site West Ham Island (WH) North Ham Island (NH) Northeast Ham Island (NEH) Marsh Elder Islands n 1998 « 1999 16 0. 1 2 ± (134 10 0±0 17 1 .00 ± 0.79* 10 0±0 14 0.21 ±0.42 10 0±0 8 0.2.=; ± 0.46 10 0±0 N indicates the number of transects conducted at each site during each year. Density values represent mean number of scallops encountered per lO-m transect + one standard error. * Represents significanUy greater scallop density lor North Hani Island site compared to others (P < 0.0001 ). 8 7 6 5 4 3- 2- 1- AprU N = 34 I = 49.2 mm 8 7 6- 5 4 3 2- 1- 8n 7. 6 5. 4 3 2 20- 15- le 0 8 !■ 6- 5- 4 3 2H 1 June N = 23 X = 51.2 mm September IN = 28 X = 55.8 mm October N = 108 \ = 59.6 mm 1998 November N = 16 X = 64.5 mm 1997 0 10 20 30 40 .SO 60 70 80 Shell Height (mm) Figure 2. Bay Scallop Size Frequency Distribution. Size distributions for 1998 collections based on scallop shell height measurements (mm). Sample size (Nl and average shell height (M given for each month. Note y-a\is scale change for October and the identified year classes for scallops represented in the sample. 92 Bologna et al. TABLE 2. Scallop reproductive potential: Visual gonad condition index. Mean shell Verv n height (mm) Undeveloped Ripe ripe Po.st-spawn 1998 April 2 43.0 2 0 0 0 June 13 52.3 0 3 10 0 July 10 50.0 1 6 3 0 August 4 55.6 0 2 2 0 September 15 55.7 1 13 1 0 October 20 58.6 0 0 2 18 November 16 64.3 0 0 0 16 1999 May 3 70.0 0 2 1 0 July 4 55.1 0 0 2 2 August 1 43.3 0 0 1 0 N indicates the number of Argopecten in-adiiius collected on each date tor reproductive assessment. Values in the table represent the number of individuals exhibiting each condition for a sample. dance data were square-root transformed and plant biomass data (Z marina and algae-detritus) were log transtormed before analy- sis. Shoot abundance, above ground biomass, and algae-detritus biomass were compared among months using a General Linear Model and significance testing was performed using an LSMEANS procedure with alpha = 0.05 (Littell et al. 1991). Genetic Stock Assessment Total DNA was extracted from 20 scallops collected in Little Egg Harbor during 1998 (PureGene extraction Kit. Centra Sys- tems. Inc) and amplified using primers specific for an 833bp frag- ment of the 12s ribosomal and NADH dehydrogenase I subunit regions of the mtDNA genome. Polymerase chain reaction (PCR) amplifications were carried out in 50|jil reaction volumes and were subjected to an initial denaturation step of 3 min at 94°C, followed bv 40 cycles consistinc of 30 sec at 94°C. 30 sec at .53°C and 1 min en -H B a OJ 0.25 0.2 .« 0.15 ■ E o 73 O 0.1 0.05 r+n rh rh rh n rh Apr/3 Jun/5 Jul/2 Aug/26 Sep/18 Oct/15 Nov/17 Date 1998 Figure 3. Seasonal Reproductive .\sse.ssment for 1998. Values repre- sent calculated mean (Gonadal Somatic Index ± Standard Error for scallops collected. See Tahle 2 for visual gonad condition index. at 72°C. Each product was digested with 9 restriction endonu- cleases {Alii I. Bun 11. Bgl I. fi.v/HKA I. Rsa 1. Hinfl. SciF I. aTa statistics Among regions Among population.s within regions Within populations 0.00553 0.00017 0.00361 59.42 1.79 38.79 <0.001 0.022 <0.00I 4)^, = 0.544 4.,, = 0.044 4),, = 0.612 Population samples were grouped into three regions (New York; New Jersey and North Carolina; Florida). 4> statistics measure haplotypic cor- relations and are analogous to the hierarchical F-statistics of Cockerhani (1969. 1973). P values indicate the probability of finding a more extreme variance component and <|)-statistic than that observed by chance alone. between these samples (Table ,3). This association is supported by the conventional interpretation of morphological data, which origi- nally described the southern sub-species from type specitnens col- lected from New Jersey. The genetic differentiation between New Jersey and North Carolina may be indicative of some restriction of genetic exchange, although the data presented here are insufficient to conclude that these populations are isolated from one another. The short larval duration of A. irradians (10-14 days) makes the probability of direct transport of larvae from North Carolina un- likely. While scallops are occasionally found in the inland bays of Maryland and Virginia, the ephemeral nature and low densities of these populations make them poor candidates as a source for larval export. The absence of evidence to support long distance larval transport in bay scallops is not uncommon. Arnold et al. (1998) found little evidence to support significant larval transport from "high" density populations (St. Joseph's Bay and Steinhatchee) in northwest Florida to "low" density populations further to the south (Crystal River and Anclote). Other studies in the sounds of North Carolina and the Peconic Bay system in New York show a similar lack of recruitment in areas decimated by toxic algal events from adjacent unaffected areas over much smaller spatial scales (Peter- son and Summerson 1992, Wenczel et al. 1993). As such, it is likely that the New Jersey population of bay scallops is a resident, albeit cryptic one. Assessment of recruitment habitat provided novel insight into the ecology of New Jersey bay scallops. During 1998, juvenile bay scallops were observed in the field during October. While this event alone was not unexpected, the fact that the two recruiting juveniles (shell height 6 and 12 mm) were both attached to adults via byssal threads is intriguing. This behavior was most likely their response to the lack of above ground eelgrass biomass (Fig. 4). This behavior has been observed for Antarctic scallops, Adamus- siuiii colln-iki (Berkman 1988). but observations of this occurring for A. irradians have not been recorded. Clearly, for an organism that is so intimately tied to seagrass as primary habitat, the loss of eelgrass signaled a significant loss of recruitment habitat. The loss of eelgrass also may have played a significant role in over-winter mortality. Field collections during November 1998 indicated that substantial passive burial of adults was occurring. Specifically, adults were frequently located within a recessed pit approximately 5-8 cm deep or adjacent to one with sediment fouling apparent on the dorsal shell. Although these observations were not quantified during this period, they suggest that sediment burial was occurring and iTiay have had significant impacts on winter survival. Tettel- bach et al. (1990) showed significant winter mortality due to burial for bay scallops in New York, and our observations of burial in 1998 and lack of adults in 1999 correspond to this trend. New Jersey Bay Scallops 95 Eelgrass not only serves as a primary recruitment habitat for bay scallops, but the beds also alter water velocity, dampen wave energy, and stabilize sediments (Fonseca et al. 1982. Fonseca & Fisher 1986). Consequently, the loss of eelgrass during the sum- mer and fall of 1998 (Fig. 4) may have had direct effects on the abundance of settling bay scallops and also indirect effects on the population, in that the lack of sediment stability may have led to increased winter mortality through burial. These losses in eelgrass biornass correspond to the reduction in bay scallop population density during 1999 (Table I ). but not their eliminatiini from the system. .Since the turn of the century, seagrasses have undergone dra- matic declines worldwide due to both natural and anthropogenic sources (Phillips 1982. Cambridge et al. 1986. Robblee et al. 1991 ). Over the la.st 25 years, several studies have investigated the distribution of seagrass from coastal New Jersey and have shown significant declines in total coverage (R. Lathrop pers. comm.. Bologna luipubl. data). These declines parallel the loss of the com- mercial and recreational fishery for bay scallops in New Jersey as well (Ford 1997). If continued loss of Zosleni marina habitat occurs, it may severely limit this population in the future. ACKNOWLEDGMENTS We would like to thank E. Duval, D. Craige, and S. Piotrowski for assistance in the laboratory and field. Genetic analysis was facilitated by the efforts of T. Jones and S. Campbell. Samples for genetic analysis were provided by W. Arnold (FL). C. Lund (FL). S. Tettelbach (NY) and T. Muiphey (NC). P. Bowers provided the base map of seagrass coverage in Figure 1 . Financial support for this project was provided through the Rutgers University Institute of Marine and Coastal Sciences (IMCS) and the Rutgers Univer- sity Marine Field Statitm. This is IMCS contribution No. 2001-10. LITERATURE CITED Able. K. W.. R. Hodcn. D. Wiiung & J. B, Duiaiid. IW2. Physical pa- rameters of the Great Bay-Mullica River estuary. Technical Report 92-06. New Brunswick. New Jersey: Institute of Marine and Coastal Sciences. Rutgers, the State University of New Jersey. 38 pp. Ambrose. W. G. & E. A. Iriandi. 1992. Height of attachment on seagrass leads to trade-off between growth and survival in bay scallop Ar- gopecten irradians. Mar. Ecol. Prog. Ser. 90:4.'i-.'i 1 . Arnold. W. S., D. C. Marelli, C. P. Bray & M. M. Harrison. 1998. Re- cruitment of bay scallops Argopecten irradians in Floridian Gulf of Mexico waters: scales of coherence. Mar. Ecol. Prog. Ser. 170:14-V 157. Barber. B. & N. Blake. 1983. Growth and reproduction of the bay scallop, Argopecten irradians (Lamarck) at its southern distributional limit. J. Exp. Mar. Biol. Ecol. 66:247-256. Berkman. P. A. 1988. Ecology of the circumpolar Antarctic scallop. Ada- mussium colbecl^i (Smith. 1902). Ph. D. Dissertation. University of Rhode Island. RI. 231 pp. Blake, S. G. & J. E. Graves. 1995. Mitochondrial DNA variation in the bay scallop, Argopecten irradians (Lamarck. 1819) and the Atlantic calico xaWop, Argopecten gil)l->us (Linneaus. 1758). J. Shellfish. Res. 14:79- 85. Bologna. P. 1998. Growth, production, and reproduction in bay scallops Argopecten irradians conceniricus (Say) from the northern Gulf of Mexico./ Shellfish. Res. 17:911-917. Bologna, P.. R. Lathrop. P. Bowers & K. W. Able. 2000. Assessment of submerged aquatic vegetation in Little Egg Harbor, New Jersey. Tech- nical Report 2000-1 1. New Brunswick, New Jersey: Institute of Ma- rine and Coastal Sciences. Rutgers, the State University of New Jersey. 30 pp. Bricelj, V., J. Epp & R. Malouf 1987a. Comparative physiology of young and old cohorts of bay scallop Argopecten irradians irradians (Lama- rck): mortality, growth, and oxygen consumption. J. E.\'p. Mar. Biol. Ecol 112:73-91. Bricelj, V.. J. Epp & R. Malouf. 1987b. Intraspecific variation in repro- ductive and somatic growth cycles of bay scallops Argopecten irradi- ans. Mar. Ecol. Prog. Ser. 36:123-137. Cambridge. M., A. Chiffings. C. Brittan. L. Moore & A. McComb. 1986. The loss of seagrass in Cockbum Sound. Western Australia. II. Possible causes of seagrass decline. Aquat. Bot. 24:269-285. Clarke, A. H. 1965. The scallop superspecies AeLjiiipecten irradians (Lamarck). Malacologia 2:161-188. Chizmadia. P., M. Kennish & V. Ohori. 1984. Physical descnption of Bamegat Bay. In: M. Kennish and R. Lutz, editors. Ecology of Bame- gat Bay. New Jersey. New York: Springer- Verlag. pp. 1-28. Cockerham. C. C. 1969. Variance of gene frequencies. Evolution 23:72- 84. Cockerham. C. C. 1973. Analysis of gene frequencies. Genetics 74:679- 700. Crenshaw. J.. P. Heffernan & R. Walker. |y9L Heritabllity of growth rate in the southern bay scallop. Argopecten irradians concentricus (Say 1822). J. Shellfish Res 10:55-63. den Hanog, C. 1987. 'Wasting disease' and other dynamic phenomena m Zostera beds. Acpiat. Bot. 27:3-14. Durand. J. 1984. Nitrogen distribution in New Jersey coastal bays. In: M. Kennish and R. Lutz. editors. Ecology of Barnegat Bay, New Jersey. New York: Springer- Verlag. pp. 29-51. Eckman, J. E. 1987. The role of hydrodynamics in recruitment, growth, and survival of Argopecten irradians (L.) and Anoinia simplex (D'Orbigny) within eelgrass meadows. J. Exp. Mar, Biol. Ecol. 106:165-191. Excoffier. L.. P. E. Smouse & J. M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: ap- plication to human mitochondrial DNA restriction data. Genetics 131: 479-491. Fonseca. M. S. & J. S. Fisher. 1986. A comparison of canopy friction and sediment movement between four species of seagrass with reference to their ecology and restoration. Mar. Ecol. Prog. Ser. 29:15-22. Fonseca, M. S., J. C. Zeiman. G. W. Thayer & J. S. Fisher. 1982. Influence of the seagrass Zo.stera marina, on current flow. Est. Coast. Shelf Sci. 15:. 35 1-364. Ford. S. 1997. History and present status of molluscan shellfisheries from Barnegat Bay to Delaware Bay. NOAA Technical Report. National Marine Fisheries Service. Vol. 127. pp. 1 19-140. Gutsell. J. S. 1930. Natural history of the bay scallop. Ball. U. S. Bar. Fish. 46:569-632. Littell. R. C. R. Freund & P. Spector. 1991. SAS System for Linear Models. Third Edition. Cary, NC: SAS Institute Inc. 329 pp. Marelli. D. C, M. K. Krause. W. S. Arnold & W. G. Lyons. 1997. Sys- tematic relationships among Florida populations oi Argopecten irradi- ans (Lamarck. 1819) (Bivalvia. Pectinidae). Nautilus. 110:31^1. Marshall. N. 1963. Mortality rates and life span of the bay scallop. Aequipecten irradians. Proc. Nat. Shellfish Assoc. 54:87-92. McElroy. D.. P. Moran. E. Bermingham & I Komfield. 1992. REAP: an integrated environment for the manipulation and phylogenetic analysis of restriction data. J. Hered. 83:157-158. Nei. M. 1987. Molecular Evolutionary Genetics. New York: Columbia University Press. Peterson. C. H. & H. C. Summerson. 1992. Basin-scale coherence of population dynamics of a exploited marine invertebrate, the bay scal- lop: implications of recruitment limitation. Mar Ecol. Prog. Ser. 90: 257-272. Peterson. C. H.. H. C. Summerson, S. Fegley & R. Prescott. 1989. The timing, intensity, and source of autumn predation on adult bay scallops. 96 Bologna et al. Argopecten inadians (Lamarck). J. Exp. Mar. Biol. Ecol. 127:121- 140. Phillips, R. C. 1982. Seagra.ss meadows. In: R. Lewis, III. editor. Creation and Restoration of Coastal Plant communities. Boca Raton, FL: CRC Press, pp. 173-201. Pohle, D. G., V. M. Bricelj & Z. Garcia-Esquivel. 1991. The eelgrass canopy: an above-bottom refuge from benthic predators for juvenile bay scallops Argopecten irrciilitms. Mar. Ecol. Prog. Ser. 74:47-59. Psuty, N. P., M. DeLuca, R. Lathrop. K. Able, S. Whitney & J. F. Grassle. 1993. The Mullica River-Great Bay National Estuarine Research Re- serve; a unique opportunity for research, preservation and management. Proceedings of Coastal Zone '93, July 1993, New Orleans, LA. Robblee. M., T. Barber. P. Carlson, M. Durako, J. Fourqurean, L. Muehl- stein, D. Porter, L. Yarbro, R. Zieman & J. Zieman. 1991. Ma.ss mor- tality of the tropical seagrass Tluilassiu lestmliniim In Florida Bay (USA). Mar. Ecol. Prog. Ser. 71:297-299. Roff, D. A, & P. Bentzen. 1989. The statistical analysis of mitochondrial DNA polymorphisms: X2 and the problem of small samples. Mol. Biol. Evol. 6:539-545. Sastry, A. N. 1970. Reproductive physiological variation in latitudinally separated populations of the bay scallop, Aequipecten irradians Lama- rck. Biol. Bull. 138:56-65. Strieb, M. D., V. M. Bricelj & S. I. Bauer. 1995. Population biology of the mud crab. Dyspanopeus sayi. an important predator of juvenile bay scallops in Long Island (USA) eelgrass beds. J. Slielljish Res. 14:347- 357. Tettelbach, S., C. Smith, J, E. Kaldy III. T. W. Anoll & M. R. Denson. 1990. Burial of transplanted bay xaWops Argopccren irradians irradi- ans (Lamarck, 18191 in winter. J. Shellfish Res. 9:127-134. Tettelbach, S., C. Smith, R. Smolowitz, K. Tetrault & S. Dumais. 1999. Evidence for fall spawning of northern bay scallops Argopecten irra- dians irradian.s (Lamarck 1819) in New York. J. Shellfish Res. 18:47- 58. Thayer, G. W. & H. H. Stuart. 1974. The bay scallop makes its bed of seagrass. Mar. Fish. Rev. 36:27-30. Waller, T. R. 1969. The evolution of the Argopecten gibhus stock (Mol- lusca: Bivalvia), with emphasis on the Tertiary and Quaternary species of eastern North America. / Paleont. 43:1-125. Wenczel. P., C. Smith & S. Tettelbach. 1993. Planting bay scallops: results of reseeding bay scallops in the Peconic Bay, New York, 1986-1992. Final report submitted to the New York State Urban Development Corporation, The New York State Department of Environmental Con- servation, and Suffolk County, New York. Wilbur, A. E. 1995. Population genetics of the bay scallop. Argopecten irradians (Lamarck): an analysis of geographic variation and the con- sequences of self-fertilization. Ph. D. Dissertation, University of Dela- ware, Lewes DE, 128 pp. Wilbur, A. E. & P. M. Gaffney. 1997. A genetic basis for geographic variation in shell morphology in the bay scallop Argopecten irradians. Mar Biol. 128:97-105. Joiinuil oj SlwIIJish Ht'M-anh. Vol. 20, No. I, 97-105. 2001. EFFECT OF MICROALGAE PROTEIN ON THE GONAD DEVELOPMENT AND PHYSIOLOGICAL PARAMETERS FOR THE SCALLOP ARGOPECTEN PURPURATUS (LAMARCK, 1819) ANA FARIAS* AND IKER URIARTE Instiniw dc Acuicidtiini. Univcisidad Austnd dc Chile. Ccisilla 1327. Pticrio Mann. Chile ABSTRACT We scaniK-cJ tiie dTccIs ot lood quality on gonadal de\clopnicnt and phyMological parameters (feeding rate, absorption efficiency, oxygen consumption rate, and excretion rate) for the Chilean scallop. Aiiiiipevlen purpiiniriis. Scallops were fed niicroalgal diets with three different protein levels, named low (L). normal (Nl and high protein (H). or subjected to starvation (C). In addition, scallops were conditioned to environmental temperature in every season of the year and in the winter two temperatures were tested (field temperature of lO'C and IS'C). The gonads of the groups receiving the normal and high protein diets matured during the conditioning experiments and responded to spawning stimuli. The gonads of scallops fed on the low protein diet did not ripen and could not be stimulated to spawn at the end of the experiments. A maximum fecundity rate and massive spawning, when subjected to spawning stimuli, was obtained in scallops fed a high protein diet in the autumn experiment. Scallops fed a high protein diet increased filtration rate, ahsoi-ption efficiency; and showed highest scope for growth (SFG) and net reproductive efficiency (K2) values. Medium values of filtration rate, absoiption efficiency. SFG and K2 were observed in the scallops fed on a normal protein diet. Scallops fed on a low protein diet had the highest oxygen consumption, lowest filtration rate, decreased absorption efficiency and negative SFG and K2 values. Considering that the low protein diet was rich m carbohydrates, the imbalance of low protein to energy (P/E) could have caused poor protein synthesis when fulfilling energy demands resulting in reduced gonadal growth. However, the poor physiological condition of scallops fed the low protein diet was still better than for starved scallops. The positive SFG observed in the winter at I5°C with the high protein diet showed that the Chilean scallop could be conditioned in the winter by both increased temperature and dietary protein. We propose that physiological responses through reproductive conditioning of the Chilean scallop may be modulated by the protein content of the diet and by the relation P/E of the microalgae. In addition, the increase of protein level of the microalgae improves the SFG and fecundity of scallop broodstock, A'^}' WORDS: Argopcclcii piirpiiniliis. mollusc nutrition, physiology, protein requirement, reproduction, scallop INTRODUCTION The Chilean scallop. Arg»pecten piirpiiratus. is being cultured in e.xperimental and commercial hatcheries in the north and south of Chile; and most often larvae are obtained from induced spawn- ing of long-line cultivated adult scallops, without laboratory con- ditioning (DiSalvo et al, 1984, Uriaile et al, 1996a). Martinez et al. (1999a) provided some data on the physiological aspects of matu- ration and the importance of nutrition during gonadal development which has been studied by Coutleau et al. (1996), Farias el al. (1997), Marrit et al. (1999). Martinez et al. (1999b). The biochemical composition of microalgal cells can be varied by altering the nitrogen concentration of the culture medium (Wik- fors el al. 1984. Utting 1986. Uriarte el al. 1993). The biochemical modification of the microalgal cultures is an important tool not only for research on the nutritional requirements of filtering or- ganisms, but also for detertnining the efficient use of microalgae during phases of high consumption for the postlarvae stage and during reproductive conditioning (Uriarte t% Farias 1999). Utting (1986) found that oyster larvae grew better when fed carbohydrate- enriched microalgae, while oyster spat grew better on protein- enriched microalgae. Farias et al. (1997) found that the fecundity of adult scallops increases when fed protein-enriched microalgae. Thus, alteration of the nutritional media could be an effective way to control the biochetnical quality of microalgae and improve ef- ficiency of production by hatchery (Uriarte & Farias 1995, 1999). The present study evaluates the effect of biochemically ma- nipulated microalgae on the physiology and gonad development of adult Argopecten purpuratiis which were introduced in the south of Chile where no natural populations exist. *Corresponding author. E-mail: afariasCs'uach.cl; Fax: -i-56-6-525-.S,S8.1 MATERIALS AND METHODS Microalgae of Different Protein Level We manipulated cultures of Isocliiysis aff. giilhaini (clon T-lso) and Chaetoceros neogracile. using the method described by Uri- arte et al. (1993) with the modifications of Uriarte & Farias (1999). Three strains pre-conditioned for two years at the different nitro- gen levels (high protein, normal protein and low protein) were selected for the bulk production of algae for our experiment on the reproductive conditioning of Argopecten purpuratiis. As all the cultures were grown using the same light, temperature and aeration conditions, the nitrogen content in the media was the only factor that varied. The microalgae of each species were harvested daily during their exponential growth phase. Three diet treatments were prepared by a mixture of both species in a proportion 1 : 1 for each protein level; high protein, nortnal protein and low protein. A fourth treatment was, starving of scallops during the entire condi- tioning period. Conditioning Experiments The scallops were collected from the long-line of the Labora- torio Biologico Pesquerti de PutemtJn (IFOP. Region X. 42°25'S). The scallops were 22 to 30 months old. The experiments were run during five periods: spring 1995. autumn 1996. winter 1996. spring 1997, summer 1997, at tem- peratures of 15. 13, 10, 16, and 15°C respectively, corresponding to the average field temperatures. Scallops were measured, weighed and induced to spawn just before each experiment. Gonad maturation was assessed using a visual index (Mason 1983). During the eight weeks of the experi- ment, one scallop per tank of 8 L was suspended in recirculated 97 98 Farias and Uriarte and filtered (0.5 jjim) seawater. The exceptions were the first and last experiment both during springtime, where tanks of 250 L with 15 tagged scallops were employed. In each test, the water was changed and the tanks and tubes cleaned every second day. The spring 1995 experiment was aimed to detect changes in the biochemical composition of scallop tissues and to determine the optimal food ration for reproductive conditioning. Scallops were fed with two diets: low protein (L) and normal protein (N). There was one tank of 250 L per diet, each tagged scallop in the sus- pended sieve was considered a pseudoreplicate. The autumn 1996 experiment examined the effect of three diets and starvation. There were three tanks for each treatment randomly distributed. The diets had three different qualities of microalgae: high (H). low (L) and normal (N) protein level. The winter 1996 experiment took place when the field tem- perature was 10°C and the study aimed at the combined effect of temperature and diet. Two temperatures were tested: 10 and 15°C in combination with high and normal protein diets and starvation. keeping three replicates randomly distributed, for each diet/ temperature combination. The spring 1997 experimeni evaluated the effect of high and normal protein diets, the pseudoreplicates were 15 tagged scallops for each diet suspended in a sieve inside of 250 L tank with recirculated water. The summer 1997 experiment evaluated the effect of high, low and normal protein diets, with three tanks for each treatment, ran- domly distributed. The microalgal diets were continually delivered to the brood- stock with a peristaltic pump. The daily rations were based on the results of the spring 1995 experiment. Biochemical Analysis At the end of the first experiment, we made biochemical analy- ses of gonad, digestive gland and muscle of scallops, as these body components are likely sites of energy storage for subsequent use in gonad maturation (Marti'nez 1991. Uriarte et al. 1996b). During the conditioning period of the autumn experiment the biochemical analysis of each microalgae species and quality was run. Samples of known concentration and volume of microalgae were filtered on pre-ashed and pre-weighed Whatman GF/C filters using a vacuum system (pressure less than 5 in. Hg). The filters were dried to constant weight at 40 C and subsequently frozen at -20 C until analyzed. The biochemical characterization of the various mixtures of microalgae was estimated from biochemical data on individual microalgal species. The biochemical analysis were carried out using the methods cited by Fari'as et al. (1998) using bovine gamma-globulin, cho- lesterol and glycogen of oyster as standards for protein, lipid and carbohydrate, respectively. The energetic content of the diet was calculated based on the energetic equivalents for protein, lipid and carbohydrates of 23.5. 39.37 and 17.08 J mg~'. respectively, re- ported by Paine (1971). Physiological A iialysis Toward the end of each conditioning period, physiological evaluations were carried out on the scallops. To determine the clearance rate (CR) a closed system was used (Widdows 1985). Each experimental scallop was placed in a 1-L glass beaker, highly aerated and with an initial concentration of 10 algae-cells p-P'. The reduction in particle concentration was measured in samples of 10 ml every 15 min for 1 h. An additional container with the same characteristics but with no scallop served as the control. The particles were counted after centrifugation and removal of 9.5 ml of supernatant. The experimental scallops were left undisturbed for at least 24 h before the clearance rate measurements started. CR (1 h"') was calculated by the exponential regression: p _ ^ (-limvM)+a)xt where C, is algae concentration after time t. C,, is initial cell con- centration, m is filtering rate of a single animal, n is number of scallops, M is volume of suspension, a = rate at which particles settle out of suspension in the control beaker, calculated from: C C The ingestion rate was estimated as the product of the clearance rate (1 h"') and the cell concentration (mg L"'). The feeding rate (FR) was estimated as the product of the ingestion rate (mg h~') and the energy of microalgae mixture (J mg"'). The absorption efficiency (AE) was measured for each individual using the method found in Conover (1966), on the basis of fecal organic content and of food consumed during the previous 8 h. The ab- sorption rate (AR) was estimated as the product of the feeding rate (J mg~') and absorption efficiency. The rates of oxygen consump- tion (VO-,) were measured at the experiinental temperature in a 1 .5-L closed chamber with a Strathkelvin Polarographic electrode following the methodology of Widdows ( 1985). The rates of am- monium excretion (VNH4-N) were determined with the method described by Widdows (1985). Scope for growth (SFGl and net growth efficiency (K2) are reference values for the production of somatic and/or reproductive tissue, depending on the age of the bivalve (Vahl 1981 ). In this case the scallops were adults subjected to a conditioning regime, then the production was considered only as growth of the gonad tissue. For this reason was used the terms scope for reproductive growth for SFG and net reproductive effi- ciency for K2. The scope for reproductive growth (SFG). the net reproductive efficiency (K2). and the stress index (O/N ratio) were calculated by incorporating all previous physiological measure- ments as described by Widdows (1985). At the end of the physi- ological evaluation, all scallops were weighed and induced to spawn. At the end of each experimeni the body components were sepa- rated and dried at 100°C for 72 h prior to weighing. All physi- ological determinations were then standardized to a mean dry body mass of 4 g (close to an average scallop. 4.06 ± 0.48 g). using the equation of Bayne el al. ( 1993) and the regression coefficients for filtration rate, oxygen uptake, and excretion rate of Navarro and Gonzalez (1998) at salinity of 30 g L~'. Statistical Analysis The protein and energy content of the diets were compared between the two microalgae species and nitrogen levels using an ANOVA. The diet quality was characterized by the ratio protein/ energy (P/E, both measured in Joules) and the total food energy availability per scallop (FE). The physiological rates and the spawning response between treatments were compared statistically using ANOVAs (Sokal & Rohlf 1981). The absorption efficiency values were arcsine transformed, because they were ratios, in ac- cordance with Sokal and Rohlf (1981). Correlation and multiple linear regressions analyses were carried out following log^, trans- formation of the physiological rates, to reduce the dependence of the sample variance on the mean and to normalize the distribution MiCROALGAB PROTEIN QUALITY FOR A. PURPUHATVS BROODSTOCK 99 of the data (Navarro and Thompson 1996). The variables included in the analyses were temperature, diet quality. FE (J h" ) and protein absorbed (J h ' ). Protein absorbed (PA) was estimated as the product of the energy absorbed (J h"') and the percentage of protein content for each diet quality. Average data were expressed as mean ± standard error. RESULTS MiiiiipiiUiUd Microalgae The protein content of the microalgae was significantly differ- ent between the qualities generated. C. neognicile showed a higher protein content than l-lsochiysis whereas the various quality cul- tures showed a decreasing protein content (F^pp = 1 1.53, d.t.= 1;18. p = 0.005; F., = 30.82. d.f. = 2; 18. p < 0.0001) as follows: H > N > L (p = 0.05). The energetic contents showed that there were no significant differences between the microalgal species used. The mixture of both microalgae (Table 1 ) was characterized by significant differences on the protein content (F = 35.18. d.f. = 2. 7, p = 0.001), carbohydrate content (F = 36.73, d.f. = 2, 7, p = 0.001 ) and ratio P/E (F = 15.56. d.f. = 2. 7, p = 0.007) between qualities. The total energy, dry mass and organic matter of mixtures were similar for the three levels of protein, with averages values of: 0.94 ± 0.04 J per million cells, 44.5 ± 5.9 |xg per million cells and 67.3 ± 4.09!-, respectively. Scallop Coiidilioiiing During the spring 1995 experiment, a daily ration of 2.5 x 10'' cells day"' scallop"' of both the low or normal protein diets was sufficient to stimulate the ripening of gonads without producing pseudofeces. Therefore, this ration value was used as constant for subsequent experiments. The absorption efficiency values did not differ significantly between diets (Table 2). the mean absorption efficiency being 57.33% (±2.16). The Mason Index did not show significant difference in the appearance of gonads between diets. The scallops fed on the L diet had a lower gonad index. Protein levels of scallop tissues did not differ between diets after conditioning (Table 3). However the differences were sig- nificant between tissues: 14.3 ± 2.5, 79.9 ± 3.4 and 90.3 ± 3.4% of dry weight of digestive gland, gonad and muscle, respectively. Lipid values of scallop tissues varied significantly between diets (Table 3), with maximum values for diet L (15.6% ± 1.27 of dry weight). Lipid levels between tissues were significantly different (p = 0.05) (Table 3), with the highest value for the gonad (17.6% ± 1 .6 of dry weight) followed by the muscle ( 1 2.5% ) and digestive TABLE L Mean values l± the standard error) for microalgal quality after manipulation of nitrogen in the culture media. H N L Protein ((xg/ 10" cells) 7.70 ± 0.63 5.83 ± 0.22 1.72 ±0.46 Lipid Ijig/lO" cells) 18.22+1.82 17.62 ± 1.34 23.11 ± 1.15 Carbohydrate ((jLg/10" cells) 1.82 ±0.14 2.97 ± 0.37 5.55 ± 0.33 Energy (J/10" cells) 0.93 ± 0.06 0.88 ± 0.(t6 1 .04 ± 0.03 P/E 0.20 ± 0.03 0.16 ±0.01 0.04 ±0.01 Dry weight (|xg/10" cells) 51.21 ±9.86 39.08 ±11.12 42.72 ± 14.65 Organic matter 0.77 ± 0.04 0.62 ± 0.03 0.59 ±0.10 TABLE 2. Argopectcn piirpiiraliis. L\alualion ol reproductive conditioning with normal and low protein diet made b> mixing N-low or N-normal cultures of C. neognicile and /. galhaim. Mean values (± the standard error) are given. There were 10 to 15 replicates per treatment. Diet .Absorption EFF (%) Final weight ig) Mason index Normal protein (N) Low protein (L) 59.9 ± 3.2 53.7 ± 3.0 63.8 : 63.6 : 1.0 1,4 4.3: 3.7: 0.2 0.2 gland (9.8%). The carbohydrate levels did not show differences between diets. The values varied significantly between tissues (Table 3). with low values in gonads and muscles (3.5 and 2.4% of dry weight, respectively) and a high level in the digestive gland (26.8% ± 0.9 of dry weight). Thus, only tissue lipids varied be- tween diets. In the autumn experiment, the clearance rate differed signifi- cantly between treatments (Table 4. F = 53.47; d.f = 3, 8: p = 0.001). The lowest rate was observed in the starved scallops, the highest rate in the scallops fed on the H diet and there were no significant differences between the N and L diets. Oxygen uptake (Table 4) showed no significant difference between treatments. The mean VO, was 2.50 ml O. h"' (±0.29). with a maximum rate (p = 0.06) seen in the scallops receiving diet L and a minimum observed in scallops fed on the N diet. The VNH4 - N (Table 4) did not show significant differences between treatments although the highest value was observed in the starved scallops. The mean excretion rate was 193.58 (±26.46) mg NH4-N Ir'. Absorption efficiency values was significantly different between diets (Table 5. F = 9.91; d.f = 2. 17; p = 0.002), with the highest efficiency in the diet H (66.4%). No significant differences were observed between the diets N and L. The stars ed scallops did not show a positive absorption. Scope for growth (SFG) was significantly higher in the broodstock maintained on the N and H diets (Table 5). (F = 14.44. d.f = 3, 8; p = 0.007), making available 23.2 and 141.3 J h"' of energy, respectively, for the production of new tissues and gametes. The starved scallops and those fed the diet L showed negative SFG. Net growth efficiency (K2) was 58% in diet N, and 72% in diet H. The negative values indicated that the starved scallops and the ones maintained on diet L did not produce new tissues (Table 5). F = 16.58; d.f = 3. 8: p = 0.005). The O/N ratio did not show significant differences between diets, al- though the higher values for O/N, that it means less stress, were observed with low protein diet associated to high VO, When induced, the scallops fed on the N and H diets spawned com- pletely, whereas the diet L scallops were obser\'ed to release sperm only. The mean fertility was 2.8 x 10'^ (±1.1 x 10") eggs scallop"' TABLE 3. Factorial ANOVA statistics of the biochemical composition of tissues of scallops fed on L and N diet. Data of percentage values of dry weight were transformed by arcsin. d.f. Protein L ipid Carbohydrate Factor F P F P F P Diet Tissue 1 0.16 99.95 0.7 0.00001 1.43 6.42 0.02 0.005 0.14 347.69 0.7 0.00001 100 Farias and Uriarte TABLE 4. Argopecten piirpiiratus. Physiological rates of standard scallops of 4 g dry tissue weight conditioned with experimental diets. Clearance Oxygen Ammonia rale uptake excretion Diet dh-') (mIO, h '» (MgNHj- N h') Autuniin C 0.91 ±0.01 2.69 ±0.1 5 273.41 ±40.97 H 34.76 ± 1.36 2.39 ±0.1 2 181.07 ±47.26 L 2.31 ±0.63 3.27 ± 0.69 17 1.59 ±40.41 N 6.0S ± 1 .54 0.95 ±0,1 4 166.54 ±21.74 Winter Temperature lO'C C 7.61 ±0.00 — 15.93 ±00.00 H 15.23 ±1.46 1 .42 ± 0.49 138.38 ±69.07 N S. IS ±0.33 1.46 ±0.23 180.41 ±60.25 Temperature 15'C C 1 .39 ± 0.00 0.24 ± 0.00 387.26 ± 00.00 H 8.69 ± 0.44 1.23 ±0.24 484.86 ± 148.66 N 6.83 ±2.22 1.38 ±0.11 2 16.24 ±87.64 Summer H 17.94 ±3.05 1.65 ±0.10 450.03 ± 58.02 L 12.70 ±2.68 2.49 ± 0.45 326.44 ±8. 13 N 19.21 ±2.28 1.80 + 0.29 464.89 ± 11.57 Spring H 3.29 ± 0.59 2.76±O..W 448.59 ±35. 17 N 2.81 ±0.67 2.98 ±0.39 560.89 ± 59.81 H. L and N refers to high, low and normal protein algae, respectively. C refers to the treatment of starvation. significant differences between diets in spite of apparently low value for the L diet (Table 4). Absorption efficiency values were the lowest during the study and did not vary with the diet (Table 3). SFG did not differ between the treatments (Table 5) and was negative for all diets with values as low as -24.6 J h"' (L diet). K2 as well was negative (Table 5). ranging between -1 .6 and -1.1 for scallops fed on diets L and H. respectively. The 0/N stress index tended to higher values with a low protein diet, showing a less stress condition of scallops, however only the scallops fed on N and H diets responded to the spawning induction. Scallops fed on the diet L did not release sperm or eggs and died 24 h from the effects of stimulating to spawn. In springtime, the last experiment of 1997 revealed no signifi- cant differences for clearance rate between treatments (Table 4). although the highest value was observed for diet H, VO, was similar between the diets (Table 4). Excretion rate in A. purpuratiis did not differ between diets N and H. The mean value of VNHj-N was 504.72 (±38.87) |xg NH4-N h"' (Table 4). Absorption effi- ciency values were not affected by the diets N and H (Table 5) and had a mean value of 52.7 (±3.5) percent. SFG was negative for both diets (Table 5) with a mean value of -54.6 J h~' and no significant difference between treatments. Similarly. K2 was nega- tive (Table 5) in both diets and was associated with a low clearance rate and a high oxygen uptake. The diets had a similar O/N ratio with a mean value of 7.6 (±0.94) (Table 5). During the spawning induction, the scallops did not respond to the stimuli and died after 24 h. Integrated Results for diet N and 13 x lO*" (±0.5 x 10") eggs scallop"' for diet H. Egg diameter was 61.3 mm (±0.4) with no signiTicant differences be- tween the treatments. The eggs were viable and developed into scallop embryos. There were no significant differences in clearance rate between temperatures or between diets during the winter experiment. Higher clearance rate values were characteristic for the H diets at both temperatures. The lowest clearance rate was observed in the starved scallops. Oxygen uptake did not present significant differ- ences between temperatures but there was a significant effect be- tween the diet (Table 4. F. 5.52. d.f. = 2.ll;p = 0.04). The minimum \ alue for VO, was observed in the starved scallops. The results of VNH4-N showed no significant difference between tem- peratures or diets and had a mean value of 237.84 (±57.61) mg NH4-N h"' (Table 4). Absorption efficiency values was also not affected by temperature and diet, ranging from 34.8% in diet N at IO°C to 48.2% in diet H at 15"C (Table 7). The starved scallops did not show absorption. The calculated SFG and K2 did not differ between temperatures and diets (Table 5). Higher values were observed for diet H. The highest value of SFG and K2 were ob- served with diet H at 10°C. The scallops fed on N and H diets at 15°C spawned when induced, whereas the scallops at 10°C did not release sperm or eggs. The eggs obtained at 15°C appeared co- agulated and were not viable. In the summer 1997 experiment, values of clearance rate did not differ between treatments (Table 4). although clearance rate was lowest in the L diet. There were no significant differences in VO2 between the diets (Table 4). The tendency of Oxygen uptake was the same as in the above experiments, with maximum values for the scallops fed on diet L. Results of VNH4-N indicated no Clearance rates showed significant differences between seasons (F.eason= 8.735; d.f = 3,48: p = 0.0001 ) with a minimum clear- ance rate of 1.61 I h"' (SE = 3.53. n = 10) during winter and a maximum clearance rate of 12.13 I h"' (SE = 2.12. n = 25) during summer. There were significant differences between diets (F ^ = 5.510-.d.f = 3.48.P = 0.0025). The minimum CR of 0.93 1 h~' (SE = 3.11. n = 4) was observed in the starved scallops while the maximum rate of 10.83 1 h~' (SE = 1.49. n = 21) was found in diet H. Feeding rate (FR) was positively correlated with the scallop weight, ratio protein/energy (P/E) of diet and seasonal temperature, while the correlation of FR with absorption rate and protein absorbed was attributed to the autocorrelation of these variables (Table 6). Multiple regression analysis between FR and the experimental variables accounted for 45% of the variation (Table 7). Dry mass of the scallops alone explained 20% of the variation. Absorption efficiency varied significantly with seasons •F.easun = '5-69: D.F. = 1, 153; p = 0.00001). with maximum values in autumn (68.25%, SE = 4.26, n = 18) and minimum values in summer (23.18%, SE = 4.26. n = 18). No significant differences were observed between the diets. The starved scallops showed a negative absorption. The overall annual mean value for absoiption efficiency, independent from the diets, was 50.52% (SE = 1.77, n = 160). The absorption rate (J h"') correlated with many of the nutritive variables, and showed an especially close correlation with the energy and protein level of the diet (Table 6). Food energy availability and ratio protein/energy (P/E) explained 59% of the variation in absorption rate (Table 7). Oxygen uptake showed significant differences between seasons (F = 8.54, d.f = 3.51. p = 0.0001). The highest value, of 2.68 ml O, h"' (SE = 0.27. n = 14), was in spring and the lowest. 1 .30 MlCROALGAE PROTEIN QUALITY FOR A. PURHUKATL'S BROODSTOCK 101 TABLE 5. Argopecleii purpuralus. Scope for growth for a standard scallop of 4 g dry tissue weight conditioned with diets. Diet Feeding rate (Jh-') ,\bsorption efficiency C/f ) .M)si>rplion rate (Jh-') Rate of 0\>gen consumption (Jh') Excretion rate (J h-') SFG (j h') K2 0:N Aulunim C H L N 318.1 + 12.5 24.2 ± 6.5 58.4 ±8.0 66.4 + 3.8 59.3 ± 3.6 57.5 ± 3.8 195.1 ±26.1 19.8 ±5.7 35.1 +4.9 54.7 ± 3.0 48,6 ± 2.5 66.5+ 14.1 19.3 ±2.8 6.8+ 1.0 4.5+ 1,2 4.3 ± 1.0 4. 1 ± 0.5 -61.5 ±2.7 141.3 ±22.3 -50.9 ± 8.0 23.2+ 16.5 0.72 ± 0,20 0.58 + 0.19 -0.21 ±0.64 13.13 + 2.81 16.54 + 3.22 27.68 ± 7.90 8.05 + 2.27 Winter Temp: IO°C C H N 72.9 ± 0.0 139.3 ±13.4 78.5 + 3.2 45.6 + 4.7 34.8 ± 3.8 61.9 ±0.8 27.7+ 1.8 -3.3 ± 0.0 28.9 ± 10.0 29.8 ±4.7 0.4 ± 0.0 3.4 ± 1.7 4.5 ± 1 .5 -7.5 ±7.1 I6.()± 11.7 -7.3 + 8,6 0.48 ±0.1 3 -0.28 ±0.31 34.5 ±21.7 11.7 ±2.3 Temp: 15' C C H N 13.2 ±0.0 79.5 ± 4.0 65.5 ±21.3 48.2 ±5.1 40.5 ± 3. 1 39.5 ± 5.8 29.8 ±9.7 4.9 ± 0.0 25.0 + 5.0 28.1 ±2.3 9.7 ±0.0 12.1 ±3.7 5.4 ±2.2 -14.6 + 0.0 2.5 ± 14.7 -8.0 ± 12.5 -0.05 + 0.37 -0.88 ± 0.78 3.4 ± 0.4 5.4 ±0.3 Summer H L N 33.6 ±2.1 50.6 + 9.1 36.7 ±5.8 20.1 +4.4 28.6 ± 6.5 19.5 ±3.8 31.9 ±6.2 34.1 ±7.8 34.3 + 6.9 33.6 ±2.1 50.7 ±9.1 36.7 ± 5.8 11.2+1.4 8.1 ±0.9 11.6+1.4 -12.9 + 7.2 -24.6 ± 10.5 -13,2 + 7.1 -1.1 ±0.5 -1.6 + 0.6 -1.2 ±0.7 5.3 ± 0.9 11.4 + 2.8 6.2 ± 1.2 Spring H N 30.1 ±5,4 26.9 ± 6.5 51.9 ±4.7 53.4 ±3.8 16.9 + 2.8 15.9 ±4.3 56.2 ± 8.0 60,7 ±7.9 11,2 ±0.9 1 3,9 ± 1 ,4 -50.4 ± 10.4 -58.8 ± 8.4 -5.3 ± 2.6 -12.9 + 6.2 7.8 ± 1.0 7.5+ 1.5 H. L anJ N. C refers to the starved scallops. ml On h-' (SE = 0.35, n = 12). in winter. Oxygen consumption rates tended to vary between diets (p = 0.06). Scallops receiving diet L tended to consume most oxygen (2.54 ml O, h" . SE = 0.28. n = II) and the starved scallops least (1.47 ml O, h-'. SE = 0.43, n = 4). VO^ was positively correlated with the sea- sonal temperature and food energy availability (Table 6). and negatively correlated with the total food cells in the tanks (FT). The multiple regression models explained 3 1 % of total variation in oxygen uptake (Table 7). with food energy availability accounting for 14<7r. The ammonium excretion rates showed seasonal ditterences (F = 3.099. d.f. = 3,55. p = 0.03) with a maximum value of 478Tl°9 mg NH4-N h"' (SE = 49.45. n = 14) in spring time and a minimum of 288.63 mg NH4-N h"' (SE = 64.27, n = 1 1 ) in autumn. The ammonium exci'etion rates did not vary with diets. VNHj-N in A. purpuratus was highly correlated with the experi- mental temperature, the seasonal temperature and the food cells in the tanks (Table 6). The best multiple regression model for VNHj-N showed that 48.5% of the variation was explained by experimental temperature, ratio protein/energy (P/E) and feeding rate, whereas experimental temperature alone accounted for 4 1 Vr of the variation (Table 7). The reproductive SFG (Fig. 1) varied between seasons (F = 26.27. d.f. = 0.3. 34. p < 0.000000 1 ) and between diets ^ Season ■ (Fdie. = '7-24. d.f. = 1. 34. p = 0.0002). SFG was negative during spring and summer, close to zero during winter and above zero during autumn SFG. There were significantly higher values of SFG in scallops fed on diet H. The seasonal difference was most conspicuous in starved A. purpuratus (Fig. 2) (F,,^,.„,„„ = 6.86. d.f. = I. 9. p = 0.03). The SFG was always negative, but ranging fi'om values as low as -62.78 J h-' (±17.05) in autumn to values very close to zero in winter. SFG for three N-diets plotted (Fig. 3). During summer. SFG was not influenced by the diet. In contrast. SFG was affected by the diet in autumn, suggesting that there was a strong positive relationship between SFG and ratio protein/ energy (P/E) in this season (F.^^.^^.p^E = '8.29, d.f. = 2, 26. p = 0.000009 ). Scope for reproductive growth was correlated with the nutritive variables, protein absorbed and total food availability, and with two physiological variables, feeding rate (FR) and VO, (Table 6). Multiple regression analysis showed that 68% of the total variance of SFG was explained by protein absorbed, food availability, VO, and FR, whereby protein absorbed alone ac- counted for 23% (Table 7). The net reproductive efficiency (K2) was highly correlated with the seasonal teinperalure, food energy availability, and ab- sorption rate (Table 6). Food energy availability alone explained 33% of the variation and when the seasonal temperature, absorp- tion rate, FR and the 0/N ratio were included, 86% of the variation was explained (Table 7). The O/N stress index was negatively coiTelated with the ex- perimental temperature and food energy availability (Table 6). The variation of O/N was best explained by the model which included the seasonal and the experimental temperature, food energy avail- ability and ratio protein/energy (P/E) as independent variables, explaining as such only 27% of the variation (Table 7). DISCUSSION The experiments on conditioning A. purpuratus showed that scallops fed on the high protein diet, had both the highest scope for reproductive growth (SFG) and net reproductive efficiency (K2) (parameters considered as indexes of success of reproductive con- ditioning, because the pioduction was not detected as somatic growth but as gonadic growth), followed by the scallops condi- tioned on the normal protein diet. The scallops fed on the low protein diet or starved had the lowest SFG and K2 values. 102 Farias and Uriarte TABLE 6. Argopecten piirpuratus. Correlation coefficients between physiological and nutritional variables. FR AR vo. VN-NHj SFG K. O/N Season ExpT DWT P/E PA FE FT FS FR 1.(10 0.50 -0.19 -0.10 0.61 -0.31 -0.06 0.24 0.18 -0.46 0.25 0.47 0.04 0.02 -0.01 (56) (56) (56) (561 (561 (561 (.561 (56) (-56) (56) (56) (56) *** (56) (56) (56) AR 1.00 0.17 -0.01 0.18 -0.33 0.13 0.24 0.13 -0.16 0.49 0.25 0.65 0.07 0.51 (56) (56) (561 (56) (56) (,56) (561 (561 (561 (-561 (56) (56) (561 (56) VO, 1.00 0.08 -0.58 -0.27 U.64 0.29 0.05 -0.08 -0.04 -0.02 0.27 -0.32 0.12 (56) (561 (56) 156) (56) (56) (56) (56) (56) (56) (-56) (-56) (56) VN-NH4 1.00 -0.22 0.08 -0.71 0.37 0.55 -0.20 0.15 -0.20 -0.14 -0.48 -0.16 (561 (561 (0.561 (56) *** (56) ** (56) *** (.56) (561 (-561 (.■^61 (561 *** (56) SFG 1.00 -0.14 -0.25 -0.13 -0.09 0.01 0.15 0.51 -0.01 0.29 0.02 (56) (56) (-56) (-56) (561 (-56) (56) (56) (56) (56) (56) K2 1 .00 -0.25 -0.59 -0.22 0.14 -0.03 -0.36 -0.56 0.23 -0.53 (56) (-56) (56) (-56) (.56) (56) (56) (56) *** (56) (56) *** ON 1.0(1 0.08 -0.39 0.09 -0.14 0.14 0.30 0.14 0.21 (56) (56) (56) (56) (561 (.561 (561 (561 (561 * p < 0.05 ** p < 0.001 *** p < 0.0001 Number of cases in parentheses. FR: feeding absoqition rate; VO, o.xygen consumption rate; VN-NHj: ammonia excretion rate; SFG: scope for growth; K,: net growth efficiency; O/N: stress index; Season: seasonal; ExpT: experimental temperature; DWT: scallop dry weight; P/E: quality diet index; PA; protein absorbed; FE: food energy availability per scallop; FT: total food cells in the tanks; celLs availability per scallop. The clearance rate (1 h"') and the absorption efficiency value.s followed the same pattern as described in the preceding paragraph, highest values on diet H, medium on diet N and the lowest values in the scallops fed on diet L. Our results suggest that improvement of reproductive conditioning on A. piirpiiraliis is the product of the combined effect of high clearance rate and increased absorption efficiency values. The same pattern to maximi/e the net energy in bivalves when food quality increases is described by Navarro et al. (1994). The respiration rates (VO,) did not follow the same pattern of SFG or K2 in relation to the protein content of the diets. The lowest respiration rates were for the scallops fed a normal protein diet, and reflected their low level of activity compared to the scallops in the other treatments. Shumway et al. ( 1988) found that metabolic rate is simultaneously affected by food, temperature and reproductive stage, and state that it is difficult to separate these effects without a good knowledge of how energy is allocated be- tween somatic growth and gametogenesis. We observed an unex- pectedly high VO, in the non-fed group in autumn, which might have been caused by the food particles consumed during the clear- ance rate test the previous day. However, the feeding activity did not provoke an increase in oxygen consumption in other pectinid genera such as Chlamys. although an increase was observed in tnytilids (Mackay & Shumway 1980). Excretion rates were not affected by dietary protein levels, contrary to the findings of Hawkins and Bayne (1991). Scallops required more proteins than mussels as previously reported in stud- ies in Argopeaen irntdians by Barber and Blake, ( 1985) and Epp et al. ( 1988). The level of energy in the high protein diet (where we TABLE 7. Argopecten piirpuratus. .Multiple regression statistics for several physiological variables or physiological index vs. subsets of nutritive variables. See Table 6 for abbreviations. Dependent Independent var. var. R- n F FR DWT 0.20 56 15.21*** P/E 0.26 56 5.55* AR FE 0.46 56 49.07*** P/E 0.59 56 19.63*** SeasonT 0.66 56 12.04** VO, FE 0.13 56 10.46** FT 0.23 56 6.94* VNHj Exp.T 0.41 56 42.45*** P/E 0.44 56 3.67 SFG PA 0.23 56 18.16*** V02 0.58 56 43.06*** FR 0.67 56 13.75** K2 FE 0.33 56 28.87*** SeasonT 0.53 56 25.93*** AR 0.66 56 20.27*** O/N Exp.T 0.14 54 ]Q TT** FE 0.18 54 3.82* SeasonT 0.26 .S4 6.27* ■ p < 0.05; **p < 0.001; ***p < 0.0001. MlCROALGAI-; PROThlN QUALIT^- FOR A. I'URPURATUS BROODSTOCK 103 o 200 150 100 -100 WINTER SPRING Season -High protein -Normal protein Figure 1. Arf-opeclen purpuratus. Seasonal changes in the scope for rei)rc groMth during conditioning experinienis in a Chilean scallop of 4-g dry tissue Height. expected a iiia\miiiiii VNHj-N) e(uild have been enough to use the available protein, producing more gametes and reducing losses due to NHj-excretion The high VNH4 observed in the starved scallops suggest the use of protein to meet metabolic demands during pro- longed starvation as previously documented in mussels and in the scallop Argopeclen irraclians (Epp et al. 1988, Hawkins & Bayne 1991). The similarity in egg size for scallops fed high and normal protein diets indicated that the additional protein in the high pro- tein diet is mainly used for the production of a larger number of eggs and not for larger eggs. A similar strategy has been observed for PUicapeclen maaellanicus. Scallops found in poor environmen- tal conditions had a reduced fecundity but the quality of the eggs was maintained (Thompson & MacDonald 1990, Napolitano et al. 1992). The maintenance costs of scallops v\ ith a low protein diet ex- ceeded the energy assimilated. Diet L scallops had lower clearance rate and absorption efficiency, as well as a high tendency to form > ;;uu AUTUMM 150 J 1 100 y^ ^ 50 ^_^ tissue weigh! 150 3 100 ? o 0) D. O ± SUMMER I I 0.05 0.10 0,15 0.20 Diet quality (Protein/Energy ratio) 0,25 Figure 3. Argopeclen piirpiiralin. Comparison between autumm and summer of the effect of diet quality upon the scope for reproductive growth for a standard scallop of 4-g dry tissue weight. pseudofeces and higher oxygen uptake. The reduction in clearance rate and absorption efficiency values together with pseudofeces production are feeding responses of bivalves predicted by Navarro et al. (19941 when food quality is poor. We observed a tendency to increase oxygen consumption during oogenesis oi' Argopcctcii ir- raJiwis indicating metabolic use of carbohydrate (Barber & Blake. 1985) and in our study could be related to a high degradation of dietary carbohydrates. On the other hand a maximum lipid value observed in the tissues of scallops fed diet L. was not related to the lower gonad index in the same scallops, but could be indicated that a carbohydrate excess of this diet was being converted to lipids which are presumably stored in developing ova. Considering that diet L was very rich in carbohydrates and showed the lowest protein content and the lowest ratio protein/energy (P/E) index, it could have caused a dietary imbalance such as documented for fish, characterized by an excess of dietary energy associated with a decreased protein intake, poor protein synthesis and reduced growth (Young Cho 1987, Tacon 1990). However, the physiologi- cal condition of the diet L scallops was still better than that of the starved scallops, and they may be diagnosed as undernourished scallops. Massive spawnings in the diets H and N were only observed in the autumn. In the other seasons, the gonads were ripe but could not be stimulated to spawn and this was associated with a negative SFG. We propose that the microalgal requirements varied season- ally and 2.5 x 10'' cell day"' scallop"' was enough to develop gametes in spring 1995. autumn 1996 and winter 1996. but did not meet energy requirements to mature or to spawn in the other sea- sons or when diet L was applied. Loss of the relationship between SFG and diet quality in summer could be attributed to insufficient food. This was suggested by the mortality after stress from spawn- ing induction. The negative SFG observed in spring 1997 could have been caused by the change in the culture system. The move- ment of the scallops was moie restricted and they decreased their consumption rate. The mortality after induction of spawning in this experiment was likely associated with the increased requirements to fuel the sudden increase in gonad growth. The positive SFG observed in winter 1996 at 15"C with diet H showed that the Chilean scallop could be conditioned over the winter by increasing temperature and dietary protein level. The net reproductive efficiency (K2) was the parameter most related to seasonal temperature and its inten.se effect on VOo. The 104 Farias and Uriarte O/N ratio (stress index) was lower than that reported for the Chil- ean scallop by Navan'o and Gonzalez (1998) and this may be related with the higher reproductive activity of the scallops in our study. The low values of O/N index associated with high excretion rates of reproductive scallops mean that gametogenesis in scallops occurs mainly at the expense of the adductor muscle protein as suggested by Barber and Blake (1985) and Epp et al. (1988). Low temperature or low energy availability might have caused high stress condition in reproductive adult broodstock conditioned at 10°C in the winter experiment. Similarly, scallops from the field in southern Chile at the end of winter, experienced massive mor- talities when they were ripe (Farias et al. 2001). The Chilean scallops in our study are below their normal distributional limit (35°S) and represent second generation individuals originating from broodstock from northern Chile (Region IV. 30°I6'S) (Uri- arte et al. 1996a). We propose that southern A. piiipuratiis and some populations of A. irradiims could be nutritionally and cli- matically stressed. (Barber & Blake 1985) since available food and glycogen reserves are insufficient to support reproduction and maintenance metabolism. In addition, the catastrophic mortalities observed at southern Chile, could be associated with the three main factors that affect the seasonal metabolism of adult scallops (Shumway et al. 1988. Uriarte et al. 1996 b): stress cau.sed by low temperature (<10°C). low food availability, and high energetic demand for gonad ripening. Martinez et al. ( 1999 b). showed that microalgae enriched with a mixture of lipids improved the conditioning of A. piiipiininis. Uriarte and Farias ( 1995. 1999) have shown that protein content of the microalgal diet affect the physiology and growth of juvenile scallops up to a size of 2 mm. and after this other dietary factors are probably more important. The present work confirms the im- portance of increasing protein in microalgae to improve the effi- ciency of massive production in the hatchery. The diet H influ- enced the SFG and the gonadic growth of scallop A. piirpiiniius by the increase of clearance rate, absorption efficiency values and the improvement of the relationship between oxygen uptake and am- monia excretion. The interaction between dietary protein and lipid enrichment on gonad ripening would be an interesting task for future investigation. ACKOWLEDGMENTS This work was supported by the International Foundation for Science (IFS A-2074 of the first author) and FONDECYT (grant 1970807 of the second author). We thank P. Varas. S. Pino. J. Gallardo and J. Santana for their assistance in the laboratory work. We also acknowledge useful coirections and comments by S. Ut- ting (England) and N. Nevejan (Belgium). Special thanks to pro- grama Acuicultura y Biotecnologi'a Marina. 1 (97), FONDAP, Chile (Subprograma Invertebrados) for the valuable discussion on the topics of this manuscript. LITERATURE CITED Barber. B. J. & N. J. Blake. I98.'i. Substrate catabolism related to repro- duction in the bay scallop Avi^opecten imulians concentriciis. as de- termined by O/N and RQ physiological indexes. Marine Biology 87: 13-18. Bayne. B. L.. I. Iglesias. A. J. S. Hawkins. E. Navarro, M. Heral & J, M. Deslous-Paoli. 1993. Feeding behaviour of the mussel, Mytilus edulis: responses to variations in quantity and organic content of the seston. / Mar. Biol. Ass. U.K. 73:813-829. Conover. R. J. 1966. Factors affecting the absorption of organic matter by /ooplaiikton and the question of supertluos feeding. Limiiol. Ocfuiioi;i: 11:346-354. Coutteau. P.. M. Caers. K. Kure & G. Gajardo. 1996. Supplementation of lipid emulsions to algal diets in the hatchery rearing of bivalves. In: G. Gajardo & P. Coutteau. editors. Improvement of the commercial pro- duction of marine aquaculture species. Proceedings of a Workshop on Fish and Mollusc Larviculture. Santiago: Impresora Creces. pp. 145- 149. DiSalvo. L. H.. E. Alarcon. E. Martinez & E. Uribe. 1984. Progress in mass culture of Chlainy.i: (Ariiopecwn) piirpiirata. Lamarck (1819) with notes on its natural history. Rev. Cliil. Hisi. Natural 57:35—45. Epp, J., M. Bricelj & R. Malouf. 1988. Seasonal partitioning and utilization of energy reserves in two age clas.ses of the bay scallop Argopecwn irradians irradian.'^ (Lamarck). / Exp. Mar. Biol. Ecol. 121:113-136. Farias, A., I. Uriarte & P. Varas. 1997. Esludio de los requerimientos nutricionales del ostion del norte Argopeclcn purpuranis (Lamarck, 1 8 19) durante el acondicionamiento reproductivo. Nutritional study on the broodstock conditioning of the Chilean scallop, Argopecien piirpu- raliis (Lamarck, 1819)). Rev. Biol. Mar. Oceanog. Valparaiso 32(2): 1 27-1.36. Farias. A., I. Uriarte & J. C. Castilla. 1998. A biochemical study of the larval and postlarval stages of the Chilean scallop Argopecien piirpii- ratus. Aquaculture 1 66:37—47. Farias, A., H. Plaza & M. Seguel. 20(11. Mitigation effect of the Chilean scallop. Argopecien purpnruiu.s. on the environmental impact of salmon farming al the South of Chile. Manuscript in prep. Hawkins. A. J. S. & B. L. Bayne. 1991. Nutrition of marine mussels: factors intluencing the relative utilizations of protein and energy. Aqua- culture ')4:m^\96. Mackay. J. & S. E. Shumway. 1980. Factors affecting oxygen consumption in the scallop Chlainys deliculatu (Hutton). Ophelia 19:19-26. Marrit. C. P. Coutteau. K. Cure. V. Morales, G. Gajardo & P. Sorgeloos. 1999. The Chilean %c3\\o\i!, Argopecien purpuranis (Lamarck. 1819). II Manipulation of the fatty acid composition and lipid content of the eggs via lipid supplementation of the broodstock diet. Comp. Bioch. Physiol. {Pan Bi. 123:97-103. Martinez, G. I99I. Seasonal variation in biochemical composition of three size classes of the Chilean scallop Argopecien purpuranis Lamarck. 18 19. VW/gc/- .34:335-343. Martinez. G.. L. Mettifogo. R. Lenoir & E. Campos. 1999a. Prostaglandins and reproduction of the scallop Argopecien purpuratus: I. Relationship with gamete development. J. Exp. Zool. 284:225-231. Martinez. G.. C. Aguilera & L. Mettifogo. 1999b. Interactive effects of diet and temperature on reproductive conditioning oi Argopecien purpura- tus brood.stock. Aquaculture 183:149-159. Mason, J. 1983. Scallop and Queen fisheries in the British Isles. Surrey: Fishing New Books. Napolitano. G. E., B. A. MacDonald. R. J. Thompson & R. G. Ackman. 1992. Lipid composition of eggs and adductor muscle in giant scallops (Placopeclen niagellanicus] from different habitats. Mar. Biol. 1 13:71- 76. Navarro. E.. J. I. P. Iglesias, M. M. Ortega & X. Larretxea. 1994. The basis for a functional response to variable food quantity and quality in cock- les Ceraslodernia cdule (Bivalvia. Cardiidae). Physiological Zoology 67:468-496. Navarro. J. M. & R. J. Thompson. 1996. Physiological energetics of the horse mussel Modiolus modiolus in a cold ocean environment. Mcir. Ecol. Prog. Ser 138:135-148. Navarro, J. M. & C. M. Gonzalez. 1998. Physiological responses of the Chilean scallop Argopecien piirpiiraliis to decreasing salinities. Aqua- culture 167:315-327. Paine. R. 1971. The measurement and application of the calorie to eco- logical problems. .Ann. Rev. Ecol. Syst 2:145-164. MiCROALGAb PROTKIN QUALITY FOR A. PURPUKATUS BROODSTOCK 105 Shumway, S.E., J. Barter. J. Stahlnecker. 1988. Seasonal changes in oxy- gen consumption of the giant scallop. Ptacopvclen magelkmicus (Gme- lin). / Shellfish Res. IJl-S,! Sokal. R. & F. J. Rohlf. 1981. Biometry. The principles and practice of statistics in biological research. 2'"' edition. New York: Freeman. Tacon. A. 1990. Standard methods for the nutrition and feeding of farmed fish and shrimp. Washington: Argent Laboratories Press. Thompson, R. J. & B. A. MacDonald. 1990. The role of environmental conditions in the seasonal synthesis and utilization of biochemical en- ergy reserves in the giant scallop. Plucopeclcn magelliiniens. Can. J. Zool. b'i-n^^l^b. Uriarte, I. & A. Farias. 199.'i. Effect of broodstock origin and postkirvae diet on postlarval growth and physiological performance of the Chilean scallop Ari>oi>eclen inapuranis . Eiimpeau Aquaculmre Siiciery Special Publication 24:69-72. Uriarte. L & A. Farias. 1999. The effect of dietary protein on growth and biochemical composition of Chilean scallop Argopeclen purpurariis (L.) posllarvae and spat. Aqiiacultiire 180:119-127. Uriarte, L, A. Farias & C. Munoz. 1996a. Cultivo en hatchery y pre- engorde del ostion del none, Argopeclen piiipnraliis (Lamarck, 1819), en el sur de Chile. (Hatchery culture and grow-out of the Chilean scallop. Argopeclen purpiiralus (Lamarck. 1819). in the south of Chile). Rev. Biol. Mar Val. 31:81-90. Uriarte, I.. A. Farias & C. Muiioz. 1996b. Growth and condition inde.x of Argopeclen purpuraiiis I Lamarck) in the south of Chile. In: G. Gajardo & P. Coutteau, editors. Improvement of the commercial production of marine aquaculture species. Proceedings of a Workshop on Fish and Mollusc Larviculture. Santiago: Impresora Creces. pp. 145-149. Uriarte. L, A. Farias. A. J. S. Hawkins & B. L. Bayne. 1993. Cell charac- teristics and biochemical composition of Diinaliella priniolecta Butcher conditioned at different concentrations of dissolved nitrogen. ./. Appl. Phvcol. .5:447-^53. Utting, S. D. 1986. A preliminary study on growth of Crussoslrea gigas larvae and spat in relation to dietary protein. Aquaculture 56:123-138. Vahl. O. 1981. Energy transformations by the Iceland scallop. Chlanns islandica (O.F.Miiller). from 70°N. I. The age-specific energy budget and net growth efficiency. J. Exp. Mar. Biol. Ecol. 53:281-296. Widdows, J. 1985. Physiological procedures. In: The effects of stress and pollution on marine scallops. B. L. Bayne, D. A.Brown, K. Bums, D. R. Nixon. A. Ivanovici. D. R. Livinstone, D. M. Lowe. M. N. Moore, A. R. D. Stebbing and J. Widdows. editors. New York: Praeger Sci- entific, pp. 161-178. Wikfors, G. H., J. W. Twarog & R. Ukeles. 1984. Influence of chemical composition of algal food sources on growth of juvenile oysters. Cras- soslrea virginica. Biol. Bull. 167:251-263. Young Cho. C. 1987. La energia en la nutricion de los peces. In: J. Espi- nosa & U. Labarta. editors. Nutricion en Acuicultura II. Plan de For- macion de Tecnicos Superiores en Acuicultura Madrid: (FEUGA). pp. 197-29(1 Journal of Shellfish Research. Vol. 20. No. I, 107-109, 2001. STEPHANOSTOMVM SP. (TREMATODA: ACANTHOCOLPIDAE), THE CAUSE OF "PIMIENTILLA" DISEASE IN CATARINA SCALLOP ARGOPECTEN VENTRICOSUS (CIRCULARIS) (SOWERBY II, 1842) IN BAJA CALIFORNIA SUR, MEXICO JUAN C. PEREZ-URBIOLA'AND SERGIO F. MARTINEZ-DIAZ- ' Centro de liivcstlgcuidiie.s Bioli)i;iciis del Naroestc. La Fa:. Baja Califontia Sur. Apdo. Postal I2S CP 23000. Mexico; "Centro Interdiscipllnario dc Ciencias Marina.^. Playa Conchalito S/N Lai Pa:. Baja California Sur CP 23000. Mexico ABSTRACT We investigated the etiology of a disease of catarina scallop (Argopecien veniricosus (circularis)) during an outbreak in Baja California Sur. B.C.S.. Mexico, in the summer of 1995. Samples from Bahi'a Magdalena and Bahfa Concepci6n were studied. Our results indicate that the disease named "piniientiUa" is caused by encysted metacercaria of the treniatode Sicphaiiosiomiim sp. The melani/.ed spots in the adductor muscle charactenstic of the disease are cysts of this treniatode. Almost all scallops analyzed had abundant cysts. The characteristics of the matacercariae are described and compared with that of other species of Slephanoslomwn. KEY WORDS: Slephoitostoimim. metacercaria. catarina scallop, disease Iti Baja California Stir (B.C.S.). Mexico, the catarina scallop (Argopecten ventricosiis (circuUiris)) fishery is an important eco- nomic activity employing about 7.000 families in extracting, pro- cessing, and marketing. The production of catarina scallop was 7,000 t of adductor muscle during 1989 and 1990. with more than 4,500 t in 1989 from Bahia Magdalena and Bahia Concepcion (Masso 1996). In recent years, the catarina scallop fishery has been affected by the collapse of natural stocks and by the presence of some para- sites, which caused a decrease in production volumes. During 1978 and 1992 the fishing areas of La Paz and Bahi'a Concepcion, re- spectively (Fig. 1 ) suffered a collapse of their natural stocks. Such events were associated with overfishing in La Pa/ (Baquciro et al. 1981) and with adverse environmental conditions in Bahia Con- cepcion (Masso 1996). During 1978, the commercial fisheries in Lagunas Guerrero Negro and Ojo de Liebre B.C.S. were closed temporarily by the Mexican Health Ministry. This was necessary because an outbreak of nematodes (Echinocepliahis pseiidounci- natus in the adductor muscle) was detected, which required an evaluation of their effect on human health (Gomez del Prado 1984). Later studies showed the presence of metacercariae and metacestodes of the families Fellodistomidae and Phyllobothriidae in the adductor inuscie. and metacestodes of Trypanorhyncha in gonads and hepatopancreas (Gomez del Prado et al. 1992). Masso ( 1996) concluded that there was no risk to human health. In pectinids, some parasites cause massive infections and mor- tality in natural populations (Moyer et al. 1993, Whyte et al. 1994), and consequently cause significant economic losses to fisheries and mariculture by a decrease in the catch volumes and by the reduction in yields and quality as human food. During summer 1995 (June and July), an outbreak of disease was detected in the commercial stocks of catarina scallop in B.C.S. The main sign of disease was the presence of abundant melanized spots in the adductor muscle, which resembled pepper granules (Fig 2a). Attempts to detemiine the aethiologic agent using stan- dard histological techniques failed. The objective of this study was to determine the etiology of the "pimientilla"" disease and ascertain if the disease represented a risk to human health. To determine the cause of the disease, 450 specimens were submitted to the Department of Experimental Biology of the In- terdisciplinary Center of Marine Sciences in July and August 1995. The specimens were collected in the two main areas of commercial exploitation of B.C.S.: Bahi'a Magdalena on the Pacific Coast and Bahi'a Concepcion on the Sea of Cortez (Fig. 1 ). After collection, the specimens were kept on ice before analysis. In the laboratory, the spots were dissected from the adductor muscle. The parasites were extracted by pressing spots between two glass slides and were fixed in AFA (70% ethanohFormali- n: Acetic acid, 80:10:10). Specimens were stored in 70% ethanol, stained with Gomori Trichromic, and mounted in synthetic resin. The description of the aethiological agent was based on nine mounted organisms. Measurements were made with an ocular mi- crometer. There was a high prevalence of diseased scallops from both sites. Ninety-eight percent of specimens analyzed show the char- acteristic signs of disease (small brownish spots in the adductor muscle). The spots in the muscle were cau.sed by the presence of embedded cysts. The number of cysts per adductor muscle varied between a few to more than 800. From each cyst, we recovered a metacercaria, which was identified as Stephanostomwn sp. The characteristics of the metacercariae (Fig. 2b) recovered were: body 764 |j.m (554-1291 jjLm) long and 280 jj.m (228-316 lj.m) width, tegumental spines, terminal oral sucker of 241 |xm (208-292 iJim) long and 235 |j.m (180-361 ixm) with the mouth located in the mid-region of the body, and two rows of 28 oral spines each (56 total) 43 p.ni (36-56 |j,m) long and 18 \x.m (12-21 jjt,m) width. The prepharynx was 234 \xm (50-476 |j.tTi) long and the pharynx 132 |jLm (98-157 jxm) long and 122 (j.m (92-146 (xni) width, continuing posteriorly to the intestinal bifurcation, that is in the posterior forebody. The caecae reach close to posterior extrem- ity and end blindly. The ventral sucker is preequatorial with diameter of 264 (jliii (201-362 (xm). There are a pair of dorsal ocelli between the oral sucker and the pharynx. There are reproductive organs in immature stage in the posterior third of the body, and two testes in line, the ovary is located anterior to the testes. The genital pore is just 107 ■ 108 Perez-Urbiola and Marti'nez-Di'az USA Guerrero Ojo de Li Pacific Ocean Bahia Concepcion fSea of C V Cortez%^^ Bahia Magdalena Figure 1. Catarina scaWaf Argopecten ventricostis (circiilaris) sampled areas. before the body middle line and the excretory vesicle in I-like form. The genus Steplumostamuin includes numerous parasites of marine teleosts. generally warm-water species (Quinteiro et al. 1993). Species determination in adult stages is based on morpho- logical characteristics, e.g.. the number, shape, and position of peristomial spines, the relative size of the oral and ventral suckers, the number and distribution of vitellaria. the size of the cirrus sac and the genital atrium, the distance between gonads, the presence or absence of uroproct, and the egg size (Stunkard 1961). In the metacercariae recovered from catarina scallop, the abun- dance of cephalic spines (56) is similar to that of 5. promicropsi (50-58), S. haccatwn (56-60), and S. fistiilariae (56). However, 5. promicropsi has been reported only from the North Atlantic and although 5. baccatiiiii has been reported in the Pacific Ocean of North America, the metacercariae has only been reported in sole (Table 1 ). S. fistidarinue has only been reported from Japan, and the life cycle has not been described yet. The number of cephalic spines in the 13 species of Sleplianoxtonuim found in the Pacific Ocean of North America (Love and Moser 1983) do not coincide with the metacercariae found in the catarina scallop. The results obtained in the present study indicate that the specimens recovered from catarina scallop have unique characteristics. Although there is not sufficient information available for the complete identifica- tion at species level, we can't discard the possibility of it being a new species. The complete cycle of SleplwnostoiniDii comprises three prin- cipal stages: redia, metacercaria, and adult, all of which require different hosts. Most species of the genus have been described from adult specimens, and intermediate hosts are unknown (Table I shows the species for which the intermediate hosts are known). The complete cycle is known for Stephanostoinuin teiiiie and Stephanostoiniiin baccaUdii and the metacercaria. and adults of Stephanostoinuin japonicnin, Steplianostoinuin l^icoronatuin. Stt'plianostoinuni hispiduni, Sli'pluiiiostonuiin ctuiiii'^iiiiii. and Stephanostomuin lopliii have been described as parasites of fish (Table 1). In general, the rediae of Steplianostoiniiin are parasites of gastropods and the other stages are parasites of marine fish. This is the first record of a bivalve mollusk that acts as an intermediate host for Steplianostoiniiin. Catarina scallops inhabit sandy bottoms, coexisting with some gastropods, and is preyed upon by fish. To date in B.C.S., gastro- pods have not been reported as hosts of rediae of Stephanostonniin nor have fish been reported as a definitive host. However, the species S. californiciiin and 5. dentatiiin have been reported in La Jolla, California from the intestinal tract of the yellowfin croaker [Uinhrina roncador) and the California halibut (Paralictis ccdifor- nicus). respectively (Manter and Van Cleave 1951 ). Nevertheless, the number of cephalic spines in the adults of these two species does not match the number on the metacercaria from catarina scallops. No metacercariae were found alive during the dissection. The method employed for transporting the scallops during our analysis is similar to that used commercially, suggesting that the parasite would not survive during commercial processing and shipping. Furthermore, the life cycle of this parasite does not include mam- mals. Thus, we can discard concerns for hmnan health. However, a bigger problem is the decrease in the quality of the market product. Because when the infestation is severe, the product will not meet the standards of quality required for international com- mercialization. The catarina fishery is an important economic activity in Baja California Sur. The recent collapse of natural stocks has been Figure 2. The "Pimieiitilla" disease of catarina scallop Argopecten ven- tricostis (circiilaris). (a) Melanized spots on a dissected adductor muscle (bar = 1 cm), (b) Stephanostomum sp. metacercaria. the causal agent, stained with (iomori Trichromie (bar = 100 ]\m). Stephanostomum sp. in the Catarina Scallop 109 TABLE 1. Stephanostomum species for h liich Ihe intermediate hosts are known. Intermediate Host Species Redia Metacercaria Adult Author StephiiiUfsliinuini Icnuc Stcphaunsltnintiu iHU'ciiluiii Stcphaui'slintiiim jtij'onniini Stephuni>su>nntiu hi\puliiin Stt'phaiu)\lnimint hufHiituitiini Sfephuuostuniuiii Ciiniui^iufH Stephantislommn lophii Stepluinosloniiiiii .sp. Nti.ssit (ihsolchi Bncclniini undaliim Neptimca decumcostalum Unknown Unkniiu n Unknown Unknown Unknown Unknown Mcniihii nuniilui imtata Marine fishes Sole Marine fishes Eiif^nuiHs japimicus Fish Lotella physis Sebasles lompsoni Fish Aciinthogohuis hasla Marine Fish Sckieiui sp. Toenokles Uicepecii Hippociiinpus coroiuuiis Marine Fish Marine fish Lophitis piscatoriiis Catarinu scallop Unknown Martin (1939) Cahallero & Caballero (1952) Schell (1970) Caballero & Cahallero (1952) Ohnishi et al. (1991) Caballero & Caballero ( 1952) Caballero &. Caballero ( 1952) Quinteiro et al. (1993) Present report associated with overfishing and adverse environmental conditions. adductor muscle. Further research is necessary to determine the Consequently, the cultivation of this species is gaining importance effect o^ Stephanostomum outbreaks on natural and cultured popu- (Castro-Orti'z 1993). During our study, we found a high prevalence lations and the risks that this parasite may impose on scallop of infected scallops and a high intensity of cysts embedded in the culture. LITERATURE CITED Batjueno, E., 1. Pena & J. A. Masso. I9S1. Anahsis de una poblucion sobree.xplotada de Argopeclen cinuUiiis (sowerby, 1835), en la Ensenada de La Paz. B.C.S.. Mexico. Ciencia Pe.squera. I.N. P. Mexico Caballero. G. & E. Caballero. 1952. Revision de los generos y especies que imegran la Familia Acanthocolpidae Line, 1909. (Treinatoda: Dige- nea). Rev. Med. Ver. y Parasitol. 1 1(1-2):1-131. Castro-Ortiz. J. L. 1993. Fishery and cultivation of the catarina clam in Baja California Sur. Mexico, Bulletin No. 25. CICIMAR-IPN. July 1993. 1 p. Gomez del Prado, R. M. C. S. Alvai'ez & J. C. Perez-Urbiola. 1992. Algunos parasites de almeja catarina Argopecten circularis en Bahia Concepcion. B.C.S. Mexico. An. Inst. Biol. UNAM 63(2):271-272. Gomez del Prado. R. M. C. 1984. Echinocephatiis pseudouncinaitis Nem- atode Parasite de Argopecten circularis (Mollusca: Bivalvia) y Het- erodontus francisci (Pisces: Elasmebranchia) en la Costa Occidental de Baja California Sur. Mexico. Msc. Thesis. Institute de Ciencias del Mar y Limnologia. U.N.A.M.. 125 p. Love, M. S. & M. Meser. 1983. A checklist of parasites of California. Oregon, and Washington marine and estuariane fishes. NOAA Tcclnii- cal Report NMFS SSRF-777: 576 (P.). Manter. H. W. & H. J. Van Cleave. 1951. Seme digenetic trematodes. including eight new species, from marine fishes of La Jolla. Calif Proc. of the United States Museum. 101:315-320. Martin. W. E. 1939. Studies on the trematodes of woods hole 11. The life-cycle of Stephanostomum lenue (Linton). Biol. Bull. 77:65-73. Masse. R. J. A. 1996. Pesqueria dc AInieja Catarma. In: V.M. Casas and D.G. Ponce (eds.). Estudio del potencial pesquero y acuicela de Baja Cahfomia Sur. SEMARNAP. GOB. BCS. FAO. INP. UABCS. CIB- NOR. CICIMAR. 1:71-85. Moyer. M. A., N. J. Blake & W. S. Arnold. 1993. An ascetosporidian disease causing mass mortality in the Atlantic calico scallop Argopecten gibhiis (Linnaeus. 1758). J. Shellfish Res. 12(2):305-310. Ohnishi, Y.. T. One & T. Kifune. 1991. Metacercariae of stephanostomum hi.spidum (Yamaguti. 1934) (Trematoda. Digenea:Acantocolpidae) found in the marine goldeye rockfish. Sebasles thompsoni. Jap. J. Parasitol. 40:432^34. Quinteiro. P., J. Toje. A. Nuiiez. M. T. Santamarina & M. L.Sanmartin. 1993. Stephanostomum lophii sp. nov. (Digenea: Acantocolpidae). intestinal parasite of Lophius piscatorius. with reference to seasonal fluctuations of metacercariae in intermediate second host (Gadidae). J. Fish Biol. 42:421-133. Schell. S. C. 1970. How to Know Trematodes. Dubuque. Iowa: WM. C. Brown Company Publishers. Stunkard. H. W. 1961 . Cercaria Dipterocerca Miller and Norlhup. 1926 and Stephanostomum dcniaiiun (Linton. 1900) Manter. 1931. Biol. Bull. 120: 221-237. Whyte. S. K.. R. J. Cawlhoni & S. E. McGladdery. 1994. Ceinfection of bay scallops Argopecten irradians with Perkinsus karlssoni (Apicom- plexa. perkinsea) and an unidentified coccidian parasite. Dis. Aquat. Org. 18(1): 53-62. JiHinnil ol Shellfish Rcseanh. Vol. 20. No. 1. 111-115. 2(X1I. INFLUENCE OF THE GONADAL CYCLE AND FOOD AVAILABILITY ON POSTMORTEM CHANGES IN GLYCOGEN, ADENOSINE TRIPHOSPHATE, HYPOXANTHINE, AND THE 26()/250 ABSORBANCE RATIO IN ADDUCTOR MUSCLES FROM SCALLOP AEQUIPECTEN TEHUELCHUS (D'ORBIGNY, 1846) N. DE VIDO DE MATTIO,' M. E. FAREDI," AND M. CRUPKIN^* ^Centra Naciomd Patagdnko (CENPAT). Alte. Brown s/n Puerto Madryn. 9120 Chiihiit. ArgciUma: -Ceniro Regiomd Stir (CEMSUR-CITEP). Marcelo T. de Alvecir 1 I6H. 7600 Mar del Plata. Argentina ABSTRACT The influence of food avaikihility and llie reproductive cycle on the postmortem levels of glycogen, adenosine tri- phosphate (ATPl. hypoxanthine, and the 260/23(1 ahsorbance ratio In extracts of adductor muscles from scallop [Aecjuipecten lehii- elchus) was investisated. Phytoplaukton production, measured as chlorophyll "a" concentration, was high in early autumn (March to April) and early sprmg (September to October). Food availability was low In winter and summer. Specimens In good and poor biological condition, according to the determination of chemical composition and energy content, were found In periods of high and low food availability, respectively. Glycogen and ATP levels were higher In muscles from specimens In better biological condition. The rate of decrease of both glycogen and ATP was not dependent on the biological condition. The values of the 260/250 absorbance ratio were higher in extracts of cold-stored adductor muscle from scallops in better biological condition. Irrespective of the biological condition hypoxanthine Increased after 24 h In cold-stored muscles. Thereafter, the highest increments were observed in muscles from specimens In the best biological condition. KEY WORDS: ATP. biological condition, glycogen, gonadal stages, hypoxanthme. Aeqmpfcwn ichiu-lihii\ INTRODUCTION A large uinount of scallop is consumed as fresh tnaterial and is used as niateriai for dried, smoked, or canned products. In addition to that use. the consumption of frozen scallop adductor muscle has been remarkably increasing in the last years. On the other hand, it is widely accepted that nucleotides influence the taste and flavor of fish and mollusks meat (Matsumoto & Yamanaka 1990). Nucleotide degradation in postmortetn fish muscle proceeds as follows: adenosine triphosphate (ATPl —> adenosine diphosphate (ADP) -* adenosine monophosphate (AMP) -^ inosine mono- phosphate (IMP) -> inosine (HxR) -> hypoxanthine (Hx) (Kas- semsarn et al. 1963, Ehira & Uchiyama 1987). In marine inverte- brates, a major route for conversion of AMP to HxR via adenosine (Ado) rather than IMP was proposed (Saito et al. I95X, Aral I960. Hitlz & Dyer 1970). However, several reports are available on the accumulation of IMP in postmortem molluskan muscle (Sakamoto et al. 1973. Nakamura et al. 1976. Suweija et al. 1989). Contrary to the results of Saito et al. ( 19581 and Aral (1960). Suwetja et al. (1989) detected a small amount of IMP in the same species of cephalopods and bivalves. The reason for these apparently contra- dictory reports could be the different biological condition (season, sex. food availability, physiological condition, etc.) of the speci- mens analyzed. De Vido de Mattio et al. (1992) reported that glycogen and ATP levels were higher in adductor muscles from the bivalve mollusk Auiacomya ater ater (Molina) in good bio- logical condition. In this work, it was also reported that the values of the 260/250 absorbance ratio in perchloric acid muscle extracts from Auiacomya specimens in good biological condition were sig- nificantly higher than the values from specimens in poor biological *Corresponding author. E-mail: mcrupkinOonidp.edu.ar condition. Because of that, it was suggested that the biological condition of Auiacomya influences the ratio among ATP and catabolite compounds of ATP breakdown in stored muscles (De Vido de Mattio et al. 1992). A poor biological condition in the bivalve mollusk Auiacomya was found during the consumption of energy reserves for metabolic requirements and the development of the gonads in winter in response to insufficient availability of food (De Vido de Mattio 1980). Therefore, more investigation is necessary to determine the real influence of either the reproductive cycle or the availability of food on postmoilem changes in ATP and its catabolites in invertebrate adductor muscles. The scallop Aequipecten tehuelchus is a rhythmic hermaphro- dite with both gonadal growth and gonadal resting periods (Chris- tiansen & Olivier 1971. Lasta & Calvo 1978). Unlike Auiacomya, the gonadal growth period of Aequipecten occurs with food avail- ability in the environment. The purpose of the present work was to study the influence of both the gonadal cycle and availability of food on postmortem changes in ATP and catabolites compounds of ATP breakdown in adductor muscles of the scallop Aequipecten tehuelchus stored at 2— 1"C. MATERIALS AND METHODS Specimens of Aequipeeleu tehuelchus (D'Orbigny. 1846) were collected in the San Jose Gulf Chubut, Argentina. Mature speci- mens, 70 mm in length, were selected. Resting and gonadal growth peiiods were determined by macroscopic observation and histol- ogy of gonads (Lasta & Calvo 1978). After cleaning the shells, adductor muscles were dissected and carefully freed from adhering pancreatic and liver tissues. Adductor muscles were stored at 2-4°C for up to 10 days. Two experimental runs with mollusks in different biological conditions were performed. At zero time and at different periods of storage, adductor muscles from about 30 speci- 111 De Vido de Mattio et al. J J Months Figure 1. Seasonal changes in plivtoplanlitonic production expressed as chloropliyll "a" concentration. Each point is a mean of three deter- minations. Average relative deviations < ±3%. mens were cut into small pieces and thoroughly mixed to ensure homogeneity. Aliquots were taken for the following studies. Proximal Chemical Composition Lipids were determined from fresh homogenized material using Soxhlet (Pearson 1971). Water content v\as measured hy drying the tissues at lOOC until a constant weight was reached. Glycogen content was determined colorimetrically using the anthrone re- agent, according to the method described by Fraga (1956). Protein content was measured by the Lowry method (Lowry et al. 1951 ). Ash content was measured by ashing the tissue in a muffle furnace at 500°C. In all cases, the results correspond to the mean of trip- licate determinations. They are expressed as percentages of the wet weight. The energy values of the muscles were determined by the per- cent chemical composition, with Rubner's coefficients (e.g.. lipids 9.45. glucides 4.20. and proteins 5.65) (Winberg 1971). ATP Content A portion of pooled, cut. and mixed muscles was homogenized with 10% trichloroacetic acid (TCA). The homogenate was cen- trifuged at 700 x g. The residue was extracted again with 5% TCA. The supernatants were combined and neutralized with a 20% KOH solution under cooling with ice. The neutralized extract was cen- trifuged and the precipitate was washed with 59!- TCA neutralized with KOH. Aliquots of supernatants were taken to measure ATP by the luciferin-luciferase assay as described by Romano and La- borde (1978). adapted for bivalve adductor muscles. A PICO- ATP luminometer (Jo\'in and Ivon. Ste. France) was employed for light peak measurement. Hypoxanthine Ten grams of muscles were homogenized in a Waring blender with 30 mL cold T7c perchloric acid for 1 min. then filtered through Whatman 1 filter paper. The extracts were neutralized with 30% KOH. The KCIO4 precipitate was removed by centrifu- gation and the volume of neutralized extracts was adjusted to 50 mL. Hypoxanthine concentration was estimated by xanthine oxi- dase activity according to the procedure reported by Jones et al. (1964). The 260/250 Absorbance Ratio The 260/250 absorbance ratios were determined as the ratio of the absorbance of perchloric acid muscle extracts at 260 nm to that at 250 nm (Korhonen et al. 1990). One gram of tissue was ho- mogenized in a Waring Blender with 25 mL cold 20% perchloric acid for 1 min. then filtered through Whatman 1 filter paper. The filtrate was diluted with three volumes of distilled water. Absor- bance was measured immediately using a Beckman DU8 spectro- photometer at wa\'elengths of 250 and 260 nm. Statistical Analysis cans were compared by analysis of variance and the Student's /-test. Differences between slopes were determined by the Stu- dent's /-test. RESULTS AND DISCUSSION Because bivalve mollusks are primary consumers, the phy- toplankton supplied with particulate organic material constitutes the main source of food. The changes during the year in the phy- toplanktonic production, expressed as chlorophyll "'a" concentra- tion, are shown in Figure 1 . High phytoplankton production can be observed in early autumn ( 1.8 mg ■ m^') and early spring (2.8 mg • m^-'). Food availability was low in winter and summer. Lasta and Calvo (1978) reported that the reproduction of the Acquipeclcn lehuekluis occurs as follows: by the end of winter and during early spring, most of the individuals show the first stages of gonad ripening. At the beginning of summer (i.e., dating December), the ovocytes increase rapidly in size and ripe animals begin to spawn naturally. The spawning season lasts for most of the summer in the southern hemisphere, from December to February. The gonadal TABLE 1. Biochemical composition and energy value of adductor muscles from Aequipecten tehuelchus at different stages of the reproductive cycle. Reproductive cycle period Lipids Biochemical composition ( % ) Glycogen Proteins Ash Water content Energy value (Kcal/g) Gonadal growth Gonadal resting^ Gonadal resting'' 0.63 0.36 0.35 -^.33 7.31 2.70 I. "^.89 17.70 17.06 1.78 1.91 1.62 74..S0 76.25 78.70 1.17 \M 1.10 Each value is the mean of three determinations. Average relative standard deviations less than . ■' Beginning of gonadal resting period. '' End of gonadal resting period. Postmortem Changes in Adductor Muscles of Aequipecten 113 c 0) en o o _>^ O i^^H Glycogen r~~iATP > = Month Figure 2. Seasonal changes in ATP and glycogen content in adductor muscles of scallop. Each value is a mean of six determinations. Average relative deviations < ±3%. resting period starts in early autumn and finishes at the end of winter. Like many other bivalves, scallops have seasonal cycles of energy use that are intimately associated with the reproductive cycle (Ansell 1974. Barber & Blake 1991 ). The biological condi- tion of scallop at different stages of the reproductive cycle is shown in Table 1. The highest values of the relative percentages of glycogen and proteins were observed in the adductor muscles of specimens at the beginning of the gonadal resting period. Corre- spondingly, the energy content was higher in these samples and therefore specimens in this period present the best biological con- dition. The relative percentages of the different components of the adductor muscle and the energy content, except the relative per- centage of lipid, were slightly lower in specimens in the gonadal growth period than in those caught at the beginning of the gonadal resting period (Table I ). Because of that, specimens caught during the gonadal growth period remain in good biological condition. Variations in body components indicate either accumulation or use of storage substrates (Barber & Blake 1991 . Martinez & Mettifogo 1998). In Chlamys septemradiata and Peclen maxinuis energetic c o o *H u >. O \ — I — 48 — 1 — 72 0 24 Time (hr) at 2-42C Figure 3. Changes in glycogen content in cold-stored muscles: (■) gonadal growth period, (A) beginning of gonadal resting period, (O) end of gonadal resting period. Glycogen content is g/10(l g muscle. Each point is a mean of si.\ determinations. Average relative standard deviations < ±3%. Time (hr) at 2-4=0 Figure 4. Changes in .ATP content in cold-stored muscles. For other details see legend to Figure 3. reserves of the adductor muscles are used for gonadal maturation, producing an important decrease in the weight of the muscles (Ansell 1974. Comely 1974). Conversely, in Aequipecten tehiiel- chiis the weight of the muscles either did not decline or declined slightly during the gonadal growth period (Lasta & Calvo 1978). The results shown in Table I and those of Lasta and Calvo (1978) suggest that little of the energetic reserves of the adductor muscles were used for metabolic and reproductive purposes during the period of highest gonadal activity. It could be suggested that the energy requirements for reproduction of A. tehuelchus are mainly satisfied by the food taken from the environment. In agreement with these results, it was reported that during artificial conditioning of the sea scallop. Placopecten magellaiuciis. the mean weight of the adductor muscle was more than twice the weight observed in the wild specimen during gametogenesis. This indicates that the scallops were accumulating reserves despite the stress of gamete production (Paon & Kenchington 1995). Glycogen constitutes about 9.^7f of the total glucides that are present in adductor muscles (De Vido de Mattio 1980). The major anaerobic energy source in postmortem muscle is the breakdown of glycogen. As it can be seen in Figure 2. high and low levels of both glycogen and ATP were found in adductor muscles in periods of high and low food availability, respectively. These results are additional evidence of those energetic reserves that were stored in adductor muscles in periods with high food availability in the environment. Glycolysis profiles of cold-stored adductor muscles TABLE 2. Changes in 260/250 absorbance ratio of muscles extracts from Aequipecten tehuchhus at different biological conditions. Hour at 2^ C Poor biological condition Good biological condition 0 24 48 72 96 120 1.22 ±0.01* 1.17±0.01* 1.12±0.0I* 1.05 ±0.01* 1.03 + 0.01** 1. 01 ±0.01** 1.26 ±0.01* 1.20 ±0.01* 1.16±0.01* 1.08 ±0.01* 1.05 ±0.01** 1.03 ±0.01** Result are means *f <0.0l. ** P < 0.05. standard deviations (;; = 4). 114 De Vido de Mattio et al. E , '.;■ 0.0 - -^ 1 L-H- ^ 1 M— ^ ' 1— t— ^ i 0 24 48 72 96 Time (hr) at 2-4«C Figure 5. Changes in hypoxanthine content in cold-stored muscles: (G) gonadal growth period. Il I heginning of gonadal resting period. (■l end of gonadal resting period. Each value is a mean of six deter- minations. .Average relative standard deviations < ±39c. from A. tehuelcluis in different biological conditions are shown in Figure 3. Irrespective of the biological conditions of the speci- mens, glycogen content decreased within the first 1 2 h of storage. Glycogen content decreased slowly thereafter up to the end of storage. No significant differences (P < 0.05) in the slope of the curves were found within the first 12 h of storage. The results shown in Figure 3 indicate that the glycolysis rate is not dependent on the initial glycogen content of adductor muscles. Similar results were observed with Aulacomya adductor muscles stored at 2— l"C (De Vido de Mattio et al. 1992). Changes in the ATP concentration in stored muscles are shown in Figure 4. The ATP content increased for a short time after death because the nucleotide is regenerated by degradation of arginine phosphate prior to destruction of ATP. As expected, the maximum ATP content in muscles was dependent on the biological condition of the specimens. After the maximum was reached, the ATP con- tent fell sharply up to 12 h. No significant differences were found in the slopes of ATP decrease. Thereafter, nucleotide levels de- creased slightly up to 72 h of storage. The ATP content was unchanged after 72 h. Based on the measurement of the postmortem conversion of adenosine nucleotides to derivates IMP. HxR. and Hx. the R value was defined as the 250/260 absorbance ratio. Adenosine nucle- otides show maximum absorption at 259-260 nm. and IMP, HxR. and Hx at 249-250 nm. The inverse of the R value, the 260/250 absorbance ratio, was successfully u.sed to determine the effects of antemortem stress on the rate of onset of rigor mortis and associ- ated biochemical changes in fish muscle (Korhonen et al. 1990). In addition, it was also successfully used to demonstrate the influence of the biological condition of the bivalve mollusk Aulacomya on the ratio among ATP and catabolite compounds of ATP break- down in cold-stored muscles (De Vido de Mattio et al. 1992). The 260/250 absorbance ratios of perchloric acid extracts of stored molluskan adductor muscles are shown in Table 2. The values of the 260/250 absorbance ratio of about 1.26 and 1.22 were obtained at zero time of storage for sainples from specimens in gonadal resting period with good and poor biological condition, respec- tively. These values were significantly different (P < 0.01 ). Khan and Frey (1971) reported initial values of about 1.2 for well-rested beef and pork, the values subsequently declining to 0.89 or lower upon the passing of the samples into rigor. Korhonen et al. ( 1990) reported initial values of about 1.07 and 0.97 for muscles from unstressed and stressed fish, respectively. The initial values found by us in adductor muscles from scallop were similar to those of beef and pork and slightly lower than those reported for adductor muscles oi Aulacomya. During the storage, the values of the 260- 250 absorbance ratio of muscle extracts from specimens in poor biological condition were significantly lower than the values of muscle extracts from specimens in good biological condition (P < 0.01 at 24. 48. and 72 h; P < 0.05 at 96 and 120 h) (Table 2). In the experiments shown in Table 2 only specimens in gonadal rest- ing periods were used. Therefore, an infiuence of the food intake on postmortem changes in the ratio among ATP and catabolite compounds of ATP breakdown in stored muscles might be sug- gested. The profiles of Hx production in cold-stored adductor muscles of scallop are shown in Figure 5. Irrespective of either gonadal stage or biological condition of specimens. Hx levels lin- early increased about 2307r during the first 24 h. Thereafter, the increment in the Hx content v\as higher in muscles of scallop in better biological condition. In agreement with these results it was reported that the Hx levels reached since day 5 in cold-stored adductor muscles from Aulacomya in good biological condition were higher than the levels from specimens in poor biological condition (De Vido de Mattio et al. 1992). The results of this paper demonstrate that glycogen and ATP levels were higher in specimens in good biological conditions. The rate of decrease of both glycogen and ATP in adductor muscles was not dependent on the biological conditions of the scallops. Food availability infiuences both the biological condition of scal- lops and the postmortem changes in the ratio among ATP and catabolite compounds of ATP breakdown in stored muscles. ACKNOWLEDGMENTS This work was supported by CONICET and CIC. In addition, this work was partially supported by PICT 98 No. 09-03794 FONCYT. The authors to wish thanks to Dr. Nestor Ciocco from CENPAT-CONICET. Puerto Madryn, Chubut, Argentina by the provision of the samples. Aral, K. 1960. Acid soluble nucleotides in muscle marine inverte- brate. Effect of storing temperature upon the content of muscular nucleotide of some seashells. Bull. Fac. Fish Hokkaido Viiiv. II: 67-72. Ansell. A. D. 1974. Seasonal changes in biochemical composition of the bivalves Chlumys septemradiata from the Clyde Sea Area. Mar. Biol. 25:85-99. Barber. B. J. & N. J. Blake. 1941. Reproductive physiology. In: S. E LITERATURE CITED Shumway. editor. Scallops: Biology, ecology and aquacullure. New York: Elsevier, pp. 377-428. Christiansen, H. E. & S. R. Olivier. 197). Sobre el hermafroditismo de Chlumys lelmelcha (d"Orb.. 1846). An. Soc. Cient. .Art;. Tomo CXCI: Comely. C. A. 1974. Seasonal variations in the llesh weights and bio- chemical content of the scallop Pecten maximus L. in the Clyde Sea area. / Cons. Int. E.xplor. Mcr. 35:281-295. Postmortem Changes in Adductor Muscles of Aeqvipecten 115 De Vido de Mattio, N. 1980. Influencia de la temperatura y de la produc- cion primaria en la variacion estacional de la conipi»ici6n quimica y peso de Aulacoinyn uler atcr en Golfo Nue\o. Cliubut. Contribucion 38. CENPAT. Argentina. De Vido de Mattio, N.. M. E. Paredi & M. Crupkin. 1992. Postmortem changes in glycogen, ATP, hypoxanthine and 260/250 absorbance ratio in extract of adductor muscle from Aulacomya ater ater (Molina) at different biological conditions. Ciiiitp. Biocheiii. Physiol. I03A:605- 608. Ehira, S. & H. Uchiyama. 1987. Determmation offish freshness using the K value and comments on some other biochemical change in relation to freshness. In: D. E. Kramer & J. Liston. editors. Seafood quality de- termination. New York: Elsevier, pp. 185-207. Fraga, F. 1956. Detemiinacion de glucogeno en moluscos con el reactivo de Antrona. Iin: Pestiiieras. 3:69-74. Hitiz, D. F. & W. J. Dyer. 1970. Principal acid soluble nucleotides ui adductor muscle of the scallop Placopeclen imifielliiiuciis and their degradation during postmortem storage in ice. / Fish Res. Bd. Can. 27:83-92. Jones, N. R.. J. Murray. E. I. Livinston & C. K. Murray. 1964. Rapid estimations of hypoxanthine concentrations as indices of the freshness of chill stored fish. / Sa. Food Agric. 15:763-774. Kassemsarn. B., B. A. Perez, J. Murray & N. R. Jones. 1963. Nucleotide degradation in the muscle of iced haddock iGadiis aeglcfimis). lemon sole (Pleuronecles microce/jluiliis) and plaice iPltiiroiu'cU-s platcssu). J. Food Sci. 28:28-37. Khan, A. W. & A. R. Frey. 1971. A simple method for follouing rigor mortis development in beef and poultry meat. Can. Inst. Food Teclmol. J. 4: 139-141. Korhonen. R. W., T. C. Lanier & F. Giesbrecht. 1990. An evaluation of simple methods for following rigor development in fish. J. Food Sci. 55:346-348. Lasta, M. L. & J. Calvo. 1978. Cicio reproductivo de la vieira Chlamys tehuelchu del Golfo de San Jose. Coinunic. Soc. Malacologica del Uruguay 5:1-45. Lowry, O. F., N. J. Rosebrough, A. L. Fan- & R. J. Randall. 1951. Protein measurement with the folin reagent. / Biol. Chem. 193:265-275. Martinez. G. & L, Mettifogo, 1998, Mobilization of energy from adductor muscle for gametogenesis of the scallop. Argopeaen purpuraiiis Lamarck. / Shellfish Res. 17:113-1 16. Matsumoto, M. & H. Yamanaka. 1990. Post-mortem biochemical changes in the muscle of kuruma prawn during storage and evaluation of the freshness. Nippon Suisan Gakkuishi 56:1 14,5-1149. Nakamura, K., Y. Fujii & S. Ishikawa. 1976. Postmortem changes in the amounts of glycogen and ATP in adductor muscle of scallop. Bull. Tokai Reg. Fish. Res. Lab. 84:21-18. Paon, L. A. & E. L. R. Kenchington. 1995. Change in somatic and repro- ductive tissue during artificial conditioning of the sea scallop. Pla- copecten muguelkmicus (Gmelin. 1791). J. Shellfish Res. 14:53-58. Pearson, D. 1 971. The chemical analysis of food. 6th ed. New York: Chemical Publishing Company, pp. 13-15. Romano. J. C. & P. Laborde. 1978. Filtration et extraction "in situ" des adenosine-5' phosphates (ATP. ADP. ."SiMP) du materiel particulare de Lean de mer. Tethys. 8:197-202. Saito. T.. K. Aral & T. Tanaka. 1958. Changes in adenine nucleotide of squid muscle. Nature 181:1127-1128. Sakamoto, M.. Y. Fujii & K. Nakamura. 1973. Changes of the components in adductor muscle of scallop during storage. Bull. Tokai Reg. Fish. Res. Lab. 75:107-118. Suwetja. K.. K. Hori. K. Miyazawa & K. Ito. 1989. Changes in content of ATP-related compounds, homarine and trigonelline in marine inverte- brate during ice storage. Nippon Suisan Gakkaishi 55:559-566. Winberg. G. G. 1971. Methods for the estimation of production of aquatic animals. New York: Academic Press. Joiirikil of Shellfish Research. Viil. 20. No. I, I 17-120. 2001. KING SCALLOP (PECTEN MAXIMUS) DEPURATION TRIALS PHILIP HEATH' * AND MARTIN PYKE" ^ Centre for Murine Resoiirees and MarkuUure (C-Mar). The Queen's University of Belfast Marine Laboratory. The Strand. Portaferry. County Down. BT22 IPF. Northern Ireland. United Kingdom: -Seafish Industry Authority (SEAFISH). Seafisli Teehnology. Seafish House. St Andrew's Doek. Hull, HU3 4QE. United Kingdom ABSTRACT Preliminary trials were undertaken to assess the potential for depuration (purification) of diver harvested king scallops (Peelen maximus) under a range of temperatures and shellfish to water ratios. Scallops taken from class B waters were held for 42 hours in standard design depuration systems. Bacteriological analysis indicated a significant decrease in E. coli. and total viable count (TVC) at all temperatures and water ratios after the 42 hour depuration period. Mortalities at temperatures above 20.5°C and increased ammonia levels at low shellfish to water ratios indicated that acceptable depuration conditions He between 6.6°C and 15.6°C at shellfish to water ratios in excess of 1:12. Scallops held upside down also depurated effectively, although TVC indicated that in this orientation depuration was less effective than in scallops held the right way up. KEY WORDS: bacteriological analysis, depuration. Peelen muMmiis INTRODUCTION In order to reduce the risks associated with eating raw or lightly cooked bivalve niolkisks the food safety (hygiene) regulations 1998 (Fishery Products and Live Shellfish) require that bivalve mollusks from class B waters be depurated (purified) to reduce bacteriological contamination before releasing for human con- sumption. The Seafish Industry Authority (SEAFISH) has estab- lished protocols for depuration, based on temperature, salinity. flow rates, oxygen levels and loading density, for a number of shellfish species. These include; cockles (Cerastoderma edule). mussels [Mytihis ediilis). oysters (Ostrea edulis and Crassostrea gigas) and razor clams (Eiuis sp.) (Boulter et al. 1994. SEAFISH 1997). While protocols for de-gritting king scallops (Pecten maxi- miis) have been set (McNamara 1996) there are currently no pro- tocols for depuration of king scallops. Currently there are three scallop farms in Northern Ireland that either have full culture licenses or are in the process of receiving a culture license. One of these farms is known to be in a "Class B, shellfish harvesting area" and at least one other farm is likely to be classified as "B" when production starts. It is therefore imperative that depuration criteria be set for scallops in Northern Ireland in the near future. An initial trial aimed at depurating king scallops was run in August 2000. using SEAFISH standard design experimental depu- ration tanks. Aims • To determine whether depuration of king scallops is possible using theoretically normal conditions (salinity 31 psu and tem- perature 10°C-I5X) • To establish the temperature range over which king scallops will depurate under noimal saline conditions (31 psu) • To establish the effect of shell orientation on king scallop depu- ration • To determine a suitable shellfish to water ratio for depuration • To determine the effect of depuration conditions on post depu- ration survival *Corresponding author. E-mail: p.heath@qub.ac.uk; Fax: -1-44-28-427- 29672 MATERIALS AND METHODS On October 17. 2000 approximately 100 kg of diver harvested king scallops were collected from Mulroy Bay, Co. Donegal and transported by road to the Centre for Marine Resources and Mari- culture (C-Mar). Portaferry. Northern Ireland, a journey lasting 3.5 hours. On October 18, the scallops were transferred to plastic trays (NW trays) (dimensions 480 mm x 480 mm x 60 mm with 4 mm square holes), in a cupped side down position, at a density of 10 scallops per layer. These trays are used for scallop cultivation in the UK and Ireland (Hardy. 1991). Eight stacks of trays, seven layers deep were then transported 2.'i miles to Bangor. County Down by road. They were then transported by boat and placed on the sea bed at a shellfish cultivation site off Greencastle. inner Belfast Lough. (50°40.50' N, 005=52.77' W) This site is known to produce "Class B" mussels. On November 28. four tray stacks of scallops were recovered from Belfast Lough. On December II. the remaining four tray stacks were recovered. The surviving scallops were removed from the trays immediately and placed cupped valve down in fish boxes and covered with damp sacking. A sample of the surviving scal- lops was packed in a cool box and sent by courier to Aqualab, Killybegs, County Donegal, for microbiological analysis. The re- maining scallops were returned to the C-Mar Laboratory. Emer- sion time for scallops used for depuration trials did not exceed 3.5 hours. On arrival at Portaferry the scallops were weighed and placed into NW trays at a maximum of 10 scallops per tray, approxi- mately 30 live scallops (5kg) for each trial (Table 1 ). The trays were then transferred to experimental scale depura- tion tanks (Seafish Industry Authority standard design) for 42 hours. Temperature, salinity, ammonia (NH3 + NH^"") and oxygen levels were monitored throughout the trials. Salinity measurements were recorded with a WTW LF325 conductivity meter, ammonia measurements were made with a Bran Leubbe (Germany) autoan- alyser (Salicylate method) and oxygen and temperature measure- ments were taken with a WTW Oxi 340 oxygen meter. On completion of the depuration trial the tanks were drained and the scallops removed. All scallops were weighed and assessed for survival. Samples of 10 scallops, randomly selected from each trial, were 117 118 Heath packed in a cool box and sent by courier for post depuration microbiological analysis at Aqualab. Samples were screened in triplicate for E. coli (MPN lOOg"' flesh), and Total Viable Count (ACC cfu g"'){30°C 48 h). Pre-depuration samples were also screened for Salmonella. The remaining scallops were returned to NW trays in a flow through system with ambient seawater (8.5°C and 3 1 psu) and held for a further seven days to assess survival post depuration. The trial aimed to assess three variables within the depuration system: 1 . Temperature temperatures of ?. 6. 10. 14. 16. 20 and 24°C were tested at 31 psu and a shellfish to water ratio of 1:15; 2. Shellfish to Water ratio calculated as I kg of animals in II of water being equivalent to a ratio of 1:1. Ratio's tested were 1:3, 1:6. 1:9. 1:12 and 1:15. Temperatures were held between I3"C and 14°C: 3. Shell orientation scallops were held either cupped vahe down (light way up) or cupped valve up (upside down) at 31 psu and I4"C and shellfish to water ratio of 1:12; Flow rates within the systems were held at a constant 750 I hr"'. TABLE 1. Pre and post depuration TVC (xKl' g') and E. cnii (100 g"') counts for scallops [Pecten maximus) depurated at temperatures between 5°C and 20'C upside doH n and in shellfish to water ratios between 1:3 and 1:15(14 C). Trial Pre dtp. T\C (xlO'g-'l Pre dep. E. coli (100 g-') Post dep. TVC (xlO'g-') Temp. 5 494.3 Temp. 6 4.0 Temp. 10 4,0 Temp. 14 4.0 Temp. 15 4.0 Temp. 20 4.0 Temp. 24 494.3 Upside down 494.3 Ratio 1 3 494.3 Ratio 1 6 494.3 Ratio 1 9 494.3 Ratio 1 12 494.3 Ratio 1 15 4,0 Post dep. E. coli (100 g-') 2733 2.8 ± 1.3 437 1910 0.6 ± 0. 1 32 1910 0.8 + 0.5 <30 1910 1.0 ±04 <30 1910 1 . 1 ± 0.6 <30 1910 2.3 + 0.7 32 2733 3.9 ± 3.8 <30 2733 1.5 ±04 <30 2733 1.5 ±04 ^\/..^<' XSN <<>■< <109 mm SL). Estimates of efficiency from the depletion e.xperiment (24,3% and 29.5% respectively) were consistently lower than those from the diver surveys (38.0% and 40.7% respectively). This difference appeared to be due to inherent variation in the efficiency of scallop dredges rather than bias from either technique. This emphasizes the need for error terms to be built into estimates of dredge efficiency. The diver survey also found that dredges were highly selective toward scallops greater than 90 mm SL. catching only 3.0% or less of individuals below this size. Consequently, the diver survey provided a much more accurate assessment of scallop size and age composition than dredge surveys. Dredge efficiency was also assessed for four species of benthic fauna commonly taken as by-catch in the local fishery. Estimates of efficiency from the depletion experiment were found to include a considerable amount of indirect fishing mortality. When efficiency was defined as total mortality due to fishing (the combination of catch and indirect fishing mortality), estimates from the depletion experiment and diver surveys were in close agreement. For two species. Liiidia ciliciris (Phillipi 1837) and Ccmcer pat^urus (L. 1758). these efficiency or total mortality estimates were approximately 45%- and 68%^ respectively, emphasizing the impact scallop dredging might have on non-target species. In summary, we recommend that if possible, depletion experiments should be combined with diver surveys when assessing scallop dredge performance. Diver surveys provided additional information on dredge selectivity along with an improved measure of the variance in dredge efficiency estimates. KEY WORDS: scallops, dredge efficiency, gear selectivity, Pecien maxiimis. depletion experiment, diver survey, stock assessment INTRODUCTION Accurate assessment of the state of fish and shellfish popula- tions requires conversion of the results from stock surveys into estimates of ""true"" abundance (Date & Palmer 1994), Scallop (Pectinidae) populations are often surveyed by towing commercial scallop dredges along the seabed. This technique enables relatively rapid coverage of large areas, increasing the relevance of results to commercial fisheries. In general, however, scallop dredges are relatively inefficient (estimates range from 2.1% (Caddy 1968) to 56'7f (Currie & Parry 1999)) and tend to be selective toward large individuals (Chapman et al. 1977. Dare el al. 1993. Fifas &. Ber- thou 1999). The selective nature of dredges provides little infor- mation to fisheries managers on the abundance of upcoming year classes of young/small scallops. Many factors are known to affect the efficiency and selectivity of scallop dredges. These include mechanical aspects such as dredge set up and design (e.g. teeth length, angle and spacing, belly ring diameter and teeth bar tension), operational factors such as towing speed, duration and warp length, and environmental conditions such as sea state and the substrate type of the seabed (for review see Dare et al. 1993). In addition, efficiency and se- lectivity may vary between scallop species according to their swimming behavior and ability to recess into the sediment (Caddy 1968, Chapman et al. 1977. Dare et al. 1993). Hence any estimate of dredge efficiency will only be relevant to the conditions under which it was measured or must be seen as an estimate of the average situation (Somerton et al. 1999). The efficiency and selectivity of scallop diedges may be mea- sured by either indirect or direct methods. One indirect method is to measure scallop density and population structure at a study site "Corresponding author. Tel.: -h44- 1 624-83 1038; Fax: +44-1624-831001; E-mail;brycebs(aHi verpool.ac.uk using a highly efficient technique such as diver surveys and then compare this to a survey of the same area using scallop dredges (e.g. Fifas 1991). Another, more commonly used indirect method is through depletion (DeLury 1947. Joll & Penn 1990, Hilborn & Walters 1992, Currie & Parry 1999). Depletion experiments are based on fitting a model to the relationship between catch rate and cumulative catch when a given area of seabed is fished repeatedly. The main advantage of depletion experiments is that they are pos- sible on a large scale, are not limited to shallow water and are therefore particularly relevant to stock surveys and commercial fisheries. However, large numbers of replicate tows may be needed to sufficiently deplete populations and take account of variations in gear efficiency and patchiness of the stock. Regression equations used to calculate efficiency in depletion experiments can therefore be highly susceptible to outliers. Depletion experiments also pro- vide only limited information on selectivity, as size-classes that are rarely caught will not be depleted sufficiently for efficiency to be calculated. Direct methods for calculating scallop dredge efficiency and selectivity include monitoring the catch rate of tagged scallops seeded on the seabed at u known density (e.g. Dare et al, 1993), Possibly the most direct technique, though, is to survey dredge tracks on the seabed and compare what has been left behind with what has been caught (e.g. Caddy 1968, Chapman et al. 1977, Mason et al. 1979). Ideally, such surveys are done by divers who can count the scallops left in the dredge tracks and collect them for later measurement and ageing. Diver surveys are generally thought to be close to 100% efficient for surveying scallop populations (Mason et al. 1979, Coleman 1998), although very small individu- als may be missed, which could lead to overestimates of dredge efficiency for this size class. The spatial patchiness of scallop populations (Brand 1991) may also necessitate that divers cover large areas in order to reduce variability in efficiency estimates. This generally restricts diver surveys of dredge tracks to relatively 121 122 Beukers-Stewart et al. shallow fishing grounds and demands a high number of diver hours. All of these methods for estimating scallop dredge efficiency and selectivity have certain advantages and disadvantages. To date, however, there have been no direct comparisons of the ef- fectiveness of these different methods when applied to the same local scallop population at the same time. In this study we aimed to compare the outcomes and effectiveness of a concurrent deple- tion experiment and diver survey of dredge tracks on the same fishing ground. The work was done in the north Irish Sea off the Isle of Man, which has supported a commercial fishery for the great scallop, Pccten maximus. for over 60 years. Over 20 years of data from detailed stock surveys of P. maximus around the Isle of Man are available and interpretation of these data would be greatly enhanced by accurate estimates of dredge efficiency and selectiv- ity. In addition to examining results for P. inaxinms. we also com- pared efficiency estimates and non-target mortality for four species of benthic megafauna that are commonly taken as by-catch in the Isle of Man scallop fishery. Accurate efficiency estimates for these species would allow further predictions to be made regarding the environmental impact of scallop dredging around the Isle of Man. a subject that has received considerable recent attention (e.g. Hill et al. 1999, Bradshaw et al. 2000, Kaiser et al. 2000, Veale et al. 2000b). of 2.45 times. Efficiency of the dredges was calculated for two size classes of scallops, 90 to 109mm shell length (SL), (under the minimum legal landing size (MLLS)) and 1 10mm SL or greater (above the MLLS). using the Leslie-Delury method (DeLury 1947, Hilborn & Walters 1992). This method plots the number of scal- lops caught during each sequential tow against the cumulative catch. The slope of the regression fitted to this plot gives a measure of the rate at which catch rates have changed during the experi- ment. When this slope is multiplied by the ratio of the total area of the experimental plot to the area covered by each tow, a measure of efficiency is given. In this experiment the ratio was 148,160 m" (the area of the plot) divided by 25,928 m" (the width of 8 dredges (7 m) X the length of the tow (3.704 m)). which equals 5.71428. Standard errors of these efficiency estimates were calculated from the standard error of the slope of each regression. Regression slopes for the two size classes of scallops (90 to 109 mm and >I 10 min) were compared by ANCOVA with the fixed factor being size class, the dependent variable number of scallops caught per tow and the covariate cumulative catch. Differences in efficiency are indicated by a significant interaction between cumulative catch and size class. Using the same method the o\erall efficiency of the dredges for catching the four by-catch species, A. nihens. L. ciliaris. P. piil- villiis. and C. pagiinis. was also calculated. MATERIALS AND METHODS Dredging for both the depletion experiment and diver surveys was done by the RV Roagan, a 24-m conveiled beam trawler, on the Bradda Inshore fishing ground off the Isle of Man on five days between July 12 and July 21, 2000. On each day of the study two tows were done for the dive surveys (one in the morning and one in late afternoon), with three tows in between for the depletion experiment. Weather conditions were very calm on survey days, with wind speeds less than 10 knots. Sea temperature was approxi- mately 15°C and underwater visibility 3 to 5 m. Water depth in the study area ranged from 29m to 35m and sediment was generally a gravelly mixture of sand, shell, and mud with occasional large stones (Jenkins et al. 2001 ). These features are typical of many of the scallop fishing grounds around the Isle of Man. A gang of four Newhaven spring-toothed scallop dredges (total width 3.50 m) was towed off each side of the boat. Mean tooth length on the dredges at the start of the study was 76 mm (± 1 mm SE). To simulate our stock surveys and local commercial practice, all tows were ap- proximately two nautical miles (3,704 m) in length, which at an average speed of 2.77 knots (5.13 km/hour) lasted 45 to 50 min- utes. At the end of each tow the catch was landed and sorted on the deck. All scallops (Pecten maximus) were counted, measured, and aged to the nearest year using clearly visible validated annual shell rings (Allison et al. 1994). Numbers of the four other most com- mon species of benthic fauna in the catch, the starfish Asterias ruhens (L. 1758), Liiidia ciliaris. Poraiiia ptilvilliis (Miiller 1776). and the edible crab. Cancer pagtinis. were also counted. Depletion Experiment An area of seabed, two nautical miles (3,704 m) long by 40 m wide (54°07'N, 4°48'W), was selected and identified for the deple- tion experiment using an onboard differential global positioning system (DGPS) linked with Microplot software (Sea Information Systems, Aberdeen). Fourteen tows were done haphazardly within this area, after which time the seabed had been dredaed an average Diver Surveys Diver surveys of dredge tracks were done on 10 tows of scallop dredges, conducted within an area of 2.74 knr, approximately 200 m inshore from the depletion experiment plot. This spatial sepa- ration was necessary to ensure the safety of the divers, as the diver survey was conducted at the same time as the depletion experi- ment. Dredge tows were done using the same configuration and procedures described above and were of as similar distance and duration as possible. Sorting and measuring of the catch was also done as above. During these tows the dive support vessel. RV Sula, dropped buoyed shot lines at two points, approximately 500 m apart, be- tween the two sets of dredge tracks. This was done on both the first and second half of tows during the course of the study. After a period of 15 to 30 minutes, between two and four pairs of divers descended to the dredge tracks. The divers then swam along the tracks as far as possible given depth and air restrictions and counted all benthic fauna encountered. AH scallops were also col- lected to allow measurement and ageing on the surface. The dis- tance covered by the divers was calculated from DGPS readings taken from the buoyed shot lines and surface marker buoys re- leased by the divers at the end of their surveys. From these dive surveys the efficiency of the scallop dredges for catching scallops and the four by-catch species could be cal- culated using the following formula: Efficiency = (Density from dredges / Density from dredges -t- dive surveys) x 100 For the purposes of the study this formula assumes the dive sur- veys to be 100% efficient, which may or may not always be the case (see discussion). Data from the dive surveys were pooled for each tow (ie. from two to four pairs of divers) to allow tows to be used as replicates. The effect of number of tows (or replicates) on the magnitude and variability of efficiency estimates for catching all scallops was plotted. Along with overall efficiency estimates. Efficiency and Selectivity of Scallop Dredges 123 TABLE 1. Comparison of nu'uii scallop (iridic t'fficiencv estimates from the depletion experiment (h for Pci'ti'ti inaxiniiis ). 14) and diver surveys of dredge tracks (h = 10 Depletion experiment Diver surveys Species R^ P Efficiency SE Efficiency SE Peclen maximiis (all) — — — 22.69 4.65 P. inaximus (<90 mm) — — — — 3.01 0.66 P. muximus (90-109 mm) 0.64 0.001 24.34 5.14 37.99 3.73 P. muximus (110 + mm) (1.6? <0.001 29.54 6.29 40.07 5.05 Cancer pagiinis U.50 0.005 6X.45 20.00 24.97 — Liiidia ciliahs 0.63 0.001 47.49 10.86 20.36 — Porunia piiJyiUus 0.18 0.01.^ 21.37 13.14 16.27 — Aslerias riihcns <().0I ().H16 — — 9.85 — efficiency was calculated tor three size classes of scallops; less than 90 mm SL. 90-109 mm SL and 1 lOmm SL or greater. A 1-way ANOVA. followed by a Tukey's test, was run on these data to determine if theie were any differences in efficiency for the different size clas.ses. These efficiency estimates were also com- pared with those from the depletion experiment. Size and age compositions of scallops collected during the diver surveys were also compared with those from the dredge catch to further investigate the selectivity of the scallop dredges. By combining data from the dive surveys and dredge catches, estimates of age and size composition, and density, at the study site were also calculated. RESULTS Depletion ExperimenI The slopes of regressions between sequential catches and cu- mulative catch for the two size classes of scallops and for Liiidia ciliaris, Poranki piilvillus and Cancer pagiinis were significant (Table 1 ). This allowed efficiency estimates to be calculated for these five groups but not for the starfish Asterias riihens. Regres- sion slopes, and therefore efficiency estimates (24.3% and 29.5%), were not significantly different for the two size classes of scallops (Table 21. For the by-catch species, efficiency ranged from 21.4% for P. piilvilhis to 68.5% for C. p(ii;nrus (Table 1 ). Diver Survey During the diver survey a total of 28 dives (one pair of divers per dive), covering a distance of 2.99 km, were done on the dredge TABLE 2. Results of ANCOVA comparing regressions from the depletion experiment for two different size classes of scallop, Pecten maximus (90-1119 mm and 110+ mm shell length). Source SS df MS F P-\AlK Model 60732.87 3 20244.29 65.74 <0.001 Intercept 132976.18 1 132976.18 431.81 <0.001 Size X Cum. catch 81,29 1 81.29 0.26 0.612 Size 14483.51 1 14483.51 47.03 <0.001 Cum. catch 8763.46 1 8763.46 28.46 <0.001 Error 7390.84 24 307.95 R- = 0.89 tracks. From these surveys the overall efficiency of the dredges at catching scallops was calculated as 22.7% (± 4.7 SE). The mag- nitude and variability of efficiency estimates was relatively stable after five tows (Fig. I ). Efficiency was significantly lower for scallops less than 90 mm SL (3.0%) than for larger scallops (38.0% and 40.7%; Tables 1 and 3. Fig. 2). Comparison of the Depletion Experiment and Diver Surveys For scallops greater than 90 mm SL the diver survey estimates of efficiency were approximately 10 to 15% higher than those from the depletion experiment (38.07f and 40.7% compared to 24.3% and 29.5%; Table 1, Fig. 2). For two of the by-catch spe- cies. L. ciliaris and C. pagurus. efficiency estimates from the diver surveys were less than half those from the depletion experiment (L. ciliaris; 20.4% compared to 47.5% and C. pagitnis: 25.0% com- pared to 68.4%; Table 1). For the starfish P. pulvilltis, however, estimates of efficiency from the two methods were similar (Diver surveys: 16.3%. Depletion experiment 21.4%). Unfortunately, due to low numbers of these by-catch species on individual dives, this dive survey data had to be pooled and results from the two meth- ods could not be compared statistically. The diver surveys also gave an efficiency estimate of 9.3% for the starfish A. riibens but there was no comparable estimate from the depletion experiment. Analysis of scallop size and age structures showed considerable differences between the diver surveys and the dredge catches. 5(1 - 40 >-. e 130^ 20 - 10 ^ 6 No. tows The interaction between size and cum. catch (cumulative catch) lest^ for difference-- in efficiency. Figure 1. The relationship between the number of tows surveyed dur- ing the diver survey and estimates of dredge efficiency (mean ± SE) for all scallops, Pecten maximus. 124 Beukers-Stewart et al. TABLE 3 45 Results of ANOVA compa ring efficiency estimates from the diver 40 surveys or th -ee different size classes of scallop. Pecleii maxiinm (<90 mm, "Jd- 1(19 mm. 110+ mm shell length). 35 - 30 Source SS df MS F P I 25 Model 1.73 1 0.S7 55.51 <0,001 l:o Intercept 7.63 1 7.63 489.01 <0.001 Size 1.73 1 0.87 55.51 <0.001 ^ 15 Error 0.42 27 0.02 10 R- = 0.80 5 Data were Niarcsine transformed to meet assumptions of normality and homogeneity of variance. Results of Tukey's HSD test comparing efficiency estimates from the diver surveys for the three different size classes of scallop. P. muxiinus: (<90 mm) < (90-109 mm) = (110+ mm). I- 0 Dredges □ Divers ■ Combined 1 1*. W-T , Size Class (nun) Scallops collected by the dredges were mainly in the 100-109 mm SL size class, while the most common size class collected by divers was 80-89 mm SL (Fig. 3a). In addition, approximately 30% of the scallops collected by divers were less than 80 mm SL. whereas these scallops were almost entirely absent from the dredge catches. This pattern was reflected by data on age composition (Fig. 3b). Most scallops caught by the dredges were 3-year-olds, while most collected by the divers were 2-year-olds. One-year-olds were not particularly common in the diver surveys (7.7%) but they were almost completely absent from the dredge catches. In both cases the combined size and age compositions were much better represented by the diver surveys than by the dredge catches. By combining the dredge catches with the diver surveys, the density of scallops. P. maxiimis. at the study site was estimated to be 4.4 per 100m- (± 0.7 SE). DISCUSSION For the scallop. Peclen ina.xiinus. both the depletion experiment and the diver surveys produced relatively precise estimates of dredge efficiency that were within the range reported in other studies (Chapman et al. 1977. McLoughlin et al. 1991. Dare et al. 1993. Currie & Parry 1999). Estimates from the diver surveys. 5U - 45 - — -o — Diving Depletion r___ 35 - ^K £30- .1 -=■ £ 20 J / ( '"" 15 - 10 - 5 ■ ^ 0 - C90 90- 109 Size Cla.ss (mm) Figure 2. A comparison between dredge efficiency estimates (mean ± SE) from (he depletion experiment (h = 14) and diver survey (h = 10) for three size classes of scallop, Pecteii maximus. .m^ 1^ - , Age Class (yrs) Figure 3. .A comparison of the percentage composition of a) size classes and bl age classes of scallops. Pecten maximus collected b> divers and dredges during the diver survey (h = 350 for divers and n = 2052 for dredges). however, were moderately, but consistently, higher than those from the depletion experiment. This could be due to biases in either method, or a combination of both. Depletion experiments are most effective when examining closed populations with little or no recruitment or natural mortality (Hilborn & Walters 1992). Although such populations rarely exist, conducting a depletion experiment over a short period of time (as in this study) will often negate the effects of recruitment and mortality (Joll & Penn 1990). Movement of the target species into or out of the study area may be more of a problem (Hilborn & Walters 1992). Given the relatively narrow width (40 m) of the depletion experiment plot, it is possible that scallops were moving into the plot during the course of the study. This would have lowered efficiency estimates. A wider plot would have reduced the potential for this to be a problem. In this .study, however, we were aiming to replicate our stock surveys, which consist of long par- allel tows. Without considerably more effort a wider plot would have necessitated shorter tows with frequent hauling of the gear. This may well have affected dredge efficiency estimates (Dare et al. 1993). making them less relevant to the commercial fishery. In addition. P. ma.xiiiui.'i is one of the most sedentary species of scal- lop (Brand 1991) and in a tagging study by Howell & Eraser Efficiency and Selectivity of Scallop Dredges 125 ( 1984). 60'^(- of individuals moved less than 30 m after 18 months. Immigration is therefore likely (o have had only had a small effect on our results. Another aspect that may have affected the results from the depletion experiment is a change in the catchability of scallops during the study (Hilborn & Walters 1992). Repeated dredging of an area may cause scallops to become unrecessed, even if they are not caught (Currie & Parry 1999). This would actually have in- creased the efficiency of the dredges, however, whereas the deple- tion experiment produced lower efficiency estimates than the diver survey. One factor that may ha\e biased the di\er survey estimates is patchiness of the scallop population (Brand 1991). Despite sur- veying almost 3 km of dredge tracks. 74 km of tracks were left unsurveyed. If. by chance, dive surveys focused on low-density patches of scallops, this would have elevated efficiency estimates. However, mean efficiency estimates in the dive survey were rela- tively stable after five tows, indicating that our level of replication and area coverage was adequate. Another possibility is that the diver survey was less than 1009(- efficient, which also would have elevated dredge efficiency estimates. Dive surveys are generally considered to be the most efficient method for assessing scallop populations (Mason et al. 1979. Coleman 1998), so it is not really possible to test their efficiency. In this study each diver carefully surveyed transects only 1.75m wide in underwater visibility of at least 3m. We are therefore confident that the dive survey was very close to 1007f efficient, at least for scallops greater than 90mm SL. It is possible that we underestimated the number of scallops less than 90mm SL (see below), but this would only have changed our efficiency estimate of 3% for this size class to a lower value and would not have altered the conclusions of the study. Finally, dif- ferences between the two methods could have been due to varia- tion in substrate type between the two study sites, although given their similarity and close proximity this is unlikely. The efficiency of scallop dredges is inherently variable (Dare et al. 1 993 ). Therefore, for the purpose of stock assessment a mean of the efficiency estimates from the two methods is probably most appropriate. The differences between them (approximately \07c) could be built in as an emir term when calculating population densities. Again it must also be emphasized that these efficiency estimates are only really applicable to fishing grounds of a sub- strate type similar to that examined in this study. Some of the fishing grounds around the Isle of Man are much more sandy than the study site, while others are stonier. Ideally, efficiency experi- ments will be repeated on those other grounds in the near future. In common with many other studies of dredge selectivity (Chapman et al. 1977. McLoughlin et al. 1991. Dare et al. 1993), we found spring-toothed scallop dredges to be very selective to- ward larger scallops. Dredges almost completely missed scallops below 90 mm SL, although in contrast to some other studies (e.g. Mason et al. 1979, Dare et al. 1993) there were no significant differences between the efficiency of the dredges for the two larger size classes (90-l(39mm and llOmm -i- SL). The implications of this selectivity for stock assessment can be seen from Figure 2. On the basis of the dredge survey, the population appeared to be dominated by 3-year-olds (100-109 mm), while in reality 2-year- olds (80-89 mm) were much more abundant. Not surprisingly, the dredge survey did not catch many I -year-old scallops either, but these were also rarely seen in the diver surveys. This could be due to poor recruitment in the previous year, but may also be due to their low visibility on the seabed (pers. obs). Diver surveys may therefore only be fully effective for P. ma.\imus of two years or older. For two of the by-catch species. Cancer pai>iinis and Liiidia ciliaris. efficiency estimates from the depletion experiment were more than twice those from the diver surveys. This was a surpris- ing result, until the effect of scallop dredging on the indirect mor- tality of by-catch species was considered (see also McLoughlin et al. 1991 ). As an extension of this study, damage to benthic organ- isms which encountered dredges without being captured, was also examined (Jenkins et al. 2001 ). On the basis of that study, 56% of C. pagunis left in the dredge tracks were expected to die (or be eaten) within a day of dredging (see also Hill et al. 1996, Veale et al. 2001). The respective figure for L. ciliaris was 31%. If it is considered that these additional individuals were effectively re- moved by dredging, values for total fishing mortality (catch plus indirect mortality) can be calculated. These work out as 67.0% for C. pafiiiriis and 44.8% for L ciliaris. very similar to the estimates of efficiency from the depletion experiment (68.5% and 47.5%, respectively). For the third species Porania pulvillus, none of the individuals left in the dredge track were expected to die (Jenkins et al. 2001 ). In this case, however, estimates from the depletion ex- periment and diver surveys were relatively similar (21.4% com- pared to 16.3%). Unlike the diver survey, the depletion experiment incorporated indirect fishing mortality into estimates of efficiency by measuring the decline in the abundance of live animals over the course of the study (10 days). It was therefore deemed appropriate to compare total fishing mortality from the diver surveys with estimates of efficiency from the depletion experiment. Indirect mortality would have had little effect on the depletion experiment results for P. iiuiximiis. as only 5% of indi\ iduals left in the dredge track were expected to die. The inability to calculate efficiency from the depletion experiment for the fourth by-catch species, Astcrias ruhcns. was probably due to this species" high mobility and tendency to aggregate on dredge tracks (Kaiser and Spencer 1996, Ramsay et al. 1998, Veale et al. 2000a). In summary, this study found that both depletion experiments and di\er surveys of dredge tracks were effective for assessing scallop dredge efficiency. Differences between the results from the two methods were probably due to the inherent variability of dredge efficiency and emphasize the need for error terms to be built into efficiency estimates. The depletion experiment was more cost-effective (in terms of person hours) than the diver surveys, but provided little information on dredge selectivity. We therefore suggest that a combination of approaches is probably most appro- priate. ACKNOWLEDGMENTS This study was supported by the Department of Agriculture, Fisheries and Forestry of the Isle of Man Government and the EU project ECODREDGE (FAIR CT98-4465). Many thanks to the boat crews of the RV Roagan (P. Crebbin, R. Georgeson and R. Gillam and G. Heaney) and RV Sula (R. Johnson and D. Kneen) and the many divers (M. Bates. C. Bradshaw. P. Collins. M. Mos- ley, C. Mullens, L. Patchell. K. Ross. T. Szebeni and L. Veale) and onboard scientists (R. BIyth, S. Buttle, G. Hughes, C. O'Donnell and T. Savage) who made the study possible. 126 Beukers-Stewart et al. LITERATURE CITED Allison. E. H.. U. A. W. Wilson & A. R. Brand. 1994. .Age determination and the first growth ring North Irish Sea populations of the scallop. Peclen imiximus (L.). / Molt. Stud. 60:91-9.^. Bradshaw. C. L. O. Veale, A. S. Hill & A. R. Brand. 2000. The effects of scallop dredging on gravelly sea-bed communities. In: M. J. Kaiser & S. J. de Groot. editors. Effects of fishing on non-target species and habitats. Tunbridge Wells. UK: Gray Publishing, pp. 83-104 Brand, A. R. 1991. Scallop ecology: distributions and behaviour. In: S. E. Shumway. editor. Scallops: biology, ecology and aquaculture. Amster- dam: Elsevier, pp. 517-584 Caddy. J. F. 1968. Underwater observations on scallop {Placopectcii ina- gellanicus) behaviour and drag efficiency. / Fi.<:h. Res. Board Can. 25:2123-2141. Chapman. C. J., J. Mason & J. A. M. Kinnear. 1977. Diving observations on the efficiency of dredges used in the Scottish fishery for the .scallop. Peclen maximiis (L.). Scottish Fisheries Research Report 10. 16 pp. Coleman. N. 1998. Counting scallops and managing the fishery in Port Phillip Bay, south-east Australia. Fish. Res. 38:145-157. Currie. D. R. & G. D. Parry. 1999. Impacts and efficiency of scallop dredging on different soft substrates. Can. J. Fish. Aqua!. Sci. 56:539- 550. Dare. P. J., D. Key & P. M. Connor. 1993. The efficiency of spring-loaded dredges used in the Western English channel fishery for scallops. Pecren ma.ximus (L.). ICES CM 1993/3:15. 17 pp. Dare. P. J. & D. Palmer, 1994, The use of efficiency factors for estimating indirectly population composition and abundance of scallops. In; N. F. Bourne. B. L. Bunting & L. D. Townsend. editors. 9th International Pectinid Workshop. Can. Tech. Rep. Fish. .-^quat. Sci. Nanaimo, BC, Canada, pp. 137-142 DeLury. D. B. 1947. On the estimation of biological populations. Biomet- rics 3:145-167. Fifas, S. 1991. Analyse et modelisation des parametres d'exploitation du stock du coquilles Saint-Jacques (Pecten ma.ximus. L.) en bale de Saint- Brieuc {Manche Ouest, France). PhD thesis, Universite de Bretagne Occidentale, Brest. 422 pp. Fifas, S. & P. Berthou. 1999. An efficiency model of a scallop iPecten ma.ximus. L.) experimental dredge: Sensitivity study. ICES J. Mar Sci. 56:489-199. Hilborn. R. & C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics and uncertainty. New York: Chapman and Hall. 570 pp. Hill, A. S., A. R. Brand. U. A. W. Wilson, L. O. Veale & S. J. Hawkins. 1996. Estimation of by-catch composition and numbers of by-catch animals killed annually on Manx scallop fishing grounds. In: S. P. R. Greenstreet & M. L. Tasker. editors. Aquatic predators and their prey. Oxford: Blackwell Scientific, pp. 111-115 Hill, A. S.. L. O. Veale, D. Pennington. S. G. Whyte, A. R. Brand & R. G. Hartnoll. 1999. Changes in Irish Sea benthos: possible effects of forty years of dredging. Estuar. Coast. Shelf. Sci. 48:739-750. Howell. T. R. W. & D. I. Fraser. 1984. Observations on the dispersal and mortality of the scallop. Peclen ma.ximtis (L.). ICES CM 1984/K. 13 pp. Jenkins. S. R., B. D. Beukers-Stewan & A. R. Brand. 2001. The impact of .scallop dredging on benthic megafauna: a comparison of damage levels in captured and non-captured organisms. Mar. Ecol. Prog. Ser. 215: 297-301. Joll. L. M. & J. W. Penn. 1990. The application of high-resolution navi- gation systems to Leslie-DeLury depletion experiments for the mea- surement of trawl efficiency under open-sea conditions. Fish. Res. 9:41-55. Kaiser. M. J.. K. Ramsey. C. A. Richardson, F. E. Spence & A. R. Brand. 2000. Chronic fishing disturbance has changed shelf sea benthic com- munity structure. / Animal Ecol. 69:494-503. Kaiser, M. J. & B. E. Spencer. 1996. Behavioural responses of scavengers to beam trawl disturbance. In: S. P. R. Greenstreet & M. L. Tasker. editors. Aquatic predators and their prey. Oxford: Blackwell Scientific, pp. 116-123 Mason. J.. C. J. Chapman & J. A. M. Kinnear. 1979. Population abundance and dredge efficiency studies on the scallop. Peclen ma.ximus (L.). Rapp. P.-v. Reun. Cons. int. Explor. Mer 175:91-96. McLoughlin. R. J., P. C. Young, R. B. Martin & J. Parslow. 1991. The Australian scallop dredge: estimates of catching efficiency and associ- ated indirect fishing mortality. Fish. Res. 11:1-24. Ramsay. K.. M. J. Kaiser & R. N. Hughes. 1998. Responses of benthic scavengers to tlshing disturbance by towed gears in different habitats. J. Exp. Mar. Biol. Ecol. 224:73-89. Somerton, D.. J, lanelli, S. Walsh. S. Smith. O. R. Godo & D. Ramm. 1999. Incorporating experimentally derived estimates of survey trawl effi- ciency into the stock assessment process: a discussion. ICES J. Mar. Sci. 56:299-302. Veale, L. O., A, S, Hill & A. R. Brand. 2000a. An in-situ study of predator aggregations on scallop (Pecten ma.ximus (L.)) dredge discards using a static time-lapse camera system. J. Exp. Mar. Biol. Ecol. 255: 1 1 1-129. Veale. L. O., A. S. Hill. S. J. Hawkins & A. R. Brand. 2000b. Effects of long-term physical disturbance by commercial scallop fishing on sub- tidal epifaunal assemblages and habitats. Mur. Biol. 137:325-337. Veale. L. O., A. S. Hill. S. J. Hawkins & A. R. Brand, 2001. Distribution and damage to the by-catch assemblages of the northern Irish Sea scallop dredge fisheries. J. Mar Biol. Assoc. U. K. 81:85-96. Jininwl ofSlwllJhh Ri-seairh. Vol. 20. No. 1, 127-134, 20(11. PROBLEMS. PREDATORS, AND PERCEPTION: MANAGEMENT OF QUAHOG (HARDCLAM), MERCENARIA MERCENRIA, STOCK ENHANCEMENT PROGRAMS IN SOUTHERN NEW ENGLAND WILLIAM CAMERON WALTON' * AND WILLIAM CHARLES WALTON" '//()/-;! Point Lahomtory. University of Maryland. Cambridge. Maryland 2161 3: -Walton Research. Inc.. 125 Tiirtlehack Road. Califon. New Jersey 07830 ABSTRACT Throughout southern New England and Long Island Sound, northern quahogs (Mercenaria mercenarla Linnel. support large commercial and recreational fisheries. Municipal managers implement a number of tools to maintain the public stock, including enhancement with hatchery-reared juveniles, or seed. In 1999, we surveyed 68 municipal managers in the region (59 in Massachusetts and 9 from Long Island. New York) to identify the extent and magnitude of current programs and any obstacles to the success of such programs, specifically including predation. Of the 36 responses. 23 had used or were using M. mercenciria seed, planting an average of 1.4 nnllion seed (+ 355.000) at an average shell length of 15.9 mm (± 1.4) on a variety of substrates. Estimates of seed loss in the first year varied widely, but averaged 44'/f (± 6.2). Estimates of survival to market size were significantly higher (P = 0,023) in Massachusetts (499^ ± 8.5) than New York (25'7r ± 4.0), the only significant difference between the two states. Losses were attributed to a variety of factors, but all respondents indicated predation was a major cause of that loss. Of predators identified as threats to M. mercenaria public seeding, green crabs. Carcinus maenas. were cited most frequently and scored as very serious predators (8.3 on a scale of 1 to 10 with 10 being the most damaging); other crabs, starfish, and gastropod predators were also perceived as serious threats. Managers were divided evenly on the effectiveness of predator trapping in reducing predator populations either within or among years in an attempt to protect seed. Our survey identified several areas where additional scientific study would benefit these municipalities: (1) assessment of seed survival to market size; (2) experimental demonstration of predators responsible for losses; and (3) evaluation of the effectiveness of trapping in both reducing predator populations and protecting seed. KEY WORDS: public stock enhancement. Menenana mencmina. Canuui<. maenas. quahog. hardclam INTRODUCTION Throughout southern New Englatid and Long Island Sound, northern quahogs, also known as hard clams (Mercenaria merce- naria Linne). support large commercial and recreational fisheries; for example, the commercial harvest in New York was over $17 million per year from 1995 to 1999 (pers. comm.. National Marine Fisheries Service. Fisheries Statistics and Economics Division, Sil- ver Spring, MD). Management of these fisheries, although subject to state and federal restrictions, is primarily carried out by munici- pal programs, which implement a variety of management tools to maintain the resource. One strategy available to resource managers is to supplement the existing wild stock with hatchery-reared juvenile "seed" sup- plied by commercial and public shellfish hatcheries. Planting seed on public grounds may enhance the fishery directly, by introducing clams that survive and grow to a legally harvestable size, and indirectly by increasing larval supply and subsequent year classes. A more subtle benefit may be a dependable interannual supply of littlenecks and cherrystones, the two smaller and most valuable market size categories. Despite these possible benefits, quantita- tive tests of the success of such efforts are rare even in the largest, most ambitious programs (e.g., Macfariane 1998). Predation. in particular, is recognized as a factor that may dramatically limit success of public seeding programs of various invertebrate species (e.g.. Stoner & Davis 1994. Peterson et al. 1995, Barbeau et al. 1996). Similarly, predation is considered a major obstacle to suc- cess of private aquaculture of M. mercenaria in the northeastern portion of the United States (Spatz et al. 1996). Despite this lack of quantitative evidence of benefits and con- *Corresponding author. Current address: Wellfieet Shellfish Department, 300 Main St.. Wellfieet. MA 02667. E-mail: williamwalton@usa.net cerns about predation. this method of stock enhancement has been widely adopted in the region, largely based on the qualitative perception that seeding benefits the fishery. Notably, the high in- cidence of the hatchery-reared subspecies M. mercenaria var. no- taia (which has a distinctive zig-zag shell coloration) in the catch is often cited by proponents of the method as evidence of success (W. Cameron Walton, pers. comm.). Although the authors encourage a long-teim rigorous assess- ment of the success of these public seeding programs, the reality is that many of these programs are driven by the perceptions of the local managers. Such perceptions are important not only because they steer management, but because they reflect many combined years of natural history observation and field work. In our study, therefore, we surveyed municipal shellfish managers throughout Massachusetts and Long Island. New York to: ( 1 ) describe the extent and nature of cunent municipal seeding programs; (2) iden- tify potential losses of seed; and. in particular. (3) determine the extent and nature of loss of seed to predation. METHODS Our three-page survey (Appendix A) consisted of four sections: use of hatchery-raised M. mercenaria seed; handling of seed; losses of seed; and predation. Most questions were multiple-choice or required a quantitative answer. We indicated that survey par- ticipants would remain anonymous in the analysis of results. Draft versions were pilot tested with several shellfish constables on Mar- tha's Vineyard. MA. We mailed the survey on April 30. 1999 to all of the 59 Mas- sachusetts shellfish constables listed with the Massachusetts Divi- sion of Marine Fisheries. One month later, we sent a duplicate of the survey with a second request to nonrespondents. Finally, two months later, we called nonrespondents requesting the return of a completed survey. 127 128 Walton and Walton In response to inquiries from municipal shellfish managers on Long Island. NY. on December 6. 1999. we sent surveys to nine municipal shellfish managers, whose names were provided by G. Rivana of the Cornell Cooperative Extension. Unlike the Massa- chusetts recipients, the New York recipients were clustered geo- graphically (toward the eastern end of the island). Follow-up re- quests were handled as they were with Massachusetts recipients. In compiling results, some responses were simplified or stan- dardized. For e.xample. an estimated loss of 40-60% of planted seed was reduced to the median of 50%. In addition, if seed size was given as mesh size, we converted this to an average shell length. For identification of predators and their percei\ed effect, we asked respondents to list up to five of the worst predators upon seed. To quantify the threat posed, we asked further for respon- dents to score each predator with a number from 1 to 10, with 10 indicating that the predator was capable of inflicting severe losses. To capture both frequency of listing and perceived magnitude of the threat, we weighted each predator "vote" by its quantitative rating to derive a summed score. To test for differences between states, we conducted paired r-tests, assuming unequal variances between the populations with a Bonferroni adjusted probability. RESULTS Survey Response Of the 68 surveys mailed, 36. or -53%. were completed and returned. Of the.se responses. 23 respondents had used hatchery- reared seed as a stock enhancement tool, of which only one Mas- sachusetts respondent had discontinued use after 1995. All nine respondents in New York currently use hatchery seed, and 13 of the 27 Massachusetts respondents currently seed. Several negative respondents (those who responded but indicated no current or past use of seed) indicated that their local shellfish beds were closed because of pollution, so they were implementing no stock man- agement programs; whereas, two others expressed interest in implementing seeding in their municipalities. Use and Extent of Hatchery Seed Among affirmative respondents (Table 1 ). managers have used seed on average for 12.6 years, with no difference between states {P = 0.99). On average. Massachusetts municipalities received just under one million seed; whereas, New York municipalities received closer to two million, although this was not significantly different (P — 0.194). Programs varied widely in magnitude, from 50.000 to six million seed received in the most recent year. Typi- cally, the size of the seed upon receipt was 5.2 mm in shell length (± 1.12./; = 21 ). with no difference between states (P = 0.512). Most respondents received seed in late spring/early summer. Although two New York municipalities planted the seed im- mediately upon receipt, most managers maintained the seed for additional grow out, placing the seed in trays, rafts, or upwellers to obtain additional growth and protect them from predation. Seeding typically occurred when the seed were 15.9 mm (± 1.44, ;; = 22). Planting of these seed occurred from August to November. Several managers noted that they would occasionally receive seed in the fall and plant these directly. In addition, several managers based time of planting on the size of the seed; one constable in Massa- chusetts would keep small seed under netting through the winter to allow an additional year of growth before planting. Most planted the seed on unvegetated sand. mud. or muddy sand substrate (Fig. 1 ). although gravel and vegetated bottom were occasionally used. Losses of Seed Managers estimated that they lost 44% of planted seed within the first season (Table I ). with no significant difference between states (P = 0.225). Notably, estimates of first year loss ranged from 9 to 90% and tended to be lower in Massachusetts than New York {P = 0.225). Based on their observations, every respondent using seed indicated that predation was a major cause of that loss (Fig. 2). Losses attributable to the stress of handling and dispersion by currents were also cited; whereas, theft by private individuals was discounted as a potential source of loss (Fig. 2). Additional suggested sources of loss were: freezing temperatures (two respon- dents); the quahog parasite QPX (one respondent); and damage from raking (one respondent). There was no relationship between planting size and estimated first year mortality (Fig. 3). Despite these initial losses, managers estimated that 39%- of planted seed survived to market size (Table I ). New York respondents, how- ever, estimated a significantly lower sur\ival rate of seed to market size than their Massachusetts counterparts (P = 0.023): 25 versus 49%, respectively. As with estimates of first year loss, estimates of survival to market size ranged widely, from 5 to 90%. Of those TABLE 1. Description of municipal Mercenaria mercenaria seeding programs in Massachusetts (MA) and New York (NY). Factor MA NY P .Ail Responses Combined Years of use 12.6 12.7 0.9S6 12.6 (± 2.40. n = 13) (± 1.55. 11 = 9) (± 1.52. n = 22) Quantity of seed received last 994.429 1.988.111 0.194 1.383.261 year (+ 422.906.2. n = 14) (± 598.288.4. n = 9) (± 3-'i4.898.6, n = 23) Average seed size upon receipt 4.6 6.2 0.512 5.2 (mm) (± 1.42, /( = 14) (± 1.85. H = 7) (± 1.12. « = 21) Average seed size upon plantmg 16.7 14.8 0.477 15.9 (mm) (± 2.35, n = 13) (± 1.05. n = 9) (± 1.44. n = 22) Loss of seed upon planting 39.3% 54.6% 0.225 44.2% (± 7.91. » = 13) (± 8.95, n = 6) (±6.19. H = 19) Survival of seed to market size 48.67r 24.6% 0.023 39.3% (± 8.48. n = 11) (± 4.03. n = 7) (±6.01,« = 18) Each table value is followed by one SEM and the responding sample size. The P-value indicated is the Bonferroni adjusted probability of differences between states. QuAHOG Stock Enhancement Programs 129 - 1 -l 1 ^H Massachusens 1 New York In Sa' lod vAud '^'^''.,W>.4 cm to create adequate separation between the bottom of the flume and the roof of the cage. Although the flows above the cage were likely influenced by the structure's proximity to the free surface, the magnitude of this interaction, and its po- tential effect on bottom shear stress, were expected to be minor. This expectation was evaluated during the experiments using water height and velocity measurements. The cage support rods were positioned >5 diameters from the sediment box and had no detect- able influence on water flow or sediment transport (Nowell and Jumars 1987). At least one representative upstream shear velocity (/( ) was measured for each flow regime tested (Table I ). For most flow regimes, ii, was calculated during each experimental trial, and averaged across the 3 or 4 replicate trials to characterize the flow treatment (note values for ii . were not recorded in some trials due to technical problems). Shear velocities were selected to fall within the range of typical estuarine and bay flows, as derived from current meter data from Barnstable Harbor. At three sites in Barn- stable Harbor, the range of velocities measured during a spring tide in 1998 corresponded to shear velocities of 0-3.8 cm s"' and, over half of the time, flows were within the range used in these trials (0.7-2.0 cm s"') (authors' unpublished data). Flow patterns around the cage were quantified by taking cen- tered velocity profiles 6 cm upstream of the cage, mid-cage, and both 10 and 20 cm downstream of the cage (Fig. 1). Velocity profiles were also taken at these locations without the cage present. This array of profiles was conducted at shear velocities of », = 1.8 and 2.0 cm s"'. These velocities were selected after conducting the clam retention trials, to investigate flow pattei-ns at and below a shear velocity that yielded significantly greater clam retention in the caged treatment (See Results. Retention section). Flow devia- tions caused by the cage were quantified in the two flow regimes by calculating the time-averaged velocity at each measurement point and comparing the difference between the cage and no-cage profiles at both velocities. Fluorescein dye visualizations were TABLE I. Mean alongstream velocities at S cm above flume bottom (ii„) and mean shear velocities (h) lor all three clam sizes and tlow treatments. Standard deviations are in parentheses. Clam category «s (cm s ') II. (cm s ') n Small 14.3 (0.3) 0.66(0.01) 8 30.1 (0.3) 1.30(0.02) 8 Medium 32.6 (0.8) 1.41 (0.02) 4 38.1 (0.5) 1.61 (0.02) 6 40.1 (0.7) 1.73(0.01) 6 43.8 (0.4) 1.83 (0.01) 6 Large 47.2(1.2) 1.97 (0.09) 6 performed to describe qualitatively the turbulence and flow sepa- ration generated by the mesh cage. Retention Trials The main objective of the retentit)n trials was to determine the effect of caging over a realistic range of field flows and first season clam sizes. Three or four replicate trials were conducted with and without the cage for each clam size over a range of selected ve- locities (n = 3 for all trials except when /( , = 0.7 and 1.3 cm s"'). The small clams were exposed to shear velocities of 0.7 cm s"' and 1.3 cm s~'; medium size clams to 1.4-1.8 cm s"'; and large clams to 2.0 cm s^'. This experimental setup did not follow an ideal 3-factor design with caged and uncaged treatments in all combi- nations of flow speed and clam size for two reasons: the number of first season clams was limited and some data were incorporated from experiments conducted as a component of another study. The large clam data and part of the medium sized elain data (». = 1 .4 and 1.6 cm s"' trials) were gathered as part of a caging study we conducted in association with flume trials in Dunn et al. (1999). In order to use the available clams most efficiently, we selected what we anticipated to be the most relevant flow speeds for each clam size. Lower speeds were used for smaller sizes because lower shear velocities were anticipated to erode smaller clams. The ve- locity ranges selected for each clam size were chosen based on preliminary fluine experiments that determined the critical veloc- ities needed for initial movement of sediment and clams, both buried and unburied. The lower velocities were selected to erode non-burrowing clams, but not significantly erode the sediment, while the higher shear velocities were intended to erode a portion of the sediment and thereby allow assessment of the burrowing clam's response to elevated flow. For each retention trial, we selected 200 active clams. This number corresponds to a density of 5,000 clams per nr (within the 20 cm X 20 cm sediment box) and approximates actual Barnstable Harbor field densities (authors'unpublished data). The clams were suspended in seawater and then rinsed through a plastic cylinder ( 10 cm diameter x 20 cm high) held 1 to 2 cm above the sediment surface. Suspending the clams ensured that they were distributed evenly on the sediment. Use of the cylinder ensured that the clams settled onto the sediment box surface, not the acrylic perimeter. The clams were given 20 minutes to burrow (sufficient time for burial based on previous observations). After the burial time, the sediment tray was removed from the seawater table, the number of non-burrowing clams was recorded and those remaining on the sediment surface were removed. The tray was inserted into the flume under conditions of very low flow ^llf^ = 3 cm s"') and made flush with the adjoining flume bottom. For the cage treat- ments, the cage and its supports were secured to the tray at this time. All trials were conducted for 40 minutes except for ii, = 1 .7 and 1.8 cm s"' which ran for 10 minutes because these were part of another study. For all runs, the sediinent remaining in the sediment box after the experiment was sieved on a 500 (j.m sieve and the fraction retained was either sorted for clams immediately or preserved in 80% ethanol, stained with Rose Bengal, and sorted later. In order to ensure that we were not losing clams during sediment transfers, we conducted procedural controls using the medium size clatns. Procedural controls followed the same procedure as the retention runs except that the tray was placed in the water table instead of the flume. These controls showed that an average of 99'7r of seeded 138 GULMANN ET AL. -* 1— =40cm/s — — * — - - ^ ^ 20 40 60 80 Distance along flume (cm) 10 6 — 1^ = lOcm/s ^ ^ -.— — ._ - — 100 0 60 80 100 Figure 2. Vertical profiles of flow velocity (h) along flume with upstream shear velocity of u, = 1.8 cm s"'. A) Profiles without the cage (h); B) profiles with the cage (h^.;,„,.(; C) deviations in velocity caused h\ the cage [i.e., h^ ,j,,. - u ]. To emphasi/e the vertical pattern, only the components of the vectors > 27 cm s"' are graphed for profiles A and B. In profile C, deviations are true differences. Distances along the flume are relative and start 20 cm before the leading edge of the cage. clams were recovered, indicating that we were not losing clams while transfening the tray from the seawater table to the flume. For retention trials, results were analyzed separately for each clam size, with cage (two levels, present and absent) and velocity as fixed factors in two-way analysis of variance ( ANOVA) (Systat version 6.1). The dependent variable was the percentage of clams retained. For the medium size clam results, two separate two-way ANOVAs were performed due to different times of exposure to flow (10 and 40 minutes) and for large-sized clam results a one- way ANOVA for cage effect was performed. Visual inspection did not reveal any pronounced deviations from a normal distribution. Variances were tested and found to be homogeneous (Cochran's test; p > 0.05). RESULTS Cage Effects on Flow A comparison of velocity profiles conducted with and without the cage indicates that the cage altered near-bed flow at both », = 1.8 cms (Fig. 2) and » . ;.0 cm s (Fig. 3). Without the cage. velocities fit the expected log-linear profile, and changed little (<8%) with dov^nstream distance (Figs. 2A. 3A). The presence of the cage disrupted this pattern by changing flows inside the cage, just above the upper rim of the cage's frame and just downstream of the trailing edge of the frame (Figs. 2B, 3B). Dye visualizations revealed that the leading face of the cage caused flow separation around the leading cross-sectional portion of the frame. The dye revealed eddies shedding off the upper edges of the cage and contributing to the observed increase in turbulence above and downstream of the cage. The velocity profiles indicated that mean velocities were reduced downstream of the cage ax it, = 1.8 cm s" ' and both within and downstream of the cage at u = 2.0 cm s"' (Fig. 2C, .3C). The effect inside the cage was different at u. = 1.8 cm s" ' than at u, = 2.0 cm s"'. Flow deviations (Fig. 2C, 3C) indicated that velocities at u, = 1.8 cm s"' were increased inside the cage (at heights of 0.9, 1.2, 1.6 cm), yet at ;/, = 2.0 cms"' these velocities were consistently reduced inside the cage. The vertical patterns of flow alteration were different at the two velocities. At (/. = 1 .8 cm s~ the flow deviations became less pronounced with height off of bottom, but at ((,. = 2.0 cm s"' the flow deviations became more pronounced w ith height off of bottom. These patterns suggest that the cage alters the flow most strongly at the sediment interface at ((. = 1.8 cm s"' and at a height above the sediment interface at «. = 2.0 cm s-'. The cage did have a small but measurable effect on water height in the flume. Upstream water heights along the flume in- creased in the presence of the cage, while downstream water heights decreased in comparison to no-cage heights at comparable target velocities. Water heights were elevated approximately 7% just in front of the cage (position C in Fig. I ) yet at the positions of our water height measurements and velocity profiles the in- crease was 27 cm s"' are graphed for profiles \ and B. In profile C deviations are true differences. Distances along the flume and start 2(1 cm before the leading edge of the cage. 80 100 cage (»l: Bl components are relative retention was not linear but appeared to have some critical velocity above which there was a notable reduction in clam retention. DISCUSSION Cage Effect on Clam Retention in the Flume The most consistent result from our Hume studies was the lack of a significant cage effect on clam retention in most flow regimes. This result was a surprise because the presence of a mesh structure along a flow path is known to alter flow patterns (Hulberg and Oliver 1980. Nowell and Jumars 1984). and natural mesh-like obstacles such as marsh grasses have been observed to reduce flow velocities by two to 10-fold (Gambi et al. 1990. Leonard and Luther 1995). A reduction in flow velocity was expected to occur in the cages and result in enhanced clam retention. The lack of significance did not appear to be due solely to low power of the statistical tests, as an increase in the number of replicate trials per treatment (from .3 to 4) did not result in higher significance (n = 4 for target shear velocities of u, = 0.7 and 1.3 cm s"'). The one treatment that did show a significant effect of the cage on retention was the combination of the highest flow velocity (/(. = 2.0 cm s"') with the large clam size (SL = 2.3 mm). We 1 1 no cage ^ cage Small clams Medium clams Large clams 1.3mmSL l.SmmSL 2.3 mm SL 1 0 1.0 y 1.0 0.9 0.9 - 0.9 T 0.8 -a 0.8 ^ [ T 0.8 I 1 0.7 r-'- 0.7 1 ^ h i 0.7 ] \m 1 0.6 1 0.5 o |o.4 CL 0.3 0.2 0.1 0.6 0.5 0.4 0.3 0.2 0.1 J 1 0.6 0.5 0.4 0.3 0.2 ^ 0.1 0.7 1.3 1. 4 1.6 :. 1.8 2.0 II . (cm s' Figure 4. Mean proportion (±1 S.D.I of clams retained over a range of shear velocities with and without cage treatments. N = 3 for all runs except u. = 0.7 and 1..3 cm s"' v^here n = 4. 140 GULMANN ET AL. suggest that the ineLhaiiisni for this cage effect is a substantial reduction in flow velocities within the cage in this particular flow regime. Reduced flow velocities should reduce the shear stress exerted on the sediment surface, thereby decreasing the rate of erosion of sediment and clams. This relationship between shear stress and erosion was observed in the present study for the me- dium size clam treatment at ;(. = 1.7 and 1.8 cm s"' (significant effect of velocity in Table 2), and has been reported by Roegner et al. (1995) and Dunn et al. (1999). We suspect that the absence of a significant cage effect on clam retention in flows slower than (/ = 2.0 cm s"' was due to the failure of the cage to reduce boundary shear stress in these conditions. Flow measurements recorded within the cage in the fastest oncoming flow treatment («. = 2.0 cm s"' ) and a slower flow {ii = 1.8 cm s"') support our explanation for the clam retention results. The cage caused a measurable decrease in velocities near the sediment only in the fastest oncoming flow (Fig. 3). One plau- sible explanation for this effect is that the cage was starting to act more as a bluff body at the higher flows, blocking flow through the mesh and forcing the water up over the upper cage surface. This effect of producing a "skimming flow" has been predicted and demonstrated for a sufficiently dense array of tubes or cylinders in boundary-layer flows (Morris 1953. Eckman et al. 1981. Nowell and Jumars 1984). We speculate that a comparable process occurs in our cage at oncoming flows of ;(. = 2.0 cm s"' and higher: in this flow regime, the velocities are high enough to produce turbu- lent wakes downstream of the mesh filaments, increasing their effective diameter and decreasing the fluid 'porosity" of the cage (Taylor 1948. Gerrard 1978). An alternative explanation for the observed significant cage effect in the tu = 2.0 cm s~' treatment is that suspension of the larger-sized clams was more sensitive than that of the smaller ones to flow alterations caused by the cage. We do not think this is a valid explanation for several reasons. Firstly, there is no evidence for an increasing sensitivity of clam suspension to caging with increased clam size; instead, the small and large clams tended to have increased retention under the cage, whereas the medium size clams showed a mixed response. Secondly, because larger clams are capable of burrowing deeper than small ones, one would expect them to be less, and not more, sensitive to caging effects. Finally, we have no independent evidence for the ""clam size" hypothesis, whereas we do ha\e independent flow measurements in the cage that support the ""flow threshold" hypotheses. For these reasons, we constrain our discussion to the latter. Behavioral Effects on Clam Retention A reduction in near-bottom flow velocity may enhance clam retention through both hydrodynamic and behavioral mechanisms. A critical shear stress is necessary to suspend clams of any par- ticular size; reducing the shear stress below this level will result in more limited transport. Reducing the shear stress will also reduce the rate of sediment erosion, potentially allowing clams more time to burrow below the unstable sediment layer. Either or both of these processes are likely responsible for the relationship between higher shear stresses and enhanced clam transport reported in tlume (Roegner et al. 1995. Dunn et al. 1999) and field (Emerson and Grant 1991) studies. Although surtlcial sediment was eroded under the cage in our trials at a,, = 2.0 cm s"'. the reduction in rate of erosion relative to uncaged flows may have allowed clams to remain buried in the sediment. In order to maintain access to the sediment-water inter- face, postlarval M. arcnaria of 1.3 to 2.3 mm SL can only burrow as deep as their siphons are long, usually to depths of 7.5 mm or less (Zwarts and Wanink 1989). We speculate that M. arenaria may actively burrow to avoid being exposed and eroded as Sakurai and Seto (1998) demonstrated for the surf clam. PseiidocanUiiin sachaVtiiensis at the reduced shear velocities but not at ambient uncaged shear velocities. Thus we suggest that the reduction in shear stress offered by the cage at u,. = 2.0 cm s~' sufficiently reduced the erosion rate so that 2.3 mm clams could maintain their position in the sediment and a\oid resuspension. Based on the siphon length-burial limitation, we expected to find that for any specific shear velocity, retention of larger clams would be greater than retention of the smaller clams. Although our data alone do not strongly suppi)rt this hypothesis, comparison with another study demonstrates a retention-size relationship. We found that 25-35% of our 1.8 mm clams were retained at ii. = 1.8 TABLE 2. Summary of ANOVAs examining the effects of cage and velocity on retention over a range of size categories. Clam Size Source Small Cage Velocity Cage*Velocity Error Medium Cage Velocity Cage*Velocily Error Medium* Cage Velocity Cage*Velocity Error Large Cage Error Sum-of-Squares df Mean-Square U.()3I) 0.072 0.008 12 0.031 0.005 0.016 0.001 8 0.008 0.0 1.=; 0,44S 0.002 8 0.007 1 0.076 4 0.009 F-ratio 0.030 0.072 0.008 0.378 0.005 0.016 0.001 0.066 0.015 0.448 0.002 0.055 0.076 0.037 0.939 0.352 2.289 0.156 0.243 0.631 0.609 0.457 1.912 0.204 0.104 0.755 2.099 0.185 64.70 <0.001 0.232 0.643 8.129 0.046 * Separate ANOVAs were performed on two sets of mid-season clams because those tested at u. 10 min (Sy.stat version 6.1). 1.7 and 1.8 cm s ' were exposed to tlow for only Caging Effects on MrA arenar/a Transport 141 cm s~'. yet Roegner et al. (1995) reported that 0% of their 0.24- 0.29 mm clams were retained at a comparable velocity (h, = 1 .7? cm s ' ). Thus smaller clams were more readily transported at this particular velocity and retention appeared to correlate with clam size. Relevance to Field Appliealions Our Hume results suggest that mesh cages used in the field may enhance retention of recently settled bivalves in flow velocities above a critical level. The extent to which our flume experiments are applicable to field environments depends on how well they represent field situations hydrodynamically. biologically, and structurally. The shear stresses used in the flume experiments fall within those measured at Barn.stable Harbor (authors' unpublished data), indicating that boundary-layer flows in the flume are a rea- sonable representation of field flows. The cultured clams burrowed actively, and their sizes covered the range observed in field popu- lations during first season growth. The cage used in the flume had the same mesh size and composition (flexible mesh), but a much smaller overall size, as mesh enclosures used in the field. Thus, we expect our qualitative prediction (mesh enclosures enhance bivalve retention only above a threshold shear stress) to be broadly appli- cable, bul the specific threshold level to depend somewhat on the geometry and mesh size of the cage. A reduction in resuspension caused by caging will notably influence population density only if loss via resuspension is a significant source of total loss. Other processes intluencing loss include inortality through predation, disease, and/or inadequate resources. Manipulative field experiments evaluating the relative influences of hydrodynamic resuspension and predation on M. aieiuiiia recruits indicate that resuspension is a significant cause of loss, especially in smaller size-classes (authors' unpublished data). Therefore, we expect that the possible increase in retention caused by caging could have a significant impact on population density by the end of the first season. Our flow and retention results are consistent with experimental field evidence suggesting that clam recruitment under cages could be enhanced at high flows. Marcotti and Leavitt (1997) tested the hypothesis that cages, in the form of suspended tents of plastic mesh (mesh size = 0.3 cm), would increase clam recruitment in field manipulations at Barnstable Harbor, Massachusetts. They found that the only site at which the cages significantly enhanced recruitment (Green Point) was one of the two sites with the highest lidal flow rates. Velocity records from (he Green Point site indi- cated that over a 10 h period during a spring tide, 409^ of the flow velocities corresponded to it, = 2.0 cm s~' or greater (authors' unpublished data). Thus velocities at Green Point commonly were high enough to produce the caging effect demonstrated in our flume experiments, suggesting that the enhanced recruitment may have been due to cage-induced retention. Because mesh enclosures are more likely to retain clam juve- niles in vigorous rather than quiescent flow environments, the hydrodynamic setting of a site is an important factor to be con- sidered in their use. This consideration should be included with others, such as the potential of the enclosures to enhance initial settlement, enhance or deplete food resources or reduce mortality through predator exclusion, when predicting whether they are likely to increase population densities. Mesh enclosures are cur- rently in use in multiple embayments along the New England coast and the results of these applications will help refine the cost- benefit analysis of their use and identify the habitats where they will substantially increase the numbers of harvestable soft-shell clams. ACKNOWLEDGMENTS We thank Rob Dunn, Anna Metaxas, and .Susan Mills for their assistance with the flume experiments. We thank Stace Beaulieu for helpful suggestions on the manuscript and assistance with the tlume work. The authors also thank Rob Jennings and an anony- mous reviewer for improving the manuscript. Beals Island Shell- fish Hatchery and Mook Sea Farm provided clams and advice. L.K.G. was supported by a National Defense Science and Engi- neering Graduate Fellowship. H.L.H was supported by a postdoc- toral fellowship from National Sciences and Engineering Research Council of Canada (NSERC). This research was supported by the NOAA National Sea Grant College Program Office, under Grant No. NA86RG0075, Woods Hole Oceanographic Institution Sea Grant Project No. R/B-142. This is contribution number 10.^90 from Woods Hole Oceanographic Institution. LITERATURE CITED Beal. B. F. 1993. Effects of initial clam size and type of protective mesh netting on the survival and growth of hatchery-reared individuals of Mya arenaria in eastern Maine. / Shellfish Res. I2( 1 ):.38-139. Beal. B. F. 1994. Effects of seagrass iZostera marina) cover on growth and survival of softshell clams {Mya arenaria). J. Shellfish Res. 13{ 1):31 1- 312. Beukema, J. J., W. D., Bruin & J. J. M. Jansen. 1978. Biomass and species richness of the macrobenlhic animals living on the tidal flats of the Dutch Wadden Sea: long-term changes during a period with mild win- ters. Neth. J. Sea Res. 12(1 l:.'i8-77. Butman. C. A. & R. J. Chapman. 1989. The 17-meter tlume at the Coastal Research Laboratory. Part I. Description and user's manual. Technical Rcpmi of rhe Woods Hole Oceanographic Institution 89-10:1-31. Dunn. R., L. S., Mullineaux & S. W. Mills. 1999. Resuspension of post- larval soft-shelled clams Mya arenaria through disturbance by the mud snail Ilyanassa ohsoleta. Mar. Ecol. Proj>. Ser. 180:223-232. Eckman. J. E., A. R. M., Nowell & P. A. Jumars. 1981. Sediment desta- bilization by animal tubes. J. Mar. Res. 39:361-374. Emerson. C. W. & J. Gram. 14')|. The control of soft-shelled clam {Mya arenaria) recruitment on intertidal sandtlats hv bedload sediment trans- port. Umnol. Oceamifir. 36:1288-1300. Gambi. M.. A. R. M., Nowell & P. A. Jumars. 1990. Flume observations on flow dynamics in Zo.stera marina (eelgrass) beds. Mar. Ecol. Prog. Ser. 61:159-169. Gerhart, P. M. & R. J. Gross. 1985. Fundamentals of fluid mechanics. Reading, MA: Addison-Wesley Publishing Co. 856 pp. Gerrard, G. H. 1978. The wakes of cylindrical bluff bodies at low Reynolds number. Phil. Trans. R. Soc. Land. 288:351-382. Giinther, C. 1992. Settlement and recruitment oi Mya arenaria L. in the Wadden Sea. J. E\p. Mar. Biol. Ecol. 159:20,3-215. Hulberg. L. W. & J, S. Oliver. 1980. Caging manipulations in marine soft-bouom communities: importance of animal interactions or sedi- mentary habitat modifications. Can. J. Fish. .Aquai. Sci. 37:1 130-1 139. Leonard, L. M. & M. Luther. 1995. Flow hydrodynamics in lidal marsh canopies. Limnol. Oceanogr. 40:1474-1484. Marcotti. T. & D. F. Leavin. 1997. The Barnstable Harbor Shelitlsh Re- 142 GULMANN ET AL. cruitment Enhancement Project (BHSREPl. a Final Report. Report to the Town of Barnstable. 35 pp. Moller. P. & R. Rosenberg. 1983. Recruitment, abundance and production of Mya arenaria and Cardiiim eJule in marine shallow waters. Western Sweden, Ophelia 22( I ):33-55. Morris, H, M, 1955. A new concept of flow in rough conditions. Trans. Am. Soc. Civ. Eng. 120:373-398. Newell, C. R. & H. Hidu. 1986. Species profiles: life histories and envi- ronmental requirements of coastal fishes and invertebrates (North At- lantiO-softshell clam. US Fish. Wildl. Sen: Biol. Rep. 82, 17 pp. Nowell, A, R, M. & P, A, Jumars. 1984, Flow environments of aquatic benthos, Ann. Rev. Ecol. Sysl. 15:303-328, Nowell. A. R. M. & P. A. Jumars. 1987. Flumes: Theoretical and expen- mental considerations for simulation of benthic environments. Oceun- ogr. Mar. Biol. Annii. Rev. 25:91-112. Pfitzenmeyer. H, T, 1962, Periods of spawning and setting of the soft- shelled clam, Mya arenaria, at Soloinons, Maryland. Clusapeake Sci. 3:114-120. Rodriguez, S. R., F. P.. Ojeda & N. C. Inestrosa. 1993. Settlement of benthic marine invertebrates. Mar. Ecol. Prog. Ser. 97:193-207. Roegner. C. C. Andre. M., Lindegarth, J, E„ Eckman & J. Grant, 1995, Transport of recently settled soft-shell clams (Mya arenaria L.) in laboratory flume flow, J. E.xp. Mar. Biol. Ecol. 187:13-26, Sakurai, I, & M, Seto, 1998, Behavioral characteristics of the juvenile Japanese surf clam Pseudocardiwn sachalinensis in response to sand erosion and deposition associated with oscillatory water flow. Fisheries Sci. 64(31:367-372. Sigurdsson, J, B„ C, W, Titman & P, A, Davies 1976, The dispersal of young post-larval bivalve inolluscs by byssus threads. Nature 262:386- 387. Taylor, G, 1, 1948, Air resistance of a flat plate of very porous material, .Aeronaulical research council reports and meitioranda. Reports and Memoranda No, 2236, Trowbridge, J, H,. W, R„ Geyer, C, A., Butman & R. J. Chapman. 1989. The 17-meter flume at the Coastal Research Laboratory. Part II: flow characteristics. Woods Hole Occanographic hisiitutioii Technical Re- port 89( 1 1): 1-37. U.S. Fisheries Report. 1998. Current Fishery Statistics No. 9800 p. 12. Zwarts, L, & J, Wanink, 1989, Siphon size and burying depth in deposit- and suspension-feeding benthic bivalves. Mar. Biol. 100:227-240, Jdiimal of Slu-llfiili Reseanh. Vol. 20. No. I. I4,VI47, :()01. (IROWTH OF BUTTER CLAMS, SAXIDOMVS GIGANTEUS DESHAYES, ON SELECTED BEACHES IN THE STATE OF WASHINGTON STUART ALAN GOONG AND KENNETH K. CHEW University of Washington School oj Aquatic and Fishery Sciences. Box 355020, Seattle, Washiniiton y.S'/y.f ABSTRACT The butter elam, Sci\icii>iiui.\ .vw"""''" Deshayes. is a highly valued recreational shellfish speeies, hut is cLirrently ot little commercial importance in Washington. Recently, a small-scale commercial butter clam harvest began, and interest in this species is expected to increase. Basic information on the biology and ecology of the clam is necessary to establish a sound management regime for recreational and potential commercial harvest of butter clams. Beaches studied were Birch Bay State Park. Double Bluffs, and Potlatch State Park. Growth rates were determined by measuring lengths-at-age for clams collected from each beach. Von Bertalanffy growth curves were produced using nonlinear regression analysis. Growth rates were significantly different among the three beaches, with the fastest growth occurring at Double Bluffs, followed by Birch Bay State Park, while clams at Potlatch Slate Park had the slowest growth. KEY WORDS: Sti.xiiliinnis gigaiileus. growth INTRODUCTION The purpose of this investigation was to assess whether growth of butter clam,s (Saxidomus giganteus Deshayes) on selected beaches in Washington differed. Beaches selected for this study were Birch Bay State Park, Double Bluffs, an Island County rec- reational area on the southwestern coast of Whidbey Island, Wash- ington, and Potlatch State Park (Fig. 1). There was no evidence to suggest that growth rates might differ within the Puget Sound Basin, although different growth rates have been observed for butter clanis from beaches within close proximity in British Co- lumbia (Quayle and Bourne 1972). The three beaches included in this study were surveyed for butter clams using methods similar to beach survey methods used by the Washington Department of Fish and Wildlife (WDFW). Clam shells were aged using the annular method, which appears to be the most reliable method of age determination for butter clams (Goong 1999). At least two researchers have independently veri- fied deposition of annuli in the shell of the butter clam using mark and recapture studies in Washington and British Columbia (Houghton 1973, Neil Bourne pers. comm.). Growth rates were calculated using von Bertalanffy methodology (von Bertalanffy 1938). MATERIALS AND METHODS Study Areas and Habitat Two hundred and forty-one butter clams were collected from three beaches for this study from August 17-19, 1997. Clams were collected from Birch Bay State Park on August 17. from Double Bluffs on August 18, and from Potlatch Slate Park on August 19. Birch Bay State Park is located near Bellingham at the southern end of the Strait of Georgia. The beach has a gradual slope, and is characterized as primarily a sand-rock beach, with numerous larger cobblestones (up to 13 cm x 15 cm) on the surface. The beach has good road access, with parking less than 50 m from the water. This allows heavy recreational use, including clam harvesting, judging from the large number of clam harvesters observed and poor clam densities found on the beach. Other clam species observed on the beach included many Manila clams (Tapes philippinarum) and a few native littleneck clams (Protothaca staminea). Green shore crabs {Hemigrapsiis sp.) were common. Water quality measure- ments, as recorded by the Washington Department of Ecology (WDOEl (Newton et al. 1997). showed surface temperatures av- eraged 7°C in winter and 10-13°C in summer. Salinity was fairly constant, ranging from 24-27 ppt throughout the year. The con- centration of chlorophyll a was about 0.6 |xg/L in winter, and about 3 p.g/L during the rest of the year. Double Bluffs Beach is located on the southwestern coast of Whidbey Island, in central Puget Sound. Substrate and topography of this beach was similar to that of Birch Bay. There was poor road access, with a parking area located approximately I km from the southern boundary of the clam bed. This beach probably has little recreational harvest, judging from the few clam harvesters ob- served, high clam densities, and large nuinbers of older clams found there. Other species of clams observed on this beach in- cluded small numbers of Manila clams, cockles {Chnocardium iiitttalli). bent nose clams (Macoiiui iiasiiia). and horse clams {Tre- sus ciipax). There were also large numbers of green shore crabs and acorn barnacles (Bakmus sp.). One juvenile cancrid crab was also observed. Sea surface temperatures averaged 8°C in winter and 10-12°C in summer, a temperature regime similar to that of Birch Bay. Salinity averaged 29 ppt throughout the year. Chloro- phyll a concentrations were about 0.4 |j.g/L in winter, and about 4.3 p-g/L in summer. These water quality measurements were simi- lar to those found at the WDOE sampling station near Birch Bay. Lower salinity measurements for the station near Birch Bay may be due to the influence of freshwater from the Eraser River in southern British Columbia (Newton et al. 1997). Potlatch State Park is at the southern end of Hood Canal near the southern end of the Puget Sound Basin. The topography of this beach was more complex than the other two beaches. Much of the beach was flat near the road, then sloped gradually upward toward the water, and then downward. However, some portions of the beach sloped gradually downward the entii'e distance to the water. Because of this, much of the flat area of the beach is flooded during high tide and had good clam populations, despite being fairly distant (50 m or more) from the water level during minus tides. Substrate of this beach was best characterized as mud-gravel, and there were no cobblestones on the surface. Sampling of this beach coincided with the final day of a three-day native tribal Manila clam harvest, during which butter clams were also col- lected. Thus, clam densities were poor, and most clams collected 143 144 GooNG AND Chew 124° Wl "^"^^x U I 12^W VANCOUVER s?BJErt|Bay' ISLAND V>c\i:'jSLSi Bellingham Everett Seattle PACIFIC OCEAN - 47° N Tacoma Figure 1. Map of the Greater Puget Sound Region showing locations of the three study sites. were in the 3+ age class or younger. Other clam species observed on this beach included a few bent nose clams and numerous Ma- nila clams. Surface temperatures averaged 7°C in winter and 14— 18°C in summer. Salinity was highly variable, due to freshwater drainage from the Skokomish River and several small creek sys- tems, but ranged from 16-26 ppt. Chlorophyll a concentrations were also highly variable, but tended to be quite low during sum- mer months, averaging only about 0.9 p,g/L. The two other beaches had steady concentrations of chlorophyll u from spring through autumn, with low chlorophyll (( concenti-ations only in w inter months. Sampling Methods Sampling was undertaken using survey methods similar to those employed by WDFW (Campbell 1996). A reference transect was established on each beach which ran perpendicular to the shoreline. The end of the transect nearest the high water line was situated approximately 30.5 m from the water line at the low tide mark. A 30.5 m length of rope was laid along the transect down to the water line. The rope was marked with a permanent marker every 1 .5 m so it could be used to measure distances on the beach. Quadrats of 0.25 m" in area were established every 4.5-10.5 m along the length of the transect to ensure that clams were collected from different tidal heights. Positions of the quadrats occasionally had to be adjusted in order to account for beach topography. Sub- sequent transects were established 9-15 m from each other and the reference transect, and were also perpendicular to the shoreline. Quadrats were established along these transects in a similar man- ner and excavated to a depth of 0.5 m. Quadrats were excavated using standard garden spades. As quadrats were excavated, rocks and other large debris were re- moved. Large butter clams were removed and collected. The re- mainder of the substrate was sifted through 1-cm mesh screens in order to collect smaller clams. Analysis of Growth Data were analyzed using the von Bertalanffy growth model (von Bertalanffy 1938): 1 (t) = L,_ - L.^e"'"' - '"'. 1 (t); e: k: length at age t base of the natural logarithm time when length is theoretically zero von Bertalanffy growth constant asymptotic maximum length of the clam. Model parameters were estimated using nonlinear regression analysis, as suggested by Gallucci and Quinn ( 1 979). in lieu of the Ford-Walford method. Gallucci and Quinn ( 1979) noted that use of nonlinear regression facilitates quantitati\'e comparisons among the parameters of the equation. Commonly, comparisons are made on the growth constant (k) only. Since it is standard practice to report differences in k as representing differences in growth, this convention was followed in the present study. An F test was uti- lized in order to make statistical comparisons among parameters Growth of S.widomus gig.anteus in Washington 145 (Neterel al. 1996). Statistical analyses were performed v\itli SPSS 6.1. IS tor the Power Macintosh. RESULTS Von Bertalanffy Growth Analysis Von Bertalanffy growth models resulting from these param- eters are plotted in Figure 2. Portions of these curves correspond- ing to the first 9.5 years of growth are redrawn in Figure 3 to exhibit greater detail. Figure 3 clearly shows that growth is different among beaches. Fastest growth occurred at Double Bluffs, followed by Birch Bay. while slowest growth was observed at Potlatch. However, although the rate of growth was slower at Potlatch. it did not appear to decrease over time as quickly as it did at the other beaches. This is probably due to an inadequate number of older specimens in the sample from Potlatch. Most clams collected from Potlatch were in the 3.5-year age class or younger. The result is a paucity of growth data beyond the first four years, possibly leading to a poorly fitting growth model. One indicator of this possibility is the L, calculated for Potlatch shown in Table 1. This is an unreasonably large value for maximum clam length. However, the early pun of the curve fits the data. Thus, the model may still adequately describe growth for the first several years. The growth constants (k) listed in Table 1 support conclusions drawn from Figure 3. Fastest growth occurred at Double Bluffs, while slowest growth occurred at Potlatch. Differences among the three beaches were significant at the a = 0.05 level. DISCUSSION Differences Among the Three Beaches Of the three beaches. Potlatch is the most diverse. Potlatch is a mud-gravel beach, which is somewhat different from the preferred butter clam substrate of sand, shell, and gravel (Quayle and Bourne 1972). The substrate at Potlatch is more suited to Manila clams (7". philippinanim). as evidenced by the numbers of Manila clams observed on this beach. Double Bluffs, where the fastest growth occurred, and Birch Bay are both sand-rock or sand-rock-shell beaches. In addition, butter clams were most abundant at Double Bluffs. Potlatch is more influenced by freshwater drainage than the other two beaches. Results of the present study indicate that butter 15 Age (years) Figure 2. Von Bertalanffy growth curves for butter tlanis from Pot- latch Stale Park, Birch Bay State Park, and Double Bluffs Beach. Washington. 0.5 3.5 6.5 9.5 Age (years) Figure .^. \ on Bertalanffy tur\es for butler clams from Potlatch State Park, Birch Bay State Park, and Double Bluffs Beach, Washington, plotted to 9.5 years. clam growth may be positively correlated with salinity. Although no other butter clam studies support this suggestion, studies on oysters suggest that oyster growth may be influenced by differ- ences in salinity (Toro et al. 1995. Mallonee 1989). Robert et al. ( 1993) conducted a study of Manila clam growth at three sites, one of which was oceanic and the others were estuarine. Clam growth was fastest at the oceanic station, and they speculated that differ- ences in growth were due in part to differences in salinity. Like butter clams. Manila clams are venerid clams. Bardach et al. ( 1972) reported that the optimum salinity range for Manila clains was about 24-32 ppt. which is the normal salinity range for Puget Sound. This salinity range may also be optimal for butter clams. Average salinities recorded at Potlatch are normally below this range, ranging from 16-26 ppt (Newton et al. 1997). As noted in Study Areas and Habitat, surface water tempera- tures at Potlatch tended to have greater fluctuations than tempera- tures at the other beaches. This may be important if butter clams are better adapted to more stable growing areas. Further, water temperatures in summer at Potlatch can be particularly high rela- tive to those at the other beaches. Irregularity in year round water temperatures and excessive summer temperatures could cause a sufficient amount of stress to have a negative impact on growth. Temperature has been shown to be an important determinant of bivalve growth in several species, including another venerid clam, the northern quahog (Mercenunu mercenaria), scallops, and oys- ters (Crockett 1988. Claereboudt 1994, Toro et al. 1995). Mann and Glomb (1978) and Mann (1979) showed a negative cortelalion between temperature and growth in Manila clams, with fastest growth occurring at 12°C. Growth was also correlated with water T.\BLE 1. Parameters of the von Bertalanff> growth equation for butter clams from three beaches in W ashlngton. calculated using all annulus lengths. Beach : SE L, ± SE t,, + SE Potlatch State Park 0.064 + 0.010 lfi.^.903 ± 20.243 -0.222 + 0.048 Birch Bav State Park 0.157 ±0-010 93.S70± 3.218 -0.083 ± O.O.-il Double Bluffs Beach 0. 1 72 + 0.006 96.905 + 1 .542 0.043 ± 0.032 Total 0.140±0.0()4 105.157 ± 1.892 -0.004 ± 0.026 146 GooNG AND Chew teinpeiature in Boiinic (1982). but the correUition was positive. Fastest growth in natural populations of Manila clams occurred in the Strait of Georgia, where surface water temperatures are high- est, commonly above 15"C in summer (Hollister and Sandes 1972). However. Mann's results were from laboratory experi- ments, and some other factor may be responsible for the observed differences in growth in British Columbia. In addition, food availability was more irregular at Potlatch. as indicated by chlorophyll a concentrations. There appears to be higher food availability during spring and autumn. Winters at Pot- latch are times of low food availability. Unlike the other beaches, there was also low food availability at Potlatch during summer. At Double Bluffs and Birch Bay. food availability was consistent from spring through autumn. This is supponed by the dissolved inorganic nitrogen levels at Potlatch. which are frequently unde- tectable from May through August in the surface water (Newton et al. 1997). Dissolved inorganic nitrogen is generally considered the limiting nutrient in marine ecosystems (Valiela 1984). Thus, maxi- mum primary productivity at Potlatch may be a growth limitation for bivalves at Potlatch. Food availability has been shown to be among the most important determinants of bivalve growth, includ- ing growth in M. merccnaria (Claereboudt 1994. Mallonee 1989, Crockett 1988). Foe and Knight (1985) found that Asiatic clams (Corhicuta flwninea) are probably food limited in the Sacramento- San Joaquin Delta, even with chlorophyll ti concentrations as high as 62.5 p-g/L. which is an order of magnitude higher than concen- trations found in Puget Sound. Finally, there is also a gradient of human activity at these three beaches. Potlatch may be a site of intense harvesting, if not for great numbers of butter clams, then certainly for Manila clams. Also as noted earlier, there was a commercial harvest by Native American harvesters. Birch Bay is not subject to commercial har- vest, but is heavily used for recreational purposes, since the road- way allows direct road access to the beach. Double Bluffs is also not available to commercial harvesting and probably has little recreational use of the butter clam resource due to distance of the clam bed from the road access. One effect of heavy use may be substrate compaction. Studies conducted in Washington demon- strated that substrate compaction may be an important cause of low productivity on clam beaches (Toba et al. 1992). Increased human activity in an area has been shown to decrease biological integrity, resulting in decreased species diversity, and may have negative impacts on growth and health in general on all taxa in a given ecosystem (Karr 1981 ). Houghton ( 1973) showed a high correla- tion between taxa richness and growing areas with high clam den- sities, strong growth, and high productivity. Thus, bivalve growth may be negatively coiTelated with a gradient of human activity. Indeed, this may be an important factor in growth on popular recreational clam beaches, although it is probably negligible on more remote beaches, such as Double Bluffs. Patchiness in age distribution may have affected growth analy- sis for Potlatch State Park. In the von Bertalanffy analysis for Potlatch. the lack of strong representation of older clams likely contributed to the computation of an unrealistic L-,_. causing the Potlatch curve to eventually cross the others. If the true asymptote for Potlatch is closer to those of the other beaches as is suspected, the curve would probably not cross the others. It should be noted that the calculated parameters are still adequate to describe the observed growth for at least the first four years. Due to poor data for subsequent years, the model does not provide an adequate description of growth for later years. CONCLUSIONS 1. Butter clams from different beaches in Puget Sound. Wash- ington have different rates of growth. Fastest growth oc- cuned at Double Bluffs, and slowest growth occurred at Potlatch. 2. The legal size limit for butter clams in Washington probably should be increased. The current size limit for butter clams is 38 mm. According to results of growth analysis, clams in this size class would be only about 2-3 years old. It is suggested that the size limit be increased to about 63 mm, the size limit in British Columbia. Clams in this size class would be about six years old, and would have had oppor- tunities to spawn. 3. Since rates of growth are different in different growing ar- eas, it is suggested that a more localized management ap- proach for the resource be utilized in order to protect the productivity of the stock. A generalized management strat- egy will not optimally manage the resource. This would require growth analyses to be performed in any area that has any commercial or intense recreational harvests for butter clams. ACKNOWLEDGMENTS The senior author wishes to express his appreciation to his Supervisory Committee for their invaluable guidance, Drs. Ken- neth Chew. Neil Bourne, and Loveday Conquest. Thanks also to my field crews for all of their hard work. Finally, special thanks to the Graduate School of the University of Washington, the Loosanoff Foundation, and the Research, Admissions, and Schol- arship Committee of the School of Fisheries at the University of Washington for their generous financial support. LITERATURE CITED Bardach, J, E.. J. H. Ryther. & W. P. McLarney. 1972. AquacuUure - The Farming and Husbandry of Freshwater and Marine Organisms. New York: John Wiley and Sons. S68 pp. Bourne. N. 14X2. Distribution, reproduction, and growth of Manila clam. Tapes phiHppmuruni (Adams and Reeves), in British Columbia. ,/. Shellfish Res. 2:47-.S4. Campbell. W. W. 1946. Procedures to Determine Interlidal Populations of Proiothuca suwiinea. Tapes philippiiianiin. and Crassostrea gif;as in Hood Canal and Puget Sound. Washington. Olympia, WA: Wash. Dept. of Fish and Wildlife. 27 pp, Claereboudt. M. R. 1994. GMDH algorithm as a tool lor bivalve growth analysis and prediction. ICES J. Mar. Sci. Sl(4):4,'?9— 445. Crockett, L. R. 198ti. Growth rate and age structure comparisons of geo- graphically isolated hard clam. Meneiuiria incnenariu. populations. /. Shellfish Res. 7(1): 1 14. Foe, C. & A. Knight. 198.5. The effect of phytoplankton and suspended sediment on the growth of Carbiciila Ihiiniiiea (Bivalvia). Hydrnhin- logia 127(2); 105-1 15, Gallucci, V. F. & T. J. Quinn, Ii. 1979. Reparameteri/mg, tiUmg, and testing a simple growth model. Trans. Am. Fish. Soc. 108:14-25. Goong. S. A. 1999. Growth and Age Determination of Butter Clams (Sa.xi- tlomus gigaiiteus Deshayes) on Selected Beaches in the State of Wash- ington, with Comments on Recruilmenl, Master's thesis. University of Washington. Seattle. 70 pp. Hollister. H. J. & A. M. Sandes. 1972. Sea surface temperatures and salinities at shore stations on the British Columbia coast 1914-1970. Growth of Saxidomus giganteus in Washington 147 Marine Sciences Directnriite. Pacific Rciiioii. Puc. Mar. Sci. Rep. Can. 93:72-113. Houghton, J. P. 1973. The Intertidal Ecology of Kikel Island, Washington. with Emphasis on Age and Growth of Pnilothaca stamiuea and Sa.xi- clomiis giganteus (Lamellibranchia: Veneridae). PhD dissertation. Uni- versity of Washington, Seattle. 1 79 pp. Karr, J. R. 1981. Assessment of biotic integrity using tlsh communities. Fisheries 6(6):2I-27. Mallonee, M. E. 19S9. Water (.|ualities associated with rapid oyster growth, / Shellfish Res. 8(2):4S6. Mann, R. 1979. The effect of temperature on growth, physiology, and gametogenesis in the Manila clam Tapes philippinanim (Adams and Reeves, 1850). ./. Exp. Mar. Biol. Ecol. 38(2):I2I-133. Mann, R. & S, J, Glomb. 1978. The effect of temperature on growth and ammonia excretion of the Manila clam Tapes japonica. Estiiar. Cuasi. Mar. Sci. 6(3):335-33y. Neter. J., M. H. Kutner, C.J. Nachlsheim & W. Wasserman. 1996. Applied Linear Regression Models. 3rd ed. Chicago: Richard D. Irwin, Inc. 720 pp. Newton, J. A.. S. L. Albeitson & A. L. Thompson. 1997. Washington Slate Marine Water Quality in 1994 and 1995, Olympia, WA: Wash. Dept. Ecol. Pub. 97-316. 71 pp. Quayle, D. B. & N. Bourne. 1972. The Clam Fisheries of British Columbia. Ottawa, Canada: Bull. Fish. Res. Bd. Canada 179, 70 pp. Robert, R.. G. Trut, i: J. L. Laborde. 1993. Growth, reproduction, and gross composition of the Manila clam Riulilapes philippinunnn in the Bay of Arcachon. France. Mar. Biol. 1 16(2 ):29 1-299. Toba, D. R., D. Thompson & K. K. Chew. 1992. Effects of substrate modification on natural recruitment, growth, and survival of hardshell clams in Washington State, / Shellfish Res. 11(1 1:207-208, Toro, J. E., M, A, Sanhueza, J. E. Winter, C. M. Senn, P. Aguila & A. M. Vergara. 1995. Environmental effects on the growth of the Chilean oyster Ostrea chilcnsis in five mariculture locations m the Chiloe Is- land, southern Chile. Aqnacullure 136(I-2):153-164. Valiela. I. 1997. Marine ecological processes. In: J. A. Newton, S. L. Albertson, & A. L. Thompson, editors. Washington State Marine Water Quality in 1994 and 1995. Wash. Dept. Ecol. Pub. 97-316. 71 pp. Von Bertalanffy, L, 1938. A quantitative theory of organic growth (inqui- ries on growth laws. II). Hitnuin Biol. 10(2): 181-213. JuHimil ,it Shellfish Re.scaith. Vol. 20, No. 1, 149-153, 2001. REGENERATION OF THE INHALANT SIPHON OF DONAX HANLEYANUS (PHILIPPI, 1847) (BIVALVIA, DONACIDAE) FROM ARGENTINA DIEGO C. LUZZATTO' AND PABLO E. PENCHASZADEH" ^Lah. Invertchnulos. Depto. Biologi'a. Facultad cle Cieiuias E.xaclas v Natitndes. Uiiivvisidad de Buenos Aires. Ciudad Universharia. Pab II. C1428EHA. Buenos Aires. Argentina: -Facultad de Ciencias Exuctas y Naturales. Universidad de Buenos Aires. MACN-CONICET. Av. Gallardo 470. 1405 Buenos Aires. Argentina ABSTRACT Doiui.x hanleyamis (Philippi. 1S47) is the soulhernmost Doiiax species in the American Atlantic. It mhabits nitertidal fine grain sandy shores in northern Buenos Aires Province. As with other species in this genus, these animals survive under pressure of "cropping" by fish that feed on their siphons. The inhalant siphon is a complex organ; its tip contains a system of branched tentacles that permit particle selection of medium and large grains and prevent their entry to the pallial cavity of the animal. To estimate the regeneration speed of the amputated siphon under laboratory conditions, a study of the regrowth sequence after an artificial cut was performed at regular inter\als for complete lO-day periods. The observations in vivo under the microscope were correlated with those made by histology at the same time intervals. Results indicate that at 24 h after amputation, rudiments of the pnmary tentacles are observed, and the siphon is fully active in selecting particles at the fifth day (i.e.. the process of regeneration of the primary, secondary, and tertiary tentacles is completed within -5 days). After this, a period of growing and tentacle ramification follows, even though the result is a siphon with tentacles less ramified than the original. REV WORDS: hi\ahc. mhulant siphon, siphonal regeneration. Dmuix hanlcxamis INTRODUCTION Beach clams of the Tellinacean genus Dona.x inhabit intertidal sandy beaches in most parts of the world. Doiui.x lianleyanus (Phil- ippi. 1847) is the only Donax species occurring throughout the Argentine littoral. These bivalves are constantly Hushed out of the sand by the pounding surf; they are adapted to live on exposed shores, and they do not occur on protected beaches or in shallow bays. D. hanleyamis require an environment in which there is a fair amount of wave action to keep the sand aerated and clean, to keep organic detritus in suspension, and to allow tidal migration (Ansell 1983, Ansell & Trevallion 1969, Mori 193S. Mori 1950. Narchi 1978. Penchaszadeh & Olivier 1975). The siphons of D. hanleyamis are very flexible and extensible, as are those of other tellinacean bivalves. When buried, D. han- leyamis extends its inhalant siphon through the sediment into the water above (-1.5 cm for a 2-cm animal). Juvenile flatfish (Tre- vallion 1971, Gillbert & Suchow 1977, de Vlas 1979), crabs (Hughes 1969), and birds (Ansell 1981) feed on tellinacean si- phons. -Some studies have included an examination of the process of wound healing or the physiological basis of the regeneration, focusing on the amount of regenerated tissue (Hodgson 1982a, Hodgson 1982b, Peterson & Quammen 1982). Siphon regeneration after amputation has been described for the tellinaceans Tellina tenuis da Costa (Trevallion 1971), Scro- biciilaria plana (da Costa). D. serra Rbding (Hodgson 1982a, Hodgson 1982b). and D. vittatus (Ansell et al. 1999). The external surface of the tip of the inhalant siphon of D. hanleyamis has six branched primary tentacles, six branched sec- ondary tentacles between two primary tentacles, twelve small ter- tiary tentacles between a primary and a secondary tentacle, and 24 smaller quaternary tentacles between two tertiary tentacles (Fig. 1 ), which are very mobile. These tentacles form an intricate siev- ing device (Narchi 1978). The siphons of D. hanleyamis are very sensitive to vibrations. They are richly innervated and rich in sense organs, at least near the tips (Hodgson 1982b, Ansell et al. 1999). It would be beneficial to the animal if damaged or lost parts of the siphons could regen- erate rapidly after injury so as to regain normal function. The aim of this study was to examine how soon the normal tip of the inhalant siphon of D. hanleyamis is regenerated, focusing on the proper function of filtration, and to study the morphological characteristics of the siphon regeneration by simulating the effect of a cropping predator. MATERIALS AND METHODS The animals were collected from sandy beaches in Punta Mog- otes (Mar del Plata), at low tide. We selected specimens that were approximately 2 cm in shell length. Previously in the lab. aquaria had been conditioned with CaCO, -saturated seavvater at ?i59,, sa- linity and pH 7.8. The experimental aquarium device consisted of two fishbowls, one of 200 I and the other of 10 1. The small one was inside the big one. Sand extracted from the sampling areas was placed inside the small one. Water was pumped from the big one Figure 1. View of the inhalant siphon tip of Donax hanleyamis show- ing tentacles and ramifications, P: primary tentacles; S: .secondary tentacles; T: tertiar* tentacles. Scale bar: 0.5 mm. (After Wade, 1969) 149 150 LUZZATTO AND PENCHASZADEH M:]\imiin water lc\ el Minimun water level Hoses - Sand level Small fishbowl Pump Figure 2. Aquaria device Mlieme to recreate the high-energ) condi- tions of the natural environment of Duiiax haiilevaniis. to a third bowl located I m above. Water from this third bowl siphoned its contents into the small fishbowl and in this way it kept the sand and water aerated and turbulent, simulating the high- energy conditions of the natural environment of these animals (Fig. 2). The animals were exposed for at least 10 days to this aquarium system, and the temperature was kept between 20 and 22°C during the experiments. After the adaptation period. 30 D. haiiU'Vinuis were placed in a shallow glass dish containing sea water and left undisturbed until the siphons were extended. The tip of the inhalant siphon was removed from un-narcotized individuals using ophthalmic scissors ( 1 to 2 mm from the tentacle crown). Five siphons, used as controls of the initial regeneration point, were fixed immediately in Bouin solution for 24 h. The remaining 25 animals were left to recover in the running sea water system described above, which contained a layer of sand deep enough to allow the animals to burrow nor- mally. These animals were sacrificed in groups of five at 1.3. 5. 7. and 10 days, always extracting the siphon tip and fixing it in Bouin solution. All of these fixed siphon tips were processed his- tologically. A dehydration and paraffin inclusion protocol was followed, and transverse sections were cut at 7 microns, after which hematoxylin-eosin tint was used. Because it is difficult to obtain histological cuts of the siphon tip always in the correct orientation, we chose the best ones for analysis and photography. The changes that followed the initial cut on the living animals were observed daily under a stereoscopic microscope to give i)i vivo corroboration. Drawings were made from observations under the stereoscopic microscope for living animals and the histology microscope for fixed siphons. Each drawing is a composit of all the information of a specific regeneration point. The qualitative regeneration curve resulted from digital images of these drawings at 600 dpi in black and white (siphon black and background white). Using a computer, we counted the number of pixels in black to calculate a percentage index of coverage as (number of black pixels)/(number of total pixels) X 100. This index was graphed against time in days. This protocol was repeated twice to provide experimental cor- roboration. All the experiments were made with the animals in starving conditions, and a control group of undisturbed non-excised ani- inals was used to test viability during the experiments. RESULTS For the histological interpretation of the complex tentacle sys- tem of the inhalant siphon of D. hanleyanus, we classified the different tentacles according to the shape and quantity of the branches. This primary filtration organ comprises 6 branched pri- mary tentacles. 6 branched secondary tentacles between the pri- mary tentacles, and 12 tertiary tentacles between a primary and a secondary tentacle. Every tentacle thinner than a tertiary tentacle was considered as a branch. The observations made under stereoscopic microscope revealed that the tentacles not only form a net that selects the grain size but also that the tentacles have the capacity to change the diameter of the siphon opening. This is shown in Figure 3. which illustrates how the siphon acts to occlude the opening and prevent the entry of large particles of sand. The structure of each tentacle and of the whole siphon, comprised of longitudinal and circular muscles, en- ables it to contract and elongate (Fig. 4). The regeneration sequence of the inhalant siphon is shown in Figures 5 and 6. The first rudiment of the new six primary tentacles is visible 24 h after the cut, and they are developed in 48 h. After the third day the secondary tentacles appear, and after the fifth day they are well differentiated from each other. On the seventh day all tentacles (i.e.. primary, secondary, and tertiary) are very well de- veloped, and at this time we can affirm that the siphon is totally functional. In the later stages its branches began to be observed in the tentacles, but from this time the reaeneration rate diminishes Figure .^. Sequence of rejection of a sand particle (arrows! by an intact inhalant siphon o{ Dnnax hanleyanus. Scale bar: 4 mm. Regeneration of Donax hanleyanus 151 Light Figure 4. Inhalant siphon transverse cut (7 microns) showing muscle structure, f: circular muscles; I,: longitudinal muscles: E: cubic cili- ated epithelium from the internal siphon surface. Scale bar: 50 mi- crons. notably. Moreover, after 10 days the complexity and ramification quantity is still not the same as the control siphon (Fig. 5). An approximation of the regeneration rate is shown in Figure 6, where the degree of regeneration point corresponds to standardized drav\ings. In the figure a lag period in the regeneration can be observed during the first 24 h. This is followed by a period of very intensive regeneration between the second and fifth day. and fi- nally the process is decelerated until the end of the regeneration. The animals used as controls (undisturbed animals) remained alive after 40 days, and no mortality was detected during the experiments in the excised animals. We did not record any differ- ence on the siphon-restoring rate during the second experiment series. DISCUSSION The morphology and internal anatomy of the siphons of Telli- nacea have been well described (Yonge 1949), although in Donax there are generally more tentacles than Yonge supposed, and in life most extend across the siphonal aperture forming a protective screen that excludes sand particles rather than all being turned back, as Yonge described. The siphons of D. hanleyanus do not differ significantly from those of the closely related European species D. rninciiliis L. (Moue/a & Fienkiel 1974) and D. villains (Ansell et al. 1999). The siphons of these three species are closely similar in many respects to other Donax species (Duval 1962, Figure 5. Regeneration of the inhalant siphon tip of Donax hanleyanus. \. B, and C: Intact siphon. D: Siphon wall immediately after amputation. E, F, G, H. and I at I, 3, 5, 7. and 10 dajs alter amputation, respectively. Scale bar: I mm. 152 LUZZATTO AND PeNCHASZADEH 10 12 Time (Days) Figure 6. Rate of regeneration of the tentacle structure of tlie inhalant siphon of Doiiax hanleyaiiiis. Each point is in correspondence with standardized drawings. Wade 1969. Ansell 1983a, Ansell 1983b). although there is varia- tion among members of the genus in the numbers of tentacles comprising each series (reflecting the number of siphonal nerves) and considerable variation in the complexity of their branching structure, especially around the tip of the inhalant siphon. The study of siphon regeneration in bivalves is relevant to the impact of predation, mainly by fish and crabs, and the adaptations that bivalves require to survive to this impact (Edwards et al. 1970, Trevallion 1971. Lockwood 1980. Hodgson 1982a. Hodgson 1982b. Petersen & Quammen 1982. Pekkarinen 1984. de Vlas 1985. Ansell & Gib.son 1990. Coen & Heck 1991. Boiisdorff et al. 1995. Ansell et al. 1999). The cropping of the siphons without killing the piey represents a way of conserving prey resources (i.e.. it relaxes the predation pi'essure because it does not cause mortal- ity) (Ansell I983u). The regeneration of the nihalant siphon in D. luiiilcyuiuis can be considered fully completed, in terms of siphonal function, between the fifth and the seventh day of regeneration. After this period the siphon continues growing in complexity and perhaps also in length, even beyond the tenth day. Hodgson (1982b) found that regenerating primary siphonal tentacles of D. serra could be de- tected as short lobes by 72 h following amputation. Within a further 24 h. these tentacles had elongated; by day 4 they had developed into tentacles, and secondary tentacles were also grow- ing; by day 5. primary and secondary tentacles were large and multilobed. and tertiary tentacles were appearing; and by 7 to 8 days after amputation, the tentacles had attained their former size and appearance. Ansell et al. ( 1999) found very similar results for D. vittatiis. TelUna tenuis da Costa. Scrobicitlaria plana, and Ma- coma balthica all lack elaborate siphonal tentacles, having only simple siphonal lobes. Hodgson (1982a) reports that these are fully lefomied by 7 days after amputation in 5. plana. In M. balthica. traces of the lobes are seen 2 days after amputation, and the lobes are fully formed again after 7 days (Pekkarinen 1984). The time course found here for the differentiation of the siphonal tentacles of D. hanleyaniis was similar to that of D. serra and D. vittatiis. However, most of these studies concentiate on the later siphon regeneration, quantifying the regenerated tissue after months of amputation, and found that the siphon continues to grow at a rate that depends on the quantity of removed tissue at the beginning of the experiments. In T. tenuis. S. plana, and D. serra. regrowth proceeds at approximately l7-259r of the (undamaged) siphon weight per week (Hodgson 1982a). whereas the observations of de Vlas (1985) indicate a somewhat lower rate of regrowth in M. balthica of approximately I09f per week, with the regrowth rate proportional to the amount lost. Pekkarinen (1984) also found a relatively slow rate of regrowth for M. balthica: when one third of the siphon was removed, it took 3 mo to regrow to its normal size. The rate of regrowth was not measured for D. hanlcyanus in this study but may be expected to be similar. These authors also found that a lag period followed the first 24 h after the cut and point out that in this period there is a reorga- nization of connective tissue in the wound area. This is also seen in D. hanleyanus during the first 24 h after the cut; that is. the amputated siphon does not undergo any lemarkable morphological change in this first period after amputation. The evidence obtained from this study, together with the results of Hodgson (1982a, 1982b) Pekkarinen (1984), and Ansell et al. (1999). confirm that inhalant siphon regeneration is a universal process inherent to species belonging to the Tellinacea superfam- ily. Moi'eo\er. all the species studied up to now suggest that the regeneration rate is very similar regardless of latitude, temperature, and intensity of wave action. ACKNOWLEDGMENTS We thank Robin Gibson and Robin Harvey from Dunstaffnage Marine Laboratoi7, Scotland, for their valuable comments of an earlier version of this manuscript. This work was supported by TX-70 UBACyT grant (Univer- sidad de Buenos Aires) and PICT 98 (0 1 -0432 1 ). Secyt, Argentina, and Fundacion Antorchas Cooperative Brazil-Argentine-Chile Program (G. Darrigran. PI) LITERATURE CITED Ansell. A. D. 1981. Functional morphology and feeding of Donax serra, Roding and Donax sordichis Hanley (Bivalvia: Donacidae). / Mollus- can Studies, 47:59-72. Ansell, A. D. 198.^a. The biology of the genus Donax. Sandy beaches as ecosystems. In: McLachlan & Erasmus, editors. Symposium on sandy beaches. Amsterdam: Junk. pp. 607-635. Ansell. A. D. 1983b. Species of Donax from Hong Kong: morphology, distribution, behavior and metabolism. In: B. Morton & D. Dudgeon, editors. Proceedings of the Second International Workshop on the Ma- lacofauna of Hong Kong and Southern China. Hong Kong, 1983. Hong Kong: Hong Kong University Press, pp. 19-47. An.sell. A. D. & R, N. Gibson. 1990. Patterns of feeding and movement of juvenile flatfislics on an open sandy beach. In: M. Bames & R. N. Gibson, editors. Trophic relationships in the marine environment. Ab- erdeen: Aberdeen University Press, pp. 191-207. Ansell, A. D.. R. Harvey & C. Giinter. 1999. Recovery from siphon dam- age in Donax vittatiis (BIVALVIA: DONACIDAE). / Molluscaii Stiulies 65:223-232. Ansell. A. D. & A. Trevallion. 1969. Behavioral adaptations of intertidal mollusks from a tropical sandy beach. J. Exp. Mar. Biol. Ecol. 4:9- 35. Bonsdorff, E., A. Norkko & E. Sandberg. 1995. Structuring zoobenthos: the importance of predation, siphon cropping and physical disturbance. J. Exp. Mar Biol. Ecol.. 192:12.5-144. De Vlas. J. 1979. Annual food intake by plaice and flounder m a tidal tlal area in ihc Dutch Wadden Sea. with special reference to consumption Regeneration of Donax hanleyanus 153 of regenerating parts ol niacmhenlhic prey. NcthcrUmds ,/. Sfu Res. 13:117-153. De Vlas. J. mS5. Secondary production hy siphon regeneration m a tidal tlat population oi Miuinna hahhuu. Ncthciiumh J. Sen Re.s. 19:147- 164. Coen, L. D. & K. L. Heck. Jr. 1991. The interacting effects of siphon nipping and habitat on bivalve {Menenaria inercenaria L) growth in a tropical seagrass (Haknlulc \\ri\;liiii Aschers) meadow. ./. £v/). Mar. Biol. Ecol. 143:1-13. Du\al. D. M. 1962. The comparative anatomy of some laniellibranch si- phons. Pvoc. Malacol. Sot. Lomioii 35:289-29?. Edwards. R. R. C. J. H. Steele & A. Trevallion. 1970. The ecology of 0-group plaice and common dabs in Loch Ewe. III. Prey-predator experiments with plaice. J. Exp. Mar. Biol. Ecol. 4:156-173. Gilbert. W. & E. Suehow. 1977. Predation by winter flounder {Psciidopleu- ronecU's aini'ricaiiiis) on the siphons of the clam Telliiiti tifiili.s. Niii{- rf/H.? 91:16-17. Hodgson. A. 19S2a. Studies on wound healing, and an estimation of the rale of regeneration, of the siphon of Sciobicnloria plana. J. E.xp. Mar. Biol. Ecol. 62:117-128. Hodgson. A. 1982b. Studies on wound healing and regeneration of the siphon of the bivalve Doiul\ scrra (Roding). Trans. Royal Soc. South Africa 44:489-498. Hughes. R. 1969. A study of feeding in Scrobicularia plana. .1. Mar. Biol. Assoc. U.K.. 49:805-823. Lockwood, S. J. 1980. The daily food intake of O-group plaice Plea- roiicctcs plalcssa L. under natural conditions: changes with age and season. J. Conseil Int. FExploration Mer. 41:181-193. Mori, S. 1938. Characteristic tidal rhythmic migration of a mussel. Donax semii>nosus Dunker. and the experimental analysis of its behavior at the flood tide. Dobutsugaku Zasshi. 50(1): 1-12. Moueza. M. & L. Frenkiel. 1974. Contribution a I'etude des structures palleales des Tellinacea. Morphologic et structure du manteau de Donnx tniiiciiliis L. Proc. Malacol. Sue. London 41:1-19, Narchi. W. 1978. Functional anatomy oi Donux hanleyanus. Phihppi 1847 (Donacidae - Bivalvia). Boletim Zoologia. 3:121-142. Pekkarinen. M. 1984. Regeneration of the inhalant siphon and siphonal sense organs of brackish-water (Baltic sea) Mucoma balthica (Lamel- libranchiata. Tellinacea). Ann. Zool. Feniiici 21:29— fO. Penchaszadeh. P. & S. Olivier. 1975. Ecologi'a de una pohlacion de ber- berecho {Donax htudeyanus) en Villa Gesell. Argentina. Malacolotiia 15(0:133-146. Peterson. C. & M. Quammen. 1982. Siphon nipping: its iinportance to small fishes and its impact on growth of the bivalve Proioihaca sia- niinea (Conrad). J. E.xp. Mar. Biol. Ecol.. 63:249-268. Trevallion, A. 1971. Studies on Tellina tenuis (da Costa). Ill: Aspects of general biology. J. Exp. Mar. Biol. Ecol. 7:95-122. Wade. B. A. 1969. Studies on the biology of the West Indian beach clam Donax denticulatus Linne. 3: Functional morphology. Bull. Mar. Sci. 19:306-322. Yonge. C. M. 1949. On the structure and adaptations of the Tellinacea. deposit-feeding Eulamellibranchia. Phil. Trans. Roral Soc. Edinbiiri^h 234:29-76. .Iininiul .-/ Shfllfisli Rccanh. Vol. 20. No. I, 155-159, 2001. SHELL GROWTH AND BIOMETRY OF THE STRIPED VENUS CHAMELEA GALLINA (L) IN THE MARMARA SEA, TURKEY MEHMET CENCIZ DEVAL Istanbul Univcrsiiesi. Su Uriinleri Fakiilresi. Onlu Cad. No: 200. Lalcli-lsianhiil. Turkey ABSTR.ACT Growth of the striped venus Cluimelea gallina (L. 1758) in the northern Mamiara Sea was studied by aging the shells. Internal growth bands with a seasonal pattern were identified on cross sections of the shell. During the period of reduced growth in October and February, a narrow hyaline band is formed. A large opaque band is formed between March and October. In the longest period of maximal growth, the increment coincides with the summer thermal maxima. C. gullina grows rapidly, reaching a length of 1 5 mm at the end of the first year and 2 1 mm at the end of the second year. At the end of the fourth and the fifth year, the length reaches 26 and 29 mm. respectively. The study of functional regressions permits us to conclude that the growth is allometric. A'£)' WORDS: bivalve. Chameleu gallina. biometry, age. growth. Marmara Sea INTRODUCTION Growth rings on shell surfaces have been widely used to make inferences about age and growth rate in bivalves. However, in some species it is difficult to recognize real seasonal growth rings from false rings that may form under different stress conditions (e.g., storms, extreme temperatures, diseases, spawning, gonad de- velopment, dredging, etc.). (Polenta 1993. Deval 1995. Caspar et al. 1995). Moreover, in older specimens the growth marks are so close to each other and to the shell margin that it may be difficult to identify them correctly. Calcium carbonates and organic matter are deposited in the shell layers with a seasonal pattern (Panella & Mac Clintock 1968. Rhoads & Lutz 1980). In cross sections, these shell layers may be recognized by their optical characteristics and therefore are used to age bivalves. Cross-sectioning has already been used successfully to age ven- erid clams (Richardson 1988. Ramon & Richardson 1992, Polenta 1993. Ameri et al. 1995). This method has also been used in the present investigation to study the growth of Chamelea gallina in the Marmara Sea. The most commonly used method to determine the rale of bivalve growth is to measure the changes in the dimen- sions of the shell. In the case of the striped venus C. gallina. the shell length along the anterior-posterior axis and the shell height measured from the umbo to the shell margin have been used. C. gallina is found in throughout the Mediterranean Sea, in- cluding the Black Sea. Along the Atlantic coast of northern Europe it is replaced by the closely related species C. striatitia (Backeljau et al. 1994). Both species are preferential, distributed on coastal fine sands (Guillou & Sariau 1985). C. gallina can be present in high densities in the Marmara Sea and in the Black Sea (Deval & Oray 1992. Deval 1995). The exploitation of C gallina in Turkish waters started 14 y ago, and the yearly catch had reached 41 thousand tons by 1994. The striped venus plays an important role in the economy of Turkey, and this species is totally exported. MATERIALS AND METHODS Between February 1993 and April 1994 at approximately monthly intervals, samples of C gallina were collected by dredg- ing (Picard 1965) in depths of 4-7 m in a site off Tekirdag (Bar- baros). in the northern Marmara Sea (Fig. I ). During the research period. 2,887 specimens were collected. Measurements of the length, height, width, total weight, shell weight, and meat weight were made with a vernier caliper, which was accurate to 0.1 mm, and with a digital balance, which was accurate to 0.01 g. The minimum and maximum values of the arithmetic mean, variance, coefficient of variance, and standard error (SE) were calculated, and the results were discussed (Valli et al. 1977. Vaili et al. 1985, Valli et al. 1986). Log y = Log a -i- b. Log x logarithmic model, which could easily be turned into y = a • \^. was used for bio- metric changes of C. gallnia (Ricker 1973. Ricker 1975). Generally, for each I -mm length class, five representative shells were sectioned. The right valve of the shell was cross- sectioned from the umbo to the ventral margin along the line of maximum growth, mounted on a glass slide, and then ground and polished on a lapidary wheel to obtain a section of approximately 20-30 |a.m. The section was then examined using a dissecting microscope under reflected light at low magnification. A total of 403 sections were prepared in this way and could be used for the interpretation of distinct annual increments. Shell sections were measured with video-analysis software, processing on a personal computer where the images were recorded through a video camera connected to the dissecting microscope. A complete record of size was obtained for each clam by measuring distance (height) from the umbo to the ventral margin of each hyaline band. Shell length was then transformed in shell height using the following regres- sion: Length = 0.49 1 + 1 .062 Height ( r" = 0.996; ;; = 536 ) In the Marmara Sea, C. gallina has a rather long spawning season, with a peak in June-July (Deval 1991 ). Therefore, to com- pute the parameters of the Von Bertalanffy growth equation June I was chosen as conventional birthday for all individuals consid- ered in the length-at-age key: L-il -Kn-l(i>-i where K is the Von Bertalanffy growth constant. L^ is the asymp- tofic maximum length, and to is the age at time zero. RESULTS During the study, the minimum measured length and the maxi- mum measured length were 3.6 mm and 34.5 mm, respectively. The mean length was found to be 20.1 ±0.12 mm (±SE). The most common repealed values of length was 20 mm (by 6.8'^). 19 mm. 155 156 Deval 26- 27- 28- 29- 30- 41- Figure 1. Location in tjit Marmara Sea where Chaiinlea galtina samples were collected. and 2 1 mm (both by 6.6%). The length frequency distiibutinn of C. iialliiui samples is shown in Figure 2. The regression relationship for length and height with weight tor C. gallina was positively linear. However, the relationship oi the dimensional and the weight parameters (length/weight) was exponential. As seen in Table I, when the regression coefficient between the two length variables (or the two weight variables) is b ^ I. there is allometry in growth; when b = I. there is isometry in growth. Likewise, when b = 3 between the length and growth variables there is isometry; when b ^ 3. there is allometry. Ac- cordingly, between all variables there is an allometric growth. Aging was carried out on 403 cross sections by two indepen- dent readers. Agreement was achieved on 297 shells (73.7%); the other shells were not further considered. To determine the period- icity in the formation of the hyaline band, the distance from the shell ventral margin to the nearest hyaline band was measured in all clams in the size range of 1 1-25 mm. When this distance did not exceed I mm, the hyaline hand was considered "at the margin band" (Fig. 3). Of the 297 shells with hyaline bands. 198 shells had margin bands. The seasonal distribution of 198 C. gallina with bands at the margin of 1 1-25 mm is given in Table 2. As seen Table 2. even though the building of the hyaline bands during the year is seen, the percentage of shells building hyaline bands in the warm summer season is very low (5.3%). However. the formation of hyaline bands appears to increase in the winter. C. gallina in the Marmara Sea spawns from May to September, with a peak in early summer (Deval 1991). Therefore, individuals in the first winter of life exhibit quite different sizes. Specimens born earlier (in May) benefit from the appropriate conditions for rapid growth during the summer. By winter they may already have attained a length of 18-19 mm. However, specimens born at the end of the spawning season (August-September) benefit from the optimal en\ ironmental conditions. For a short period and in winter TABLE 1. Regression and confldence intervals of allometric growth for biometrics parameters of Chamelea gallina. Variables (Dependent- •JS'r Confidence Independent) H R Log a b Intervals H-L 536 0.998 -0.069 1.024 0.937.VI.1|()7 W-L 232 0.984 -0.303 1.005 0.9895-1.0206 SW-L 322 0.992 -3.544 2.900 2.8598-2.9393 MW-L 42 0.720 -4.771 3.265 2.6229-4.2676 TW-L 536 0.993 -3.383 2.902 2.8728-2.9314 SW-TW 367 0.995 -0.137 0.976 0.9666-0.9858 MW-TW 42 0.807 -1.039 1 . 1 80 0.9056-1.4547 H = height; L = length: W = width; SW = shell weight: MW = meat weight; TW = total weight, they may be only 2-3 mm in length. L'sually in the following summer these specimens have an above-average growth rate. The growth rate has a high interindividual variability. A specimen of 19 mm in length may be 1-3 y old (Fig. 4; Table 3). The routine included in the software package FISAT (Gayanilo et al. 1996) was used to estimate the parameter of the Von Berta- lanffy growth equation, together with their asymptotic standard enor. L.,. = 33.46 ± 1 .32 mm K = 0.37 ± 0.067 t„ = -0.69 ± 0.30 y The theoretical mean length at age and the annual growth com- puted according to the above parameters of the Von Bertalanffy growth equation are listed in Table 4. In the northern Marmara Sea. C gallina has a rather rapid growth during the first 2 y of life (Table 4) and reaches the mini- mum marketable length of 2i mm in less than 3 y. The live-weight annual increment is highest during the third year of life (Fig. 5). Growth rates, both in length and weight, are much lower in older individuals, and the maximum size attained in the Marmara Sea is significantly lower than in the Adriatic Sea. The mean life span in the C. gallina population of the northern Marmara Sea seems not to e.\ceed 5 or 6 y (shell length 29-30 mm). Only 2.29c of the sampled specimens were larger than 30 mm. indicating a rather high exploitation level in the investigated area. DISCUSSION This research is the first analysis on the growth and age of C. gallina in the Marmara Sea. giving the most detailed biometric parameters of C. gallina in Turkey. I |l|l|l|l|l|l|l|l|l|l|l|l|l|l| Length ( mm ) Figure 2. The length-frequency distribution of Chamelea gallina Figure 3. "At the margine band" on the coss-section of the shell of *'""P''^*- Chamelea gallina. Shell Growth and Biometry of C. gallina 157 TABLE 2. Position of the hyaline band nearest to the ventral margin observed in samples caught in Narious seasons. Shells Milh H> aline Date Sections (H) Band "at the Margin" (/ (")<-) June 2}. \99} OLtoher 4. 1993 February 6. 1994 75 44 79 4 (5.3 1 8(18.2) 44(55.7) Figure 4. Thin section of Chamelea gallina of the Northern Marmara with arrows showing the annual growth (hyalin band) increments (length 27.9 mm, age 4+; data from the October sample). TABLE 3. The lengths of Chamelea gallina in winter the formation of the hyaline growth bands. Number Minimum Maximum of .\nnual Records Length Length .Average Hyaline Bands (H) (mm) Imni) Length* 1 3(X) 2.3 19.3 8.79 ±4.258 2 252 12.6 25.6 18.73 + 2.530 3 161 19.1 29.7 23.58 ±2.1 17 4 92 22.0 32.0 26.37 ± 1 .775 5 55 24.1 32.9 28.46+ 1.699 6 27 25.3 33.8 29.54 ± 1.898 7 12 28.4 34.3 30.28 ± 1.736 8 4 30.0 32.2 31.45 ±0.998 ■ Average ± SD. TABLE 4. The average length at age and weight at age and the annual growth of Chamelea gallina calculated according to the \ On Bertalanffy growth equation and the length weight exponential equatiim. Age Mean .Annual Mean Annual (y) Length (mm) Growth (mm) Weight (g) (irowth Ig) 1 15.3 15.3 1.14 1.14 1 20.8 5.5 2.77 1.63 3 24.7 3.5 4.56 1 .69 4 27.3 2.6 6.09 1.63 5 29.2 1.9 7.40 1,31 6 .30.5 1.3 8.40 1.00 7 31.4 0.9 9.14 0.74 0 12 3 4 5 6 AGE (YEAR) Figure 5. V. Bertalanffy grovith curve of Chamelea gallina in the Northern Marmara Sea. The investigations conducted on the biometric parameters in Tmkey reported the maximum length of C. galliiui as 36 iDm in the Dardanellas (Alpaz & Onen 1989). 35.5 mm in northern Marmara Sea (Cebeci 1992), and 32.2 mm in western Black Sea coast (De- val 1991). Many studies performed on C. gallina showed a higher maximum length of the specimens sampled in Europe and in the MeditoTunean. The )iiaximum length of C. gallina was 41 mm in the coast of Spain (Vives & Suau 1962), 49 mm in the middle Adriatic Sea (Polenta 1993). and 46 mm in the southern Adriatic (Marano et al. 19X2). By analyzing the relationships between the weight and the di- mensional parameters, calculated fro)ii regi'ession and correlation analyses, the growth was determined to be allometric. The growth of C. gallina studied on the coast of Spain and in the sea of Adriatic was also allometric (Pogiani et al. 1973, Marano et al. 1982. Cano & Hernandez 1988). The hyaline and the opaque growth bands were clearly evident in cross sections of valves from C. gallina from the Marmara Sea. These growth bands appeared to be formed seasonally and were used to age claiDs and to establish a length-at-age key. The pa- rameto's of the Von Bertalanffy growth equation calculated from these data, together with the size fiequency distributions obtained from the monthly sampling program, show that C. gallina in the Mainiara Sea has a lower growth rate and attains a smaller size in comparison with other MediteiTanean areas. In the Marmara Sea the annual growth pattern, characterized by reduced or no growth during the winter months with sea tempera- tures below 10°C. is similar to that reported for the Adriatic Sea (Froglia 1975, Polenta 1993. Arneri et al. 1995). In both areas a hyaline band is formed in the shell during winter inonths. In the western Mediterranean, off the southern coast of Spain, reduced growth and formation of the hyaline band were observed late in summer (Ramon & Richardson 1992), and apparently tempera- tures above 27°C negatively affect the growth of C. gallina. Table 5 lists the average lengths at age reported in the literature for different Mediterranean locations. Although different authors used different methods to study the age and growth of C. gallina. the results obtained in different Adriatic regions are in agreement that, in the Mediterranean Sea. the )nost growth occurs in the first four age classes. The maximum sizes recorded by Adriatic work- ers are well above those reported for Spanish and Turkish C. gallina populations. The differences in the growth rates of C. gallina off the Span- ish. Adriatic, and Marmara coasts could be related to environmen- tal factors such as annual sea temperature and salinity. C. gallina 158 Deval TABLE 5. Average lengths at age of Chamelea gallina reported in the hterature for different Mediterranean localities. .\ge (V) Reference Region Method 1 ■> 3 4 5 6 7 8 Length Poggiani et al. (1973) Middle Adriatic Surface rings 15 24 30 33 35 4(1 Froglia (1975) Middle Adriatic Length distribution 17 25 Polenta (1993) Middle Adriatic Internal hands 17 25 30 34 38 39 42 45 49 Maranoel al. (1982) South Adriatic Surface rings 15 23 31 35 38 42 46 Vives and Suau ( 1962) West Meditter. Surface rings 18 24 28 31 Ramon and Richardson (1992) West Meditter. Internal hands 20 25 28 31 Present investigation North Marmara Internal hands 15 21 25 27 29 30 34 has the highest growth rate in the Adriatic eiitrophic waters. The lowest growth rate was observed in the population of the notlhern Marmara Sea (off the Tekirdag coast). Water masses of low sa- linity flow from the Black Sea to the Mediterranean Sea through the Marmara Sea (Kocatas et al. 1993). Therefore, the salinity over the Chamelea fishing grounds in the Marmara Sea (17-24%() is significantly lower than in Adriatic and Spanish coastal waters (32-36%f). At the moment no experimental data are available to relate the different halide regimes to the observed differences in growth. In conclusion, C. galliiui reaches 15 mtii at the end of the first year and 21 mm at the end of the second year. The regulations on the fishery of this species restrict the catch of individuals less than 22 mm (conesponding to 24 mm in length), and C. gallina reaches this height around the end of the third year (Fig. 5). The estimated growth coefficient and asymptotic maximum size were 0.37 and 33.46 mm, respectively. ACKNOWLEDGMENTS My thanks are extended to Dr. Carlo FROGLIA and Dr. Enrico ARNERI (I.R.PE.M.-C.N.R./ Ancona - Italy) who supported this research by providing their research facilities and the TUBITAK/ NATO which financially supported this research. This research was partially up supported by TUBITAK/NATO International Research Scholarship. REFERENCES Alpbaz, A. & M. Oneii. 1989. Tiirkiye'den Ihrai^' Edilen Kuni .Midjesi (f. gailiiia L.) Uzerine Arastirmalar. Ege Universitesi Su Uriinleri YUksek Okulu Dergisi. pp. 78-86. Arneri, E.. G. Giannetli. G. Polenta & B. Antoloni. 1995. Age and growth of Chamelea gallina (Bivalvia: Veneridae) in the central Adriatic Sea obtained by thin sections. Men Medit. 34:17. Backeljau T.. Bouchet P.. Gofas S.. de Bruyn L. 1994. Genetic \ariation. sistemates and distribution of the venerid clam Chamelea i;allina. J. Mar. Biol. Ass. U.K. 74:21 1-233. Cano. F. V. & J. M. Hernandez. 1988. Biometria y Reproduccion de C. gallina L. en la bahia de Mazarron (SE de la Peninsula Iberica) y Alcunas Consideraciones sobre su Pesca, FAO Rapp. Pvshe.s 395:107- 112. Cebeci. M. 1992 Kuzey Marmara Denizi'nde C. gallina L.'nin Biyometrisi ve Avciligi, I.U. Fen Bilimleri EnstitUsii, Doktora Tezi. Deval. M.C. 1991. 1990 Yilinda Marmara Denizi'nin Kuzey Batismda ve Bati Karadeniz'in Bazi Bolgelerinde Beyaz Kum Midyesi'nin (C. gallina L.) Yumurta Birakma Siiresinin Tespiti (The spawning of the striped venus C. gallina in the northern Marmara Sea and m the western Black Sea. Turkey) I.U.Fen Bil.Enst.. M.Sc. tesis. Deval. M.C, 1995. Kuzey Marmara Denizi'nde Chamelea gallina L.,1758'nin Yas ve Kabuk Gelisimi (The age determination and the shell growth of Chamelea gallina L., 1 758 in the Northern Parts of Sea of Marmara) I.U.Fen Bil.Enst., Doktora Tezi (Ph.D. thesis). Deval. M. C. & I. K. Gray. 1992. Marmara Denizi ve Kara Deniz'de C. gallina L.'nin Yumurtlama Diinemi \e Biyometrisi. l.U.Su Uriinleri Dergisi. 1(6):127-142. Froglia. C. 1975. Osservasioni sulfacrescimento di C f>allina L. ed £". minor (Chenu) nel Medio Adriatico. Quaderni del Laboratorio di Tec- nologia della Pesca. 2( I ):37-48. Kocatas, A., T. Koray, & M. Kaya. 1993. Fisheries and environmental studies in the Black Sea. Part 3: A review of the fisheries resources and their environment in the Sea of Marmara, Studies and Reviews. Gen. Fish. Conn. Medil. 64:87-143. Maiano, G., N. Casavola, C. Saracino & E. Rizzi. 1982. Riproduzione e Crescita di C gallina L. e V. verrucosa L. nel Bassso Adriatico. Mem. Biol. Mar Ocean. X1I:93-1I0. Gaspar, M. B.. M. Castro, & C. C. Monteiro. 1995. Age and growth rate of the clam, Spisula solida L., from a site off Vilamoura, south Portugal, determined from acetate replicas of shell sections. Sci. Mar. 59(Suppl. l):87-93. Gayanilo, F. C. P. Sparre, & P. Pauly. 1996. The FAO-ICLARM stock assessment tools (FISAT) user's guide, FAO computerized information series (Fisheries). 8:126. GuiUou, J, & P. G. Sariau. 1985. Some observations on the biology and the ecology of a Venus strianda population in the bay of Douamenez, Brittany. J. Mar. Biol. .\ss. U.K. 65:889-900. Pannella, G. & C. Mac Clintock. 1968. Biological and environmental rhythms in molluscan shell growth. J. Paleontol. 42:64-80. Poggiani. L., C. Piccinetti & G. Manfrin-Piccinetti. 1973. Ossevazioni sulla Biologia dei Molluschi Bivalvi V. gallina L. e Adriatico.Note Lab.Bi- ol.Mare e Pesca. Fano 4(8 1:189-212. Polenta, R. 1993. Osservazioni suU'Acrescimento della Vongola C gallina L. nel Medio Adriatico, Univ. degli studi di Bolog..La Tezi di Laura. Ramon M. & C. A. Richardson. 1992. Age determination and shell growth of C. gallina L. in the western Mediterranean. Mar. Ecol. Progr. Ser. 89(0:15-73. Rhoads. D.C. & R. A. Lutz, editors. 1980. Skeletal growth of aquatic organizms. Biological records of environmental change. Topics Geo- hiol. 1:750. Ricker. W. E. 1473. Linear regression in fishery researh. J. Fish. Res. Hoard Can. 30:409-434. Shell Growth and Biometry of C. galuna 159 Ricker, W. E. 1975. Computation and interpretation of biological statistics Valli, G.. D. Zardini & P. Nodari. 1985. Cycle Reproductif et Biometiie of fish populations, bulletin 191. Ottawa: pp. 202-235. Chez C. gallina dans le Golfo de Trieste. /?40 mm). All results were compared with data from populations in North Atlantic and adjacent waters. KEY WORDS: Arctica islandica. ocean quahog. distribution, abundance, growth, size, meat yield. Baltic Sea. Mecklenburg Bight INTRODUCTION The ocean quahog. Arctica islandica. is an arctic-boreal species that occurs in North Atlantic and adjacent waters. In the North Sea region, the species" range extends into the Western Baltic Sea and reaches its eastern limit of distribution in the Arkona basin (von Oertzen and Schulz 1973). The largest populations reside in Kiel and Mecklenburg Bights. The ocean quahog is the largest bivalve in the Baltic Sea and lives for more than several decades (Amtz and Weber 1970. Brey et al. 1990). The long living species is an important indicator for environmental conditions. Beside salinity and sediment structure, o.xygen concentration has a strong influ- ence on the composition of Baltic Sea fauna and flora. Since the I960"s, oxygen deficiency below the halocline (16 to 20 m) has resulted in destruction of the bottom fauna of Mecklenburg Bight (Gosseick et al. 1987. Schulz 1968). Although A. islandica is highly resistant to oxygen depletion, the population is diminished by mortality caused by frequent and long lasting periods of anoxic conditions. In the deeper parts of Liibeck and Mecklenburg Bights a severe decrease of the Baltic Sea population of .4)rr/av islandica has resulted during recent decades (Gosseick et al. 1987. Schulz 1968). Long living species like the ocean quahog {Arctica is- landica) decreased in abundance and were replaced by communi- ties of short living polychaetes (spionids. capitellids). The geographical variability in growth rates of A. islandica is well documented for the North Sea and adjacent waters (Witbaard et al. 1999) and for US and Canadian east coasts (Kennish et al. 1994, Kraus et al. 1992. Murawski et al. 1982). Much is known about the reproduction, ineat yield and age mostly for Atlantic populations off North America and Iceland coasts (e.g. Chaisson and Rowell 1985. Fritz 1991. Steincrimsson and Thorarinsdottir *Corresponding author. E-mail: michael.zettler@io-warnemuende.de; Fax: +49-381-5197-440. 1995. Thorarinsdottir and Einarsson 1996). Little is known about populations of this species in the vicinity of the Baltic Sea {Arntz and Weber 1970, Brey et al. 1990). The purpose of this study was tt) investigate the distribution, frequency and biomass of the recent population of Arctica islandica in Mecklenburg Bight as the first extensive study on the population characteristics of this important indicator species near its limit of distribution. A further aim was to compare these results with existing data from the I960's compiled by Schulz (1969a) and to include an 1 1 -year time series of one monitoring station. Area of Investigation The Mecklenburg Bight is part of the Belt Sea and belongs to the transition area between the North Sea and Baltic Sea (Fig. I). It is connected with Kiel Bight via Fehniarnbelt and with Kattegat via the Belts. To the East, the Kadet trench crossing the Darsser rise connects it with the Baltic proper. The location in the southwestern part of the Baltic with its relatively high salt content is decisive for benthic colonisation. Fauna and flora of the area are determined by hydrological and morphological factors, with salinity being the formative influence. Another important factor is the sediment structure, which is formed by residual sediments and mud. During 1999 a macrozoobenthos survey was made in the Meck- lenburg Bight. In total 95 stations were sampled between March and September (Fig. 1). The depth interval was 5 to 29.6 m. the sediment varied from fine sand at the shallowest stations to sand mixed with silt and clay at deepest stations. The sediment charac- teristics and current data within the area were published by Lange et al. (1991). MATERIALS AND METHODS Profiles of salinity were recorded throughout the water column using a CTD (conductivity/temperature/depth probe) system. 161 Zettler et al. 54 5 54.0 Gedser • • • 7S ® ( . •. * *: '(^r Rostock Lubeck 1 1 0 12 5 13 0 1 ! 5 12 0 Longitude (E) Figure 1. Investigation area with 95 stations in the Mecklenburg Bight (encircled stations refer to text and figures). Samples for bottom water oxygen were taken with a 5 I water sampler (moLinted on the CTD) at 0.5 m above the bottom and were determined with Wini\ler titration. Benthic samples were taken with a O.I wc Van Veen grab. At each station three parallels of grab samples were carried out. Due to sediment conditions grabs of different weights were used. The samples were sie\ed through a 1 mm screen and animals preserved with 4% formaldehyde in the field. For sorting in the laboratory, a stereo microscope with 10- 40x magnification was used. A dredge haul (net mesh size 5 mm) was taken in order to obtain A. islandica specimens for the study of population size structure. For the characterisation of the habi- tats, i.e. assessment of sediment structure, epibenthos and current, an underwater video-system was used which was mounted on a sledge. The sledge was towed over the bottom by a drifting vessel at lowest possible speed (< 1 knot). The camera was installed on a pan and tilt head. Scaling was accomplished by four crossed laser beams projected into the picture. In addition bivalve siphon open- ings detectable at the sediment surface were counted to assess patchiness and distribution of adult (approximately >30 mm shell length) ocean quahogs. At each station the shell length of all collected individuals oi A. islandica were measured with a vernier calipers to the nearest 0. 1 mm for the length-frequency distribution and the length-meat weight relationship. In total about 2300 specimens were measured. The valves and the wet meat of the specimens were weighed separately. Furthermore, the dry weight (DW) and ash free dry weight (AFDW) were determined to the nearest 0.01 mg. Because of very low weights, quahogs with shell lengths <3 mm were weighed in groups of 5 or 10 indi\iduals sorted in size classes of 0.1 mm. Length-frequency distribution for each station was cal- culated for 10 mm size classes. The age of each A. islandica was estimated by measuring internal growth bands in the shell. This method is only applicable for specimens younger than 40 years (nearly all observed individuals are younger). For comparison and calibration, several shells were processed for observation of inter- nal growth lines in acetate peels (see Ropes 1985). The distribution map of the ocean quahog in the Mecklenburg Bight was made using Surfer (Win32) version 6.04 program of Golden Software Inc. The recent distribution was compared with the results from Schulz ( 1969a), whose data were transformed into the Surfer program to obtain a comparable map. Quahog distribu- tional data from the monitoring station 12 of the Baltic Sea Re- search Institute Warnemuende and the Institute of Marine Re- search Kiel were used for the 1 1-year time series (1988-1999). RESULTS Bottom Water Variables Salinity ranged between 7.5 and 22 psu in the water column. Bottom water salinity of areas inhabited by Arctica islandica var- ied between 12.5 and 22 psu in 1999. No oxygen deficiency was observed during our survey. Up to a depth of 16 m more than 7 mg/1 were measured. Oxygen content decreased to a minimum of 4.5 mg/1 only in the deepest parts of the Bight. Distribution, Abundance and Biomass In 1999 Arctica islandica was distributed between 15.6 and 29.6 m in the Mecklenburg Bight. Most stations within these depths were colonized (Fig. 2). Only in the deepest parts of the innermost area (Lubeck Bay) were no quahogs found. Further- more, the outer Kadet trench with strong currents and stony sub- strates was not inhabited. The highest abundance was found in the south-eastern part of the Bight with a maximum of 571 ind./m" and an AFDW of 34 g/nr at station 103 (water depth 17.4 m). Thirty years ago (in the mid 1960's). A. islandica reached an abundance between 10 and 100 ind./m' with a maximum of 200 ind./m" in the central part of the Bight (Schulz 1969a) (Fig. 2). At depths below 20 m of the innermost area (Lubeck Bay), no A. islandica were found. The highest biomass (AFDW) was observed in deeper parts of the central Bight at station 54 (water depth 23 m) with 120 g/m" and an abundance of 236 ind./m" (Fig. 3). The Mecklenburg Bight had an estimated colonized area of 5200 km" within water depths of 15 and 30 m. a mean abundance of 91 ind./nr and a biomass (AFDW) of 15 g/m". Based on these values, the whole population Ocean Quahog (Arctica island/ca) in the Baltic Sea Arctica islandica abundance ind./m^ in 1999 163 Longitude (E) Arctica islandica abundance md./m^ in 1965 ■ubecK I 110 115 12 0 12 5 130 Longitude (E) Figure 2. Distribution ot Arctica islandica (ind-Zm") in 1999 and during the investigation period in the mid 1960's of Schulz ( 1969a) (same scale for better comparison). of the Bight was estimated at approximately 4.7E + I 1 individuals with a hiomass ( AFDW) of 7.8E + 04 t ( 1 .04E + 06 t wet weight). In 196? the mean abundance reached 20 ind./nr. In the last 1 1 years the abundance of A. islandica at station 12 varied between 20 and 120 ind./m~ (Fig 4). A. islandica occurred in mean density of 40 to 75 ind./m". However, due to patchy distribution of this species, the standard deviation was very high. The same was ob- served in 1999. when the highest abundance occurred with 120 ± 70 ind./m~. Similarly, records by the video showed a similar patchy distribution of the quahog siphon openings. If only adults, specimens were taken into account, the average counts of openings were comparable with the estimated abundance by grabs at several stations. However, due to patchiness. the maximum abundance was much higher and ranged between 400 and 700 ind./m". Be- tween these colonized centers, big patches of unsettled sediments were visible. In this study the highest abundance was observed in depths between \5 and 20 m. whereas the maximum of biomass occurred between 20 and 2.5 m (Fig. .5). With increasing water depth, the abundance decreased from an average of 155 ind./ni" in 15-20 m, to 85 ind./nr in 20 to 25 m. and 35 ind./nr in 25 to 30 m. respectively. In comparison the biomass (AFDW) was low in both the 15 to 20 m. and 25 to 30 m interval (7-9 g/m"), but averaged 28 g/nr at 20 to 25 m. Population Structure, Meat Yield and (irowlh During the benthos survey, shell lengths of Arctica islandica were between 1.5 and 64 mm (Fig. 6). Quahogs in the size class 0 to 10 mm composed 409^ of the population. Particularly at shal- lower stations ( 15-20 m). the one- or two-year size classes (<10 mm) dominated. In deeper water (>20 m) individuals >20 mm were the dominant group. Generally only one size class dominated the population structure of an area. At shallow stations (stns WT5, MB3 and 103. see fig. 1 and 6) the most successful settlement took place during the last two years (the size class 0 to 10 mm domi- 164 Zettler et al. Arctica islandica biomass AFDW in g/m^ in 1999 11 0 12 5 115 12 0 Longitude (E) Figure 3. Biomass distribution o( Arclica islandica (g AFDW/m") in 1999. nated with over 509c). whereas httle or no recruitment was de- lected in stations deeper than 20 m. The dominant size class varied from 1 1 to 20 mm at stations 1 2 and Z3 via 2 1 to 30 mm at stations 63 and 07. to 31 to 40 mm at stations 75 and 54 (Fig. 6). The wet meat yield (percentage of total wet weight) increased with increasing size up to shell length of 55 mm (Fig. 7). The mean meat yield of all individuals (1.5-55 mm) was 18.3% and varied between a minimum at 5*^ (2.7 mm shell length) and a maximum at 38* (47 mm shell length). The organic content (all organic material of a quahog. i.e. meat, periostracum and ligamentum) was calculated as the relation of individual ash free dry weight to dry weight (Fig. 8). With increas- ing shell length a significant decrease in organic content was ob- served. The highest amount ( 14. 3'^'r) appeared in the shell si/e class <10 mm. The lowest organic content was 9.49c in the size class >40 mm. The mean organic content of DW of all individuals (1.5 to 55 mm shell lengths) ranged from 7.1% to 22.3% with a mean value of 1 1%>. Relationship between shell length and individual weight (meat 200 180 -- 160 Stn. 012 .'i4° 18.58 N 11' 3:.sie deplh=25 m 1965 1<)88 1990 1991 1992 199.1 1994 1995 1997 1998 1999 Figure 4. The development of mean abundance (± SD) of .Arclica is- landica at station 12 in the last 11 years in comparison with data of (Schuiz 1969a) from the mid 1960's. wet weight and ash free dry weight) is shown in Figure 9. In these graphs all measured specimens (1.5-64 mm) of all stations are included. The smallest meat wet weight was 0.095 mg at a shell length of 1.5 mm and the largest was 14.7 g at 64 mm (Fig. 9a). The average meat yield per unit shell length of .4)c7/90'7r of the A. is- landica population of the Mecklenburg Bight. This population structure is completely different from populations of the Atlantic Ocean, where the smaller size classes are mostly absent (e.g. Fritz 1991. Murawski et al. 1982. Thorarinsdottir and Einarsson 1996. Steingrimsson and Thorarinsdottir 1995). We attribute this to suc- cessful recruitment in the Mecklenburg Bight during the last 10 to 15 years. In contrast in Atlantic waters, ocean quahog populations have only sporadic recruitment and strong recruitment events oc- cur only every twenty (or more) years (Murawski et al. 1982. Steingrimsson and Thorarinsdottir 1995). The mean wet meat yield of A. islandica of 18.3% (5% to 38%) is similar to other reported values. For a population in the Kiel Bay. Arntz and Weber ( 1970) reported a percentage of 19.3 to 22.2 with minimum values in winter. The meat yield of A. islandica off Iceland coasts ranged between 17% and 40%- and averaged 30% (Thorarinsdottir and Einarssson 1996). If only the smaller size classes (<40 mm) of quahogs off Iceland are considered, then the mean wet yield of approximately 20% is similar to the result of this study. The meat yield observed from grounds off the coast of the United States and Canada ranged between 20.7% and 23% (Bakal et al. 1978. Chaisson and Rowell 1985). In the Mecklenburg Bight, the mean individual ash free dry weight of A. islandica decreased from 14.3% to 9.4% with increas- ing size classes. The increasing mean wet yield and the decreasing 1 1 T 4- 1 I 1 ■ --- T T + + i 1 n=79 n=74 n=78 n=44 n=35 n=310 1 ' 40 mm I meLin Shell length (mm) Figure 8. Mean individual percentage of ash free dry Height for dif- ferent size classes of .4;r//(a islandica from the Mecklenburg Bight in 1999 (± SD). Ocean Quahog (Arctica islandica) in the Baltic Sea 100000 10000 167 -I 1 1 1 1 1 1 r- 20 30 40 50 Shell length in mm 10000 01 o i b) 1000 100 10 0.1 0.01 ^ ,>« _ .£^ > «** .J^ igfff- ^ ^ ^ / w / / y = 0,01x='^°»' R- = 0,9933 n=312 7 . , 1 ^ — — 1 1 1 — 10 20 30 40 50 Shell length in mm 60 70 Figure 9. a) Estimated shell len)>tli-nieat wet weight relationship for .-1;<7(V« islandica. b) estimated shell length-ash free dry weight relationship for Arctica islandica. In these graphs all measured specimens of the whole Mecklenburg Bight populations are included. The corresponding equations for the calculated best tit regression lines is given in the figure. ash tree dry weight with increasing shell length indicate relatively higher water content of bigger quahogs. The length specific meat wet weight (shell-length to meal-weight relationship) in this study was similar to that reported for ocean quahogs off United States coasts. In the Mecklenburg Bight, the calculated meat wet weight for an individual of 95 mm shell length w as 39 g. The meat wet weight of A. islandica off New York reported by Murawski et al. (1982) were 36 and 38 g and off New Jersey 37.9 and 5 1 .3 g (Fritz 1991 ). The biggest meat wet weights oi A. islandica ever reported (47.6, 55.5 and 70 g) were found off Iceland by Thorarinsdottir and Johannesson ( 1996). The calculated .AFDW for an individual of 95 mm shell length reached 5.4 g. Several authors reported on geographical differences in growth rates of A. islandica in its distribution area and between field and laboratory populations (e.g. Brey et al. 1990. Fritz 1991. Josefson et al. 1995, Kennish et al. 1994, Kraus et al. 1992, Steingrimsson and Thorarinsdottu- 1995, Witbaard 1996, Witbaard el al. 1999). Ocean quahogs are among the slowest growing and long-living marine bivalves (Murawski et al. 1982, Thompson et al. 1980). Growth rates for this species vary with respect to location (tem- perature, food supply, salinity, and abundance). Kraus et al. (1992) observed the fastest growth in the laboratory where ,4. islandica reached 54 mm from 9.6 mm within three years. In the laboratory the specimens grow to 37 mm within two years. In the field they need 26 to 33 years to achieve the same shell length. The com- parison of the growth curve of the Mecklenburg Bight population with growth curves calculated by Witbaard et al. (1999) for 12 populations from the North Sea and adjacent waters shows similar growth increments for the first 40 years of life. For the first five years, the Mecklenburg Bight population had a growth rate very similar to populations of the Kattegat with approximate linear increase in size with age (Josefson et al. 1995). In comparison w ith 168 Zettler et al. 60 50 40 -- ^ 30 C/3 10 -- 0 \jjjj. o o " a 0 ^ « =s ^ a § i o8 '^J^'^i T s TtUrtTS ' 0 «B^ <1^' iUf«" J 4 f 1 _ 1 y= 14.416Ln(x)-l, 14561 R - n-12 72 r , , , , ' ' ' ' 1 1 1 1 1 1 ' ' 1 ' ' ' ' 1 ' ' ' ' 1 0 10 30 35 40 15 20 25 Age (years) Figure 10. Growth curve of .4;(7iVa islandica from the Mecklenburg Bight for the first 40 >ears of hfe. the growth rates of a Kiel Bay population calculated by Brey et al. (1990), the growth in the Mecklenburg Bight was slower. A. is- landica in the Kiel Bay reached about 84 mm after 40 years; whereas, the specimens in the Mecklenburg Bight had a mean shell length of 52 mm in the same time. These differences are difficult to explain. The lower salinity may be one reason for the observed slower growth of quahogs in the Mecklenburg Bight. A possible further explanation might be that the Kiel Bay higher bottom cur- rents (closer openings to the Small and Great Belt), enriched in suspended or re-suspended material, would supplement A. is- landica, as a filter-feeder, with advective food supply (see Wit- baard et al. 1999). The present study shows the peculiarities of the ocean quahog in the Mecklenburg Bight near its eastern distributional boundary within the Baltic Sea. Further investigations must deal with the differences in observed growth rates of quahogs within the Baltic and adjacent waters. The comparative population dynamics of dif- ferent water depths and/or sediment structure and the dispersion and settlement patterns of the larvae in this "border" area are of interest. Two of the main questions following this study are: Is the short life expectancy of quahogs in the Baltic due only to the salinity and which conditions cause the success of strong recruit- ment in its geographical distribution. ACKNOWLEDGMENTS Parts of this work were supported by the German Umwelt- bundesamt and the Bundesministerium fiir Forschung und Tech- nologie. The authors are indebted to Volker Schroeren (LANU Schleswig-Holstein) and to Dr Heye Rumohr (IfM Kiell for sup- plying of data. We would like to thank Dr Herbert Siegel and Monika Gerth for introducing into the Surfer program. Thanks are also to Christine Peters and Toralf Hoth for helpful technical as- sistance. LITERATURE CITED Al-Hissni. Z. 1989. Saisonale und annuale Fluktuationen des Makrozoob- enthos in der Liibecker und Mecklenhurger Bucht in den Jahien 1985- 88. PhD Universitat Rostock. Arntz, W. E. & W. Weber. 1970. Cypnmi islandica L. (Mollusca, Bivalvia) als Nahrung von Dorsch und Kliesche in der Kieler Bucht. Ber. Dt. Wiss. Komm. Meeresforsch. 21:193-209. Bakal, A.. W. F., Rathjen & J. Mendelsohn. 1978. Ocean quahog takes supply spotlight as surf clam dwindles. Food Prodtut Development 12:70-78. Boll, E. 1852. Cyprina islandica in der Ostsee (vor Niendorf) in Meck- lenburg. A /r/i. Fi: Ver. Nalurgesch. Mecklb. 6:pl25. Brey, T., W. £., Arntz, D., Pauly & H. Rumohr. 1990. Arctica {Cyprina) islandica in Kiel Bay (Western Baltic): growth, production and eco- logical significance. J. exp. mar. Biol. Ecol. 1 36:2 1 7-235. Chaisson, D. R. & T. W. Rowell. 1985. Distribution, abundance, popula- tion structure, and meat yield of the ocean quahog (Arclica islandica) and Stimpson's surf clam {Spisula polynyma) on the Scotian Shelf and Georges Bank. Can. Ind. Rep. Fish. Aqu. Sci. 155:pl34. Fritz. L. W. 1991. .Seasonal condition change, morphometries, growth and sex ratio of the ocean quahog, Arctica islandica (Linnaeus, 1767) off New Jersey, U.S.A. / Shellfish Res. 10:79-88. Gosselck. F.. F., Doerschel & T. Doerschel. 1987. Further developments of niacrozoobenthos in Liibeck Bay. following recolonisation in 1980/81. Int. Revue ges. Hydrohiol. 72:631-638. Hagmeier, A. 1930. Die Bodenfauna der Ostsee im April 1929 nebst eini- gen Vergleichen niit April 1925 und Juli 1926. Ber. Dt. Wiss. Komm. Meeresforsch. 5: 1 56- 1 73. Josefson. .\. B.. J. N.. Jensen. T. G.. Nielsen & B. Rasmussen. 1995. Growth parameters of a benthic suspension feeder along a depth gra- dient across the pycnocline in the southern Kattegat, Denmark. Mar. Ecol. Prog. Ser. 125:107-115. Kennish. M. J., R. A.. Lutz. J. A., Dobarro & L. W. Fritz. 1994. In situ growth rates of the ocean quahog. Arctica islandica (Linnaeus, 1767), in the middle Atlantic Bight. / Shellfi.sh Res. 13:473-478. Kohn. J. 1989. Zur Okologie sandiger Boden der Mecklenhurger Bucht. PhD Universitat Rostock: 170pp. Kraus, M. G.. B. F.. Beal, S. R., Chapman & L. McMartin. 1992. A comparison of growth rates in Arctica islandica (Linnaeus. 1767) be- tween field und laboratory populations. / .Shellfish Res. 11:289-294. Ocean Quahog (Arctica islandica) in the Baltic Sea 169 Landers. W. S. 1976. Reproduction and early development of the ocean quahog. Arctica isUiiulica. in the laboratory. Naiilihis 90:88-92. Lunge. W.. E. Mittelstaedt & H. Klein 1991. Stromungsdaten aus der Westlichen Ostsee. MePergebnisse au.s deni Zeitrauni 1982 bis 1986 sowie hislorische Beobachtungen vom Feuerschift" "Fehmarnbelt". Dr. Hydmgr. Z. Erg. Heft Reihe B 24:1-129. Lenz. H. 1875. Die wirbellosen Thiere der Travemunder Bucht. Theil 1. Komm. wiss. Unlers. dlsch. Meere Kiel.\-24. Mann. R. & C. C. Wolf. 1983. Swimming behavior of larvae of the ocean quahog Arctica islandica in response to pressure and temperature. Mar. Ecol. Prog. Ser. 13:211-218. Malthaus. W.. G. Nausch, H.U. Lass, K. Nagel & H. Siegel 1999. Hydrog- raphisch-cheniische Zustandseinschatzung der Ostsee 1998. Meere- .swi.s.s. Ber. 35:69pp. Muravvski. S. A., J. W., Ropes & F. M. Serchuk. 1982. Growth of the ocean quahog. Arctica islandica. in the mid-Atlantic bight, f/.v/i. Ball. 80:21- 33. von Oenzen. i.-.\. 1972. Cycles and rates of reproduction of si,\ Baltic Sea bivalves of different zoogeographical origin. Mar Bid. 14:143- 149. von Oerlzen. l.-A. & G. Schlungbaum. 1972. Experimentell-okologische Untersuchungen iiber 0,-Mangel- und H,S-resistenz an marinen Ever- tebraten der westlichen Ostsee. Beitr. Meereskd. 29:79-91. von Oert/en. J. -A. & S. Schulz. 1973. Beitrag zur geographischen Ver- breitung und okologisclien Existenz von Bivalviem der Ostsee. Beitr. Meereskd. 32:75-88. Oeschger. R. & K. B. Storey. 1993. Impact of anoxia and hydrogen sul- phide on the metabolism of .4/rr/aj islandica L. (Bivalvia). J. exp. mar. Biol. Ecol. 170:213-226. Persson, L.-E. 1987. Baltic eutrophication: A contribution to the discus- sion. Ophelia 27:31—12. Prena, J., F., Gosselck. V., Schroeren & J. Voss. 1997. Periodic and epi- sodic benthos recruitment in southwest Mecklenburg Bay (western Bal- tic Sea). Helgol. Meeresunters. 51:1-21. Ropes. J. W. 1985. Modern methods used to age ocean bivalves. The Nautilus 99:53-57. Schulz. S. 1968. Riickgang des Benthos in der Luhecker Bucht. Monatsher. Dtsclt. Akad. Wiss. Berlin 10:748-754. Schulz. S. 1969a. Benthos und Sediment in der Mecklenburger Bucht. Beitr. Meereskd. 24(25):I5-55. Schulz. S. 1969b. Das Makrozoobenthos der siidlichen Beltsee (Mecklen- burger Bucht und angrenzende Seegebiete). Beitr. Meereskd. 26:21^6. Sleingrimsson, S. A. & G. Thorarinsdottir. 1995. Age structure, growth and size at sexual maturity in ocean quahog, Arctica islandica (Mollusca: Bivalvia). off NW-Iceland. ICES Council Meeting K 54:1-11. Theede. H., A. Ponat. K. Hiroki & C. Schlieper. 1969. Studies on the resistance of marine bottom invertebrates to oxygen-deficiency and hydrogen sulphide. Mar. Biol. 2:325-337. Thompson. I.. D. S.. Jones & D. Dreibelbis. 1980. Annual internal growth banding and life history of the ocean quahog Arctica islandica (Mol- lusca: Bivalvia). Mar. Biol. 57:25-34. Thorarinsdottir. G. & S. T. Einarsson. 1996. Distribution, abundance, population structure and meat yield of the ocean quahog. .Arctica is- landica. in Icelandic waters. / Mar. Biol. .4i.s. UK 76:1 107-1 1 14. Thorarinsdottir. G. & G. Johannesson. 1996. Shell length-meat weight relationships of ocean quahog. Arctica islandica (Linnaeus. 1767). from Icelandic waters. / Shellfish Res. 15:729-733. Weigelt, M. 1991. Short- and long-term changes in the henthic community of the deeper parts of Kiel Bay (Western Baltic) due to oxygen deple- tion and eutrophication. Meeresforsch 33:197-224. Witbaard. R. 1996. Growth variations in Arctica islandica L. (Mollusca): a reflection of hydrography-related food supply. ICES J Mar Sci 53: 981-987. Witbaard. R., G. C. A., Duineveld & P. A. W. J. de Wilde. 1999. Geo- graphical differences in growth rates of Arctica islandica (Mollusca: Bivalvia) from the North Sea and adjacent waters. J. Mar. Biol. Ass. t'A' 79:907-915. Jiiiinml of Shclljhli Research. Vol. 20, No. I, 171-17b, 2001. GROWTH, SURVIVORSHIP, AND NUTRIENT UPTAKE OF GIANT CLAMS {TRIDACNA) IN AQUACULTURE EFFLUENT MARIA SPARSIS.' - JUNDA LIN,' * AND RANDOLPH \V. HAGOOD" 'Dcpiirtiiu'iii (if Bidldglcal Sciences. Florida lustiiiile of Technology. Melbourne. Florida 32901: 'Proteus Consulting Inc.. Vera Beach. Florida ABSTRACT Experiments were conducted to evaluate the feasibility of using juvenile giant clams [Itutucna) to remove dissolved inorganic nutrients (nitrate and phosphate) from aquaculture effluent. Three {T. derasa Roding. T. niuyima Roding. and T. squamosa Lamarck) of the four species tested during a two-month experiment all had very high survivorship in both eftluent and control seawater, but T. gigas (Linne) experienced 50% mortality. T. maxima and T. squamosa showed virtually no growth in both effluent and control waters. T. gigas had the fastest growth rate among the four species. T. derasa grew significantly faster in effluent than in seawater. In a 24-h experiment (12 h dark followed by 12 h light), all the four species absorbed similar and significant amounts of nitrate and phosphate from effluent during the light period. Our study demonstrates that although it may not be possible to rely on giant clams alone to remove all excess dissolved nutrients from intensive food fish aquaculture effluent, it is entirely feasible lo use giant clams to remove all excess nutrients in effluent of marine ornamental species. Giant clams can be either incorporated into a polyculture system with other marine aquarium trade species, or they can be grown in effluent in separate tanks. KEY \yORDS: aquaculture effluent, giant clams, growth, juvenile, nutrient uptake, survivorship. TriJacna INTRODUCTION In recent years, environnientul impact and sustainable devel- opment have been central issues for aquaculture (e.g.. Chua 1992, Pillay 1992. Shpigel et al. 1993, Shpigel & Neori, 1996). Many solutions have been attempted to remove the particulate and/or dissolved nutrients in aquaculture effluent. One way of alleviating aquaculture's environmental impact is to utilize recirculating sys- tem technology, thus reducing the amount of effluent introduced into the environment and allowing facilities to be located inland. Water exchange is most commonly used to prevent nutrient build- up in a recirculating system. The obvious drawback to this tech- nique is the production of effluent, which must be disposed of under permitting restrictions. To completely eliminate effluent dis- charge, water from recirculating systems must be treated to remove particulate and dissolved waste before reintroduction to the culture system. One of the techniques employed to treat effluent is to utilize micro- and macro-algal primary production to take up dissolved nutrients (e.g. Subandar et al. 1993). Suspended solids are usually removed by mechanical filtration. Biofiltration by bivalve mol- lusks has also been used to remove particulate waste from aqua- cultural effluent. In China, there is a long tradition of using poly- culture and integrated culture to produce additional crops and re- duce negative impacts on the environment (Yang et al. 1992). Integration of bivalve and/or seaweed culture to remove suspended particles and/or dissolved nutrients from aquaculture effluent has also been shown to be effective (e.g. Jones & Iwama, 1991. Shpi- gel et al. 1993). Giant clams (family Tridacnidael are unique among bivalve mollusks in that they possess symbiotic algae {Syiiihiocliniimi ini- croariaticum) in their mantle tissue. In contrast to reef-building corals, the giant clams cultivate the symbiotic zooxanthellae in a special tubular system (Norton et al. 1992) which enables them to keep a substantially higher number of symbionts per unit area (Knop 1996). These algae contribute a variety of photosyntheti- *Corresponding author. E-mail: jlin&fit.edu; Fax: +1-407-674-7238. cally produced compounds to the host's metabolic needs (Lucas 1988). The single-cell thick tubes are bathed in haemolymph. al- lowing the efficient supply of nutrients to and photosynthate from the zooxanthellae (Belda & Yellowlees 1995). Klumpp et al. (1992) describe the tridacnids as "trophically opportunistic," ca- pable of deriving nutrition using autotrophic or heterotrophic means. Giant clams are endemic to the Indo-Pacific waters. The nine described species. Hippopus hippopus Linne, H. porcellanus Rose- water. Tridacna crocea Lamarck. T. derasa. T. gigas, T. maxima, T. rosewateri Sirenko and Scarlato. T. squamosa, and T. tevoroa Lucas, Ledua, and Braley, range from 10 cm to over 1 m in shell length ISL) as adults (Lucas 1988, Lucas et al. 1991, Knop 1996). They have formed a significant part of human diets in the Indo- Pacific and Southeast Asia for thousands of years. Clamshells have also been marketed. In recent years, giant clams have been suc- cessfully farmed for food, for restocking over-fished tropical reefs and, more recently, for the aquarium trade. The juvenile clams are popular in the aquarium trade due to their brightly colored mantle tissue, which is often exposed to allow light entry for algal pho- tosynthesis. Their retail price ranged from S2() to over $200 per animal in the US. The brownish background coloration of the mantle is caused by the zooxanthellae. but the iridescent pigmen- tation creating different patterns in a variety of colors is not due to the algae (Knop 1996). This bright pigmentation may protect the clams against UV light or too much light in general and confuse potential predators (Knop 1996). Tridacnids are ideal candidates for treating mariculture efflu- ent. They have been successfully reared in captivity and are fast growing, reaching the 8-10 cm SL aquarium market size in only 2-3 years. Because of their unique trophic style, they could be successfully incorporated into recirculating mariculture systems to remove both nutrient and particulate waste. In addition, tridacnids are popular as aquarium pets and in the sushi/sashimi markets and would therefore be a valuable by-product. We conducted a series of experiments to evaluate the potential of using the tridacnids for tertiary effluent treatment. The growth, survivorship, and nutrient uptake rate of four giant clam species growing in aquaculture effluent were compared. 171 172 Sparsis et al MATERIALS AND METHODS Growth and Survivorship In November 1996. juveniles of four giant clam species, T. derasa, T. gigas. T. maxima and T. squamosa, were obtained from a local aquarium dealer. The sizes were similar within a species, but considerably different among the species (Table 1), as con- strained by the availability. The clams were imported from the Solomon Islands where they were farm-raised. Upon arrival, the clams were acclimated to filtered seawater over a period of one hour. Clamshells were brushed briskly under running fresh water to remove any epiphytes and potential para- sites. The clams were then introduced to two indoor raceways (245 cm long and 30 cm wide) filled with 15-cm deep seawater (1 10 L). Two metal halide lamps. 60 cin above the raceways, were used to provide sufficient light intensity and appropriate wavelength com- position (Knop 1996). The light intensity maintained at 350—100 (iE/s/nr and photoperiod was set at 12 hours light and 12 hours dark. The water temperature was maintained at 25-29°C. salinity at 35 ppt. and pH at 8.3. within the range for optimal growth for the giant clams (Lucas 1988, Price & Fagolimul 1988, Knop 1996). After one week the clams were removed from the system, measured (SL) and tagged. The 48 clams were then randomly divided into the two raceways again (24 clams, six for each of the four species, in both the control and treatment raceways). For the following two months. 10'7f of the water in the raceways was replaced daily. In the control raceway, water was replaced with fresh filtered seawater; while the tieatment raceway received ef- fluent water from a fin fish aquaculture facility. Dissolved nutrient concentrations of the added effluent varied and were tested weekly, as were those of the filtered .seawater. control and treat- ment raceways. In the course of the two-month study period, biofouling became a problem, particularly in the treatment raceway. Two different species of parasitic snails were also found preying on the clams. Cymatium miiriciiim feeds on the mantle of clams and the parasitic pyramidellid snails feed on clams" lymphatic fluid (Knop 1996, Boglio & Lucas 1997). To mitigate these problems, the clams were periodically removed from the raceways and brushed under run- ning freshwater and returned to the raceways. Since most of the biofouling appeared to be algae. 25 herbivorous gastropods {Astrea tecta) were added to each raceway to control algal growth. Both techniques proved to be very effective. After two months of moni- toring, all the clams were again removed from the raceways and measured (SL). Dissolved Nutrient Uptake A separate experiment was conducted to compare the dissolved nutrient uptake rates among the four species in January 1997. All the clams used in the experiment had been acclimated to effluent water for at least two weeks prior to the study. Only healthy looking clams with quick shell closing responses were used. Each clam was immersed in a 500-ml beaker with l-|j.m fil- tered effluent water of known nutrient concentrations of dissolved nitrate (NO,) and phosphate (PO4). Five replicate clams were used for each species. Again, the sizes were similar within a species, but considerably different among the species. Mean (±s.d.) SL for T. derasa. T. gigas. T. maxima, and T. sqiianuisa are 60.2 (±2.2), 84.8 (±5.4), 32.4 (±1.7) and 36.6 (±2.3) mm, respectively. Two controls (each with a pair of empty giant clamshells) were set up. Nitrite and ammonia levels in the effluent were below detection limit (0.01 ppni) at the beginning and end of the experiment. The clams remained in darkness for 12 hours after which a sample of the effluent was removed from each beaker and tested for nitrate and phosphate concentrations. The clams remained immersed for an additional 12 hours under light with intensity of 350-400 [iE/s/m". Samples of the effluent were again removed and tested during the 1 2-hour light period (every two hours for nitrate, at the end of the 12-hour light period for phosphate). The water temperature was mauitamed at 2-29°C. salinity at 35 ppt. and pH at 8.3. Measurement of dissolved nutrient (ammonia, nitrite, nitrate, and phosphate) concentrations followed the techniques described in Menzel and Corwin (1965). D'Elia et al. (1975), and Greenberg et al. (1985). Ammonia was determined using the direct nessler- ization method, nitrate the cadmium reduction method, nitrite the azo-dye formation technique, phosphate the persulfate oxidation method and the resulting orthophosphate be measured using the ascorbic acid method. Color development in all cases was mea- sured using a spectrophotometer. RESULTS Growth and Survivorship All the indi\iduals of T. maxima and T. squamosa survived. One T. derasa in the treatment tank and three (50%) T. gigas in both treatment and control tanks died. T. derasa and T. gigas displayed some growth during the study, whereas T. maxima and T. squamo.ui showed virtually no growth (Table 1 ). Only T. derasa showed significant difference in growth (P = 0.008. /-test) be- tween the treatment and control tanks (Table 1 ). Concentrations of dissolved nutrients in the seawater and con- trol tanks remained low (Fig. 1). Except for the first two weeks, concentrations of ammonia and nitrite in the effluent and treatment tanks were also low (Fig. 1). probably due to the nitrification process carried out by the biofilter in the fin fish culture facility. The concentrations of nitrate (Fig. Ic) and phosphate (Fig. Id) in the treatment tanks were lower than those in the effluent tanks, suggesting that the clams were removing the nutrients throughout the two-month period. TABLE 1. Mean (±s.d.) initial SL and SL increase of the four clam species after the iwo months of immersion in control (C) and treatment (T) raceways. T. derasa T. gigas T. na.xima T. s quamosa C(n = 6) T(n = 5) C (n = 3) T (n = 3) C (n = 6) T (n = 5) 32.3 ± 1 .9 0.0 + 0.0 C In = 6) 35.6 ± 2.y 0.0 ± 0.0 T (n = 6) Initial SL (mm) SL Increase (mm) 53.8+ 11.0 0.2 ± 0.4 59.2 ± 1.6 1.8+1.1 86.0 + 6.1 1.3 ±2.3 86.7 ± 7.5 2.7 + 2.1 33.2 ± 7.5 0.2 ± 0.4 36.0 + 3.0 0.2 ± 0.4 Growing Giant Clams in Effluent 173 6"^s-0 — o Time (week) Figure 1. Conctntrations of ammonia (a), nitrite (b), nitrate (c). and phosphate (d) in filtered seawater (S), effluent (K), treatment water (T), and control water (Cl over the two-month study period. Dissolved Sutriinl Uptake The concentrations of nitrate and phospliate did not change in the two controls over the 24-h period and in the treatments over the 12-h dark period (Fig. 2). All four species showed rapid uptake of both nitrate and phosphate during the 12-h light period (Fig. 2). 0\er the 24-h period, the nitrate concentration decreased from 12.6 ppm to an average (±s.d.) of 9.0 (±0.3) ppm for T. derasa. 8.0 (±0.2) ppm for T. gigas. 10.0 (±0.2) ppm for T. niaxinia. and 9.9 (±0.7) for T. squamosa (Fig. 2). Taking clam size into consider- ation, all four species had similar rates of nitrate uptake. The phosphate concentration decreased from 0.23 ppm to an average (±s.d.) of 0.15 (±0.01) ppm for T. derasa, 0.14 (±0.01) ppm for T. gigas. 0.14 (±0.00) ppm for T. maxima, and 0.18 (±0.00) for T. squamosa (Fig. 2). Phosphate uptake rates were also proportional to clam size, except that T. maxima seems to be more efficient. DISCUSSION Klumpp et al. (1992) describe the tridacnids as "trophically opportunistic," capable of deriving nutrition using autotrophic and/ or heterotrophic means. The ability of the giant clams to assimilate inorganic nutrients via photosynthesis has been well-documented (Wilkerson & Trench 1986. Fitt 1988, Mingoa 1988, Solis et al. 1988. Braley et al. 1992. Haslie et al. 1992. Klumpp et al. 1992, Belda et al. 1993b, Fitt et al. 1993, Klumpp & Griffiths 1994, Klumpp & Lucas 1994. Hawkins & Klumpp 1995). Autotrophy under high light intensities may be able to satisfy the basic nutri- tional needs of the clams (Griffiths & Streamer 1988, Klumpp et al. 1992, Klumpp & Griffiths 1994, Klumpp & Lucas 1994). Increased growth in the giant clams with increasing dissolved nitrogen concentrations was reported in several studies (Wilkerson & Trench 1986, Solis et al. 1988, Braley et al. 1992, Hastie et al. 14 12 10 - 8 6 4 2 0 Isquamosa 12 14 16 18 20 22 24 0 12 24 Time (hr.) Fij-ure 2. Mean nitrate and phosphate concentrations (ppm) over the 24-h period (12 h dark followed by 12 h lightl for 7'. derasa (mean ± s.d. SL= 6((.2 ± 2.2 mm). T. gigas (mean ± s.d. SI. = 84.5 ± 5.4 mm), T. maxima (mean ± s.d. SI. = .^2.4 ± 1.7 mm), and /'. squamosa (mean ± s.d. SL = 36.6 ± 2.3 mm). 174 Sparsis et al 1<-W2. Fitt et al. 1993). In a 60-day nutrient enhancement study with T. denisa (Hastie et al. 1992). the clams grown in elevated dissolved inorganic nitrogen had both higher SL and tissue weight increments over untreated controls. Ammonium and nitrate were equally effective and no detrimental physiological effects of en- richment were detected. The clams (initial average size; 27.6 mm SL) in the elevated nitrate treatment grew an average of 0.1-0.16 mm SL/day, as compaied to 0.06 mm SL/day for the control sea- water (Hastie et al. 1992). In our two-month study, the much larger sized (average initial SL were 53.8 and 59.2 mm for control and effluent tanks, respectively) T. derasa grew an average of 0.03 mm SL/day in the effluent, but had negligible growth (0.004 mm SL/ day) in the control seawater. A large-scale grow-out study at 1 1 village farms in Solomon Islands (Hart et al. 1998) found that among the juveniles (20-30 min SL) of giant clam species tested. T. derasa had the best growth (0.14 to 0.21 mm SL/day) and survival (80.1 to 99.1%), and therefore the highest estimated rev- enue for the aquarium trade, than those of the other two species tested.- T. maxima (growth: 0.07-0.14 mm SL/day, survivorship: 1 6.0-72. 7Sf) and T. crocea (growth: 0.04-0.07 mm SL/day, sur- vivorship: 1 1.8-81.9%). In a similar 8-nionth grow-out study on T. squamosa in the Solomon Islands (Foyle et al. 1997), the clams (initial size about 24 mm SL) grew an average of 0.08-0.29 mm SL/day with the survivorship ranged from 7-83%-. Growth rates of juvenile T. squamosa found in other studies ranged from 0.05-0.23 mm SL/day (Klumpp & Griffiths 1994. Foyle et al. 1997). T. gigas has the highest growth rate (0.13-0.37 mm SL/day) among the giant clam species tested in our experiment and other studies (see Table 7 of Klumpp & Griffiths 1994 for a review summary). The average daily growth rates of T. gigas in the control seawater (0.02 mm SL) and in effluent (0.05 mm SL/day) in our study are much lower than those of the other studies. Interestingly, while effluent greatly enhanced the growth of T. derasa and probably T. gigas. it did not have an effect on the other two species: T. maxima and T. squamosa. In fact, these two species showed virtually no growth during the two-month period, in effluent as well as in control .seawater (Table 1 ). Giant clams are hardy animals and have excellent survivorship when grown in effluent in a recirculating system. Among the four species tested in both seawater and effluent in our two-month study, T. gigas was the only species that suffered significant (50%) mortality. However, available literature does not indicate that T. gigas is less hardy than the other species. Ranellids (e.g.. several species of Cymatium) and pyramidellid snails are common preda- tors/parasites of tridacnid clams in the wild and can cause serious mortality (Knop 1996. Boglio & Lucas 1997). The pyramidellid snails are obligate parasites with various degrees of host specific- ity. They use their proboscis to penetrate the hosts and gain their nutrition by feeding purely on the hosts" body tluid (Boglio and Lucas, 1997). The snails are small (a few mm) nocturnal feeders and can be difficult to detect and remove. Therefore, it is critical to carefully examine imported clains for parasites. The four species tested in our study showed similar capabilities of absorbing both nitrate and phosphate, after taking clam size into consideration. The nitrate uptake rates found in our study are simi- lar to those found in the studies of Juvenile T. derasa grown on seawater spiked with artificial nutrients (Hastie et al. 1992, Fitt et al. 1993). Fitt et al. (1993) also found that the rate of NO, uptake was lower than that of NH,. In fact, uptake of NO, was repressed in the presence of NH,. In the present study, low NH, concentra- tion in the effluent may have facilitated the uptake rate of nitrate by the clams. In the study of Fitt et al. ( 1993). no detectable uptake of ammonia and much lower uptake rate of nitrate in the dark as compared to under light were found. Among the four clam species we tested, virtually no uptake of nitrate or phosphate took place during the 12-h dark period (Fig. 2). There has been limited study of phosphate uptake by giant clams. Large juvenile T. gigas (22-26 cm in SL) can take up significant amount of phosphate, with the individuals previously held in control seawater deplete phosphate at a much greater rate (11.5 p-M P/hr.) than the ones previously held in phosphate- supplemented (10 |jlM) raceways (4.3 p.M P/h) (Belda & Yellow- lees 1995). The four species we tested, especially T. maxima, can rapidly absorb phosphate from seawater in the light period (Fig. 2). They are also capable of continuously removing phosphate for an extended period of time (Fig. Id). Zooxanthellae do not have ready access to inorganic phosphate in seawater and within the clam (Belda et al. 1993b. Belda & Yellowlees 1995). The clam host may actively regulate P availability to its symbiotic algae (Belda et al. 1993b). Juveniles of T. derasa exposed to elevated phosphate (2 (jlM) alone actually grew less than control clams and phosphate additions to elevated dissolved inorganic nitrogen treatments had little apparent effect on the clam's growth rates (Fitt et al. 1993). High concentrations of phosphate inhibit calcification (Simkiss 1964) and may have a negative effect on the growth of giant clams. Further studies are needed to examine the effects of dissolved phosphate concentration on clam's growth if aquaculture effluent is to be used for culturing giant clams. Elevated nutrients of inorganic nitrogen and phosphorus, how- ever, may lead to weakening shell structure in the giant clams, resulting in thinner and more transparent shells (Belda et al. 1993a). Ammonium, as well as phosphate, may inhibit or depress calcification in clams (Simkiss 1964, Belda et al. 1993b). These clams may be more vulnerable to predation if released to the wild, but would not have any negative impacts when used in the aquarium trade. Enhanced nutrients also increase biofouling in the clams, but this can be removed by gentle scrubbing with a tooth- brush (Hastie et al. 1992: present study) or by introducing appro- priate grazers, such as herbivorous snails (present study). The relative importance of autotrophy in giant clam nutrition has been the subject of speculation for many years (Trench et al. 1981. Fisher et al. 1985). The relative contributions of phototrophy and heterotrophy toward the carbon requirements vary among the tridacnid species and among different sized clams within a species (Klumpp et al. 1992. Klumpp & Griffiths 19'-)4. Klumpp & Lucas 1994). T. gigas was the most efficient in uptake of C via both photosynthesis and filter feeding among the four species compared {T. gigas. T. crocea. T. squamosa, and Hippopiis luppopus). The interspecific differences, however, declined with clam size (Klumpp & Griffiths 1994). Filter feeding is able to provide over half of the total carbon needed both for respiration and growth in small individuals of T. gigas (Klumpp et al. 1992). However, filter feeding is a relatively ininor source of energy in other species (T. crocea. T. derasa. T. squamosa, T. tevoroa. and Hippopus hippo- pus) and its importance decreased with increasing clam size (Klumpp & Griffiths 1994, Klumpp & Lucas 1994). In the present study, the degree of nutritional contribution to the clams by small particles in the effluent is unknown. Among the four species we tested, T. derasa is probably the best candidate for culturing in aquaculture effluent. In addition to growth, survivorship, and nutrient uptake parameters, seed avail- ability and ease of cleaning to remove epiphytes and parasites also Growing Giant Clams in Effluent 175 need to be tiikeii into consideration. All the four species are equally susceptible to epiphyte infestation, but those with smooth shells, such as T. f;igcis and 7". deiasa. are easier to clean by scrubbing. The highly fluted shells of T. squaiuosa and T. nia.xinui are more likely to harbor parasites and other epifauna. which are harder to detect and remove due to the shell's high relief. Seed of T. gigas is becoming difficult to obtain as most hatcheries are concentrating on the more colorful species. All the four species are popular in the aquarium trade, with the more colorful 7'. lua.xima commanding higher prices. It may not be possible to rely on giant clams alone to remove all the dissolved nutrients from intensive food fish aquaculture effluent at a commercial scale. However, it is entirely feasible to use giant clams to treat aquaculture effluent of other marine orna- mental species. Most of more than 3,000 species of marine fishes and invertebrates marketed in the aquarium trade industry in the world are collected from coral reef systems. Concerns over the impacts of wild harvest, with expanding popularity of coral reef animals in the aquarium trade, have spurred interest in developing or improving cultivation technology for marine ornamental fish (Fletcher et al. 1999, Hoh 1999), shrimp (see Lin et al. 1999 for a review), corals (Carlson 1999), giant clams, and other inverte- brates. Either giant clams can be incorporated into the other aquarium species in a polyculture system, if the environmental requirements are compatible between (among) the species; or they can be grown in effluent in separate tanks. The clams can be grown not only in tropical and subtropical regions, but also in greenhouses in tem- perate areas, as currently practiced by the aquarium industry. In colder areas and/or months, the effluent may be diverted to an indoor greenhouse to be treated by the giant clams (Braley et al. 1992). Growing juvenile giant clams in recirculated water would also pave the way for inland culture (Braley et al. 1992), thereby eliminating the possible negative consequence of introducing ex- otic species. ACKNOWLEDGMENTS Financial support was jointly provided by the Sea Grant (Grant number: NA36RG-0070) and the Atlantic Aquaculture Technolo- gies, Inc. Butch Almberg of Reef Rascal Aquarium provided clams and advice. Dick Peirin of Tropicorium provided valuable techni- cal advice. Mary Clark and LeRoy Creswell of Harbor Branch Oceanographic Institution shared their clam references. The late Dr. Kerry Clark of Florida Institute of Technology helped in for- mulating the initial concept of the project, Steve Ameson, Jeremy Baker, Morris Bevis, Amy Riedle. Tiffany Williamson, and Harry Yamalis provide various assistance dtiring the study. Dong Zhang produced the figures. Helpftil comments were provided by anony- mous reviewers. Belda, C. A., C. Cuff & D. Yellowlees. 1993a. Modification ol shell formation in the giant clam Tridacna gigas at elevated nutrienl levels in sea water. Marine Biology 117:251-257. Belda, C. A., J. S. Lucas & D. Yellowlees. 1993b. Nutrienl Iniiitation m the giant clam-zooxanthelae symbiosis: effects of nutrient supplements on growth of the symbiotic partners. Marine Biology I 17:655-664. Belda. C. A. & D. Yellowlees. 1995. Phosphate acquisition in the giant clam-zooxanthelae symbiosis. Marine Biology 124:261-266. Boglio, E. C. & J. S. Lucas. 1997. Impacts of ectoparasitic gastropods on growth, survival, and physiology of juvenile giant clams {Tridacna gigas). including a simulation model of mortality and reduced growth rate. Aquaculture 150:1-2, 25— 13. Braley, R. D., D. Sutton, S. M. Mingoa & P. S. Southgate. 1992. Passive greenhouse heating, recirculation, and nutrient addition for nursery phase Tridacna gigas: growth boost during winter months, Aquaculture 108:29-50. Carlson, B. A. 1999. Coral culture as an alternative to harvesting from the reef. Book of Abstracts, Sydney: World Aquaculture '99, April 26- May 2, 1999, p. 132. Chua, T.E. 1992. Coastal aquaculture development and the environment: The role of coastal area management. Marine Pollution Bulletin 25: 98-10. D'Elia. C. F., P. A. Streudler & N. Corwin. 1975. Determinalion of total nitrogen in aqueous samples using persulphate digestion. Lininol. Oceatiogr. 22:760-764. Fisher. C. R.. W. K. Fitt & R. K. Trench. 1985. Photosynthesis and res- piration in Tridacna gigas as a function of irradiance and size. Biol. Bidl. 169:230-245. Fitt, W. H. 1988. Role of zooxanthellae in the mariculture of giant clams. In: J. W. Copland & J. S. Lucas, editors. Giant clams in Asia and the Pacific. Canberra: Australian Centre for International Agricultural Re- search, pp 166-167. Fitt, W. K., G. A. Heslinga & T. C. Watson. 1993. Utilization of dissolved inorganic nutrients in growth and mariculture of the tridacnid clam Tridacna derasa. Aquaculture 109:27-38. Fletcher, D. J., E. Wood & D. A. Jones. 1999. Manne ornamental culture LITERATURE CITED and reef conservation. Book of Abstracts. Sydney: World Aquaculture ■99. April 26-May 2, 1999, p.265. Foyle. T. P.. J. D. Bell. M. Gervis & I. Lane 1. 1 997. Survival and growth of juvenile tluted clams, Tridacna squamosa, in large-scale grow-out trials in the Solomon Islands. Aquaculture. 148:2-3, 85-104. Greenberg. A. E.. R. R. Trussell & L. S. Clesceri. 1985. Standard methods for the examination of water and wastewater. 1 6th ed. Washington D. C: American Public Health Association. Griffiths, D. J. & M. Streamer. 1988. Contribution of zooxanthellae to their giant clam host. In: J. W. Copeland & J. S. Lucas, editors. Giant Clams in Asia and the Pacific. Canberra: Australian Centre for International Agriculture Research, pp. 151-154. Hart, A. M., J. D. Bell & T P. Foyle. 1998. Growth and survival of the giant clams, Tridacna derasa. T. maxima and T. crocea. at village farms in the Solomon Islands. Aquiuulturc. 165:203-220. Hastie, L. C. T. C. Watson, T Isamu & G. A. Heslinga. 1992. Effect of nutrient enrichment on Tridacna derasa seed: dissolved inorganic ni- trogen increases growth rate. Aquaculture 106:41—49. Hawkins, A. J. S. & D. W. Klumpp. 1995. Nutrition of the giant clam Tridacna gigas (L.). II. Relative contributions of filter-feeding and the ammonium-nitrogen acquired and recycled symbiotic alga towards to- tal nitrogen requirements for tissue growth and metabolism. J. E.xp. Mar. Biol. Ecol. 190: 263-290. Holt, G. J. 1999. Marine ornamental fish: is larval culture the bottleneck to production? Book of Abstracts. World Aquaculture '99. Sydney. -Aus- tralia, Aprd 26 to May 2, 1999. 337 pp. Jones, T. O. & G. K. Iwama. 1991. Polyculture of the Pacific oyster, Crassoslrea gigas (Thunberg), with chinook salmon, Oncorhynchus tshawylscha. Aquaculture 92:3 l.V ,122. Klumpp, D. W., B. L. Bayne & A. J. S. Hawkms. 1992. Nutrition of the giant clam. Tridacna gigas (L). 1. Contribution of filter feeding and photosynthates to respiration and growth. ./. £v;). Mar. Biol Ecol. 155:105-122. Klumpp. D. W. & C. L. Griffiths. 1994. Contributions of phototrophic and heterotrophic nutrition to the metabolic and growth requirements of 176 Sparsis et al four species of giiint clam (Tridacnidae). Mm: Ecoi Pro. Ser. 115: 103-115. Klumpp. D. W. & J. S. Lucas. 1994. Nutritional ecology of the giant clams Trutacim tevoroa and T. cienisa from Tonga: influence of light on filter-feeding and photosynthesis. Mar. Ecol. Prog. Ser 107:147-156. Knop, D. 1996. Giant Clams: a comprehensive guide to the identification and care of tridacnid clams. Ettlingen, Germany: Dahne Verlag. 255 Lin. J.. D. Zhang & R. L. Creswell. 1999. Marine ornamental shrimp: status and prospects. Aquaculture Magazine. May/June:52-55. Lucas, J. S. 1988. Giant Clams: Description, distribution and life history. In: J. W. Copland & J. S. Lucas, editors. Giant clams in Asia and the Pacific. Canberra: Australian Centre for International Agricultural Re- search, pp. 21-32. Lucas, J. S„ E. Ledua & R. D. Braley. 1991. Tridacna tevoroa: a recently- described species of giant clam (Bivalvia; Tridacnidae) from Fiji and Tonga. Nautilus 105:92-103. Menzel. D. W. & N. Corwin. 1965. The measurement of total phosphorous in seawater based on the liberation of organically bound fractions by persulfate oxidation. Limnol. Occ>anof>r. 10:280-282. Mingoa, S. S. M. 1988. Contribution of zooxanthellae to their giant clam host. In: J. W. Copland & J. S, Lucas, editors. Giant Clams in Asia and the Pacific. Canberra: Australian Centre for International Agriculture Research, pp. 145-150. Norton, J. H„ M. A. Shephard. H. M, Long & W. K. Fitt. 1992. The zooxanthellal tubular system in the giant clam. Bml. Bull. 183:503- 506. Pillay, T. V. R. 1992. Aquaculture and the en\ironment. New York, To- ronto: Halsted Press, John Wiley and Sons, Inc. Price, C. M. & J. O. Fagolimul. 1988. Reintroduction of giant clams to Yap stale. Federal States of Micronesia. In: J. W. Copland & J. S. Lucas, editors. Giant Clams in Asia and the Pacific. Canberra: Australian Centre for International Agricuhural Research, pp. 41—13. Shpigel, M. & A. Neori. 1996. The integrated cuUure of seaweed, abalone, fish, and clams in modular, intensive land-based systems: I. Proportions of size and project revenues. Aquaculture Engineering 15:313-326. Shpigel, M.. A, Neori, D. M. Popper & H. Gordin, 1993. A proposed model for "environmentally clean" land-based culture of fish, bivalves and seaweeds. Aquaculture 1 17:1 15-128. Simkiss, K. 1964. Phosphates as crystal poisons of calcification. Biol. Rev. 39:487-505. Soils, E. P., J. A. Onate & M. R. A. Naguit. 1988. Growth of laboratory- reared giant clams under natural and laboratory conditions. In: J. W. Copland & J. S. Lucas, editors. Giant clams in Asia and the Pacific. Canberra : Australian Centre for International Agricultural Research, pp. 201-206. Subandar, A., R, J. Petrell & P, J. Harrison. 1993. Luminaria culture for reduction of dissolved inorganic nitrogen in salmon farm effluent. Journal of Applied Phycology 5:455-463. Trench, R. K., D. S. Wethey & J. W. Perter. 1981. Observations on the symbiosis with zooxanthellae among the Tridacnidae (Mollusca, Bi- valvia). Biol. Bull. 161:180-198. Wilkerson, F. P. & R, K, Trench. 1986. Uptake of dissolved inorganic nitrogen by the symbiotic clam Tridacna gigas and the coral Acropora sp. Marine Biology 93:237-246, Yang, H., Y'. Fang & Z. Chen. 1992. Integrated fish farming systems in China, and the allocation of resources. World Aquaculture Magazine, March 1992. pp. 61-68. Jaiinwl of Shellfish Research. Vol. 20. No. I. 177-1S4. 2001. REPRODUCTIVE CYCLE AND BIOCHEMICAL COMPOSITION OF THE ARK SHELL SCAPHARCA BROUGHTONII (SCHRENCK) IN A SOUTHERN COASTAL BAY OF KOREA MI SEON PARK,* CHANG-KEUN KANG, AND PII -YONG LEE National Fisheries Research & Development Institute. Sliirant^ Kijang-Gun, 619-WU Piisan. Republic of Korea ABSTRACT The reproductive cycle and biochemical composition of the ark shell Scapharca hroughwnii (Schrenck) in Chinju Bay, a southern coastal bay of Korea, were studied from July 1999 to August 2000 in relation to environmental conditions (temperature, salinity and available food). Animals were cultured in the bottom. S. hwughloiui from Chinju Bay was characterized by a spawning period during July to September due to a monocyclic gametogenesis throughout the year. Clear seasonal cycles in biochemical composition of the animal tissues were recorded. Summer spawning of S. hroughtonii was followed by a clear accumulation period of reserve materials in autumn (from October to December). A great increase in animal tissue weight was attributed to the rapid accumulation of protein and carbohydrate during that penod. A new cycle of gametogenesis began at the lowest water temperature in Febmary. The gonadal growth during winter to spring when the available food was poor took place at the expense of reserves (mainly glycogen) accumulated previously in autumn. Lipid biosynthesis during gametogenesis was associated with glycogen breakdown. Blooms of phytoplankton in the ambient waters took place in the low saline summer period when water temperature was highest, depending on the addition of excessive nutrients via streams. Spawning of S. broughtonii occurred at the time when the available food was very abundant. This seemed to enable larvae to be produced at the most opportune moment with regard to food availability, as previously proposed. Our results indicate that this bivalve may be considered a typical conservative species in gametogenic pattern. KEY WORDS: Sccipharca broughumii. reproductive cycle, bivalves, biochemical composition, seston. food index INTRODUCTION The southern coastal bay systems in Korea have been widely exploited for shellfish cultures. In the shallow semi-enclosed coastal bays, uncontrolled increases in stock density often result in reduced growth rates and environmental disruption thereby caus- ing mass mortality of cultured bivalves (Park el al. 1998. Park et al. 1999). The ark shell Scapharca broughtonii (Schrenck) is cul- tured in the bottom of the inshore bay systems. Thus, this shell, which is a common benthic suspension feeder from intertidal to subtidal zone along the northwestern Pacific coast, is one of the most important commercial bivalve species in Korea. Morphologi- cal and genetic differences between the ark shell populations of Korea. China and Japan have been found (Lee et al. 1997. Yokogawa 1997). Annual production of the ark shell S. brough- tonii in Korea was maximized in 38.000 tonnes in 1986. However, since the 1990s production has decreased, depending largely on local shortages of healthy seed (Rho & Pyen 1977. Park et al. 1998). Knowledge of reproductive strategy in a given area could have essential importance for collecting ark shell seed from nature. Up to now. biological studies on the ark shell have been limited to natural and artificial seed collection and growth of spat (Kim & Koo 1973. Yoo & Yoo 1974, Pyon et al. 1976, Rho & Pyen 1977. Kim et al. 1979. Kim & Yoon 1980). Little attention has been paid to the physiological ecology and gametogenic cycle (Park et al. 1998). The reproductive cycles of marine bivalves are strongly related to energy storage-utilization cycles and environmental pa- rameters such as water temperature and food availability (Giese 1969. Gabbott 1975. 1983. Bayne 1976). Therefore, seasonal cycle and frequency of the reproductive activity in marine bivalves ex- hibit geographical variations within a single species (Bayne & Worrall 1980. Newell et al. 1982, Rodhouse et al. 1984. Bricelj et al. 1987). Gametogenesis is an energy-demanding process. The use of energy required for gametogenesis differs between species. Ac- cording to the classification of Bayne (1976). one group, called opportunistic species, uses the recently ingested energy from seston and another group, conservative species, uses the energy stored in various organs through feeding prior to its gametogen- esis. The seasonal cycles of gametogenesis and reserve storage in S. broughtonii may reveal its reproductive strategy in relation to local environinental conditions. This study presents the reproductive cycle and biochemical composition of 5. broughtonii cultured in the bottom of Chinju Bay. a southern coastal bay system of Korea. The objective of the study was to observe the seasonal cycles of energy storage- consumption in the ark shell, to examine the use of reserve mate- rials in relation to the gametogenic cycle and the environmental parameters, and to establish the gametogenic pattern. MATERIALS AND METHODS Stiidv Site *Corresponding author. E-mail:parkms(9nfrdi.re.kr; Fax: +82-51-720- 2498. Chinju Bay is a small shallow bay system (average depth 4 m) situated on the southern coast of Korea (Fig. I ). It is well protected by Namhae and Changsun Islands to the south. The bay is about 25 km long and 13 km wide. It is a well-mixed bay and the tide is semidiurnal with a maximum tidal range of 3.6 m and strong tidal currents. Freshwater pulses occur irregulariy from an artificial dam of the Nam River on the north of the bay. Sediments of the bay are composed mostly of mud. A broad area (about 8 km^) of its sub- tidal zone has been developed for bottom culture of the ark shell S. broughtonii. The sampling station for the study was located in the southwestern part of the bay. It was 3 m deep at the low tide and 6 m at the high tide. Sample Collection and Preparation for Analysis Fifty to 100 specimens of S. broughtonii were collected ran- domly by diving at monthly intervals from June 1999 to August 177 178 Park et al 34 58 N 34 50 128 00 E 128 20 Figure 1. Location of the site for arlv shell collection and environmental measurements in C'hinju Bay, Korea. 2000. Mean shell length was 66.5 (±6,2) mm for one-year class. These animals were seeded on the bottom in April 1998. All ani- mals were carried to the laboratory quickly and kept alive over- night to evacuate gut content. After biometric measurements (shell length, width, height, and total weight) they were dissected care- fully. The flesh was weighed for wet tissue weight and freeze-dried for the determination of dry tissue weight and biochemical com- position. Shell valves were rinsed with distilled water and weighed after drying in a furnace at 50°C for 48 h. Temperature and salinity of the surface and bottom waters were recorded at the time of sampling by a CTD meter (Seabird Electronics, Inc.). We first investigated the characteristics and the seasonal variability of .seston of the bay. Waters for seston measurements were pumped for the surface and bottom waters at the same time and screened through a 180-|jim Nitex mesh to eliminate zooplankton and large debris. Water samples for suspended particulate matter (SPM) were immediately filtered through a preweighed Whatniann GF/C glass-fiber filler. For pigment and chemical analysis. 0.5 to 3-L aliquots were concentrated on GF/F filters on board in situ and the filters were carried to the laboiatory. All filters were then kept frozen (-80 C) until analyzed. Seston Analysis and Food Index Filters for dry weight were dried at 80°C for 24 h and re- weighed after cooling in desiccator. Chlorophyll a concentration was determined on acetone extracts using fluorometric method as modified by Parsons et al. (1984) with 10 AU Fluorometer (Turner Designs). Biochemical components of seston (particulate protein, carbohydrate and lipid) were analyzed by the same methods as described later for animal tissues. An evaluation of the nutritional value of the seston throughout the year was done using the bio- chemical composition. Food quantity was defined as the sum of the concentrations of these components and a food index was calcu- lated as the percentage of food in the seston [{Food/Total Seston) xlOO] (Widdows et al. 1979). Condition Index and Reproductive Activity Temporal variation of condition index reveals either the onset of accumulation of organic matter for reproduction or release of gametes. The condition of the ark shells was estimated from an index that was expressed as dry tissue weight:dry shell weight (Walne 1976. Mann & Glomb 1978). Thirty individuals from each sample were used for microscopic examination of histological smears. A transverse cut was made across the body of the ark shell and a section ,3-mm thick was fixed in Bouin's solution. It was then routinely processed for histology and 5-p.m paraffin-imbedded sections were stained with iron hematoxylin-eosin (Humason 1979). The stage of gonadal development was classified and scored on a 0 to 4 scale according to a scale of maturity (Ivell 1979). At grade 1, sexes are just identifiable but very little gonad is de\eloped. Grade 2 represents a stage of moderate gonadal development. At this grade, oocytes are beginning to develop. At grade 3, gonads are distended and ripe. At this stage, gametogen- esis is complete but spawning has not yet occurred. Oocytes are liberated freely on dissection and free sperm are manifested as streaks. Grade 4 is the spawning stage of animals and in half-spent to completely spent condition. Grade 0 represents both the com- pletely spent stage and the immature stage. At this grade, sexes are unidentifiable. The arithmetic means of the individual scores of the whole sample was recorded as the Gonadal Maturity Index (GMI) for each sampling date (see details in Duiamani 1987). Biochemical Measurements of Animal Tissue The dry tissue of 20 individuals was pooled and homogenized to determine average biochemical composition. Protein was deter- mined by the colorimetric method of Lowry et al. (1951) after extraction with normal sodium hydroxyde. Carbohydrate and gly- cogen were extracted in 15% trichloroacetic acid, and precipitated with 99% ethanol. They were analyzed using the phenol-sulfuric acid method as described bv Dubois et al. (1956). Extraction for Reproduction and Composition of Scapharca broughtonii 179 total lipid was performed in a mixture of chloroform and methanol (Bligh & Dyer 1959) and lipid content was determined using the method of Marsh and Weinstein ( 1966). Ash content was obtained by igniting a subsample (30 to 80 mg) of homogenized tissue at 45()'C for 4S h in a muftle furnace. Stamkiid Aiiimul To e\aluate the physiological state of 5. hriiKglnnnii indepen- dent of growth, absolute values for tissue weight and biochemical composition of a standard animal of 66.5 mm in shell length were compared for each sampling date. Allometric equations of log,,, dry tissue weight against log,,, shell length at each sampling date were determined by linear regression analysis. The results of the biochemical analysis were then expressed in milligrams per stan- dard animal. All regressions were statistically significant (p < 0.01 ). A similar method has been used in the studies for the bar- nacles, Balanus balanoides and B. haliuis (Barnes et al. 1963), and for the tellin, Tetlina tenuis (Ansell & Trevallion 1967). RESULTS Environmenlal Parumelers Water temperature increased from March to August and de- creased from September to February (Fig. 2). The maximum water temperature was in August 2000 (25.8°C) and the minimum in February 2000 (5.3°C). Salinity was generally variable and lower than 30 psu in summer (June to September), and constant and higher than 30 psu in the other period. The maximum salinity was recorded from January to June 2000 (around 33.6 psu) and the minimum in September 1999 (26.7 psu). Available Food Seasonal variations in the concentration and composition of SPM in the water column are presented in Figure 3. SPM concen- tration recorded several peaks throughout the year (Fig. 3a). The concentration of chloriiphyll a exhibited a clear seasonal pattern characterized by a marked peak in summer when salinity was very low (Fig. 3b). The peak in 1999 was higher (7.5 |j.g ■ P') than in 2000 (around 4.0 |jLg • I"'), but the former was recorded in Sep- tember and the latter was during the June to August period. During the rest of the year, the chlorophyll a concentration was relatively low. The food material present in SPM, as represented by the sum of protein, carbohydrate and lipid concentrations, showed a marked summer peak (over 400 pig ■ L"' ) during the phytoplank- ton bloom of each year, although another peak was recorded in February 2000 (Fig. 3cj. Peak values of the food index, calculated Figure 2. salinity ( J A S O N D J F 1999 2000 Month Monthly variations of water temperature (black circles) and while circles) in Chinju Bay. Figure i. Monthly variations of suspended particulate matter (SPM) (a). chloroph\ll a (Chi a) (b), food material (a sum of particulate protein, carbohydrate and lipid) (c) and food index [(Food material/ SPM) X 100] (d), as a percentage of the SPM, were related to the phytoplankton bloom (Fig. 3d). Peak value of ca. 4% was also recorded in winter months December 1999 to February 2000. During the last of the year, the food index decreased to values lower than 2%. Condition and Reproductive Cycle The reproductive cycle of S. broyglitnii is characterized by a clear seasonal pattern, as summarized in Figure 4. Maximum val- ues in the condition index were recorded in the early summer months of July 1999 and June 2000. respectively (Fig. 4a). The condition index then decreased sharply to the minimum in late summer of September 1999, suggesting a summer spawning. This index increased progressively from October to June. Since the gametes exhibited the same developmental stage in both feiuale and male follicles, and differences between sexes were not con- sidered in the analysis of biochemical composition, GMl is pre- sented as the means of pooled data from both sexes (Fig, 4b). The developirtent of gonadal tissue started in Februar\ and peaked during July to August when the condition index dropped suddenly. Ripe gonads were found from June to August (Fig, 4c), According to the dates of appearance and disappearance of grade 4, which is used as criteria for the beginning and end of spawning, spawning began in July and terminated in September. After spav\ning, there was a sexual resting period from October to January. Tissue Weight Seasonal change in tissue weight for a standard animal (shell length = 66.5 mm) in Chinju Bay is presented in Figure 5. The dry 180 Park et al JASONDJ FMAMJ 1999 2000 Month Figure 4. Scapharca broughtonii. Monthly variations of condition (CI) (a), gonad maturity index (GMI) (b) and gonadal development (c(. tissue weight was maximal in July 1999 and June 2000 (ca. 7 nig ■ standard animal"' for both cases). The tissue weight fell abruptly in summer when spawning occurred. Then, there was a great increase in tissue weight at the timing of sexual resting stage and reserve buildup (see later) during the autumn period (October to December). Thereafter, only a little tissue weight increased steadily during the winter to spring period (February to May). re E 'E re re C re c/) E J F 2000 Month Figure 5. Scapharca broughtonii. Monthly variations of dry tissue weight in a standard animal (shell length = 66.5 mm). Gross Biochemical Composition The gross biochemical composition of a standard animal was calculated from the percentage compositions of each component and the dry tissue weight in Figure 5 and Figure 6 summarizes seasonal cycles of the biochemical components in absolute values as mg dry weight per standard animal. A sudden fall in all the biochemical components of a standard animal occurred during the summer spawning. Thereafter, a great increase in protein and car- bohydrate was synchronous with the increase in tissue weight dur- ing the autumn period (October to December). At this period, the greater increase in carbohydrate reflected the storage of glycogen reserve in the soft tissue of animals. However, lipid kept lower during this period. During the winter to spring period (February to May), the patterns of variation in each component were different. Protein content continued to increase until May but carbohydrate (also glycogen) content, which was maximal in December, de- creased progressively from February to May. Inversely to carbo- hydrate, lipid content increased steadily during this period. A sharp decrease in all the components was then recorded from June to the end of the study. Ash content showed relatively steady values. Association between Different Parameters The Kendall's rank correlation coefficients between the envi- ronmental and Siciphiircd measures are presented in Table 1. Chlo- rophyll (/ concentration was correlated positively (p < 0.01) with water temperature but negatively (p < O.O.'i) with salinity, standard Reproduction and Composition of Scapharca broughtonii 181 J A s 0 N D J F 1999 2000 Month Figure 6. Scapharca hrniighlonii. Monlhly variations of biochemical component levels in a standard animal: protein, larbohvdrate, glyco- gen, lipid and ash. animal protein and carbohydrate (also glycogen) values. Both the condition index and the standard animal dry weight were corre- lated positively (p < 0.01) with its protein value. The GMI corre- lated positively (p < 0.01) with temperature and lipid value but negatively (p < 0.01) with carbohydrate (also glycogen) value. With the exception of carbohydrate-glycogen pair, no significant correlation between the absolute values of biochemical compo- nents was found. DISCUSSION Rcpnidiictirc Cycle Changes in temperature, salinity or photoperiod have been con- sidered physical environmental factors, inducing spawning of bi- valves (Mann 1979. Dohmen 1985. Lubet et al. 19S7). The game- togenesis of 5. hroiighlonii in Chinju Bay was characterized by a unimodal cycle, as has been reported for other north Temperate populations of the species in Korea (Park et al. 1998) and Japan coasts (Numaguchi 1996). Positive correlation between GMI and water temperature (t = 0,658. p < 0.01 ) indicates that temperature might play an important role in inducing the gonadal development and spawning of 5. broughtonii. A new cycle of gametogenesis began at the lowest water temperature (ca, 5.3°C) in February. .Similar results were also observed by Park et al. ( 1998) for other populations of the same species from the southern coast of Korea. However, the lowest water temperature, at which gametogenesis in the kilter populations was initiated, was 10°C. This difference indicates that initiation of gametogenesis in 5. hnmglnonii occurs at different temperatures in different places, as reported for the oysters Ostrea edidis and Crassostrea gigas (Ruiz et al. iy92a, b). for the cockle Cerasioilcniui cdiile (Navarro et al.. 1989), for the mussel Myliliis echili.s (Newell et al. 1982) and for the manila clam Tape philipiiiiiuniiu (.Sbrenna and Campioni 1994). In these other species, even spawning also took place at different water tempera- ture. Salinity of Chinju Bay was relatively constant throughout the year with the exception of low salinity in the rainy summer season. No significant coirelation between the GMI and salinity was found (Table 1 ). Food availability also may have an effect on initiation of spawning. Starr et al. (1990) demonstrated that phytoplankton lev- els during blooms should be sufficient to induce spawning in sea urchins Strongylocentrotits droelxichiensis and mussels M. edidis. Ruiz et al. (1992a) also postulated that the stimulation by phy- toplankton blooms could explain the spawning of C. gigas at the low temperature during late October. It is difficult to explicate the summer spawning event by an environmental factor that may trig- ger spawning. In Chinju Bay. spawning of S. broughtonii took place at the moment when water temperature was highest, salinity lowest and photoperiod longest, throughout the year. Also, the blooms of phytoplankton occurred in summer. Kang et al. (1999) reported for the southern coastal bay systems of Korea that the addition of excessive dissolved inorganic nutrients (particularly nitrogen) via streams enhances phytoplankton biomass greatly in the low saline summer period. Therefore, spawning in summer enables larvae to be produced at the most opportune moment with regard to food availability (Newell et al. 1982, see also Navarro et al. 1989). Numaguchi (1996) observed that spawning of S. Ivoughtonii stock cultured in Kasado Bay (Japan) throughout the year took place in insufficient maturation degree from July to September. Spawning of another slock transplanted to Kasado Bay three years after culture in Kagawa Prefecture with the similar latitude and thermal condition was expanded in sufficient maturation from May to October. He concluded that the difference in maturation be- tween two ark shell stocks resulted from nutritive and growth condition. In the present study, the GMI conelated negatively with standard animal carbohydrate and glycogen levels (t = -0,632 and -0.579. p < 0.01 for both). These results indicate that tem- perature is not the only determinant of gametogenic cycle. Many authors have taken it into account that nutrient accumu- lation — depletion cycle in marine bivalves due to variations in ambient food availability may play a critical role in determining their gametogenic cycle. The effect of temperature may be direct by affecting bivalves' metabolic rate, or indirect by affecting the availability of food which increases in summer when temperature is high (Taylor & Venn 1979, Pazos et al. 1997). Both such effects of temperature may be relative to the seasonal cycle of storage and utilization of energy reserves. Gametogenesis and Biochemical Composition The gonadal growth in winter to early spring took place at the expense of the stored reserves (mainly glycogen). This was re- flected to a negative correlation between the GMI and the glyco- gen level (Table 1 ). The gametogenisis in marine bivalves occurs at the expense of the recently ingested food and/or energy stored in various tissues. In the mussel Mytdus eduiis, glycogen reserves are stored in the mantle (Lubet 1976, Bayne et al. 1982) and are used in the gametogenesis (Gabbott and Bayne 1973). However, in the oyster Ostrea eduiis. the storage reserves and the gametogenesis take place concurrently (Ruiz et al. 1992b). The relationship be- tween energy storage and gametogenesis within a single species 182 Park et .^l T.4BLE 1. Matrix of Kendall's rank correlation, t, between the environmental and Scapharca measures S SPM Chlfl FM FI CI GMI DW P CHO G L Ash T (1.410 (I.16S 0.564** 0.308 0.256 -0.231 0.658** -0.308 -0.231 -0.872*** -0.821*** 0.462* -0.667** S -0.219 -0.538* -0.385 -0.231 0.513* -0.211 0.538* 0.513* 0.487* 0.590** -0.128 0.590** SPM 0.168 0.194 -0.219 -0.142 0.093 -0.116 -0.142 -0.116 -0.168 0.039 -0.219 Chi a 0.282 0.231 -0.410 0.237 -0.590** -0.513* -0.590** -0.641** 0.128 -0.795*** FM 0.590** -0.256 0.395 -0.333 -0.308 -0.333 -0.282 -0.026 -0.333 H -0,205 0.263 -0.333 -0.308 -0.282 -0.231 -0.077 -0.282 CI 0.000 0.615** 0.692** 0.308 0.410 0.308 0.462 GMl 0.000 0.079 -0.632** -0.579** 0.605 -0.316 DW 0.923*** 0.385 0.436* 0.231 0.641** P 0.308 0.359 0.308 0.564** CHO 0.897*** -0.385 0.692** G -0.282 0.744*** L -0.128 T = water temperature. S = salinity. SPM = suspended particulate matter. Chi <; = chlorophyll a. FM = food material. FI = '7c food index, CI = condition index, GMI = gonad maturity index, DW = standard animal dry weight, P = protein, CHO = carbohydrate, G = glycogen, L = lipid). * 0.01 < p < 0.05, ** 0.001 < p < 0.01. *** p < 0.001. may depend also upon whether glycogen sufficient to meet the energy requirement for gonadal development, is stored under the feeding condition prior to the gametogenesis. Navarro et al. ( 1989) reported inter-annual differences in the timing of gametogenesis and the storage cycle of carbohydrate reserves in a population of the cockle Cerastodenna echile from Mundaca Estuary (Spain). In fact, gatnete proliferation of the cockle population occurred during the expense of glycogen reserves or simultaneously with the ac- cumulation of energy reserves, depending on prior feeding condi- tions. Such a difference within a species was also observed geo- graphically in the oyster Crassostrea gigas. While the glycogen stored is used in the gametogenesis of an oyster population in El Grove (Spain) (Ruiz et al. 1992a), the accumulation of reserve materials and the gametogenesis commenced simultaneously in Korean Waters (Kang et al. 2000). Recently, Luna-Gonzalez et al. (2000) also found that the scaWop Argopecten venthcosus uses the available food in the environment tiiore than tnuscle reserves for the gonadal maturation when the food is abundant, but they use the muscle reserves when the food abundance is poor. Summer spawning of S. brmightonii was followed by a subse- quent accumulation period of reserve materials in autumn (from October to Decetnber). A great increase in animal tissue weight in autumn was attributed to the rapid accutnulation of protein and carbohydrate during that period (Figs. 3 and 6). Glycogen level fell slowly during the course of gonadal development in the winter- spring period. In the present study, the low chlorophyll a. food inaterial and food index indicated that both quantity and quality of food were poor during the winter-spring period (from February to May), although a tetnporary peak in food material was recorded in February (Fig. 3). Therefore, our results indicate clearly that the gametogenesis of 5. hrtnigluonii in winter-spring depend largely on the glycogen stored during the previous autumn. It is generally accepted that glycogen reserves are the main source of energy in bivalves (Reid 1969, De Zwaan & Zandee 1972) and also may be utilized for the formation of gametes under conditions of nutrient stress (Gabbott and Bayne 197.3. Newell and Bayne 1980, Barber and Blake 1981, Beninger and Lucas 1984). Protein and lipid levels increased steadily during gametogen- esis, with maxima prior to spawning. Lipid is converted from glocogen reserves and biosynthesized during the formation of ga- metes (Gabbott 1975, Lubet 1976). Lipid level may be a good index of gonad tnaturity (Pazos et al. 1997), because lipid has been considered one of the principal energy sources used during non- feeding embryonic and eariy larval stages of the bivalves (Holland 1978, Gallager et al. 1986). A positive con'elation between the GMI and the lipid level (t = 0.605, p < 0.01) supports these hypotheses. Protein constitutes the major organic cotnponent of bi\al\e oocytes (Holland 1978). Therefore, protein tnaxitna prior to spawning may suppoil this hypothesis. Protein also supports reproduction after carbohydrate and lipid reserves were depleted (Mann & Glomb 1978, Barber & Blake 1981) and constitutes the main reserve during periods of energy imbalance (Gabbott & Bayne 1973, Beninger & Lucas 1984, Navarro et al. 1989). How- ever, the lack of evidence for such roles of protein reserves in the present study may be attributed to the biochemical analyses for pooled tissues. Another possible explanation for the steady in- crease of protein level in winter-spring is that part of the energy necessary for gametogenesis and maintenance may be supplied by the recently ingested food without intervention of the energy re- serves. Analyses for various organs such as striated adductor tnuscle, digestive gland and gonad will reveal roles of reserve materials in detail, as proposed by Barber and Blake (1981) and Pazos et al. (1997). In conclusion, S. hroughtniiii from Chinju Bay, a southern coastal bay of Korea, was characterized by a spawning period in summer due to the monocyclic gametogenesis throughout the year. Gatnetogenesis took place in winter-spring at the expense of re- serves accumulated previously in autumn. An inverse relationship between glycogen levels and GMI suggests that carbohydrate played an important role as the gametogenesis fuel. Therefore, this bivalve may be considered a typical conservative species in ga- metogenic pattern (Bayne 1976). ACKNOWLEDGMENTS This work was supported by the Ministry of Maritime Affairs and Fisheries - Special Grants for Fisheries Research and Devel- opment Project in Korea. The authors thank anonymous reviewers for helpful comments on an earlier draft of this manuscript. We are also grateful to Ki-Ryung Kim and Mi Kyoung Son for their as- sistance in hiochcniical analvsis. Reproduction and Composition of Scapharca broughtonii 183 LITERATURE CITED Ansell. A. D. & A. Trevaillion. 1967. Studies on Tellina teiuiis Da Costa. I. Seasonal growth and biochemical cycle. / Exp. Mar. Biol. Ecol. 1:220-235. Barber. B. J. & N. B. Blake. 1981. Energy storage and iitili/ation in relation to gametogenesis in Argopecten irridians concentrictis (Say). / Exp. Mar. Biol. &o/ 52:121-134. Barnes, H.. M.. Barnes & D. M. Finlayson. 1963. The seasonal changes in body weight, biochemical composition and oxygen uptake of two com- mon boreo-arctic cirripedes. Bananus halanoidcs and B. Inilantis. J. Mar. Biol. As.s. U.K. 43:185-211. Bayne, B. L. 1976. Aspects of reproduction m bi\alve molluscs. In: M. L. Wiley, editor. Estuarine processes. New York: Academic press, pp. 432-448. Bayne. B. L. & C. M. Worrall. 1980. Growth and production of mussels Mytilus edulis from two populations. Mar. Ecol. Prog. Ser 3:317-328. Bayne. B. L., A., Bubel, P. A., Gabbott. D. R., Livingstone. D. M.. Lowe cS: M. M. Moore. 1982. Glycogen utilisation and gametogenesis in Mytihts edulis L. Mar. Biol. Lett. 3:89-105. Beningcr, P. G. & A. Lucas. 1984. Seasonal variations in condition, re- productive activity and gross biochemical composition of two species of adult clam reared in a common habitat: Tapes decussalus L. (Jef- freys) and Tapes philiphinarum (Adams and Reeve). J. Exp. Mar. Biol. Ecol. 79:19-37. Bligh. E. G. & W. F. Dyer. 1959. A rapid method of total lipid extraction and purification. Cati. J. Biochem. Physiol. 37:91 1-917, Bricelj. V. M., J., Epp & R. E. Malouf. 1987. Intraspecific variation in reproductive and somatic growth cycles of bay scallops Argopecten irradians. Mar. Ecol. Prog. Ser. 36:123-137. De Zwaan, A. & D. 1, Zandee. 1972. Body distribution and seasonal changes in the glycogen content of the common sea mussel Mxiiliis edulis. Coinp. Biochem. Physiol. 43A:53-58. Dinamani, P. 1987. Gametogenic patterns in populations of Pacific oyster. Crassostrea gigas. in Northland, New Zealand. Aqiiaciilture 64:65-76. Dohmen, R. M. 1985. Gametogenesis, In: N. H. Verdonk & J. A, M. Van der Biggeelaar, editors. The mollusca (development). London: Aca- demic Press, pp. 1-37. Dubois. M.. K. A.. Gilles. J. K., Hamilton, P. A., Rebecs & F. Smith. 1956. Colorimetric method for the determination of sugars and related sub- stances. Anal. Chem. 28:350-356. Gabbott, P. A, 1975. Storage cycles in marine bivalve molluscs: an hy- pothesis concerning the relation between glycogen and gametogenisis. In: H. Barnes, editor. Proceedings of the Ninth European Marine Bi- ology Symposium. Aberdeen, Scotland. UK: Aberdeen University Press, pp. 191-211. Gabbott. P. A. 1983. Developmental and seasonal metabolic activities in marine molluscs. In: A. S. M. Saleuddin & K. M. Wilhur. editors. The mollusca, vol. 2, New York: Academic Press, pp. 165-217. Gabbott. P. A. & B. L. Bayne, 1973, Biochemical effects of temperature and nutritive stress on Mytilus edulis L, J. Mar. Biol. Assoc. U.K. 53:269-286. Gallager, S. M., R., Mann & G. C. Sasaki. 1986. Lipid as an index of growth and viability in three species of bivalve larvae. Aquaculture 56:81-103. Giese, A. C. 1969. A new approach to the biochemical composition of the mollusk body. Oceanogr. Mar. Biol. Ann. Rev. 7:175-229. Holland, D. L. 1978. Lipid reserves and energy metabolism in the larvae of benthic marine invertebrates. In: D. C. Malins & J, R. Sargent, editors. Biochemical and biophysical perspectives in marine biology. London: Academic Press, pp. 85-123. Humason. G. L. 1979. Animal tissue techniques. 4th edition. San Francis- co: W. H, Freeman and Company. 661 pp. Ivell, R. 1979. The biology and ecology of a brackish lagoon buvalve, Cerastoderma glaucum Bruguiere, in an English lagoon, the widewa- ter, Sussex. / moll. Stud. 45:383^00. Kang, C.-K., M, S. Park, P.-Y, Lee, W.-J. Choi & W,-C. Lee. 2000. Seasonal variations in condition, reproductive activity and biochemical composition of the Pacific oyster, Crassostrea gigas (Thunberg). in suspended culture in two coastal bays of Korea. / Shellfish Res. 19:771-778. Kang, C.-K., P.-J. Kim, W.-C. Lee, P.-Y. Lee. 1999. Nutrients and phy- toplankton blooms in the southern coastal waters of Korea: I. The elemental composition of C. N. and P in particulate matter in the coastal bay systems. J. Korean .Soc. Oceanogr. 34:86-94. Kim, J. D. & J. H. Koo 1973 Studies on the seedling production of the ark, Anadara hroughtimii (Schrenck) m lank ll). Bull. Fish. Res. Dev. Agency 22:55-65. Kim, J. D., S. C, Cheong. Y. J., Chang & H, W. Kang. 1979. Studies on the artificial mass seed production of the ark shell Anadara broughtonii (Schrenck) (1), On the spat collection in tank and transpon of attached young. Bull. Fish. Res. Dev. Agency 22:55-65. Kim, Y. & D. S. Yoon. 1980, On the spat collection of ark shell, Anadara broughtonii Schrenck. I. Water depth for effective spat collection and effects of intermediate culture. Bull. Fish. Res. Dev. Agency 23:219- Lee, J. M., J, W, Park. M. S. Yoo & Y. K. Hong. 1997. Morphological characteristics and genetic diversity using the RAPD technique in the arkshell. Scapharca broughtonii (Schrenck) from Korea and China. J. Korean Fish. Soc. 30:297-304. Lowry, O. M., N. I., Roseborough, A. L., Farrand & R. J. Randall. 1951. Protein measurement with the folin phenol reagent, i. Biol. Chem 193: 263-275. Lubet, P. 1976. Recherches sur le cycle scxuci cl remission des gametes chez les Mytilidae et les Pectinidac (Moll. Bivalves). Rev. Trav. Off. Peches Mar. 23:387-548. Luhet, P., M., Mathieu & F. Lenoir. 1987, Controle endochrinien de la reproduction chez les mollusques bivalves. Oceanis 13:291-304. Luna-Gonzalez, A. C. Caceres-Martinez. C. Ziiiiiga-Pacheco. S. Lopez- Lopez & B. P. Ceballos-Vazque/-.2000. Reproductive cycle of Ar- gopecten ventricosus (Sowerby 1842) (Bivalvia: Pectinidae) in the Rada del Puerto de Pichilingue. B.C.S., Mexico and its relation to temperature, salinity, and food. J. Shellfish Res. 19:107-112. Mann, R. 1979, Some biochemical and physiological aspects of growth and gametogenesis in Crassostrea gigas and Ostrea edulis grown at sus- tained elevated temperature. J. Mar. Biol. As.mc. U.K. 59:95-110, Mann, R. & S. J. Glomb. 1978. The effect of temperature on growth and ammonia excretion of the Manila clam Tape japonica. Est. Coast. Mar. Sci. 6:335-339. Marsh, J. B. & D. B. Weinstein. 1966. Simple charring method for deter- mination of lipid. J. Lipid. Res 7:574-576. Navarro, E., J, I. P., Iglesias & A. Larranaga. 1989. Interannual variation in the reproductive cycle and biochemical composition of the cockle Cerastoderma edulc from Mundaca Estuary (Biscay, North Spain). Mar. Biol. 101:503-511. Newell. R. I. E.. T. J., Hilbish, R. K., Koehn & C. J. Newell. 1982, Temporal variation in the reproductive cycle of Mytilus edulis (Bi- valvia, Mytilidae) from localities on the east coast of the United States, Biol. Bull. Mar. biol. Lcib. Woods Hole 162:299-310. Newell, R. I. E. & B. L. Bayne. 1980. Seasonal changes in the physiology, reproductive condition and carbohydrate content of the cockle Cardium (= Cerastodenna) edule (Bivalvia: Cardiidae). Mar. Biol 56:11-19. Numaguchi. K. 1996. Gonadal development of the ark shell Scapharca broughtonii (Schrenck) broodstock in farming grounds of Japan. Nip- pon Suisan Gakkaishi 62:384-392. Park. M. S.. H. J., Lim & P. J. Kim. 1998. Effect of environmental factors on the growth, glycogen and hemoglobin content of cultured arkshell. Scapharca broughtonii. J. Korean Fish. Soc. 31:176-185. Park. M. S.. H. J.. Lim, Q.. Jo. J. A.. Yoo & M. J. Jeon. 1999. Assessment of reproductive health in the wild seed oysters. Crassostrea gigas. from two locations in Korea. J. Shellfish Res. 18:445^50. 184 Park et al Parsons. T. R., Y.. Maita & C. M. Lalli. 1984. A manual of chemical and biological methods tor seawater analysis. Pergamon Press. New York. Pazos. A. J.. G.. Roman. C. P., Acosta. M.. Ahad & J. L. Sanchez. 1997. Seasonal changes in condition and biochemical composition of the scallop Pecten maximiis L. from suspended culture in the Ria de Arousa (Galicia, N.W. Spain) in relation to en\ ironmental conditions. / £v/>. Mar. Biol. Ecol. 211:169-193. Pyon. C. K., Y. G., Rho & Y. K. Yoo. 1976. Studies on the spat collection and rearing of the larvae. Anaditra hraughlonii {Schtsnck) in tank. Bull. Fish. Res. Dev. Agency 15:7-18. Reid, R. G. B. 1969. Seasonal observations on diet, and stored glycogen and lipids in the horse clam. Tresi(.s capa.x (Gould. 1850). Veliger 11:378-381. Rho, Y. G. & C. K. Pyen. 1977. Experimental studies on the seed collection of Anadara hroiiglittnui (Schrenck). Bull. Fi.sh. Re.s. Dev. Agency 16: 135-150. Rodhouse. P. G.. C. M.. Roden. G. M.. Burnell. M. P.. Hensey. T.. Mc- Mahon. B.. Oltway & T. H. Ryan. 1984. Food resource, gametogenisis and growth of Mytilus edulis on the shore and in suspended culture: Killary Harbour. Ireland. J. Mar. Biol. Assoc. U.K. 64:513-529. Ruiz. C. D. Martinez. G. Mosquera. M. Abad & J. L. Sanchez. 1992b. Seasonal variations in condition, reproductive activity and biochemical composition of the tlat oyster. Oslrea ediili.s. from San Cibran (Galicia. Spain). Mar. Biol. 112:67-74. Ruiz, C. M. Abad. F. Sedano. L. O. Garcia-Manin & J. L. Sanchez Lopez. 1992a. Influence of seasonal environmental changes on the gamete production and biochemical composition of Cra.s.sosrreii gigas (Thun- berg) in suspended culture in El Grove. Galicia. Spain. J. Exp. Mar. Biol. Ecol. 155:249-262. Sbrenna. G. & D. Campioni. 1994. Gametogenic and spawning patterns of manila clams Tapes philippinanmi (Bivalvia: Veneroida) in two la- goons of the River Po Delta. Italy. J. Shellfish Res. 13:37-46. Starr, M.. J. H.. Himnielman & J. C. Therriault. 1990. Direct coupling of marine invertebrate spawning with phytoplankton hloonis. Science 247:1071-1074. Taylor. A. C. & T. J. Venn. 1979. Seasonal variation in weight and bio- chemical composition of the tissues of the Queen scallop. Chlaniys operciilaris. from the Clyde Sea area. J. Mar. Biol. Ass. U.K. 59:605- 621. Walne. P. R. 1976. Experiments on the culture in the sea of the butterfish Venenipis deciissala L. Aijiiaculture 8:371-381. Widdows. J.. P.. Fieth & C. M. Worrall. 1979. Relationships between seston, available food and feeding activity in the common mussel Myii- liis edulis. Mar. Biol. 50:195-207. Yokogawa. K. 1997. Morphological and genetic differences between Japa- nese and Chinese red ark shell Scapharca hroughronii. Fish. Sci. Tokyo 63:332-337. Yoo. M. -S. & S. K. Yoo. 1974. Spat collection and the growth of .Anadara broiighlonii Schrenck. Bull. Korean Fish. Soc. 7:79-86. Joiirmil of Shellfish Research. Vol. 20. No. 1, 185-195, 2U0I. A STUDY OF THE ARKSHELL CLAMS, NOETIA PONDEROSA (SAY 1822) AND ANADARA OVALIS (BRUGUIERE 1789), IN THE OCEANSIDE LAGOONS AND TIDAL CREEKS OF VIRGINIA KATHERINE A. MCGRAW,' * MICHAEL CASTAGNA,' AND LOVEDAY L. CONQUEST' ^National Oceanic & Atmospheric Adminislnition. National Marine Fislieries Service. SSMC-3 (F/HC3). 1315 East-West Hww Silver Spring. Maryland 20910: -Virginia Institute of Marine Science. College of Williaw and Maty. Eastern Shore Laboratory. Wachupreagiie. Virginia 23480: ^School of Fisheries/Center for Quantitative Science. Box 352515. University of Washington. Seattle. Washington 98195 ABSTRACT Two species of arkshell ("hlood") clams. Nocria ponderosa and Anadam oralis, have recently been targeted by watermen on the eastern shore of Virginia for sale to both East and West Coast markets in the United States. Until 1991. fishermen caught both species in the harvest of oysters and hard clams, and discarded them as bycatch with little value. Very little is known about either species of blood clam, and preliminary data from a pilot study in 1993 indicated that they were being over-fished. We conducted a survey in September 1994 in the oceanside lagoon system along the eastern shore of Accomac and Northampton Counties. Virginia, and collected data on density, abundance, habitat preference, age-size and morphometric relationships, and mortality rates for both species of blood clams, as well as some ancillary data on the hard clam. Mercenaria mercenaria. The study provides baseline data for establishing management practices and regulations for the blood clam fishery. The total estimated abundance in the study area was about 16 million N. ponderosa and 6.4 million A. oralis. Of the clams taken in commercial catches on the oceanside of the eastern shore. M, mercenaria constitutes about 849r. N. ponderosa 15%. and A. oralis 1%. In our field survey. M. mercenaria was the most abundant species (72% of the total catch), followed by N. pondero.w (17%) and A. oralis ( 1 1%). Densities for blood clams averaged 0.35 clams m"", or 3.500 clams per hectare, and were highest in shell and shell/mud substrate (1.1 and 1.2 clams m"-. respectively). Growth studies and age-size data show that A. oralis grows about twice as fast as N. ponderosa and that market-size N. ponderosa (about 56 mm in shell height) may be 8+ years old. We also present information on mortality rates and morphometric relationships for both species of blood clams, and recommendations for maintaining and enhancing the fishery. KEY WORDS: Noetia ponderosa. Anadara oralis, arkshell. blood clam, growth rate, density, substrate INTRODUCTION Since 1991 two species of arkshell or "blood" clams, Noetia ponderosa (Say 1822) (ponderous ark) and Anadara oralis (Bru- guiere 1789) (blood ark), have been harvested by watermen on the eastern shore of Virginia for sale to markets in Washington D.C., New York City, Los Angeles, and Chicago. Long considered a useless incidental catch in the harvest of the hard clam, Merce- naria mercenaria (Linnaeus 1758), and the eastern oyster, Cras- soslrea rirginica (Gmelin 1791), arkshell clams now constitute a rapidly growing fishery with potential for future development. However, there is very little published information on the life hi.story of either of these species. Chanley (1966) and Chanley and Andrews (1971) described the larval stages of N. ponderosa and Anadara transversa (Say 1822) in Virginia waters and reported spawning periods of June to December and May through Septem- ber, respectively. In addition, they stated that A. transversa was common in Chesapeake Bay and tributaries with salinities above 15. Loosanoff and Davis ( 1963) and Loosanoff et al. (1966) also described the larval development of A. transrersa. McGraw and Castagna (1994) conducted preliminary investigations on growth rates of A. ovalis and N. ponderosa on the eastern shore of Virginia concurrent with the growth of a blood clam fishery there. Ander- son et al. ( 1985) also reported the potential for a viable blood clam fishery along the South Carolina coast but could not find a market for the clams. Walker ( 1998) studied the growth and survival of A. ovalis in suspended pearl nets in Wassaw Sound. Georgia. The intensive harvesting of blood clams and paucity of data on impor- tant factors such as distribution, densities, growth rates, and sur- *Corresponding author. E-mail: kay.mcgraw@noaa.gov vival present a problem for management of the fishery in Virginia waters. Blood clam landings in Virginia (Fig. 1) from 1993 to 1998 (National Oceanic and Atmospheric Administration 2000) show that about 8.9 metric tons of blood clams were harvested in 1993, with a decline to 2.5 metric tons in 1995, and an increase to 10.8 metric tons in 1997. The most logical explanation for the resur- gence in tonnage appears to be a change in gear type, from mostly mechanical tongs during the early phase of the fishery to clam dredges after 1996. Landings are reported as wet-meat weights, whereas the clams are usually sold whole, in the shell. Using a conversion factor supplied by the state of Virginia (Iverson pers. comm., Virginia Marine Resources Commission, February, 2000), we estimated that the number of clams range from about 340,000 clams landed in 1995 to 1.5 million clams in 1997. Most of the blood clams harvested along the eastern shore of Virginia are N. ponderosa: however, some A. 1 (Sokal and Rohlf. 1969). Therefore, we transfonned density data using log (X -i- 1 ) transformation according to the method discussed in Zar (1974) and Sokal and Rohlf (1969) before we applied analysis of variance (ANOVA) or other statistical tests (e.g., to test for differences between mean densities among substrate types). After transforma- tion, we calculated clam densities by dividing the number of clains FIsjure 3. Noetia ponderosa (a) and Antuiuru inalis (b) showing length (L) and height (H) dinicnsuins; (c) Anadara ovalis showing depth (D) dimension and periostracuni layer (/'). caught in each sample by the area covered by the mechanical tongs (i.e., 1.12 ni"). Then we tested mean densities of clams (both species combined) found in the various substrates using ANOVA (Sokal and Rohlf 1969) and the Student-Newman-Keuls (SNK) test (Zar 1974) to determine which ineans, if any, were signifi- cantly different (a = 0.05). We analyzed data further according to substrate type and water depth (i.e., deep water/channels or mud flats/shallow areas). Where substrate type was the same for two or three samples from a station, we averaged those for each station: if substrate types differed, we treated thein as separate samples. Therefore the number of samples for comparing species by substrate types was 169 instead of 119. Shell Aging We used acetate peels on different sizes of blood clams to determine age more precisely and to estimate the maximum Ion- 188 McGraw et al. gevity of the clams. The technique has long been used by paleon- tologists (Rigby and Clark 1965). but has proven effective in age determination for several species of bivalves (Ropes and O'Brien 1479. Ropes 1984. Ropes 1987. Kennish 1980). We specifically employed Farrow's (1971) method, which eliminates the step of embedding shells in epoxy resin. In addition, we modified Far- row's (1971 ) method by using thick ( 1.56 mm) acetate and cutting it into 7.6 cm x 2.54 cm pieces (i.e.. the size of microscope slides) that could fit easily onto the mechanical stage of a compound microscope. Several authors have discussed shell microgrowth patterns in detail, including Pannella and MacClintock (1968), Rhoads and Pannella ( 1970). and Lutz and Rhoads (1980). Age and size data can be applied to size distribution data through back-calculation procedures to create age frequency distributions, thus providing a better understanding of the population structure (Robson and Chapman 1961. Giilland 1966. Ricker 1975). We determined ages for all blood clams from the field survey and increased sample size and accuracy by supplementing these with data from previous studies and additional clams purchased from local watermen. Then we constructed Von Bertalanffy growth curves (Ricker 1975) for both species of blood clams. Mortality Rates We used articulated clam shells from survey samples to esti- mate annual age-specific and instantaneous mortality rates for both species. Some bivalves remain articulated for a time after death before the hinge ligament deteriorates and the valves separate. The.se can be used to help estimate natural mortality (Dickie 1955, Buckner 1984) by dividing the number of articulated shells by the sum of live clams plus articulated clams for different size/age groups. There are no data documenting the length of time for deterioration of the hinge ligament for either N. ponderosa or A. ovalis in Virginia waters, and all mortality rate calculations are based on an assumed time period of one year for disarticulation. We computed instantaneous mortality rates using the equation (Z) = -logE (1 - A), where A = the number of living clams in an age group and E = 2.718. the base of natural logarithms (Ricker 1975). First, we applied age-length data from acetate peels to the size distribution of living and articulated clams from the survey and determined the number of clams in each age-size class. Then we divided the number of articulated shells in a given age- size category by the number of live clams to arrive at an age specific or annual mortality rate expressed as a percentage. Commercial Catch Data Mr. David Bishop (Oakhall. VA. Accomac County) collected and recorded data on the proportion of blood clams versus hard clams from many of his tong catches over a three-month period, from September through November, 1994. We used height- frequency data from his catches for comparison with data from 1993 fisheries samples to help determine if average clam size in commercial catches was decreasing and also to more accurately assess the percentage of blood clams in clam harvests. RESULTS AND DISCUSSION Field Survey Results Clam Densii\ and Abundance The 1 1 9 stations included random stations, mostly in Hog Island Bay. as well as non-random, or systematic stations, in the northern part of the study area, taken mostly in channels or creeks. Before combining data from all stations, we first ascertained (using trans- formed data and ANOVA) that there was no significant difference (a = 0.05) in mean clam densities (total densities or by species) between random and non-random stations. For the non-random (mostly channel) stations, we also tested to see if there was any correlation between clam densities and station location within particular waterways, as defined by distance (nau- tical miles) from the entrance of a channel (Scheaffer et al. 1996). We found no correlation or trends (maximum r" = 0.06 in the Great Machipongo Channel) and combined all non-random chan- nel stations for further analyses. Average density (mean ± standard error = .Y ± S.E.) at all stations (/i = 1 19) was 0.21 ± 0.06 clams m""^ for N. ponderosa and 0. 14 ± 0.07 clams m"" for 4. oralis. Total average blood clam density was 0..^5 ± 0. 10 clams m"". Compared to hard clams, 17% of the total catch was N. ponderosa and 1 1 % was A. ovalis. To compare densities by substrate, we averaged densities for each substrate at each station; if station samples contained two or more different substrates, those were considered separate samples. We did this to minimize any pseudoreplication (Hurlbert 1984) and to avoid combining samples across substrates. Therefore, sample size for comparing substrates was 169 (i.e.. at some sta- tions not all samples were of the same substrate). In addition, because the mean densities are weighted differently when catego- rized by substrate type, they are slightly different from those cal- culated for stations, where the sample size was 119. The propor- tion of different substrate types among stations was as follows: mud = 357f. /; = 59: sand = l\%. n = 18: sand/mud = 21%, n = 36: shell/mud = 17%, n = 29: shell/sand = 6%, « = 11; and shell = 10%, n = \6. Mean clam densities varied among substrate types (Fig. 4) and were highest in shell (1.24 ± 0.4 clams m""^) and shell/mud sub- strates (1.12 ± 0.49 clams m""^). Noeliu po)nlerosa accounted for most or all clam densities in those two substrates. For example, mean density of N. ponderosa was highest in shell/mud substrate (1.12 clams m' ±0.49) and shell (0.79 ±0.36), whereas A. ovalis densities were highest in shell substrate (0.45 ± 0.24 clams m"~) and mud substrate (0.19 ± 0.15). The highest density of M pon- dero.sa we observed in a single sample was 13.4 clams m"" at a station in the Great Machipongo Channel (intracoastal waterway): the mean density at that station was 4.46 clams in"~. The highest 20 1 5 \N ponderosa SSA ovalis HZ Mean ± S E A Mean density of A ovalis ° Mean density of N ponderosa E i5 10 CJ 05 0.0 0 08 T 0 0 0 03 0 " ill -A i 1 ' A U We took a total of 355 individual samples at 119 stations, which yielded adjusted totals of 86 N. poudeidsa. and 55 A. ovalis. IVIud Sand Sand/mud Shell/mud Sheli/sand Shell Substrate Types Figure 4. Comparison of A. ovalis and A', ponderosa densities (clams m ■) in various substrates from the eastern shore of Virginia. Arkshrll Clams in Virginia Co.vstal Waters 189 density of A. oralis was 15.0 clams m"" fiDm a sample in The Deeps Channel, a hranch of the Great Machipongo Channel. After transforming density (i.e.. log [X + 1|). we compared mean blood clam densities for the six substrate types using ANOVA (a = 0.05) and the Student-Newman-Keuls test (Zar 1974) to detennine which, if any. means were different. Theie was a significant difference in mean total blood clam densities among substrates [P < 0.001). There were no significant diffeiences in mean total blood clam densities between shell and shell/mud sub- strates, or between mud substrate and shell/sand, sand/mud. or sand (Table 1 and Fig. 4). However, there were significant differ- ences in inean densities between both shell or shell/mud substrates and all other substrate types (P < 0.001 ). Results are summarized as follows: (shell = shell/mud) ^ (mud = shell/sand = sand/mud = sand). Next, we compaied densities for both species by substrate type. Results were the same for analyses of N. ponderosa densities by substrate types as for both species combined. However, density of A. ovatis in shell substrate was significantly different from those in ail other substrates, i.e., shell ^ (shell/mud = mud = shell/sand = sand/mud = sand). The higher densities of clams in shell and shell/mud substrates suggests that shell is impoilant either for at- tachment, protection from predation. or both. We estimated the abundance of clams (Table I ) by using sub- strate data from Haven et al. (1981 ) and multiplying the density of clams found in various substrates by the number of hectares of that substrate for the study area. That is. A, = S (Ds x ha), where A, = the total abundance of clams in the study area; Ds = the total mean density of clams in a given type of substrate; and ha = number of hectares of a given substrate in the study area. Haven et al. ( 1981 ) estimated the following amounts (converted to hectares here): shell = 116 ha; shell/sand = 838 ha; shell/mud = 1.002 ha; sand = 1,523 ha; mud = 2,933 ha; sand/mud = 1.057 ha. The combined areal totals from Haven et al. (1981) are for Burton's Bay. Bradfoid Bay. Swash Bay. Upshur Bay. Major Hole Bay. Revel Island Bay. Hog Island Bay (above and below North Chan- nel), Ramshom Bay, and Sand Shoal Channel. Total estimated blood clam abundance in the study area (Hog Island Bay, Burton's Bay, and Bradford Bay and contiguous waterways) was 22 million blood clams. Total estimated abundance by species is as follows: N. poiulcrosa. about 16 million; A. oralis, about 6 million. The proportions are based on those from the field survey (Table 1 ) in which N. poudcrosa had an average density of 0.31 clams m~' ± 0.09 ()i = 169). and/1. (n((//.v of 0.12 clams m"- ±0.06 (/! = 169) o\er all substrates. We examined clam densities in relation to water depth (i.e., channels or mudflats). Of the 119 stations sampled, 67 were in channels or locations with a water depth >2 meters, and 52 were over mudflats, or in shallower water. Using log transformation and Student's /-test (Zar 1974), we determined that there was no sig- nificant difference (a = 0.05) in mean total blood clam densities (Fig. 5) between channel/deep water stations (0.48 ± 1 .3) and mud flat stations (0.19 ± 0.10). even though channel stations had twice the density of shallower ones. The relatively high variances for both means affected results and P = 0.052. just slightly more than the stated level of significance. Although mean clam densities were higher at channel/deep water stations, there were no significant differences (P > 0.10) between mean densities in channel and mud flat stations for either species (Fig. 5). One explanation for the higher density of clams in the channels/deep water stations might be that clams are sloughed off or eroded from the sides of the channels, along with substrate, and aggregate in the bottom of the channel. In some of the tidal creeks currents may expose areas of shells, providing more attach- inent sites for blood clams, particularly N. pouderosa. Size-Frequency Average shell height for blood clam species (Fig. 6) were: N. pouderosa. 42.6 mm (± 2.2, /; = 43) and A. oralis. 25.1 mm (± 0.2, II = 29). There were lelatively few small N. pouderosa (i.e.. <25 mm. or about 2 years old) taken in survey samples. This could simply be the result of sampling variability, but could also indicate that recruitment may be low or that moilality rates may be high during the first year after settlement. Most of the hard clams (M. iiwrcenaria) in samples were also larger, 60 to 100 mm in height (75.3 ± 1.2. II = 146). and sizes were more normally distributed (Fig. 6c). In contrast, most of the A. oralis (Fig. 6a) were O-i- to 2 years old. with very few older, larger clams in samples. Articulated Shells and Mortality After calculating Von Bcrtalanffy growth curves for both spe- cies of blood clams, we applied age-length data to articulated shells, put them into age categories, and estimated annual and instantaneous mortality rates (Table 2) as previously described. In the absence of published or other data on the length of time for TABLE 1. .Mean densities (m"") of clams (±S.E.) by species and substrate types, areal estimates of substrate types (from Haven l'>81), and clam abundances. Substrate type Shell Shell/Sand ShellAIud Sand Mud Sand/Mud Total avg. density Species (N = 16) (N = III (N = 29) (N = 18) (N = 59) (N = 36) (N = 169) N. pouderosa 0.79 0.16 1.12 0 0.08 0 0.31 (±0.36) (±0.16) (±(1.49) (±0.06) (±0.09) A. ovcilis 0.45 (±0.24) 0 0 0 0.19 (±0.15) 0.03 (±0.03) 0.12 (±0.06) Total 1.24 0.16 1.12 0 0.27 0.03 0.42 (±0.45) (±0.16) (±0.49) (±0.15) (±0.03) (±1.49) Hectares 116 838 1,002 1,524 2.934 1.057 7.470 Clam abundance (millions) 1.4 1.3 11.2 0 7.9 0.3 T> ■) 190 McGraw et al. 1.00 0,75 I I N ponderosa ^H A ovalis ^H Total blood clam density Zn Mean t S E D Mean density of N ponderosa Mean density of A ovalis Mean total blood clam density Mud Flat Channel Location Figures. Comparison of blood clam densities (clams m "I on mudflats and in channels from the eastern shore of \ irginia. disarticulation of shells for either species of blood clam, we as- sumed a period of one year. Because blood clam density in the study area was relatively low, there were some age groups for which no articulated clams appeared in samples and thus mortality rates were zero. Despite the gaps in data, we think that the mor- tality estimates provide some insight into basic mortality trends for N. ponderosa and A. ovalis. The annual mortality rate (Fig. 7) for O-i- to 1 year A. oralis, calculated using articulated shells in samples, is 86%. then de- creases to 30% for the l-i- to 2 year class. Because there were only four articulated clams in the age 3+ category, all less than 48 mm. we pooled the data for a better estimate of mortality rates. The average was 80% for age 3-1- clams. Distributed over the estimated maximum life span of si.x years for A. ovalis. the annual mortality rate would be about 27% per year for clams over age 3-I-. There were no articulated shells in the 2-1- to 3 year size range in our samples, so the annual mortality rate for that year class of A. ovalis was zero. The increase in mortality rates for older (>i+) clams may be due to senescence or other factors such as disease, or synergistic effects involving spawning and higher water temperatures in the summer. Toyo et al. (1978) (as cited by Broom 1985) reported a sudden mass mortality of a species of Anadara (probably A. 15 I Mean height = 25.1 mm 10 5 A. ovalis (J) £ "o 10 20 30 40 50 60 70 80 90 100 110 120 130 U 1 5 Mean height = 42 6 mm (U 5 N. ponderosa 0 10 20 30 40 50 60 70 80 90 100 110 120 130 M mercenaria Mean height = 75 3 mm c) 0 10 20 30 40 50 60 70 80 90 100 110 120 130 Height (mm) Figure 6. Height frequency distribution for A. ovalis, N. ponderosa, and M. mercenaria. TABLE 2. Mortality rates (%) and instantaneous mortality rates (-Z) for A. ovalis and N. ponderosa, based on articulated clam shells. Year class No. articulated shells No. live Total Mortality (%) («) A. ovalis (1-1 lH-2 2-H-3 >3-i-'' N. ponderosa 0-1 1-1-2 2-1-3 3-t-^ 4-1-5 S-H-IO" l(l-i-l,V 70 7 0 4 11 81 16 23 1 5 1 9 T 5 9 9 3 3 3 3 11 17 14 19 86.4 2.0 30.4 0.4 0.0 0.0 80.0 1.6 88.9 ~t ~i 60.0 0.9 0.0 0 0.0 0 0.0 (1 35.3 0.1 26.3 0.1 •■ Pooled because of low number of clams in age category. Mortality is estimated to be about 27% per year and (-Z) is estimated to be about 0.54 per year for age 3 -t— 6 clams. •^ Mortality is estimated to be about 1% per year and (Z) is estimated to he about 0.1 per year for age 5 -i— 10 clatiis. '' Mortality is estitnated to be about 5.3%- per year and (-Z) is estiinuted to be about 0.1 per year for age 10 +- 15 clams. hroiighloni) in Japan due to a rapid rise in water temperature above 25 'C. Although our estimates are based on relatively few articulated clams, we feel that the data reflect the actual situation, because very few large ,4. ovalis (>50 mm in height) are taken in commer- cial clam catches and most seem to die before reaching six years of age. A similar phenotnenon has been observed for the bay scallop. Argopecten irradians. in which about 80% die between months 13-16 (Castagna 1975. Castagna and Duggan 1971 ). Ob- servations of large A. ovalis (i.e.. -53 mm in height) held in flow- ing water tables indicate that they are sensitive to water tempera- tures and begin to die above about 27°C, whereas large N. pon- derosa in the same water tables are not affected. Sotne watermen have told us that sotne A. ovalis in their catches seem to gape much more readily than A', ponderosa during warm weather. It appears that few A. ovalis live longer than six years in the Virginia lagoon system, and that mortalities of older clams are exacerbated by high water temperatures. 1+-2 2+-3 3-^-4 4+ - 5 5+-10 10+-15 Year Class Figure 7. Comparison of mortalit) rates for /V. ponderosa and A. ova- lis. Arkshell Clams in Virginia Coastal Waters 191 Walker ( 1998) reported mortality rates of about 557r for small and medium (<20 mm in length) A. o\(ili\ held in pearl nets in Georgia (i.e.. from an initial group of 78 clams down to ."^S) during the first year; average mortality rate for the same groups was about the same during the second year of growth (i.e.. 16 out of 33 clams survived). Presumably the pearl nets decreased mortality rates from siltation and predalion. and therefore rates were lower than those calculated in our field study. Walker ( 1998) also stated that the life span for A. oralis in Georgia waters is about three years, whereas we estimated that the life span is six years in Virginia. Annual mortality rates for N. ponderosu (Table 2 and Fig. 7) were highest for the 0+ to 2-year-old groups (89* and bWi . re- spectively). For convenience sake, and because of small sample sizes, we grouped the 5+ to 10-year-classes and lO-i- to l.'>-year-old clams to determine mortality rates. The cumulative annual mor- tality rate for the 5+ to 10-year-old group was 35% over the five year period, or about 1% per year; for the 10+ to 1.5-year-old clams, it was 26%, or about 5% per year. There were no articulated N. ponderosa for the 2+ to 5-y age classes in samples, suggesting a decreasing mortality rate for those age classes. This trend could be a result of increasing size and lower densities during the first two years, as a function of com- petition for food and space. Morphometries The data used for morphometric relationships are from several sources, including growth studies, fisheries samples, survey samples, and e.xtra clams purchased for shell aging studies. We examined the relationships of several variables: shell height, length, depth, whole weight, and wet weight. The relationship between valve height and length for A', pon- derosa (Fig. 8a) is described by the regression equation: L = 1.22H + 1.73 (r- = 0.99) where L = length in mm. H = height in mm. and r" = coefficient of determination. Height and shell depth were also linearly related (Fig. 8b) as described by the regression equation D = 0.98H - 2.57 (r"' = 0.99). where D = depth of the clam in mm. Relationships between height and whole (i.e.. shell -i- meat) or wet meat weights were nonlinear (Figs. 8c and 8d). For example. the relationship of height and whole weight is described by the allometric equation of the form W = aH" (Fig. 8c) where W = whole weight of the clam in grams. H = height in mm. and a and b are allometric coetficients (a = 0.0(J06 and b = 3.06). Trans- formed to the linear form, this equation is; log W = b\og H + loga. Wet meat weight and height were likewise nonlinearly related by the equation M = aWb. where M = meat weight, a = 0.001. and h = 2.52 (Fig. 8d). The coefficient of determination, r" = 0.88. is slightly lower than that for height and whole weight, since meat weight determinations are subject to more sampling error, mostly because of varying amounts of water loss. Mean shell weight was 46.8 g (± 36.3. n = 132). and mean whole wet weight of N. ponderosa was 60.5 g (± 44.06. n = 132). or about 77% of the total weight. Mean meat weight (13.6 g) was about 23% of the total weight of the clams sampled. Relationships of shell height to length or depth for A. oralis were linear (Figs. 9a and 9b). Unlike A', ponderosa, shell length in A. oralis changes little in relationship to the height, and most clams are. as the name {"oralis") implies, oval or nearly round. By comparison, increase in shell depth in A. oralis per increase in shell height is proportionately smaller than that in N. ponderosa. in which shell depth is almost equal that ol' height. For A. oralis, shell depth is about 70% of shell height. The relationship of whole weight and height in A. oralis (Fig. 9c) is best described by the curvilinear equation; W = aH^, where W = whole weight in grams, a = 0.0003. and h = 3.14 (r^ = 0.97). As with N. ponderosa. the correlation (M = 4E H' ') between height and wet meat weight (Fig. 9d) was more variable than with whole weight, and the coefficient of determination was slightly lower (r'^ = 0.89) than for the regression of height and whole weight. Mean whole weight for the A. oralis sample was 34.2 g (+ 21.2, n = 139), mean shell weight was 22.7 g (± 14.2, 3 o s 250 n r c) 200 - W = 0.0006H'" / 150 - r^ = 0 98 «f 100 ^ J^ 50- ^^' n - f-w-*"^^! 1 1 60 70 1 b) 60- J^' E 40 - D = 0 98H - 2 57 ^^^ l' = 0.99 oJ^* ^ t30- Q 20 - ^y^' 10 - 0 - " <^ 1 1 1 1 40 Height (mm) 60 45 40 d) 35 m=o.ooih'" 30 r^ = 0 88 25 .. 5 20 J 15 10 5 ^ ( ) 20 40 Height (mm) Fijjure 8. Resjrc'ssion of height versus length (a), depth (b). whole weight (c). and wet meat v\eight (d) for ,V. ponderosa. Data are from field surveys, commercial fisheries samples, and growth studies (« = 540 for a and b; n = 132 for c and d). 192 McGraw et al. 40 60 80 Height (mm) Figure 9. Regression of lieiglit versus length (a), depth (b), whole weight (c) and wet meat weight (dl for A. ovalis. Data are from field surveys, commercial fisheries samples, and growth studies in = 778 for a; ii = 641 for b; n = 139 for c and d). II = 139), and mean wet meat weight was 1 1.5 g (± 8.2, n = 139). Therefore, shell weight in A. ovalis constituted about 66% of total weight, and meat weight 34%. or an average of about 10% more meat weight for a given size than in N. pondeiosa. Age-size Relationships We prepared acetate peels from clams taken in the field survey, augmented with clams purchased from fishermen. Some data from growth studies (1992 to 1994) were incorporated as baseline data points for 1- and 2-y-old .4. ovalis and N. poiulciosii. We fit age-length data to the Von Bertalanffy growth equation; L,„ = U ( 1 - e-'") where L,,, is length at time t; L^ is the asymptotic, or maximum theoretical length; and K is a growth constant indicating the rate at which L-, is approached. The L.^ and K estimates for N. ponderosa (Fig. 10a) are; L, = 71.5 mm and K = 0.24 {r = 0.94). Both of these parameters are very similar to those given by Cahn ( 1951 ) for A. gninosa bisenensis in Japan. Mean length of N. ponderosa from a commercial fishery sample in 1992 was about 70 mm. Our data indicate that clams >70 mm in length would be 10+ years old. Samples of commercial catches (taken in 1994) from the same vicinity (Parting Creek) showed that the average length of N. ponderosa was about 56 mm, a decrease of about 14 mm, or the size of clams of about age 5+. In addition, mean length for M ponderosa in field survey samples was about 54 mm, which also indicates that the older clams are being depleted, and smaller, younger clams are now being har- vested. It is possible that some of the clams harvested are older, stunted clams; however, the size ranges for different age groups in the growth model suggest that this is not the case. The Von Bertalanffy model also provides a good fit for age and size data for A. ovalis (Fig. lOb). Values for L^ and K are 57.5 and 0.45. respectively (r^ = 0.83). Mean length for A. ovalis taken in the field survey was 27 mm, or about 1+ year-class clams, while mean length for A. ovalis in 1994 commercial fisheries samples was 56.5 mm, or 5+ years old. We obtained only five A. ovalis over five years in age in survey samples, which suggests that this species does not live very long in this geographic area. This is also corroborated by the moilality data from articulated A. ovalis shells, and from laboratory observations where large A. ovalis held in water tables died as ambient water temperatures approached 27°C. Commercial Fisheries Catch Data Catch data from a commercial fisherman on the eastern shore (D. Bishop, unpubl.) showed that M. mercenaria constituted an 90 1 a 80 . 70 " ■ *L4— 1 E" 60 ■ ^ •^4^ "fiTi ' ' £50 a. ^ ■f!^ ! * - |40 « 30 : •/ 1 20 - ■ j\ 10 - 0 - r , , , — , — 1 — , — ^ — ^ — . — 1 — . — , — , — ^ 10 20 0 12 3 4 5 6 7 Age (Years) Figure 10. \'on Bertalanffy growth curve for ,V. ponderosa (a), n = .^79, and A. ovalis (b), 11 = 211, from the eastern shore of Virginia. See text for equations. Arkshell Clams in Virginia Coastal Waters 193 average of about 78% of his catch with mechanical tongs, and blood clams (M ponderosa and A. oralis combined) 22% for the period of September to November 1994. The percentages of blood clams in catches for those three months were about 1 8% for Sep- tember. 21% for October, and 26% for November. The species distribution was similar to that from the field survey samples (i.e.. about 72% Mercenaria and 28% blood clams). Average daily catches (Fig. 11) for September, October, and November, 1994 were; 4.373. 3.873. and 3.642 clams per day. respectively. The estimated average catch per unit effort (i.e.. clams in one tonging effort, covering 1.12 m") during two days of fishing was 9.72 (II = 540) and 3.6 clams (/( = 720), respectively, for an overall mean of about 6 clams per tonging effort (« = 1,260). This equates to an average density of 3.4 clams m'~ in the harvest area. The average heights of A. oralis and N. ponderosa from sub- samples of September to November catches were 34.0 mm and 44.7 mm, respectively (Figs. 12a and 12b). The mean height for N. ponderosa was less than that for the commercial fishery sample taken in 1992 (70 mm) and the Parting Creek sample (56.0 mm) taken in 1993. The decrease in average size may indicate that overfishing is occurring. During December 1994 and January 1995, Mr. Bishop worked in an area just north of Wachapreague, VA (Gargatha Creek) and reported catching almost all A. oralis, the highest percentage catch of that species of which we are aware. His observation is note- worthy because of the norniall\ small percentage of A. oralis in catches, and indicates that small, dense beds of A. oralis exist in some isolated locations. The average height for A. oralis from the January sample in Gargatha Creek was 40.2 mm (Fig. 12c). Most of the clams were in the 30-40 mm size range ( = 33 to 44 mm in length), or >3-i- years old. The absence of O-i- year-class A. oralis in commercial fisheries samples may simply reflect the difficulty in seeing and collecting very small clams in the mud and debris which accompany catches, or it may also indicate low recruitment and/or high mortality rates. Another, more probable explanation is the fact that small A. oralis are epiphytic. During our study, the easiest place to find O-i- year-class A. oralis was enmeshed in bryozoan and hydrozoan colonies attached to commercial mollusc floats near Quinby or Wachapreague Inlets. CONCLUSIONS Survey data provide information about the species composition of the blood clam fishery on the Eastern Shore of Virginia. The hard clam. M. mercenaria. constitutes the majority of the catch (72%), with N. ponderosa accounting for about 1 7%. and A. tnalis 50 4,5 4.0 3.5 3.0 2,5 2,0 1 5 1 0 0,5 0,0 0,8 0,8 3 1 3 6 0.9 27 SEPT 94 OCT 94 NOV 94 D M. mercenaria D Blood clams Figure II. .\verage daily catch of blood and hard clams (A/, merce- naria) by a nsherman on the eastern shore of \ irginia (Sept.-Nov. 1994). W5 " 15 20 25 30 35 40 45 50 55 60 65 E 03 30 r - ^ A', ponderosa (J 20 E 15 20 25 30 35 40 45 50 55 60 65 30 20 Mean height = 40 2 mm | c) ^^ p A. jvalis 10 n rr. — — ii ■■ 15 20 25 30 35 40 45 50 55 60 65 Height (mm) Figure 12. Height frequency distributions for A. oralis (a I and (cl and A', ponderosa (b) from Quinby and Wachapreague. Virginia. 1 1 %. However, age-length relationships clearly show that N. pon- derosa. even though it is more abundant than A. oralis, is a rela- tively slow-growing species, and may not be suitable for a com- mercial fishery with high exploitation rates. Anadara oralis grows faster but appears to have a high mortality rate in most areas, as indicated by the relatively few. small ones that were taken in survey samples, the low percentage taken in commercial catches, and field experiments (McGraw et al. 1998). We estimated abundance for N. ponderosa and A. oralis to be about 16 million and 6.4 million clams, respectively, in the general area sur\'eyed. Catch estimates range up to 1 .5 million blood clams harvested annually in the survey area. Although this is about 15% of the estimated abundance, we are concerned that if fishing con- tinues unabated, blood clam populations could be decimated within a few years, particularly in light of data showing that N. ponderosa grows so slowly and observations that settlement is very sporadic. Survey data as well as samples of commercial land- ings from the Great Machipongo River and other areas show a decrease in average size of N. ponderosa. indicating that N. pon- derosa is currently being overfished and that there is a need to re-evaluate current policies governing the blood clam fishery on the eastern shore. Given the distinct possibility of overfishing, we think one of the best conservation measures with regard to blood clams is the cul- tivation of A. or(dis. It has a comparatively fast growth rate, should be relatively easy to spawn under hatchery conditions, and could be grown in conjunction with M. mercenaria on existing leases. We have anecdotal evidence that natural set of blood clams is sporadic and undependable. but hatcheries could provide a reliable source of seed. In addition, effective, feasible methods of floating culture (i.e.. "Taylor floats") are already in use for hard clams on the eastern shore (Luckenbach and Taylor): similar use for A. oralis would substantially reduce predation and enhance survival rates. If A. oralis can be successfully cultured, this may augment or supplant the harvest of blood clams by other means and. possibly, lessen some of 1^)4 McGraw et al. the fishing pressure from them. Otherwise, the blood eliim population on the eastern shore may dechne rapidly over the next several years. ACKNOWLEDGMENTS This research was funded by the National Oceanic and Atmo- spheric Administration. National Marine Fisheries Service (NOAA Grant # NA46FD()339), and we thank Ms. G. Faye (NMFS/ NOAA) for her helpful advice during the project. The NOAA Restoration Center provided funds for publication costs. We grate- fully acknowledge the help and technical assistance of Radford University faculty, staff, and students, especially Dr. S. Dennis. We are indebted to Captain D. Bishop for his willingness to un- dertake the field survey and for providing valuable catch informa- tion, and to his wife. Elisa. for compiling and forwarding the information. We also thank Dr. M. Luckenbach. Ms. J. Watkinson. Ms. N. Lewis, Mr. R. Bonniwell. and Mr. R. Cashwell. and other technical staff of the College of William and Mary, Virginia In- stitute of Marine Sciences. Eastern Shore Laboratory, Wachapreague. VA for their hospitality and invaluable assistance during various portions of the project. Dr. G.H. Johnson. Depart- ment of Geology, College of William and Mary, graciously helped us with shell aging techniques, and Ms. G. Arnold. Virginia Insti- tute of Marine Science. Glouster Point Laboratory, prepared a map of the study area. We also appreciate the helpful review comments and suggestions of Dr. C. Byerly. Dr. M. Carriker, Mr. J. Ewarl. Dr. J. Kraeuter. Dr. C. Langdon, and anonymous reviewers. LITERATURE CITED D. Princeton, NJ: Van Nostrand Abbott, R. T. I y-'i4. American Seashell Co. 541 pp. Abbott, R. T. 1968. Seashells of North America. New York: Golden Press. 280 pp. Anderson. W. D.. W. H. Lacey & A. G. Eversole. I9«5. Arks— is there a resource and a market? / Shellfish Res. .5:31 (Abstract). Bae, S. W. 1986. Origin and developing process of ark-shell culture in- dustry in Korea. Bull. Korean Fish. Soc. 19:72-82. Baqueiro, E. 1980. Population structure of the mangrove cockle A. iiiher- culosa (Sowerby, 1833) from eight mangrove swamps in Magdalena and Almejas Bays, Baja California Sur. Mexico. Proc. Nail. Shellfish Assoc. 70:201-206. Broom. M. J. 1983. Mortality and production in natural, artificially seeded and experimental populations oWAiuulani gninosa (L.) (Bivalvia:Ar- cidae). Oecologia 58:389-397. Broom, M. J. 1985. The biology and culture of marine bivalve molluscs of the genus Anadaia. International Center for Living Aquatic Resources Management. Manila, Phillipines. ICLARM Studies and Reviews. 12. Contribution No. 263. 37 pp. Buckner, S. C. 1984. Aspects of the population dynamics of the hard clam, Mercenaria mercenaria L., in Great South Bay, New York. Ph.D. thesis. State University of New York at Stony Brook. 217 pp. Cahn. A. R. 1951. Clam culture in Japan. U.S. Department of the Interior, Fish and Wildlife Service. FL-399. Castagna. M. 1975. Culture of the bay scallop. .Argnpecteii irraJians. in Virginia. Mar. Fish. Rev. 37:19-24. Castagna, M. & P. Chanley. 1973. Salinity tolerance of some marine bi- valves from inshore and estuarine environments in Virginia waters on the western mid- Atlantic coast. Malacologia 12:47-96. Castagna. M. & W. Duggan. 1971. Rearing the bay scallop, Aequipeclen irradians. J. Shellfish Res. 61:80-85. Chanley, P. 1966. Larval development of the large blood clam. Noelia ponderosa (Say). Proc. Natl. Shellfish .Assoc. 56:53-58. Chanley, P. & J. D. Andrews. 1971. Aids for identification of bivalve larvae of Virginia. Malacologia 1 1:45-1 19. Dickie, L. M. 1955. Fluctuations in abundance of the giant scallop, Pla- copecteii magellanicus (Gnielin) in the Digby area of the Bay of Fundy. J. Fbh. Res. Board Can. 12:797-857. Farrow, G. 1971. Periodicity structures in the bivalve shelkExperiments to establish growth controls in Cerasloderma ediile from the Thames estuary. Palaeontology 14:571-588. Gosner, K. L. 1978. A Field Guide to the Adantic Seashore Invertebrates and Seaweeds of the Atlantic Coast from the Bay of Fundy to Cape Hatleras. Boston, MA: Houghton Miftlin Co. 693 pp. Gulland. J. A. 1966. Manual of Sampling and Statistical Methods for Fisheries Biology. Part 1. Sampling methods. (FAO) Food Agric. Or- gan. New York: United Nations. 87 pp. Guo, X., S. E. Ford & F. Zhang. 1999. Molluscan aquacallure in China. J. Shellfi.'ih Res. 18:19-31. Haven. D. S.. J. P. Whilcomb & P. C. Kendall. 1981. The present and potential productivity of the Baylor Grounds in Virginia. Vol. II. James River, Pocomoke and Tangier Sounds, The Bayside and Seaside of the Eastern Shore, and the Virginia Tributaries of the Potomac River (Coan and Yeocomico Rivers, and Lower Machodoc and Noniini Creeks). VIMS Spec. Rpt. No. 243. 220 pp. Hrs-Brenko. M. 1980. Preliminary survey of populations of the bivalve Noah's ark (Area iioae. Linne) in the Northern Adriatic Sea. Acjiiaeul- tiire 21:357-363. Hurlbert. S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monogr. 54:187-211. Ismail, W. 1986. Preliminary study on sea water quality of Kamal waters. J. Mar. Fish Res. 35:89-94. Kayombo, N. A. 1993. Substrate grain-size analysis in culliired and natural populations of the edible ark clam Anadara antiquata (Linnaeus, 1758) on the Tanzanian coast. World Aqiiaciiltnre 24:68-71. Kennish, M. J. 1980. Shell microgrowth analysis:yWc/re/u7rio mercenaria as a type example for research in population dynamics, pp. 255-294. //i:Skeletal Growth of Aquatic Organisms, D.C. Rhoads and R.A. Lutz (eds.). New York and London: Plenum Press. Loosanoft. V. L. & H. C. Davis. 1963. Rearing of bivahe molluscs. .4i/i'. Mar. Biol. 1:1-1.36. Loosanoff. V. L., H. C. Davis & P. E. Chanley. 1966. Dimensions and shapes of larvae of some marine bivalve mollusks. Malacologia 4:35 1- 435. Luckenbach, M. & J. Taylor. (N.D.). Oyster gardening in Virginia: An overview of techniques. Publications of Virginia Institute of Marine Science, School of Marine Science. Gloucester Point, VA: The College of William and Mary. 1 1 pp. Lutz, R. A. & D. C. Rhoads. 1980. Growth patterns within the molluscan shelkan overview, pp 203-254. //rSkeletal Growth of Aquatic Organ- isms, D.C. Rhoads and R.A. Lutz (eds.). New York: Plenum Press. McGraw. K. A. & M. Castagna. 1994. Life history studies of the arkshell clams, N. ponderosa and A. ovalis and implications for fishery man- agement. Final report for Virginia Sea Grant College Program. Uni- versity of VA. Chariottesville. VA. Grant # R/MG-92-7. McGraw. K. A., M. Castagna & S. D. Dennis. 1996. Population structure of the arkshell clams Noelia ponderosa and Anadara oralis in the oceanside lagoons and tidal creeks of Virginia and implications for fisheries management. Final report for the National Oceanic and At- mospheric Administration, National Marine Fisheries Service, Glouc- ester. MA, NOAA Grant # NA46FD()339. 62 pp. McGraw, K. A., M. Castagna & S. D. Dennis. 1998. The arkshell clams, Noelia ponderosa and Anadara ovalis. in the oceanside lagoon system of Virginia:a study of predation, reproductive biology, and condition index. Final report for the National Oceanic and Atmospheric Admin- istration. National Marine Fisheries Service, Gloucester, MA, NOAA Grant # 66FD001(). 59 pp. Narasimham. K. A. 1969. Studies on some aspects of biology and fishery Arksheix Clams in Virginia Coastal Waters 195 of the cockle A. granosa (Linnaeus) from the Kakinada Bay. Pmc. Symi>. Motlusca. Men: Biol. Assn. India 2:407-417. Narasimhani, K. A. 1988. Biology of the blood clam /^. gninuM l Linnaeus l in Kakinada Bay. / Mar. Biol. Assn. India 30:137-150. National Oceanic and Atmospheric Administration. National Marine Fish- eries Service. March 2000. Fisheries Statistics. Nie. Z. Q. 1990. The culture of marine bivalve mollusks ni Chma. In: R. W. Menzel, editor. Estuarine and niaruic hi\alve culture. Boston. MA: CRC Press. Inc. pp. 261-276. Pannella. G. & C. MacClintock. 1968. Bioloj;ical and environmental rhythms reflected in molluscan shell growth. / Fali'onlol. 42:64-8(1. Rhoads, D. C. & G. Pannella. 1970. The use of inolluscan shell growth patterns in ecology and paleoecology. Letluiia 3:143-161. Ricker. W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191. 382 pp. Rigby. J. K. & D. L. Clark. 1965. Section E. Casting and molding. P. 407. In: B. Kummel & D. Raup, editors. Handbook of Paleontological Tech- niques. San Francisco. CA: W.H. Freeman and Co. Robson. D. S. & D. G. Chapman. 1961. Catch curves and mortality rates. Tran.s. Am. Fish. Soc. 90:181-189. Ropes. J. W. 1984. Procedures for preparing acetate peels and evidence validating the annual periodicity of growth lines formed in the shells of ocean quahogs. Arclica islandica. Mar. Fish. Rev. 46(2):27-35. Ropes, J. W. 1987. Preparation of acetate peels of valves from the ocean quahog. Arctica islandica. for age determinations. NOAA Tech. Rpt. NMFS 50. 5 pp. Ropes. J. W. & L. O'Brien. 1979. A unique method of aging surf clams. Bull. Am. Malacol. Union Inc. pp. 58-61. Sahavacharin. S.. A. Chindanond, S. Amornjaruchit, J. Nugranad, K. Si- lapajarn. V. Chawivanskorn & S. Limsurat. Hatchery techniques for tropical bivalve molluscs. 1988. In: E. W. McCoy & T. Chonpeepien, editors. Bivalve Mollusk Culture Research in Thailand. ICLARM Tech. Rep. No. 19. pp. 19-30. Scheaffer. R. L.. W. Mendenhall III & L. Otl. 1996. Elementary survey sampling. Fifth edition. Belmont. CA: Duxbury Press. 278 pp. Sokal. R. R. & F. J. Rohlf. 1969. Biometry. San Francisco, CA: W.H. Freeman and Co. 776 pp. Terry, K. S. H.M. Terry Co. 1991. Letter to Mr. William Pruitt, Chairman. Virginia Marine Resources Commission, Newport News, VA, 23607. Ting, Y. Y. 1984. Shellfish culture in Taiwan. In: I. C. Liao, R. Hirano, editors. Proc. of Roc-Japan Symposium on Mariculture, Dec. 14-15, 1981, Taipei, Taiwan, Roc. TML Conf. Proc. 1:129-142. Toyo, T.. I. Tesuji & N. Inoue. 1978. The mass culture of ark — Anadara and their problems in Yamaguchi Prefecture. Cull. Res. 7:51-66. (as cited in Broom, 1985). Walker, R. L. 1998. Growth and survival of the blood ark. Anadara oralis (Bruguiere, 1789), in coastal Georgia. Georgia J. Sci. 36:192-205. Wong, T. M. & T. G. Lim. 1985. Cockle (A. granosa) seed produced in the laboratory, Malaysia. ICLARM Newsl. 8(4): 13. Yonge, C. M. & T. E. Thompson. 1976. Living Marine Molluscs. Glasgow: Wm. Collins and Sons, Ltd. 287 pp. Zar, J. H. 1974. Biological Analysis. Englewood Cliffs, NJ: Prentice-Hall. Inc. 620 pp. .hmrihil i>t Shellfish Research. Vol. 20. Nci. 1, 197-206, 2001. A DIGITAL PRESENTATION OF THE MARYLAND OYSTER HABITAT AND ASSOCIATED BOTTOM TYPES IN THE CHESAPEAKE BAY (1974-1983) G. F. SMITH, K. N. GREENHAWK. D. G. BRUCE, E. B. ROACH, AND S. J. JORDAN Cooperative Oxford Laboratory. Maryland Departiuent of Natural Resources, 904 S. Morris St., Oxford, Man-land 21654 ABSTRACT Between 1975 and 1983. Ihe Maryland Department ol Natural Resourees conducted a survey of Maryland's portion of the Chesapeake Bay to reassess the extent and condition of oyster bottom (bars) that were initially surveyed in 1912. A variety of methods were employed to assess bottom condition, including bottom grabs, sounding poles, and dragged microphones. The survey used six categories to describe substrate type. Three of these represented oyster bottom: cultch (exposed oyster shell), cultch with sand, and cultch with mud. The other three categories were non-oyster bottom: sand, mud. and consolidated hard sediment. Bottom characterizations were drawn on 37 transparent Mylar sheets. Although the resulting charts were used to generate new legal oyster bar boundaries, the original bottom characterizations were never published. We present here the digitized Mylars in a Geographic Information System (GIS) polygon format, which is now available as a digital file. A comparison of results with more recent habitat assessments indicate the survey was a generally valid representation of the true bottom condition. The digitized survey results, however, could be misapplied if used as an exact description of current bottom conditions, because in many areas, cultch categories are actually shell covered by varying depths of sediment. This work provides a demonstration of how GIS-based analysis can be employed to interpret historical surveys of oyster habitats and provide a useful management product to resource managers. KEY WORDS: oyster, habitat. Chesapeake Bay. GLS. cultch INTRODUCTION Fisheries management and ecological restoration of oyster habitat are generally obliged to rely upon old original bottom surveys, often over l(X) years old. These surveys covered broad areas but relied upon crude methods. In the Chesapeake Bay. which lies within the boundaries of Maryland and Virginia, broad scale maps of oyster {Crassostrea virginica) habitat were con- structed in the late 19th and early 20th centuries. In Virginia. oyster habitat was mapped by Baylor (1894). The first official survey of Maryland oyster bars was referred to as the Yates survey. Completed in 1911 (Yates 1911. Smith 1997). it was authorized in 1906 (Kennedy & Breisch 1983) to delineate oyster bottom to be reserved for public oyster harvest. A dragged chain was employed as the principal sampling method for the survey. The survey was in response to declining oyster harvest which went from 686,700 m' of live oysters in the shell in 1884 to 201.432 m' in 1906. despite increased harvest effort (Graves 1912). In the years fol- lowing the Yates survey, much bay bottom was added to the public oyster ground by judicial decree. By the 1970s, there was increas- ing opinion that much of the original Yates-designated public bot- tom had become unproductive and should no longer be reserved exclusively for harvesting of oysters from public beds. Alternative uses were proposed for the public bottom, which by 1974, pro- duced only 128.184 m' (MD-DNR 1989) of oysters annually (8,294 metric tons. NMFS 2000). Fishers interested in harvesting soft clams {M\a areiieria). in particular, called for a re- examination of the original survey to determine which areas were still productive public oyster beds, which areas could be desig- nated for clatnining. and which could be available for private oyster leases. In 1974. the Maryland State Legislature authorized the Maryland Bay Bottom Survey (MBBS). The original mandate for the survey was to delineate the natural oyster bars of the State, the soft clam areas of the State, and what will be termed barren bottoms of the State, which could be leased to private individuals for oyster culture or other aquaculture purposes. The challenges presented by the need to map oyster habitat and track historical changes arc formidable. In Maryland, survey meth- ods have varied greatly since the Yates Survey. The MBBS em- ployed a combination of mechanical grab samples and hydro- acoustic techniques to classify bottom types. The output from the MBBS were bottom-type renditions, represented as polygons, drawn on a series of transparent Mylar sheets. Each transparency matched the scale of one of the 37 official State of Maryland Natural Oyster Bar (NOB) charts (MD-DNR 1983) that depicted the boundaries of Maryland's public oyster bars. Although they were drawn at the scale of Maryland oyster bar charts, no reference information, such as latitude and longitude or shoreline, was de- picted on the sheets. As such, they were extremely difficult to utilize in their original format. Because these renditions were not geographically referenced, and because they were gross simplifi- cations of the infomiation generated in the survey, they fell short of providing a comprehensive tool to tiianage the resource and track historical change. Instead, the principal use of the MBBS transparencies was to produce new legal oyster bar boundaries. In conjunction with the older Yates survey, the primary criteria for designation of new oyster bar boundaries was the distribution of cultch (exposed oyster shell )-bearing bottom. Because of the physical complexity of the bay bottom, new legal oyster bar boundaries were negotiated gross simplifications of bottom de- picted by the MBBS. The first of the new NOB boundaries became effective in 1 982. After serving as a basis for the generation of new official NOBs. the 37 Mylar sheets depicting the MBBS were archived. In 1994. the Chesapeake Bay Stock Assessment Committee provided funding to digitize Ihe original MBBS data into a GIS format for future public dissemination. GIS technology allows these data to be integrated with other spatially related datasets, such as bathymetry, for purposes of resource assessment. This paper was prepared in association with the introduction of the MBBS digitized dataset to the public. Because no detailed and integrated documentation of this survey was ever prepared, this review serves as a description and analysis of survey methods and results. In an effort to track historical changes in the oyster re- 197 198 Smith et al. source and habitat, and to provide for more effective management and restoration activities, it can be seen from tfiis effort that there is a need to develop quantitative tools for comparing historical and current or future surveys of oyster habitat. From this process, we have identified approaches to applying quantitative Geographic Information System (GIS) analysis and display to plan for new approaches to oyster habitat mapping. METHODS Survey Overview The survey began in 1975 and was expected to require approxi- mately 3 years to complete. The original objectives stated that the entire Maryland area of the Chesapeake Bay would be surveyed, a departure from the original Yates Maryland oyster bar survey (Fig. 1 1, where local "experts" identified potential oyster bottom for survey (Graves 1912). Because of the slow pace of work, the scope of the survey was later changed to assess only bottoms within and adjacent to the public bars identified by the Yates survey. Navigation was principally by Raydist radio receivers (Tele- dyne Hastings. Hampton. VA.) Transmitting stations were estab- lished especially for the MBBS. and Raydist control sheets were prepared for boat use by overprinting existing natural oyster bar charts with overlapping hyperbolic Raydist lanes. For most sam- pling cruises, tracks directly followed Raydist lanes, thus transect bearings varied throughout regions of the bay. Sextant triangula- tion was used when electronic navigation failed. Bottom sampling methods varied throughout the survey and included combinations of patent tong grabs (Chai et al. 1992). sounding poles, echo sounder tracings, and sounds from a dragged microphone. Patent tongs are a commercial oyster harvest device employing toothed jaws and a steel-framed enclosure capable of retaining oysters of commercial size. Patent tongs collected mate- rial from 1 m' surface area. Individual sample depth was variable depending upon substrate composition. It is estimated that approxi- mately \5'/'r of the survey was conducted with patent tongs, 10% with sounding poles, and the remaining 75'7r with microphones. Table I provides phases of the survey in which particular sampling methodologies were employed. Maximum MBBS survey depth was about 9 m because of limitations of patent tong sampling; boat draft limited minimum surveying depth to about 2 m. All field surveying, with the exception of the Potomac River. Figure 1. Comparison of the area covered by the original Yates oyster bars (black) to the region of the Chesapeake Bay covered by the Maryland Bay Bottom Survey (gray). Yates bars were surveyed in 1911 to delineate grounds for public oyster harvest. The boxed region is the lower Choptank River repre.sented in Figure 2. Chesapeake Bay Oyster Habitat 199 TABLE 1. A chronology of sampling methods used by Maryland Bay Bottom Survey". Method Transect lane spacing Station spacing along lanes Data collected Estimated km'' sampled/day Phase I-early 1975 to late 1976 Patent tongs (I nr). bottom profiling recording falhometer. and sounding poles Phase ll-late 1976 to early 1977 Recording fathometer only'' Phase Ill-summer 1978 to late 1979 Dragged microphone' with shipboard transcription, recording fathometer, and sounding poles Phase IV-February 1980 to May 1980 Dragged microphone recorded shipboard with transcription on shore Phase V-June I9S() to .No\ember 1983 Dragged microphone analog output recorded on fathometer paper trace 137-183 m 137-183 m 183 m 183 m 1 S3 m 137-183 m 137-183 m I S3 ni Continuous Continuous Bottom type and presence of benthic organisms Three bottom types: hard. soft, unknown Seven bottom types: sand. mud. shell/sand, shell/mud, shell, clay, and gravel Six bottom types: sand. mud. shell/sand, shell/mud. shell, hard bottom Six bottom types: sand. mud. shell/sand, shell/mud. shell, hard bottom 1.42 2.02 1.62 11.33 11.33 ■' Adapted from an anonymous and undated table titled "History of B.B.S. in Maryland" found in Maryland Department of Natural Resources files. •^ Survey speed was to be increased by eliminating surveying in "nonproductive areas" and to use echo-sounding as an "exploratory/sample investigation." It was hoped that rapid preliminary reconnaissance with the bottom sounder and sounding pole could identify cultchless substrate that would not require exhaustive patent long sampling. Reconnaissance data were apparently used as a basis to focus where more exhaustive sampling would be employed at a later date. ' This instrument was a waterproofed microphone dragged along the bottom that transmitted sound to an amplifier and speaker on the survey boat. Although specifications for the exact design are not available, ihe configuration described by Haven et al. ( 1979) would have operated in a similar fashion. Event marks were placed in the location record where there were marked changes in the sound produced. was completed b> October 1982. Survey of the Potomac began in July 1983 and was completed in November of 1983. This portion of the survey was conducted independently of the MBBS. because the Potomac River fishery is jointly managed by the States of Virginia and Maryland. Oyster bars on both shores of the river were surveyed, but oyster bars in Virginia tributaries were not charted. Preparation of bottom charts continued until the end of 1983. The total urea surveyed by the MBBS was 2.^91.4 km' and covered 42% of the Maryland portion of the Chesapeake Bay. A total of 7.543 polygon objects were used to depict bottom type (Fig. 2). Of the total area surveyed, 1,858 km" was outside of the original charted Yates bars. In the total survey. 477.7 km" of TABLE 2. Summary of bottom-type polygons digitized from original Maryland Ba\ Bottom Survey Charts. Bottom type No. polygons Area (knr) Sand and shell 1.999 201.9 Shell 1.958 477.7 Sand 1.177 838.5 Mud and shell 1.155 205.2 Mud 845 733.4 Hard bottom 409 135.2 Total 7543 2591 » bottom was cultch. 201.9 km-" sand v>,ith cultch. and 205.2 kni" mud with cultch (Table 2). The CIS generated results included 0.575 km" of designated leased bottom, an attribute characteriza- tion for which habitat classification is not available. Digitization We digitized MBBS MyUir charts indi\ idually with Maplnfo® software and then merged them into one datafile. The Maplnfo projection and datum used was Longitude/Latitude (NAD27). Twelve to 15 control points were used to register the map to Earth coordinates. Coordinates of control points were georeferenced by overlaying the MBBS Mylar over Ihe coiTesponding Maryland Natural Oyster Bar chart that had a graticule of latitude and lon- gitude. Control point error values were all less than 0.05 cm root mean square in true chart dimension. Bottom type on MBBS my- lars was represented as inked shapes, shaded to depict six discrete bottom categories. These categories were: cultch. sand with cultch, mud with cultch, sand. mud. and consolidated "hard" sediment (believed to be clay). Hand-drawn, smooth-sided delineations on the Mylar sheets were transformed to straight-sided polygons in the digitization process. We attempted to duplicate the original curved-sided objects as closely us possible by utilizing enough segment nodes to make the straight-sided polygons appear curved at normal magnifications of the CIS image (Fig. 2). Final data cleaning was performed to identify sliver polygons where two objects overlapped. This and other cleaning procedures were per- formed in Arclnfo® utilizing the "build" and "clean" functions. 200 Smith et al. Kilometers ^^ Figure 2. An example of the digitized rendition of tlie Maryland Bay Bottom Survey in the Choptaniv River region. The shoreline and the borders of the original Mylar charts are layered upon the survey bottom themes, but are not included in the digital file. Original Mylar transparencies were 70 x 111 cm, drawn at a scale of 1:2II,(IIM( and projected in V. S. State Plane NAD27. The general North-East / South-West orientation of bottom themes depicted here is the result of radio beacon navigation. One of the original 37 Mylars, (southern East Bay) was missing from the set before digitization. Summary statistics do not inckide this lost Mylar, and the charted area it occupies in the digital dataset is empty. Validation of Habilal Characlerizalioii Comparative Analysis To determine the relationship between categorical bottom types of the MBBS depictions and raw patent tong data from which they were generated, we obtained original field data from 1975 for two oyster bars. Cook's Point and Sandy Hill (Fig. 3), in the Choptank River. These oyster bar locations were chosen from a limited set of raw data still available from the original survey. We georeferenced and then superimposed this raw categorical bottom type data over the MBBS habitat depictions. Categorical bottom types identified on the patent tong datasheets included shell, sand and shell, mud and shell, sand, mud, mud and sand, .stone, and clay. These are consistent with the MBBS bottom objects for these same locations. In addition to categorical bottom type data, survey sheets gave the percentage of the patent tong sample by volume, which was cultch. To validate further how final survey bottom classification were made, we evaluated the association between cultch quantity in patent tong samples and MBBS habitat categories present at the two oyster bars. Data for individual samples from both bars were pooled, and a square root transformation on cultch percentages was perfomied after an evaluation of residuals. Data were subjected to a one-way analysis of variance ( ANOVA). with bottom type as the experimental variable and percentage cultch as the response vari- able (a = 0.05). A Tukey's mean comparison test (HSD) was then performed at a = 0.05. We also compared 1975 MBBS patent tong samples from Sandy Hill and Cook's Point to unpublished (M. Homer pers. comm.) 1993 patent tong samples obtained from the same bars by the Maryland DNR Oyster Stock Assessment Program (OSAP). Unlike MBBS datasets. OSAP sampling was uniform and com- plete within the entire Yates bar boundary. OSAP data were re- corded as the nuinber of liters of surface shell in each sample; whereas MBBS data were recorded as the percentage of shell in each sample. Because the two sampling methods were not in com- parable units, nor convertible, cultch densities for both percentage Chesapeake Bay Oyster Habitat 201 Fifjurt 3. Selected diver observation and sediment analysis sites located on Maryland Bay Bottom Survey ciiltch designated bottoms in the Choptank River. Results of bottom analyses from sites A-H arc presented in Table 3. Yates bar boundaries outline Cook's Point and Sandy Hill oyster bars. of sample and liters per sample were each standardized to fre- quency distributions with a mean of 0 and an SD of 1 . Patent tong data from OSAP sampling at Cook's Point and Sandy Hill were also used to validate the shape of MBBS cultch objects. We created a continuous surface from OSAP data by importing point datafiles into an Arclnfo grid cell model. Cultch density patterns in the grid were then compared to the shape of cultch objects in the MBBS. We used SCUBA dives in 1998 to perform qualitative assess- ment of the abundance of surface shell, and subsurface composi- tion by probing with a diver knife, at nine locations designated as cultch in the MBBS (Fig. 3). At each site. SCUBA divers also collected one to three two-liter sediment samples to a depth of 13 cm. Samples were weighed and then sieved through 33-mm, 10- nim, and 5-mm mesh to determine percentage of shell by weight in each sample and the composition of nonshell material. Material passing through the 5-mm mesh was then subjected to traditional grain size analysis down to 16 |j.m (ASTM 1990). An underwater benthic sled with color and B/W cameras was towed across the nine locations from where we collected the sedi- ment samples. The direction of the video observation was variable, largely based upon wind and current condition. But, the exact path traversed was logged directly on the video tape by differential GPS signal. Sub-Bottom Profiling We used acoustic sub-bottom profiling systems to determine the relationship between MBBS cultch objects and Yates bar boundaries to oyster bar geomorphology. Transects were taken in January 1998 with an X-star® sub-bottom profiling system (Edgetech Inc.. Milford, MA. USA) on and adjacent to Sandy Hill and Cook's Point oyster bars. Sub-bottom profiling output was a GPS logged gray-tone trace that identified bottom depth and the presence and depth of sub-bottom density discontinuities. Because output was GPS logged, we were able to associate profile features with MBBS objects. RESULTS Comparative Analysis At Cook's Point and Sandy Hill oyster bars, only four of the six MBBS bottom classifications were recorded (Figs. 4a. b). Survey bottom polygon characterizations generally corresponded to raw patent tong bottom-type categorical data at Cook's Point and Sandy Hill bars. Seventy-eight percent of patent tong sample points were bounded by corresponding bottom type polygons at Cook's Point and 71% of points corresponding to like-bounding polygons at Sandy Hill. One apparent discrepancy in MBBS bot- tom characterization at Cook's Point was that patent tong datapoints characterized as sand with cultch or mud with cultch were depicted within bottom- type objects defined as cultch. Simi- lar inconsistencies were found at the Sandy Hill site. Mean percentage cultch varied significantly among the five categorical bottom types at the two sites. Tukey's mean compari- son test indicated that mean percentage shell was significantly lower in the mud and sand bottom designations than in the cultch and sand, cultch and mud, and cultch categories (Fig. 5). Cultch density differences among the cultch, sand with cultch. and mud with cultch categories were not significant. These results indicate that the MBBS used qualitative assessment methods, and not strictly quantitative, to categorize cultch bottom types. Analysis of OSAP data at Cook's Point and Sandy Hill pro- vided little insight on the accuracy of MBBS bottom classifications and shapes of chart objects. Standardized cultch densities from the two surveys had differently shaped frequency distributions (Fig. 6). Cultch distributions at Cook's Point and Sandy Hill oyster bars 202 Smith et al. ■ (C) - cultch ^ (M-C) - mud and cultch ^ (S-C) - sand and cultch M □ (M) - mud M (S) - sand N 1 Figure 4. Relationships between patent tong point data and final bottom-type polygons of the Maryland Bay Bottom Survey. Point data are layered over bottom themes within Yates Bar boundaries at (a) Cook's Point and (b) Sandy Hill oyster bars. Characters represent substrate type: C = cultch, SC = sand and cultch, MC = mud and cultch, S = sand, and M = mud. Irregular spacing of datapoints within the Sandy Hill \ ates bar boundary are attributable to lost data sheets. from 1975 (MBBS) indicated a majority of low-density samples, with some high-density samples being present. Later 1993 (OSAP) data was highly skewed to low-density samples, with no high- density samples collected. Differences in shell density distribu- tions indicate that either MBBS patent tong sampling was impre- cise, relative to OSAP sampling, or that there has been a temporal decline in cultch density at the two oyster bars o\er the inter\ening two decades. We did not observe any relationships between shell density patterns from OSAP grid cell models and the shape of MBBS cultch polygons. Multiple displays at various shell density levels used to depict cultch did not match MBBS cultch object shapes at either of the two oyster bars. Results of diver observations, video footage, and sediment analysis at selected sites in the Choptank River indicate that much of the bottom described as cultch in the MBBS currentlv has little surface shell (Table .^). Divers, however, noted dense shell below surface sediments at all but one location. These results indicate that much of the bottom depicted as cultch in the MBBS is cunently degraded by sedimentation or that some of the methods used in the MBBS were unable to discern sedimented shell from clean shell. Sub-Bottom Profiling We found that the physical perimeters of oyster bars generally occur at points where hard sub-bottom terraces emerge from sur- face layers of soft sediment or at main channel margins. These features are identified on sub-bottom profiles by acoustic density discontinuities or rapid bathymetric change. Borders of MBBS cultch objects and Yates bar boundaries at Cook's Point and at Sandy Hill generally coincide with these features (Fig. 7). In most cases MBBS cultch boundaries were more accurate in depicting Chesapeake Bay Oyster Habitat 203 C SC MC M S Bottom Category Figure 5. Mean percentage cultch from patent tong samples relative to corresponding bottom classiflcations. Patent tong data were pooled from collections made at Cook's Point and at Sandv Hill ojster bars. We used analysis of \ariance (ANO\ A) and ruke>"s multiple com- parison lest {a - (1.(15) on square root transformed data to assess the \aliditj of categorical bottom designations. Mean percentage cultch at bottom t>pes with the same letter were not signillcanlly dilTerent. Bottom categories are: C = cultch. .SC = sand and cultch. MC = mud and cultch, M = mud, S = sand. true oyster bar margins than the original Yates bar charts. The generality of Yates bar boundaries results from the purpose of the survey, which was charting protected public oyster bottom rather than identifying bottom types. DISCUSSION The 1475 to 1483 Maryland Bay Bottom Survey was the first bay wide assessment of oyster habitat since the early 19()0s. The area of bottom covered was 342% larger than the portion of the bay designated as public oyster bar by the Yates survey of 1911. Unlike the Yates survey, the MBBS is a categorical description of bottom that is relevant to oyster ecology in addition to regulation of harvest. Our digital rendition of the MBBS survey increases the facility of analysis and display at local and regional scales. The digital format will also allow resource managers and interested parties to identify historic regions of oyster habitat and areas for rehabilitation. This dataset is available for distribution from the Cooperative Oxford Laboratory (Mapping and Analysis Project. Cooperative Oxford Laboratory. 904 S. Morris St.. Oxford. MD 216-54. (410) 226-0078). An understanding of the strengths and weaknesses of the MBBS should precede any serious application of these data. Figure 7 is an example of how the MBBS may be a more accurate de- piction of the true morphological character of oyster bars than the previous Yates survey. In many cases, the MBBS polygon bound- aries corresponded to oyster bar edges identified by sub-bottom acoustic profiles; whereas. Yates bar boundaries provided a more generalized representation. Comparison between bottom category and percentage cultch indicated that bottom classifications, at least where patent tong sampling was employed, were systematically determined from cultch volume and that there was a relationship between MBBS bottom category and cultch density. Little documented information exists about the methodology and consistency of the inethods used to translate field data to the final six bottom classifications. The limited information available from raw patent tong data at Cook's Point and Sandy Hill oyster bars (Figs. 4a. b) showed a relationship of sample data to the de- rived objects, but there were some discrepancies. These inconsis- tencies suggested that additional information was incorporated into developing bottom classification. Patent tong field sheets from Cook's Point and Sandy Hill show that data were recorded for bottom type and percentage cultch for upper, middle, and lower strata within each patent tong sample. Although bottom type and percentage cultch were, in many cases, different between strata from the same sample, there was no consistent pattern indicating whether sub-surface information was used to generate final bottom categories. The graphic representation of the MBBS was affected by data collection and translation procedures. A linear pattern of northeast&ndash'.southwest orientation can be seen throughout the charts (Fig. 2). This is an artifact of transect orientation along radio beacon lines, and the width of these patterns provides the scale of transect spacing. Because each of the 37 Mylar sheets were drawn 40 20 in 0) a. E 10 CO (1) o 5) 50 a. Cooks Point Sandy Hill -0 505 15 30 45 60 -0 505 15 30 45 6 0 MBBS OSAP Standard Deviations of Cultch Density Figure 6. A comparison of cultch densities from 1975 patent tong samples recorded on Maryland Uav Bottom Survey (MBBS) field sheets to patent tong samples collected in 1993 at Cook's Point and Sandy Hill oyster bars. Recent patent tong samples were collected for the MD DNR Oyster Slock Assessment Program (OSAP) and were recorded as liters of shell per sample; whereas, MBBS data were re- corded as percentage shell per sample. To compare absolute and rela- tive units, both datasets were fit to standard frequency distributions with a mean of II and an SI) of 1. 204 Smith et al. independently of each other, objects at chart edges did not merge precisely, and we were forced to adjust most junctures when we integrated individual charts into the final GIS data file (Fig. 2). For some information requirements, the MBBS may not be a realistic depiction of the conditions of the Chesapeake Bay bottom in the 1970s and 1980s. Most bottom areas observed in the me- sohaline portion of the bay are transitional, grading between the si.\ generic categories of the MBBS. Intermediate bottom types often predominate over discrete bottom types, and many areas of the bottom, particularly where cultch is present, exhibit extreme het- erogeneity. The categorical polygon data model used does not represent transitional areas, because it depicts an abrupt change between bottom types. Furthermore, depictions of bottom, are for the most part, characterizations assembled from sampling points along a transect line, thus fine scale heterogeneity is not adequately represented. Our diver and video observations on the bottom in the vicinity of Cook"s Point and Sandy Hill oyster bars show extreme patchiness of cultch and sediment at centimeter and larger scales. An additional complication is that we do not know exactly what the '"hard bottom" MBBS bottom habitat classifications means. The tenn "deep water" has been associated with it in some refer- ence notes, although the survey did not extend into waters deeper than 9 m. Our visual observations in hard bottom areas often show a clay-like substrate. Current Relevance Most of the MBBS survey was conducted mainly with a dragged underwater microphone, and identification of cultch was generated by contact and abrasion on hard objects. Little informa- tion, however, is available to determine the accuracy of MBBS microphone data. Verification sampling via patent tong and di\'er observation to validate microphone acoustics was initiated toward the end of the survey, but results cannot be located. A principal source of error associated with the microphone method is that oyster shell is not the only hard substrate on the bottom. We have observed pebbles, cobbles, and occasionally boulders from video and diver observation in the Choptank River region. Some of these stony bottoms often include scattered or dense shell and were characterized as cultch in the survey. We knov\ that the micro- phone distinguished smooth from rough bottom and could identify transition zones between the two. However, the presence and den- sity of cultch may be confounded by other hard objects. The use of an underwater microphone for surveying was not novel. From 1978 to 1981 the public oyster grounds of the James River in Virginia were similarly surveyed (Haven et al. 1979. Haven & Whitcomb 1983), with data being reduced to six bottom charac- terizations identical to those later employed by the MBBS. Acoustics were also employed in the mapping of Galveston Bay. Texas beginning in 1991 (Powell et al. 1995). In this case, dual frequency transceivers (27 and 300 kHz) were utilized to compare results to data collected in the 1970s by a manual polling technique. Polling was additionally employed as ground truthnig during the recent survey. Principal oyster habitat distinctions em- ployed were between hard consolidated reef and unconsolidated shell. GIS data generation were employed principally to generate polygon areas of these categories. Unlike the Maryland Bay Bot- tom Survey, wholesale habitat type mapping was not a component of the published results. Our analysis of sediment samples, qualitative diver observa- tions, and video footage, at areas characterized as cultch by the MBBS showed small amounts of surface cultch and an almost universal presence of shell beneath surficial sediments. Trace quantities of exposed cultch or no exposed cultch was the domi- nant surface character. This situation is true for almost the com- plete extent of the large MBBS cultch object depicted at the top of Cook's Point oyster bar (Fig. 7). Video transects over tens of meters did not yield any surface cultch. These observations seem to agree with results of the OSAP patent tong-sampling program from the early 1990s. From the OSAP data, we calculated a mean shell volume of 1.6 x L x m""" from samples collected within all MBBS cultch, sand with cultch, and mud with cultch objects within Yates bar boundaries for the entire Choptank River region. When completely filled, patent tongs sample 52.0 Im' of substrate in each grab. The standardized rank comparisons in Figure 6 also indicate that much of the bottom depicted as cultch in the MBBS TABLE 3. Recent substrate examinations at sites identifled as cultch bottom by the Maryland Bay Bottom Survey 1975-1983. Site Bottom category Qualitative visual estimates Mean shell abundance (<7f bv weight) Shell abundance i.'7<) Nonshell" substrate type Subsurface composition A B C D E F G H 1 Cultch Cultch Cultch Cultch Cultch Cultch Cultch Sand/cultch Mud/cultch 7 5" 6 g 9 0 16 0 0 0 50'' 1 15 1 1 0 1 I) 4 2,3,5,6,7 1,2,3,5,6,7 4,6 3,4,7 7 4 3,7 4 Dense shell Shell/rock Shell/rock Dense shell Dense shell Light shell Dense ^hell Dense shell Mud ■' Substrate codes: I = boulder. 2 = cobble. 3 = grit. 4 = mud, 5 = pebble. 6 = shell hits. 7 = sand. '' Difference in shell abundance due to extreme bottom heterogeneity at site. Sites were selected to be representative of the widest variety of bottom types observed \n the Choptank River estuary of the Chesapeake Bay. One to three replicate sediment samples were collected by divers at each site in summer of 1998. Shell abundance was measured as percentage by weight of shell >30 mm from 2 L samples taken to a depth of 15 cm. Visual estimates were recorded from qualitative diver observations of surface shell abundance, nonshell substrate composition, and subsurface composition determined from probing with a diver's knife. Site locations are depicted in Figure 3. Chesapeake Ba\ Oyster Habitat 205 BBS2 YB2 f *.... YB1 f OBM Emergent Terrace BBSl "fe BBS2 YBI'^ 1 YB2 W -- E OBM Emergent Terrace 5 m 0.1 km West East West Fisurt 7. Aidu^Ul sub-lxitUmi prulllfs taken at Cook's Point (al and Sandy Hill (b) oyster bars. Arrows indicate where boundaries of Maryland Bay Bottom Survey cultch objects (BBSl and BBS2) and Yates bar perimeter boundaries (VBl and Mil) intersect sub-bottom profiling transects. Inserts show where transects cross MBBS cultch objects and Yates bar boundaries. Actual oyster bar margins (OBM) occur where hard sub-bottom terraces emerge from soft bottoms or at the edges of existing channels. The western margin of Sandy Hill occurs at the edge of a sediment-fdled paleochannel of a tributary of the Choptank River. The dark feature in the channel is the signature of gas-charged sediments. is cuiTenth degraded by an oxerburden of sediment. We cannot determine whether this is a recent phenomenon or existed in the mid-1970s to 1980s when the MBBS survey was performed, and methods used by the MBBS may have been biased against detect- ing sediment on shell. Furthermore, the underwater microphone may have been unable to distinguish clean surface cultch from cultch covered by a layer of sediment. Also, sampling the bottom with patent tongs allows any overlying sediment to be washed from shell as the sample is brought up through the water column. Caution must be used in how the results of the MBBS are applied. Comparison between the Yates survey and MBBS charts have been used to estimate temporal decline of oyster habitat (Rothschild et al. 1994). Although this study implicates that over- harvest has an impact on the oyster habitat, comparative methods between the Yates boundaries and MBBS polygons did not include the MBBS bottom category mud with shell as oyster bottom. Aerial summations were only made with the shell and sand with shell categories, thus potentially greatly inflating habitat loss es- timates (Table 3). In addition, we have shown that Yates bar boundaries may provide coarser descriptions of the distribution of shell resources than the MBBS. This is because the Yates survey was intended to map legal boundaries of public oyster bottom and not necessarily shell distribution. MBBS cultch objects reveal the distribution of shell on the bottom, but say nothing about the quality of oyster habitat. Because oyster veligers require a clean surface for settle- ment, sediment-free shell is one of the most important components of quality oyster habitat (MacKen/ie 198.^1. Patent tong and un- derwater microphone methods used to assess oyster habitat may also be unable to identify clean shell from that can'ying a sediment load. To assess temporal changes in oyster habitat accurately, new methods must be developed to differentiate between clean shell and shell covered by an overburden of silt (Smith and Greenhawk 1998). The digitization of survey results and GIS analysis of data has shown the value of GIS as a tool for the examination and analysis of benthic data. This complete dataset of over forty chart size Mylars can now be readily examined at ease on a computer screen in association with such related datasets as shoreline, bathymetry, and recent or planned sonar surveys. Because of the recent integration of acoustic habitat assessment technology with extremely accurate satellite positioning informa- tion and GIS data capture and display, new possibilities are avail- able to chart cost effectively and computer generate extensive bottom surveys. The cumbersome nature of data reduction per- formed in the MBBS generation can now be eliminated. Benthic habitat charting \ ia remote acoustic methods now pro- vides three principal technological approaches for the assessment and representation of the bottom. Traditional single beam sonar can be used to assess general surface characteristics or sub-surface features. Habitat classification with such systems is still largely subjective. Side scan sonar technology can provide for strip-like textural images of the bottom that can now. by means of computer mosaicking, be seen as large, two-dimensional images. .Although 206 Smith et al. often highly informative as to bottom character at a fine resolution, these side scan systems are still imaging reflectivity of sound only and often require extensive ground truthing. Finally, new technol- ogy referred to as Acoustic Seabed Classification Systems (ASCS) has been developed that statistically classifies acoustic echo re- turns by waveform characteristics into definable habitat types. With ASCS. rapid bottom classification can be made for each sonar ping transmitted. Because reflected acoustic waveforms carry a multitude of substrate information, it may not originally be known just what parameters of the substrate are being used in classification. As with other acoustic techniques, these systems require extensive ground truthing. Despite technological advances, almost twenty years after its completion, the MBBS is still the best spatial depiction of over-all oyster habitat in the Mar>land portion of the Chesapeake Bay. The scale and effort of the undertaking was so great that it will prob- ably remain for some time as our best baseline information source on the distribution of shell resources. Despite weaknesses in docu- mentation of methodology, and sediment overburden in areas de- scribed as cultch, we feel comfortable that the survey is a good guide to the location of areas most suitable for oyster bar restora- tion and a general guide to sediment composition in Maryland's portion of the Chesapeake Bay bottom. ACKNOWLEDGMENTS This work was partially funded by the Chesapeake Bay Stock Assessment Cominittee. Maryland Oyster Fishery-Dependent. Fishery-Independent, and Rehabilitation Data: Integration of Da- tabase and Sampling Programs. (Submitted April 1993). S. J. Jor- dan and M. L. Homer, principal investigators. G. F. Smith, asso- ciate investigator. We thank Kim Insley. Dorothy Jensen, Leslie Lyons, and Ann McManus of the Cooperative Oxford Laboratory. Mapping and .'\nal\sis Project, Fisheries Service. Maryland De- partment of Natural Resources, Oxford. MD for data preparation, digitization, and related tasks, and Howard Wienberg of the NOAA Chesapeake Bay Program Office, Annapolis, MD, for as- sistance in data verification. A special thanks to Roger Newell of Horn Point Laboratory. University of Maryland, for his persistence in review s of several drafts of the manuscript. LITERATURE CITED American Society for Testing and Materials (ASTM). 1990. Standard test method tor particle size analysis of soils. American Society for Testing and Materials (ASTM I. annual book of ASTM standards. Philadelphia: ASTM. Baylor, J. B. 1 894. Method of defining and location of natural oyster beds, rocks, and shoals, oyster records. Richmond: Board of Fisheries of Virginia. Chai, A.. M. Homer. C. Tsai & P. Goulletquer. 1992. Evaluation of oyster sampling efficiency of patent tongs and an oyster dredge. .V. Ant. J. Fisheries Manag. 12:825-8,^2. Graves, C. 1912. Fourth report of the Shellfish Commission of Maryland. Baltimore: King Bros. Inc. Haven. D. S.. J. P. Whitcomb. J. M. Zeigler & W. C. Hale. 1979. The use of sonic gear to chart locations of natural oyster bars in Lower Chesa- peake Bay. Proc. Nut. Shellftsheries Assoc. 69:11-14. Haven. D. S. & J. P. Whitcomb. 1983. The origin and extent of oyster reefs in the James River. Virginia. J. Shellfish Res. 3:141-I.'S1. Kennedy. V. S. & L. L. Breisch. 1983. Sixteen decades of political man- agement of the oyster fishery in Mary land's Chesapeake Bay. / Env. Manag. 16:53-171. MacKenzie, C. 1983. To increase oyster production in the Northeastern United States. Mar. Fish Rev. 45:1-22. Maryland Department of Natural Resources.(MD-DNR). 1983. Maryland State natural oyster bars. (MSNO) 1983. Maryland Department of Natural Resources- 1961. Base map prepared hy the Coast and Geo- detic Survey, date of final reissue-1985. MD-DNR. 1989. Maryland Department of Natural Resources. In-house harvest statistics. National Marine Fisheries Service. 2000. Webstlt.MF_ANNUAL_ LANDINGS. RESULT Powell, E. N.. J. Song. M. S. Ellis, and E. A. Wilson-Ormond. 1995. The status and long-term trends of oyster reefs in Gaheston Bay. Texas. J. Shellfish Res. 14:439-457. Rothschild. B. J.. J. S. Ault, P. Goulletquer & M. Heral. 1494. Decline of the Chesapeake Bay oyster population: a century of habitat destruction and overtlshing. Mar. Ecol. Prog. Ser. 1 1 1 :29-39. Smith. G. F. 1997. Maryland's historic oyster bottom: a geographic rep- resentation of the traditional named oyster bars. Maryland Department of Natural Resources. Fisheries Service. Cooperative Oxford Labora- tory. Mapping and Analysis Project. Smith. G. F. & K. N. Greenhawk 1998. Shellfish henthic habitat assess- ment in the Chesapeake Bay: progress toward integrated technologies for mapping and analysis. G. F. Smith, editor. J. Shellfish Res. 17 (5). Yates. C. C. 1911. Survey of the oyster bars [by county of the State of Maryland]. Department of Commerce and Labor-Coast and Geodetic Survey. Washington. DC: U.S. Government Printing Office. Ji,iinuil v) Shellfish Rfsearch. Vol. 20, No. 1. 207-213. 2001. DIFFERENTIAL DIAGNOSIS OF MIXED HAPLOSPORIDWM COSTALE AND HAPLOSPORIDIUM NELSONI INFECTIONS IN THE EASTERN OYSTER, CRASSOSTREA VIRGINICA, USING DNA PROBES NANCY A. STOKES AND EUGENE M. BURRESON* Viri;iiilii Instititte of Marine Science. Coilei^e ofWillicini and Mary. Gloucester Point. Viri^iiiia 23062 ABSTRACT Huplospundiiim aistulc and Haplospondnii}i ncl.siiiii are morphologically similar palhogens of the eastern oyster Crassostreci virginka. In the absence of the spore stage, infections of the two species are extremely difficult, if not impossible, to distinguish using traditional light niicro.scopy of stained tissue sections. Species-specific molecular diagnostics were developed for H. cosHile from the small subunit ribosomal DNA (SSU rDNA) sequence. The polymerase chain reaction (PCR) primers amplified a 557 base pair (bp) region of the H. costale SSU rDNA. but did not amplify DNA from oyster (C. virginica) or from six other haplosporidans (W. nelsuni. H. loiiisuimi. H. Iiisilanicum. Minchinia leredinis. M. chitonis. or M. tapeti.'i). The DNA probe was used with in situ hybridizations of oyster tissue sections to visualize H. costale Plasmodia and prespore stages; it did not hybridize with oyster (C. virginica) or other haplosporidans {H. nelsoni. H. louisiana. or Minchinia teredinis). DNA-based diagnostics for H. costale, in conjunction with molecular tools previously developed for H. nelsoni. have overcome limitations of histological examination. From in siiK hybridizations using both probes, some Virginia oysters previously diagnosed with H. ciisl(de were found to have mi.xed infections consisting of approximately 80 to Wi H. costale plasmodia and 10 to 20''* H. nelsoni plasmodia. Plasmodia of H. costale were not found in epithelial tissue, only in connective tissue. In addition, use of the DNA probe confirmed the presence of H. costale Plasmodia in Virginia oysters collected in the fall, an unprecedented seasonality for an advanced H. cosiale infection. KEY WORDS: in situ hybridization, small suhuiiit nhosomal DNA, Haplospondmm nelsoni. Haplospoiidium costale. eastern oyster, Crassostrea virginica. parasites INTRODUCTION Hapliispiiriiliuni nelsoni Haskin. Stauber, atid Mackiti (MSX disease) atid Haplnsporidiuin eostide Wood and Andrews (SSO disease) arc tiiorphologically similar pathogens of the eastern oys- ter. Crassostrea virginica Gmelin. that occur along the East Coast of the United States, Haplosporidium costale is generally thought to be restricted to high salinity bays (>25 ppt) along the open coast from Virginia to Maine; it is rare in the Delaware Bay and in the Chesapeake Bay (Andrews & Castagna 197S; Andrews 1988). Haplosporidium nelsoni occurs from Florida to Maine in both estuarine and oceanic habitats where the salinity is greater than about 10 ppt (Haskin & Andrews 1988). Thus, the distribution of the two pathogens overlaps in high salinity areas frotn Virginia to Maine. If spores are present the parasites are easy to distinguish be- cause H. nelsoni sporulates only in the epithelium of the digestive diverticula, whereas H. costale sporulates throughout the connec- tive tissue of most organs (Couch 1967. Andrews & Castagna 1978J, Moreover, spores of H. nelsoni are about twice the size of H. costale spores (Couch 1967). However, in the absence of spores, differentiation of the two parasites is very difficult, if not impossible. According to Couch (1967) plasmodia stages of both H. Jicl.soni and H. costale occur in epithelial and connective tissues in both mi.xed and single infections, so location of plasmodia is not helplul. Haplosporidium costale has a very restricted seasonality, with Plasmodia present from March through June and spore stages present during May and June (Andrews et al. 1962. Andrews & Castagna 1978). However, plasmodia stages of H. nelsoni may also be common during the spring (Andrews 1982). Morphology of Plasmodia has apparently been used to distinguish the species, with sotne difficulty. Couch and Rosenfield (1968) conducted a comparative study of H. costale and H. nelsoni in Chincoteague *Corresponding author. Bay, Virginia, They state that diagnoses of the two parasites in living oysters was based on recognition of the Plasmodium, but they do not give any criteria used to distinguish the plasmodia of the two species. Mixed infections of H. nelsoni and H. costale were observed during the same study (Couch 1967). but they were based on the presence of spores of both species. However, criteria for distinguishing plasmodia of H. nelsoni and H. costale were provided (Couch 1967). They included: nuclear membranes of H. costale usually not as sharply defined or distinct as those of H. nelsoni and nucleoli (endosomes) of H. costale nuclei proportion- ately larger, less distinct, more diffuse, and more central than nucleoli of H. nelsoni. Andrews and Castagna ( 1978) stated that all stages of H. costale average smaller than those of H. nelsoni. but they went on to say that no definitive characters have been found with Harris hematoxylin and eosin (HHE) stain to distinguish H. costale and H. nelsoni plasmodia. The specificity of molecular diagnostic tools, especially DNA probes used in ;;) situ hybridizations, inake thetii ideal for distin- guishing morphologically similar species. Such tools are invalu- able in elucidating certain ecological aspects of parasites that are difficult using traditional techniques (Burreson et al. 2000). Mo- lecular diagnostic tools have been developed for H. nelsoni (Stokes ct Burreson 1993. Stokes et al. 1995a). Specific polymerase chain reaction (PCR) primers have been developed for H. costale (Ko et al. 1995). but a DNA probe for that species has not been devel- oped. Here we develop molecular diagnostic tools for H. costale and use the DNA probe in conjunction with an H. nelsoni DNA probe to identify mixed plasmodial infections of the two species. In addition, the molecular tools provided unexpected new infor- mation on the seasonality of H. costale in Virginia. MATERIALS AND METHODS DNA Sequences and Oligonucleotide Synthesis The SSU rDNA sequences of H. costale. H. nelsoni, and C. virginica (GenBank accession AF387122. U19538. and X60315. 207 208 Stokes and Burreson respectively) were aligned using the MacVector software package (Oxford Molecular Group) and regions unique to H. costale were identified. PCR primers SSO-A (5'-CACGACTTTGGCAGT- TAGTTTTG-3') and SSO-B (5'-CGAACAAGCGCTAGCAG- TACAT-3') and DNA probe SS01318 (same sequence as SSO-B. 5' end labeled with digoxigenin) were commercially synthesized (Genosys Biotechnologies). PCR Amplification PCR reaction mixtures contained reaction buffer ( 10 niM Tris. pH 8.3; 50 mM KCl; 1.5 mM MgCi; 10 jjLg/mL gelatin). 400 p.g/mL bovine serum albumin. 25 pmoles each of SSO-A and SSO-B. 200 jxM each of dATP. dCTP. dGTP. dTTP. 0.6 units AmpliTfli^ DNA polymerase (Perkin-Elmer), and template DNA in a total volume of 25 |xL. The reaction mixtures were cycled in a GeneAmp PCR System 9600 thermal cycler (Perkin-Elmer) 35 times at 94"C for 30 sec. 59°C for 30 sec. and 72X for 1 .5 min with a final extension at 72°C for 5 min. PCR reaction mixtures and cycling conditions for H. nelsoni were identical, except the primers were MSX-A' and MSX-B (Renault et al. 2000. Stokes et al. 1995a). An aliquot (10% of reaction volume) of each PCR reaction was checked for amplification product(s) by agarose gel electrophoresis and ethidium bromide staining. PCR Specificity and Sensitivity Primer specificity was tested in PCR reactions using cloned SSU rDNA from H. costale, H. nelsoni. Haplosporidiiim louisiana Sprague. and Minchinia teredinis Hillman. Ford, and Haskin. and genomic DNA from Haplosporidinni lusitanicum Azevedo. Min- chinia chilonis (Lankester). Minchinia tapetis (Vilela). and unin- fected C. virginica. Preparation of the cloned SSU rDNAs were described previously (Stokes et al. 1995a). Hatchery-reared juve- nile C. virginica were collected in July 1999. and genomic DNA was tested for the presence of H. nelsoni by PCR. as described previously (Stokes et al. 1995a). Limpets. Helcion pelhicidiis. were collected from Cap de La Hague, near Cherbourg. France in September 1998 and screened for the presence of H. lusitanicum spores. Chitons. Lepidochitona cinereiis, were collected from Wembury Bay. near Plymouth. England in September 1996 and screened for the presence of M. chitonis spores. Minchinia rapetis- infected clams. Riuiitapes decussatus (L.). collected from Vila- longa in the Ria de Arousa. Galicia. Spain, in 1997 were kindly supplied by Antonio Villalba. Spores were concentrated from in- fected tissues and DNA extractions from spores and from C. vir- ginica were performed with mechanical grinding followed by de- tergent lysis, as described previously (Stokes et al. 1995b). Primer sensitivity to homologous target DNA was determined with ten- fold serial dilutions from 100 pg to 1 fg of cloned H. costale SSU rDNA. Histology Tissue samples were preserved in Davidson's AFA for at least 24 h. Fixed tissues were embedded in paraffm. sectioned 5-6-p.m thick, and placed on positively charged slides (Fisher Scientific) for in situ hybridization or hematoxylin and eosin (H&E) staining. Tissue sections were kept in order as they were cut. and the con- secutive sections were numbered on the slides. The microtome blade and forceps were cleaned with xylene between samples to prevent carry-over DNA contamination. In Situ Hybridization (ISH) Tissue sections for ISH were processed as described previously (Stokes & Burreson 1995), except hybridization solution contained 5 ng/p.L SSOI318 DNA probe or 2 ng/|iL MSX1347 DNA probe and the addition of Bismarck Brown Y counterstain after the ni- troblue tetrazolium and 5-bromo-4-chloro-3-indoyl phosphate (BCIP) color development. Slides were washed with TE buffer (lOniM Tris, pH 8.0; ImM EDTA). then with dH,0 to stop the NBT/BCIP color development. Tissue sections were stained with 1% Bismarck Brown Y (Sigma Chemical) for I min. then rinsed three times with dH.,0. The slides were coverslipped with GVA Mounting Solution (Zymed Laboratories) and examined by light microscopy. Negative control ISH consisted of dH^O instead of DNA probe in the hybridization solution. Consecutive tissue sec- tions of all samples were processed in the following order: section 1. stained with H&E: section 2. ISH with SSOI318: section 3. ISH with MSX1347. section 4. ISH with no probe. DNA Probe Specificity In situ hybridization with both DNA probes SSOI318 and MSX1347 were performed on four C. virginica that had been diagnosed by histological examination as infected only with H. costale (Table 1 ). The Virginia Marine Resources Commission (VIMS) Oyster Disease Archive reference numbers for these oys- ters, embedded in paraffm. are 177.822. 181.676. 181.677. and 196.774. All of these oysters were collected at Wachapreague. VA. on the sea side of Virginia's Eastern Shore, the type locality (Wood and Andrews 1962) for H. costale. To demonstrate the TABLE I. Samples tested with in situ hybridizations using DNA probes SS01318 and MSX1347. Archive Sample Number Collection Date 177.822 May 1988 181.676 Mav 1989 181.677 Mav 1989 196.774 October 1994 Diagnosis by Histological Examination Diagnosis by in situ Hybridization H. costale heavy (spores pre.sent) H. costale heavy (spores present) H. costale heavy (spores present) tentative H. costale heavy* H. costale heavy; H. nelsoni heavy H. costale heavy H. costale heavy; H. nelsoni rare H. costale heavy; H. nelsoni light * Diagnosis of 196.774 was uncertain. The infective agent appeared to he H. costale. however, such seasonality of an advanced infection was unprecedented. Histological examination column indicates parasites identified in tissue sections and infection levels of original diagnoses. In siiu hybridization column indicates parasites identdled in tissue sections and infection levels with DNA probes. All samples were collected from the vicinity of Wachapreague. Virginia. Differential Diagnosis of Haplosporidans 209 specificity of probe SS01318. ISH was performed on sections of C \ii\vinicci tissues infected with H. nelsoni and Perkiiisiis mciri- luis (Mackin. Owen, and Collier), of shipworm (Teredo sp.) tissue infected with M. leredinis. and of mud crah iPanopcus sp.) tissue infected with H. loidsiana. RESULTS Specificity and Sensitivity of PCR Primers The H. costale PCR primer pair SSO-A and SSO-B amplified a 557 bp region of the H. costale small subunit rDNA (Fig. 1 A), targeting bases 784 to 1.^40 of that gene. The primers did nut amplify DNA from oyster or from the haplosporidans H. nelsoni. H. hndsiana. H. hisitanicum. M. teredinis. M. chilimis. or M. tape- tis (Fig. lA). The PCR product was readily detected after ampli- fication of 100 fg to 100 pg of cloned H. costale SSU rDNA; 10 fg of template DNA was amplified, but the product band was very faint in the asarose sel (Fia. IB). M 10 11 M12 345 67M B Figure 1. Specificil) and sensitivity of Haplosporidinm costale PCR primers. ( Al Specitkit). Lanes M, 100 bp ladder size marker, arrow indicates 600 bp; lane L cloned H. costale SSI' rDN.\; lane 2, cloned H. nelsoni SSI rDN.V; lane 3, H. lusilanicum genomic DNA: lane 4. cloned H. loidsiana S.SU rDN.\; lane 5, cloned Minchinia teredinis S.SL' rDNA; lane 6, M. chitonis genomic DN.\; lane 7, M. tapetis genomic DN.A; lane 8, uninfected Crassostrea virginica genomic DN.A; lane 9. no DN.\ control: lane 10, H. nelsoni-intected C. virginica genomic DN.\; lane 11, //. i(«ro/('-intected C. virginica genomic DNA. (B) .Sensitivity. PCR amplil'icalion products using H. costale primers SSO-A and SSO-B against serial dilutions of cloned H. costale SSI' rDNA template. Lanes M: 100 bp ladder size marker, arrow indicates 600 bp; lane 1: 100 pg template DNA; lane 2; 10 pg: lane 3: 1 pg; lane 4: 1(«) fg; lane 5: 10 fg; lane 6; I fg; lane 7; no DNA control. Specificity of DNA Probe One of the candidate H. costale probes, designated SS01318, was found to be sensitive and specific for H. costale in in situ hybridizations of tissue sections. Optimal hybridization required 5 ng/(iL SS01318 and incubation at 42°C. The SSO probe readily detected H. costale plasmodia and immature spores in sporocysts in ISH of oyster tissue with virtually no background, as indicated by the cells that stained dark purple to black (Fig. 2). DNA probe SS01318 did not hybridize with oyster tissue (C. virginica). the oyster pathogen P. marinns. or the haplosporidans H. nelsoni. H. loidsiana. and Minchinia teredinis {Fig. 3). Differential Diagnosis using D,\.A Probes Four oysters previously diagnosed by routine histological ex- amination of H&E-stained paraffin sections as infected v\ith H. costcde. but not H. nelsoni (Table 1) were subjected to ISH using separate DNA probes for H. costale and H. nelsoni. Plasmodia and immature spores in tissues of all four oysters hybridized with the SSO probe (Figs. 4-6). thus supporting the histological diagnoses. However, some plasmodia in three of the four oysters did not hybridize with the H. costale probe, but instead hybridized with the H. nelsoid probe (Figs. 4-6). These mixed infections of H. costale and H. nelsoni were not distinguishable nor detectable by histological examination in part because only plasmodial stages of H. nelsoni were present, but they were readily apparent by in situ hybridization (ISH) using the species-specific DNA probes. Even a light infection of H. nelsoiu plasmodia, scattered among a heavy infection of H. costale was easily detected using a DNA probe (Fig. 4E). In addition, the H. costale probe enabled discrimination of early and maturing plasmodia, the latter, which have vacuolated cytoplasm (Wood & Andrews 1962). The vacuoles within the stained plasmodia are easily seen at low power in ISH with the H. costale probe (Fig. 3B). Haplosporidinm nelsoni plasmodia were found in both connective tissue and epithelia (Figs. 4E. F; 5C; 6C); H. costale plasmodia were located throughout the connective tis- sue but not in the epithelium of the four oysters examined (Figs. 4C. D; 6B). One oyster collected in October 1994 seemed to be infected with H. costale based on the presence of plasmodia and appropri- ately sized immature spores within sporocysts scattered throughout the connective tissue as determined by stained paraffin sections (Table 1 ). However, the diagnosis was recorded as tentative, be- cause advanced infections of H. costale were known only from April to June (Andrews & Castagna 1978) and none had ever been reported from the fall season. In situ hybridizations using both H. costale and H. nelsoni DNA probes confirmed a mixed infection of the two parasites (Fig. 6). thus documenting unprecedented timing of an advanced H. costale infection. Plasmodia and sporocysts of H. costale were abundant in connective tissue (Fig. 6B). but not in epithelium; plasmodia of H. nelsoni occurred in epithelium (Fig. 6C) but not in connective tissue. DISCUSSION The PCR primers SSO-A and SSO-B and the DNA probe SS01318 were sensitive and specific for the target organism, H. costale. Another set of PCR primers for H. costale was previously reported by Ko et al. (1995); however, we chose to target a dif- ferent region of the SSU rDNA. The two regions targeted by the probe and primers described here are highly variable within the phylum Haplosporidia. accessible for probe hybridization and have f . « .% -'.A i^ ■• '?:* i.i- ' > , ^ . • * i-"-.. ♦ ^«^ • 1 «, **••* > f *^*. «» *« * 1 y 0 • * f ..* i * • » , m ( Figure 2. Consecutive histological sections of H. fo.s7a/f-infected C. virginica tissue (#181,6761 shouing Plasmodia and sporocysts containing immature spores in the connective tissue. (A) Hematoxylin and eosin (H&E) stain. Bar = 101) )im and also applies to B. (B) In silii hybridization (ISH) with H. cDstale DNA probe. Arrov> points to Plasmodium enlarged in C (C) ISH at higher magnification, arrow points to same Plasmodium indicated in B. (Dl ISH with H. neisuni DNA probe. Bar = 10(1 pni and also applies to C y / y x^ y B y y :'■? ^_^^B m m Figure 3. Lack of in situ hybridization I ISH I of various other parasites in histological sections demonstrating specificity of the H. costale DNA probe. (Al H. Hc/vo/d-infected ('. virginica tissue, arrows indicate some of the plasmodia present. Bar = 100pm. (B) I'erkinsiis mflni/Hs-infected C. virginica tissue, arrows indicate some of the cells present in the epithelium. Bar = 60 pm. (C) A/. r<'rt'(/m/,s -infected Teredo sp. tissue, with immature (i) and mature (m) spores. Bar = 100 (im. (D) H. /oHisiaiia-infected Panupeus sp. tissue, with immature (i) and mature (m) spores. Bar = 100 pm. Differential Diagnosis of Haplosporidans 211 B V \ i^ * \' « ' ^"^ "iT * _ V 1 • ^ 4 v.,A- •' * » "^ f'^.K^'' ■ ^«^V A\; ' .. * • ■, •*• •••*,-.■ * " ', \ » : - •• ^ • r ^- \ -1^: ^» Figure 4. /;; utii hybridization (ISH) of constcutivt liistolofjlcal seclions of oyster tissue (#181.677) collecled in May 1989 from Virginia's Eastern Shore with a mixed haplosporidan infection. A, C, E all show the same area: B. D. F are higher magnifications of A. C. and E. respectively. Asterisk in lower right of each figure indicates the same epithelial lobe. (A) Hematoxylin and eosin (H&E) stain. Bar = 60 pm and also applies to C and E. (Bl Hematoxylin and eosin (H&El stain showing H. nelsoni plasmodia (arrowsl in the epitheliium and some of the H. costale sporocysis (arrowheads) in the connective tissue. Bar = .V) (im and also applies to D and F. (C. D) ISH with H. costale UNA probe of same region in A and B showing positive reaction with H. costale. but not H. nelsoni. Arrowheads in D point to same H. costale sporocysts as in B. (E, F) ISH with H. nelsoni DN'A probe of same region in A and B showing positive reaction with H. nelsoni plasmodia. but not with H. costale sporocysts. Note ability of DNA probe to identify rare H. nelsoni plasmodia in a heavy H. costale infection. Arrows in F indicate hybridization of H. nelsoni Plasmodia in the epithelium as shown in B; arrowhead in F indicates lack of reaction with H. costale sporocyst shown in B and D. been used successfully for H. nelsoni-specific diagnostics (Stokes & Burreson 1995. Stokes et al. 1993a). The H. costale probe hybridized with H. costale plasmodia and immature spores, but not with mature spores, the same hybridization pattern as with the MSX probe with H. nelsoni (Stokes & Burreson 1995). In ISH of oyster samples from France, the SS01318 probe did not hybridize with the Pacific oyster C. gigas nor with a haplosporidian infecting that host (Renault et al. 2000). Mixed infections of H. costale and H. nelsoni thai ha\e not advanced to sporulation can now be diagnosed with confidence using these new tools. The plasmodia that hybridized with the H. nelsoni probe were not the same plasmodia that hybridized with the H. costale probe: although, these plas)nodia were indistin- guishable by traditional histological examination of stained tissue sections. The mixed parasite infections described here were origi- nally diagnosed as being only H. costale. This diagnosis was un- doubtedly made because of the preponderance of H. costale plas- modia and immature spores as compared to the relatively light infections of//, nelsoni and also because spores of//, nelsoni were absent. Couch (1967) reported finding mixed infections of//, cos- 212 Stokes and Burreson B it ♦ C- ^' m Figure 5. In situ hybridization (ISH) of consecutive histological sections of oyster (#177.822) collected in May 1988 from Virginia's Eastern shore with mixed haplosporidan infection illustrating ease of plasniodia differentiation with DNA probes. (A) Hematoxylin and eosin (H&El stain. Bar = 50 pni and applies to B and C. (B) ISH with H. costale DNA probe of same area shown in A. (C) ISH with H. nelsoni DNA probe of same area shown in A and B. telle and H. nelsoni. but they were based on the presence of spores of both species. In oysters tested to date with DNA probes, we have not observed H. costale plasmodia in the epithelium. The inability to distinguish nonsporulating mixed haplosporid- ian infections by traditional histological examination may have skewed epizootiology data for high salinity regions in the past. It is possible that H. nelsoni has been more common in Virginia oysters in high salinity than previously reported. If so. this may raise questions about past disease data and oyster mortality attrib- uted to H. costale. Results of diagno.ses using DNA probes have revealed an un- precedented seasonality of H. costale infections. The original di- agnosis of//, costale infection in oyster #196.774 in October 1994 was uncertain. The plasmodia and immature spores throughout the connective tissue looked like //. costale. but the timing of this advanced infection was unusual. Epizootiological studies of //. costale had established the annual infection cycle as quite predict- able. Clinical infections appear in the spring, as early as March, with sporulation and oyster mortality primarily in May and June. New infections occur before August 1st but remain subclinical 'A •" -.:^ ■ •-•• m^ ■ ''f f ■ . ^ vs^ *-. Y » ! ^ *" V Figure 6. In situ hybridization (ISH) of consecutive histological sections of oyster (#1%,744) collected in October 1994 from Virginia's Eastern Shore with mixed haplosporidan infection. (.\) Hematoxylin and eosin (H&E) slain. Bar = 1(10 fini and applies to B and C. (B) ISH with H. costale DNA probe of same area shown in A. Note H. costale plasmodia in connective tissue, but not in epithelium. (C) ISH with H. nelsoni DNA probe of same area shown in A and B. Note H. nelsoni plasmodia in epithelium, but not in connective tissue. Differential Diagnosis of Haplosporidans 213 iinlil the following spring (Couch & Rosent'ield 1968. Andrews & Castagna 1978. Andrews 1988). Andrews and Castagna (1978) reported that numerous samples of seaside Virginia oysters from summer through winter revealed no H. costalc infections. Diag- nosis by DNA probes of oyster 196.774 confirmed the H. costale diagnosis, as about 80% of the plasmodia hybridized with the SS01318 probe, but also revealed it as a mixed infection, because about 20% of the plasmodia hybridized with the MSX I .^47 probe. This H. costale infection, where the parasite's identity was con- firmed bv DNA-based diagnostics, did not meet historical criteria for SSO disease suggesting that the seasonality and epizootiology of this pathogen must be re-examined. ACKNOWLEDGMENTS The authors thank Juanita Walker for diagnoses by histological examination and Rita Crockett for technical assistance in the his- tology laboratory. This work was funded in part by the A. Marshall Acuff Oyster Disease Research Endowment at VIMS. VIMS con- tribution 2378. Andrews, J. D. 1982. Epizootiology of late summer and fall infections of oysters by Haplosporidium nelsoni. and comparison to annual life cycle oi Haplosporidium costalis. a typical haplosporidan. / ShcUfish Rex. 2; 15-23. Andrews. J. D. 1988. Haphisporidiitm costale disease of oysters. In: C. J. Sindermann & D. V. Lightner. editors. Disease diagnosis and control in North American marine aquaculture. Amsterdam; Elsevier, pp. 296- 299. Andrews. J. D. & M. Castagna. 1978. Epizootiology of Mincliinici cosliilis in susceptible oysters in seaside bays of Virginia's eastern shore. 1959- 1976. / Invert. Pathol. 32:124-138. Andrews, J. D.. J. L. Wood & H. D. Hoese. 1962. Oyster mortality siudies in Virginia: III. epizootiology of a disease caused by Huplospoiidium costale Wood and Andrews. J. Insect Pathol. 4:327-343. Burreson. E. M., N. A. Stokes & C. S. Friedman. 2000. Increased virulence in an introduced pathogen: Haplosporidium nelsoni (MSX) in the east- ern oyster Crassostrea virginica. J. Aquat. Anim. Health. 12:1-8. Couch. J. A. 1967. Concurrent haplosporidian infections of the oyster. Crassostrea virginica (Gmelin). J. Parasitol. 53:248-253. Couch, J. A. & A. Rosent'ield. 1968. Epizootiology of Minchinia costalis and Minchinia nelsoni in oysters introduced into Chincoteague Bay. Virginia. Proc Nat. Shellfish Assoc. 58:51-59. LITERATURE CITED Haskin. H. H. & J. D. Andrews. 1988. Uncertainties and speculations about the life cycle of the eastern oyster pathogen Haplosporidium nelsoni (MSX). Amer. Fish. Soc. Spec. Publ. 18:5-22. Ko. Y.-T., S. E. Ford & D. Fong. 1995. Characterization of the small subunit ribosomal RNA gene of the oyster parasite Haplosporidium costale. Mol. Mar. Biol. Biolechnol. 4:236-240. Renault. T.. N. A. Stokes. B. Chollet. N. Cochennec. F. Benhe & E. M. Burreson. 2000. Haplosporidiosis in the Pacific oyster, Crassostrea gigas. from the French Atlantic coast. Dis. Aquat. Org. 42:207-214. Stokes, N. A & E. M. Burreson. 1995. A .sensitive and specific DNA prohe for the oyster pathogen Haplosporidium nelsoni. J. Euk. Microbiol. 42:350-357. Stokes. N. A., M. E. Siddall & E. M. Burreson. 1995a. Detecdon of Haplosporidium nelsoni (Haplosporidia: Haplosporidiidae) in oysters by PCR amplification. Dis. Aquul. Org. 23:145-152. Stokes. N. A.. M. E. Siddall & E. M. Burreson. 1995b. Small subunit ribosomal RNA gene sequence of Minchinia teredinis (Haplosporidia: Haplosporidiidae) and a specific DNA probe and PCR primers for its detection. / Invert. Pathol. 65:300-308. Wood, J. L. & J. D. Andrews. 1962. Haplosporidium costale (Sporozoa) associated with a disease of Virginia oysters. Science 136:710-711. JoKiViil oj Shellfish Research. Vol. 20, No. 1. 215-220. 2001. ANALYSES OF GONADAL CYCLING BY OYSTER BROODSTOCK, CRASSOSTREA VIRGINICA (GMELIN), IN LOUISIANA JOHN E. SUPAN' AND CHARLES A. WILSON" ^ Office of Sea Grant Development. Louisiana State University. Baton Roiii;e. Louisiana 70803: 'Coastal Fisheries Institute. Louisiana State University. Baton Rouge. Louisiana 70803 ABSTRACT Oy.sters held near-shore in Caminada Bay. Louisiana during the summer, exhibit hypertrophic gonads with prominent genital canals beneath transparent mantle tissue about four weeks post-hatchery spawning, indicating recycling. Broodstock (N = 200) were analyzed histologically over a two-year period to document such gametogenesis, using Gonad/Body Ratios (GBR) and devel- opmental stages. Ten oysters were randomly selected from a broodstock pool prior to each spawning attempt, and monthly during the winter-spring. As expected, the mean GBR before successful spawning attempts was significantly greater (P =£ O.O."!) than the mean GBR before unsuccessful attempts. A dramatic drop in the percent occurrence of the advanced spawning and regression stage from May to June, a >40% spawning stage occurrence from May to October, and fluctuations in the percent occurrence of early and late developmental stages during the summer months dlustrates gonadal recycling. KEY WORDS: Cmssostrea virginicii. oyster, gonad, development, cycling, Louisiana INTRODUCTION Broodstock gonadal cotidilioti Is fuiidainenlal to consistent pro- duction of oyster larvae (Lannan et al. 1980) and successful trip- loid inductioti (Downing & Allen 1987). Although temperature requirements for gonadal development and spawning have been reported for the southern variety of the eastern oyster (C virginica) (Hopkins 19.^1. Hopkins et al. 1953. Loosunoff 1969. Hayes and Menzel 1981. Gauthier and Soniat 1989) documentation of the relationship between gatnetogenesis and induced spawning in the Gulf of Mexico region is limited. This is of particular interest, since oysters from southern populations are multiple spawners indicated by repeated gametogenic development throughout the spawning season (Ingle 1951. personal hatchery observations). An understanding of gonadal redevelopment is critical for collection of large numbers of viable eggs to produce cotnmercial-scale broods, especially triploids (Allen et al. 1989) Oysters held near-shore at the hatchery on Catninada Bay. Louisiana redevelop gonad and can be spawned repeatedly in the summer. Hypertrophic gonads with prominent genital canals be- neath transparent mantle tissue can occur about 4 weeks post- spawning. Such gonadal activity stimulated interest in histological examination of gonadal cycling in the hatchery's broodstock. The purpose of this study was to document oyster broodstock game- togenesis, including winter development leading up to the spawn- ing season, redevelopment (or recycling) between spawnings, and to determine the relationship between gonadal coiidilioii and spawning success. MATERIALS AND METHODS A broodstock of oysters was collected frotn oyster reefs in Louisiana and maintained in containers near-shore at the Sea Grant Grand Isle Bivalve Hatchery (29^15' 12"N. 90°0.V26"W) on Cami- nada Bay. Ten oysters were randomly selected over a two-year period for histological analyses when brought to the hatchery, prior to each spawning attempt, and bimonthly from the remaining broodstock during the winter-spring, A 4-5 mm cross-section of each oyster was removed just pos- terior to the labial palp-gill junction and preserved in Davidson's fixative for histological processing (Howard & Smith 1983). A 4 (jtin section was obtained approximately 1.000 |a.m from the junc- tion, mounted and stained with Gill's hematoxylin and eosin. This allowed the use of a standard cross section from each oyster for comparison. The sections were characterized by two H-shaped structures (large appendix of the stomach cecutn) ventrally located in the histological sections as described by Morales-Alamo and Mann (1989), Qualitative Descriplian Individual sections were microscopically exatiiined to deter- tnine the sex and stage of gamete de\eloptnent for each oyster. Classifications included Early Developtiient. Late Development. Spawning, and Advanced Spawning and Regression (ASR) after Kennedy and Krantz (1982). Quantitative Analyses Gonad/Body Ratios (GBR) were generated from histological sections similar to Kennedy and Battle (1964) using equidistant transects across each section to deteiniine the gonadal width rela- tive to body width (Fig. I). Only transects 3 to 8 were used. However, preliminary statistical comparison of mean GBR's among transects indicated that transects from the dorsal and ven- tral regions of the gonads were different from the rest, likely a result of the shape of the right and left gonads (Fig. 1 ). Histological sections of oysters (N = 200) were analyzed to determine GBR differences among developmental stages and spawning attempts. GBR was also measured in groups of oysters used during nor- tiial hatchery operation to determine how GBR varied between spawning and non-spawning populations. During each spawning attempt, appro.ximately 150 oysters were retiioved from near-shore containers. 10 were randomly selected for histological analysis and the remainder were placed in a spawning table and exposed to atnbient (30'-35°C) seawater for about 30 minutes. If necessary, the broodstock were then exposed to 18°C for approximately 1 hour, then re-exposed to ambient conditions for approximately 2 hours to stimulate spawning. If temperature stimulation did not !15 216 SuPAN AND Wilson Gonad Digestive Diverticula Mantle Figure 1. Image of histological transverse section througli the niid- hod) region of an oyster. Measurements from transects 3 to 8 were used in determining the gonad/bodj ratio. lead to spawning, the broodstock were exposed to a sperm sus- pension. Successful spawning was noted when a majority of the oysters had spawned, while unsuccessful spawning was character- ized by little or no spawning activity. The GBR's were analyzed using analysis of variance (ANOVA) with a two-way fixed factorial model (SAS Institute 1991) to test the difference among spawning attempts. The model included GBR as the dependent \ariable. spawning result as a fixed effect, and deselopmental stage and its Interaction with the spawning result as random effects. The GBR's met the assump- tions of normality and variance homogeneity after angular trans- formation (Dowdy and Wearden 1991 ). Tukey's Honestly Signifi- cant Difference Procedure was used to test the difference la = 0.05) among de\elopmental stage and spawning result. RESULTS Quantitative Analyses As expected, all developmental stages had significantly differ- ent mean GBR's (P < 0.05)(Table 1) and the mean GBR during successful spav\ning attempts was significantly greater (mean = 0.51. P < 0.05) than the mean GBR during unsuccessful attempts (mean = 0.45)(Table 2). TABLE 1. Results of analysis of variance: comparing mean gonad/bod> ratio of Crassostrea virginica by developmental stage. Developmental Stage Ratio* Mean .SD Comparisons* Early Development 0.31 (1.14 A Later Development 0.46 0.12 B Spawning 0.52 0.13 C Advanced Spawning & Regression 0.38 0A5 D * Ratio = Arsin (\/ (Gonad Width/Body Width)). ** Tukey's Honestly Significant Difference (a = 0.05). SD = Standard Deviation. TABLE 2. Results of analysis of variance: comparing mean gonad/body ratios at' Crassostrea rirgiiiica b\ spawning result. Spawning Result Ratio* Mean SD Comparisons** Successful Unsuccessful 0.51 0.45 0 10 0. 1 2 A B * Ratio = Arcsin (v (Gonad Width/Body Width I). ** Tukey's Honestly Significant Difference (a = 0.05J. SD = Standard Deviation. Qualitative Analyses After evaluating the histological sections for developmental stage (Kennedy and Krantz 1982). it became apparent that oyster gonads were recycling. Histological sections representing Early Development (Fig. 2. A & B). Late Development (Fig. 3. A & B). Spawning (Fig. 4. A & B) and Advanced Spawning-Regression (ASR)(Fig. 5, A & B) where noticeably different than sections from recycling (Fig. 6, A & B ). ASR gonads were typically atretic, with significant amounts of cellular debris and amoebocvtes in Ay 5 •^ V "^ •s -rf^S- '-%, %3$SW M 'St t GC '^^ ^'^ V* Figure 2. A iV H. I'holoinicrdi^raphs il histological sections showiiij; early gonadal development (.March) of C. virginica. posterior-ventral region (A: bar = 2t)l)//: B: bar = 30// 1. LG = left gonad, DD = digestive diverticula. M = mantle, GI = gill, G = gonad, F = follicles. CT = connective tissue, CG = genital canal. Oyster Gonadal Development in Louisiana 217 /H'f K^Wff^H Figure 3. A & B. Photomicrographs of histological sections showing later gonadal development (April) of C. r;>?/;i(ta.posterior-dorsal re- gion (A: bar = 2(1(1//: B: bar = 30//). RG = right gonad, 1)1) = digestive diverticula. M = mantle. G = gonad. F = follicles, CT = connective tissue, nearly enipt\ t'olliL'les and sunounding connective tissue. In con- trast, recycling gonads were characterized by proliferation of de- veloping follicles typically found in Early and Late Development sections (i.e., enlarged germinal cells and young pendant priniaiy oocytes in fonales [Kennedy & Battle 1964]). Figure 7 illustrates the gametogenic cycle produced by the descriptive stage characterization conducted during this study. A high percentage of the oysters examined were in Early Develop- ment at the beginning of the year; this stage decreased from May to November, rose during September, and increased again in De- cember. A short occurrence of Late Development in the spring during the sharp rise in the occurrence of the Spawning stage shows how oysters held in the near-shore waters of Grand Isle rapidly developed toward spawning condition. The percent of oys- ters in Later Deselopment fluctuated during the summer. The oc- currence of ASR reached 60'/f during April to May. decreased to approximately 25% during May to June, remained stable during the summer, and increased dramatically beginning in September to October. DISCUSSION Quantitative Results The difference between the mean GBR of oysters that spawned and those that did not was as expected (Table 2). Oysters in spawn- GI \ I imuL 4. \ \ 1!. riicitDiiiitrdgraphs of histological sections shciwidg spawning gonadal development (,lune) of ( . nrgm/co. (A: bar = 200//; B: bar = 30//, posterior-dorsal region). R(; = right gonad, LG = left gonad. DD = digestive diverticula. M = mantle. GI = gill. S = stomach, G = gonad, CT = connective tissue, ing condition should have higher GBR's than non-spawning oysters. Follicles anastomose through the surrounding connec- tive tissue during gonadal development (Figs. 2-6) increasing the mean GBR. The creamy-white appearance of a sexually mature oyster is due to the hypertrophy of the gonad (Kennedy & Battle 1964: Fig. 1). Fecund oysters are prolific spawners. Such spawning activity is reflected in the mean GBR's in Table 1, which depict gonadal attrition between spawning (0.52) and advanced spawning and regression (0.38). Hormonal control (Morse et al. 1978). genetics (Lannan 1980) and food availability to broodstock are examples of the many sources of variation in obtaining successful oyster spawning. Qualitative Results The occurrence of oysters in Spawning stage in Figure 7 de- picts an expected curve of over 60% occurrence during May to October, with reductions in the winter and spring. A similar au- tumn gonadal state has also been docuiitented in Louisiana (Gauthier & Soniat 1989). One would think that a line depicting an ASR gonadal stage would sharply increase from June to December, but this is not evident in Figure 7. The line depicts the highest ASR occurrence 218 SuPAN AND Wilson Figure 5. A & B. Photomicrographs of histological sections showing advanced spawning and regression gonadal development (August) of C. virginica. posterior-ventral region (A: bar = 200//; B: bar = 30//). l.G = left gonad. DD = digestive diverticula. GC = genital canal. M = mantle. GI = gill. G = gonad. CT = connective tissue, F = follicle. Figure f>. A & B. Photomicrographs ol histological sections showing gonadal recycling (October) of C. virginica (A: bar = 200//; B: bar = 30//: ventral region). RG = right gonad, IXi = left gonad. DD = diges- tive diverticula, S = stomach. M = mantle, GI = gill, CT = connective tissue, F = follicle. in May (609r) and December (807f ) with lower occurrences (24'^J-- 39%) during June to July. The dramatic drop in ASR occurrence and the rather stead) occurrence of the Spawning stage, as well as fluctuations in the percent occurrence of Early and Later Developmental stages dur- ing the summer months, illustrates recycling during June to Octo- ber. This may or may not be evidence of spawning in the wild since the samples were taken from suspended broodstock coerced to spawn at the hatchery during the study. Nevertheless, comparison of Figures 2 to 5 with Figures 6. S80»" B. Figure 8. A & B. Photomicrographs of histological sections showing gonadal recycling (August) of C. virgiiiica (A: bar = 200//; B: = 200//. dorsal region). Rti = right gonad, l.G = left gonad. I)D = digestive diverticula, M = mantle. GI = gill. CT = connective tissue. S = stomach, GC = genital canal. GC / DD Figure 9. A & B. Photomicrographs of histological sections showing gonadal recycling (August) of C. rirginka (A: bar = 200//; B: bar = 200//. dorsal region). RG = right gonad. LF = left gonad. DD = diges- tive diverticula. M = mantle. (;i = gill. CT = connective tissue, S = stomach, GC = genital canal. Histological sectioning is expensive and. therefore, not a valu- able tool for evaluating hatchery broodstock. Gonadal recycling of broodstock held near-shore is \isually apparent, however, by the observance of transparent mantle tissue occurring with a spotty or hypotrophic gonad during early recycling, followed by a full or hypertrophic gonad after complete recycling. ACKNOWLEDGMENTS We are grateful to Jordan Bradford, Lee Hanson, the late Tony Venterella and his wife, Gayle, and the late Carlo Venterella for their logistical support during the project, and Mr. Ron Becker for financial support. We thank Cheryl Crowder for the histological preparations and Stephanie Ehrett for assistance with section analyses. We are also grateful to Dr. Stan Allen for his review of the manuscript. Thanks is extended to the following for providing broodstock; Peter Vunjnovich. Wilbert Collins, Jules Melancon. Jordan Bradford, and James Cheramie. Financial support was pro- vided by grants from the Louisiana Board of Regents [LEQSF (93-96)-RD-B-08] and the Louisiana Sea Grant College Program INA89AA-D-SG226]. LITERATURE CITED Allen. S. K.. Jr. S. L. Downing & K. K. Chew. 1989. Hatchery manual for producing triploid oysters. Wash. Sea Grant Publ. No. WSG-89-3. Seattle. WA: Univ. of Washington. 27 pp. Dowdy. S. & S. Wearden. 1991 . Statistics for Research. New York: Wiley. 629 pp. Downing. S. L. & S. K. Allen. Jr. 1987. Induced tnploidy m the Pacific oyster. Crassoslrea gigas: optimal treatments with cytochalasin B de- pend on temperature. A(jiiaci p n nj inon a ■ i .• ■ u i ■ i ^ - 5 ' ' Morales-Alamo, R. & R. Mann. 1989. Anatomical teatures in histological sections of Cras««(/£'a i7/?/i;iai (Gmelin. 1791 ) as an aid in measure- Kennedy, V. S. & L. B. Krantz. 1982. Comparative oametoeemc and . r a c j , m ./,- , n ' r ^ ^ ments or gonad area tor reproductive assessment. J. Slicllfish Res. spawning patterns of the oyster Cras.wsrrca virginlca (Gmelin) in cen- >, . 1.-7107 tral Chesapeake Bay. / 5/iW//?.T/i /?f.r. 2:133-140. Lannan. J. E. 1980. Broodstock management of Cras.'inia 23062 ABSTRACT We examined survival, growth, and disease susceptibility of triploid Cnissoslrea ariakensis ( = rivuhiiis) and compared results with that ot diploid Crassostrea virginicci. Two hundred and fifty oysters (age = 2 yr, mean shell height = 60-64 mm) of each species were deployed at duplicate sites. (Chesapeake Bay, and the Atlantic Coast of Virginia) within low, medium, and high salinity regimes respectively (< 15%, 15-25%, > 25%). Over the course of the study, from June 1998 to September 1999, C. virginica exhibited low survival, modest growth and high disease susceptibility. In contrast. C ariakensis exhibited high survival, high growth rate, and low disease susceptibility. At low salinity sites, final mean cumulative mortality of C. virginica (81%) was significantly higher than that of C. ariakensis (14%). At medium and high salinity sites, all C. virginica died before the end of the study whereas final mean cumulative mortality in C. ariakensis was 1 3 to 1 6%. After 1 year of deployment, mean shell height of C. virginica at low. moderate, and high salinity sites was respectively 70. 80 and 73 mm. In comparison, mean shell height ofC. ariakensis was respectively 93. 121 and 1 37 mm. At low salinity sites, mean growth rate of C. virginica was not significantly different from that of C. ariakensis. At medium and high salinity sites, mean growth rate of C. virginica was significantly lower than that of C. ariakensis. Prevalence and intensity of Perkinsus marinus infections were significantly higher in C. virginica than in C. ariakensis. During the second summer of disease exposure, prevalence in C. virginica was 100% at all sites whereas in C. ariakensis it ranged from 0 to 28%. Heavy intensity of infections were prevalent in C. virginica whereas infections in C. ariakensis were limited to light intensity. Haplosporidium nelsoni (M.SX) was present in C. virginica. but absent in C. ariakensis. Mud worms (Polydora spp. ) were present in both oyster species, but infestations were low and did not appear to affect condition or growth. In summary, wide salinity tolerance and low disease susceptibility were associated with high sur\ ival and growth of C ariakensis in Chesapeake Bay and the Atlantic Coast of Virginia. KEY WORDS: Chesapeake Bay. aquaculture, species intrududion. Crassostrea ariakensis. C virginica INTRODUCTION In contrast to extensive information available on the eastern oyster Cnissoslrea virginica, and Pacific oyster Crassostrea gigas. reports on the Siiminoe oyster Crassostrea ariakensis ( = C, rivu- laris). are scarce, Suminoe oysters have been reported to be natu- rally distributed from southern Japan along the south China coast through southeast Asia to the western coast of the Indian subcon- tinent, but the taxonomy is tenuous in some areas and its actual distribution not clearly known (Carriker & Gaffney 1996). Larval settlement for C. ariakensis is reported to occur primar- ily in estuarine areas with low salinity, but juvenile and adult oysters grow within a wide range of salinity (Guo et al. 1999, Ahmed et al. 1987, Cai et al. 1992). Cultivation is important in southern China using seed oysters collected from the wild (Guo et al. 1999). On the West Coast of USA, where C. ariakensis was introduced with shipments of C. gigas and kumamoto oysters i.Crassostrea sikaniea. Anemiya 1928) from southern Japan in the 1970s (Breese and Malouf 1977). its aquaculture potential has been established (Langdon and Robinson 1996), Using t"ield ex- periments to compare the growth of C. ariakensis and C. gigas. Langdon and Robinson ( 1996) found that both species had similar growth and meat condition at various locations along the West Coast. To the best of our knowledge, no parasitic diseases have been reported in Suminoe oysters within its native range. However, in Zhanjiang Bay, southern China, mass mortality of C. ariakensis has been associated with outbreaks of toxic phytoplankton blooms (Yongjia et al. 1995). In Marennes Oleron, France, mortality in association with Bominiia-Wke parasites was observed in quaran- *Corresponding author. E-mail: calvo®' vims.edu lined C. ariakensis animals exposed to Bonamia ostrcae endemic waters (Cochennec et al. 1998). Studies on the potential performance and disease susceptibility of Suminoe oysters are not available for the Atlantic Coast of USA. However, as native eastern oyster stocks collapsed through- out much of the mid-Atlantic seaboard due to over harvesting, disease, and water quality deterioration, interest in the potential use of non-native oyster species has grown. Following a Virginia program to examine the suitability of non-indigenous oyster spe- cies to the local environtnents (VIMS 1996), C. gigas was the first species to be evaluated in Chesapeake Bay and the Atlantic Coast of Virginia (Calvo et al. 1999). Over the course of that study, from May 1997 to 1998. C. gigas had lower disease susceptibility than C. virginica. but survival and growth were equal or superior in native oysters than in C. gigas within Chesapeake Bay. Based on its close resemblance to the native oyster and its tolerance of temperate to sub-tropical environments. C ariakensis was the sec- ond candidate species selected for testing in Virginia (VIMS 1996), Considering its documented ability to grow in a wide range of salinity, we hypothesized that C. ariakensis would perform better relative to C. virginica than had C. gigas in Chesapeake Bay. The objectives of the present study were to compare survival, growth, disease susceptibility and infestations by shell boring polychaetes in C. ariakensis and C virginica deployed over a range of salinity. MATERIALS AND METHODS Study Sites Six sites were selected based several criteria including salinity regime, geographic location, available information on oyster grow- ing conditions and water quality, safely, logistics, and relevance Calvo et al. Atlantic Ocean Figure 1. Location of study sites in Chesapeake Bay and the Atlantic Coast of Virginia: ▲ = Low salinity (<15'7r | sites, • = Medium salinity ( 15-25% ) sites, ■ = High salinity (>25'7, ) sites. for the oyster industry. Sites were established at diiplieate locations within low salinity (<15%), medium salinity (15-259^). and high salinity (>25%) areas (Fig 1 ). Low and medium salinity sites were established near the margins of rivers (Coan, Great Wicomico, and York) or in shallow creeks surrounded by marshes (Woodas Creek. a tributary of the East River). High salinity sites were located in well-tlushed narrow channels surrounded by marshes and mudflats in the coastal lagoon system of the Atlantic Coast of Virginia. Temperature and salinity were measured during monthly site visits with a stem thermometer and a refraclomeler. To further characterize environmental variables, hourly temperature, salinity, and turbidity were measured with Hydrolab-Minisonde® datalog- gers deployed at the same sites for weekly to monthly intervals. Oysters Individually certified triploid C. ahakcnsis were produced and maintained in quarantine first at the Haskin Shellfish Research Laboratory. Rutgers University (HSRL) and then at the Virginia Institute of Marine Science's (VIMS) Aquaculture Genetics and Breeding Technology Center. Crassoslrea ariakensis brood stock, originating from an established line maintained in quarantine at HSRL and derived from sources on the West Coast of USA. was spawned in July 1996. Triploidy was induced by treatmenl oi fertilized eggs with cylochalasin-B using the methods described by Downing and Allen (1987) and Allen et al. (1989). Juvenile C. ariakensis were transferred to flow -through York River water with quarantined effluents at VIMS, where oysters were maintained until they were individually examined for triploidy. as described later, before field deployment in early 1998. Crassastrea viri;iniva brood stock, collected from Mobjack Bay. VA was spawned by a local commercial hatchery (Middle Peninsula Aquaculture) in July 1996. Prior to deployment, diploid C. virginica juveniles were maintained by Mobjack Bay Seafood in the Ware River. VA. Experimental Design Between May 29 and June 2. 1998. adult oysters were dis- pensed into replicate 9.5-mm mesh bags and placed within indi- vidual floating trays at the study sites. There were two replicate sites within each of three salinity regimes (Fig. 1). Each floating tray contained two bags with 100 oysters and one bag containing 30 individually labeled oysters, to follow growth, as described later. Floating trays (2.3 m x 0.3 m x 0.3 m) were constructed by fitting wire mesh trays (25-mm square 16 gauge mesh) into float- ing frames built with 4-inch (10.16 cm) PVC pipe, following the design of Luckenbach and Taylor (1997). Floating trays and bags were cleaned of fouling organisms at least once a month during regular site visits and more often if necessary. All sites were vis- ited monthly (±15 days). Mortality, Growth and Conilition All live and dead oysters within each float were counted monthly to determine survival. Monthly mortality was calculated as the number of oysters that died during each month interval, divided by the number of live oysters at the beginning of the month, corrected for oysters removed by sampling. Cumulative mortality was calculated as the sum of interval mortality (Barber and Mann 1994. Krebs 1972). Mortality data was examined for normality and homogeneity of variance using plots of means ver- sus standard deviations and Bartlett's chi-square test (Zar 1974). A two-way ANOVA was employed to examine the effects of species and salinity on arcsin-transformed cumulative mortality. Statistical analyses were performed using Statview® and Statistica® soft- ware. To follow growth. 30 oysters within each float were individu- ally labeled and shell height was repeatedly measured to the near- est 0. 1 mm using calipers, once monthly except January. February, and April 1999. Monthly growth rates for individual oysters were calculated as the overall shell height increment during the growing period while live oysters of both species were still available at all sites. June 1998 to May 1999. and divided by the deployment time in days standardized to 30 days. When oysters died measurements were taken from the remaining individuals without replacement. Growth rate data was examined for normality and homogeneity of variance using the same tests specified above for mortality data. The effects of species and salinity regime on arcsin-transfonned mean growth rate were examined using a two-way ANOVA fol- lowed by a Newman-Keuls test. At the end of the experiment, in September 1999. whole weight, shell weight, tissue wet and dry weights were measured on the same oysters collected for disease diagnoses. Following Lawrence and Scott (1982), Condition Index (CI) was calculated by the formula: CI = tissue dry weight * 100/ (whole weight - shell weight). (1) Oysters were allowed to air-dry for 13 to 20 minutes before weigh- ing and whole oyster weight was recorded to the nearest 0.0 Ig. Oysters were then shucked, shells weighed to the nearest 0.01 g. and wet tissues were gently rolled on a paper towel and weighed on pre-tarred vessels to the nearest 0.00 1 g. Wet tissues were dried at 80 C o\ernight and tissue dry weight was measured the next day FiFXD Study of C. ariakensis and C.virginica 223 to Ihe nearest O.OOlg. Condition nide\ dutu was examined for normality and liomogeneity of variance using the same tests speci- fied above for mortality data. Non-parametric statistics were em- ployed because means and standard deviations were still highly correlated (r = 0.952) after transformation. Mann-Whitney tests were used to examine differences in mean condition index and mean ranked body weights between species. Kruskal-Wallis tests were employed to examine differences in the above parameters among salinity regimes. Diseasex and I'dlydura A baseline sample of 25 oysters was taken to assess the disease status of each species prior to deployment in May 1998. Subse- quent samples of each species at each site were collected in August and September 1998. and in May. August, and September 1999. Pfikiiisiis mariiuis was diagnosed using Ray's fluid Thioglycollate medium (RFTM) assays (Ray 1952) on combined mantle, gill, and rectal tissue. Infection intensity was rated based on Ray ( 1954) and Mackin (1962). For the calculation of weighted prevalence infec- tion intensity was ranked, following Paynter and Burreson ( 1991 ). as: 0 = negative. 1 = light, 3 = moderate and 5 = heavy. Light-moderate infections were ranked as 1 and moderate-heavy infections were ranked as 3. Weighted prevalence was calculated by the formula: Weighted prevalence *n,/N (2) Where I = infection intensity rank n, = number of oysters w ithin I N = total number of oysters examined in the sample. Prevalence and weighted prevalence data were examined for nor- mality and homogeneity of variance using the same tests specified above for mortality data. Two-way ANOVAs followed by New- man-Keuls tests were employed to analyze the effects of species and salinity regime on arcsin-transformed prevalence and un- transformed weighted prevalence. Haplosporidium nelsoni. the causative agent of MSX disease, was diagnosed using standard paraffin histology procedures with oysters preserved in Davidson's AFA and 6 fjim tissue sections stained with Hams' hematoxylin and eosin (Burreson et al. 1988). Infection intensity was rated as light, moderate or heavy based on Burreson et al. ( 1988). Histology sections were also used to docu- ment the presence of other parasites and to examine development of oyster gonads. Disease diagnoses and histology were performed by the VIMS Shellfish Pathology Laboratory. The spionid polychaetes Polydora websteri and P. ligni are commensal with bivalves, including oysters. These suspension- feeding worms do not feed on the oyster, but the mechanical irritation caused by their presence causes the oyster to lay down additional layers of conchiolin over the worin's tube in what are often termed mud-blisters. At sufficiently high levels of infestation this can severely limit the growth of oysters and reduce their condition index. Examination for mud-blisters associated with Polydora spp. was conducted on the same oysters collected for condition and disease diagnoses in Septeinber 1999. Worms were not identified to species, but Polydora websteri is the most com- mon species affecting oysters in the northeast coast of the United States (Blake and Evans 1972, Wargo and Ford 1993). The internal surface of right valve shells was visually inspected and rated ac- cording to the presence and extent of mud-blisters. Examination was restricted to right valves as in Wargo and Ford (1993) who reported that infestations by Polydora spp. were equally found in right and left valves. Following the methods of Handley and Bergquist (1997). infestation was rated as: (0) no visible mud- blisters or any evidence of boring by Polydora spp.: (1) mud- blisters affecting less than 25% of the valve; (2) 25-50% of the valve affected; (3) 50-75% of the valve affected; or (4) more than 75% of the valve affected. Weighted prevalence was calculated similarly to equation (2) using the five categories above. Preva- lence and weighted prevalence data was examined for normality and homogeneity of variance using the same tests specified above for mortality data. Non-parametric tests were employed because zero variances precluded computation of Bartlett's test for deter- mining homogeneity of variance. Mann-Whitney tests were used to examine differences in mean prevalence and mean weighted prevalence between species. Kruskal-Wallis tests were employed to examine differences in the above parameters among salinity regimes. Repruduclive Stains and I'loidy A baseline sample of C. ariakensis was collected to ascertain the percentage of triploid individuals in the lot of animals exposed to Cytochalasin-B as described above. Prior to field deployment, all C. ariakensis animals were individually certified as triploids following tlow-cytometric methods (Allen 1983). Briefly, experi- mental animals were notched on the dorsal side of the right valve using a Dremel® rotary tool equipped with a fiberglass-cutting wheel. A 1- ml syringe fitted with a 23-gauge needle was inserted into the adductor muscle and 0.05 ml of hemolymph was removed. A 10 jjLg/ml DAPI-10% DMSO staining solution was added to the hemolymph and the sample was vortexed, aspirated, and filtered through a 23p.m Nitex® screen. DNA content of prepared samples was determined on a PARTEC® Cell Cycle Analyzer via ultra- violet light excitation. Histograms of relative DNA content were used to identify diploid cells with modal DNA values 1.5 times lower than that of triploid cells. Individuals with triploid and dip- loid cells were categorized as mosaics. A 2mm x 2mm piece of gill tissue, as well as a cross-section of gonad tissue were also sampled from mosaic individuals and examined for DNA content as above. The remaining gonad from mosaic individuals was processed by histology. Over the course of the experiment, samples of C. ariakensis (n = 16-35) were collected from each site in July and August 1998 and in May. June, and July 1999. Ploidy assays were conducted at HSRL and the VIMS Aquaculture Genetics and Breeding Tech- nology Center. A two-way ANOVA was employed to examine the effects of salinity and time on mean arcsin-transformed percentage of mosaics. Significant effects were further examined using a Newman-Keuls test. RESULTS En vironmcntal Parameters Means of monthly salinity measures at the two low salinity sites were below 10% only during June and July 1998. Drought conditions prevailed throughout much of the study and salinities above 15% were recorded at the low salinity sites from November 1998 to March 1999. Medium salinity sites experienced relatively low salinity (<15%) during June 1998, but were between 15 to 257c on all other sampling dates (Fig. 2). Salinity fluctuations in 224 Calvo et al. high salinity sites were within the expected range (25-35%). Tem- perature followed similar seasonal trends at all sites with a maxi- mum of 28° to 32°C in July and a minimum of 0° to 5°C in March. High salinity sites experienced overall cooler temperature with monthly means 2° to 4' C lower than medium or low salinity sites (Fig. 2). Turbidity was low (<70 Nephelometric turbidity units at all sites and times tested). Dissolved oxygen was relatively low (60% air-saturation) at Woodas Creek in July 1999 compared to all other sites and times measured (68-87% air-saturation). Mortality As the experiment progressed. C. virginica percent cumiilatixe mortality rapidly increased whereas C. ariakensis mortality re- mained low. Tlie highest increase in mean cumulative mortality, from 5% to 78%\ was observed in C. virgiiiicci at medium salinity between July and October 1998 (Fig. 3). At the end of the experi- ment, mean cumulative mortality in C. virginica (81%- 1 00%) was significantly higher (p <0.0005) than that in C. ariakensis (13%- 16%). Salinity had a significant (p = 0.006) effect on final cu- mulative mortality. The interactive effect of species and salinity was also significant (p = 0.011) and may be attributed to the increase in mortality between low and higher salinities observed LOW SALINITY REGIME 35-, O ^25- = 20 m 0) 15 a. E 10 - 80 n 60 ^ > 40 (0 F 20 3 ( ) 0 _100-, >^ 80 S 60 o 40 ili I I S JJASONDJFMAMJJAS MEDIUM SALINITY REGIME ,11 ^ ^ I S ^ ■■ F^n-ILItllilL . . Ih. ILILiyyi JJASONDJFMAMJJAS HIGH SALINITY REGIME I 1 I A JJASONDJF 1998 M A M J J A S 1999 Figure 3. Mean cumulative mortality by salinity regime (N = 2 sites, + SDl from May 1998 U> September 1999. Open bars = C virginica. Solid bars = C. ariakensis. ** = Significant al a = 0.01. NS = Not sampled. MEDIUM SALINITY REGIME 35 „30 O £25 = 20 « 15 a. E10 0) I- 5 .-^i-*-. I ' 1 \ \ \ ] ] ' 1 : 1 1 1 \ \ — JJASONDJFMAMJJAS 35 -30 25 2 a. 20 ~ >. 15 I 5 0 for C. virginica whereas low mortality was similarly observed for C. ariakensis at all salinities (Fig. 4). GroHih Growth varied with species and salinity regime (Fig. 5 and Table 1 ). At the start of the experiment mean shell height was 60 mm in C. virginica and 64 mm in C. ariakensis. After 1 year of deployment, in May 1999, mean shell height of C. virginica at low, medium, and high salinity sites was respectively 70, 80 and 73 mm. In comparison, mean shell height of C. ariakensis at low. HIGH SALINITY REGIME 35-, — 30 O ^25 = 20 C3 O 15- Q. E 10 ^- 5 04 -•- *- S o 1998 N D J F M A M 1999 p35 -30 25^ a 20- ISE ■ 10 *o 5 0 J J A S Figure by salinity 1999. • = Means of monthly temperature and salinity measurements regime (N = 2 sites, ± SDl, from .Fune 199X to September Temperature using stem thermometer, ■ = Salinity using low medium Salinity regime high refractonieter. Figure 4. Interaction between oyster species and salinity on final cu- mulative mortalitv. Means of 2 sites (± SD), • = C. virginica. ■ = C ariakensis. Field Study of C. ariakensis and C.virginica 225 _125- E ^ 75- i ^^ ^ 25 0 LOW SALINITY REGIME I-X-i-i-J 150 _125 E £100 ff 75 0) X ^ 50 ^ 25 0 JJASONDJFMAMJJAS MEDIUM SALINITY REGIME Market size 3"i ST JJASONDJFMAMJJAS HIGH SALINITY REGIME Fisure 5. Mean shell height by salinity regime (N = 2 sites, ± SD) of 5() in(li^i(Ulal oysters repeatedly measured from May 1998 to September 1999. % \ = C. virginica. ■ = C. ariakensis. TABLE 1. Effects of species and salinity regime on mean growth rate. A. Two-way ANOVA Effect df MS Species Salinity Species*Salinity Error 32.293 3.441 3.225 0.526 61.3X2 6.536 6.124 <0.nO(15 0.052 0.076 B. Multiple comparison (Newman-Keuls test) moderate, and high salinity sites, was respectively 9.3. 121 and 137 mm. Oysters continued to grow until July 1999 when mortality of C. viiginica approached 100% in all medium and high salinity sites. At that time mean shell height of C. ariakensis at low. me- dium and high salinity sites was respectively. 96. 125 and 140 mm. No growth was observed for either species during July to Septem- ber 1999. Most of the growth occurred during fall 1998 and spring 1999. Species and salinity regime had a significant effect on mean growth rate (Table I A). Similar growth rates were observed for C. virginica at all salinities in contrast to increasing growth rates with increasing salinity observed for C. ariakensis (Fig 6). At low sa- linity sites, mean growth rate of C. virginica (1.1 mm mo" ) was not significantly different than that of C. ariakensis (2.6 mm mo"'). At medium salinity sites, mean growth rate of C. virginica (1.7 mm mo"') was significantly lower than that of C. ariakensis (4.9 mm mo"'). At high salinity sites, mean growth rate of C. virginica (1.0 mm mo"') was also significantly lower than that of C. ariakensis (6.2 mm mo"' ). For C. virginica. growth rate did not differ significantly among salinity regimes. For C. ariakensis. however, growth rate at low salinity was significantly lower than that at medium and high salinity regimes (Table IB). Disease At the beginning of the study, there was no P. mariniis and a 4% prevalence of W. nelsoni in C. virginica and 12% prevalence of P. marinus and no H. nelsoni in C. ariakensis (Fig. 7). In October 1998, prevalence and weighted prevalence of P. marinus were significantly higher (p <0.0005) in C virginica than in C. ariak- ensis. In September 1999. when no live C. virginica remained at the medium salinity York River site, prevalence in C. virginica at all other sites was 100% whereas pre\alence in C. ariakensis ranged from 0 to 75% and did not differ (,p >0.250) among salinity regimes (Fig. 7). Heavy infections were prevalent in C. virginica whereas only light infections were observed in C. ariakensis (Table 2). Maximum prevalence of H. nelsoni (25%) was observed in C. virginica at the York River site in May 1999. H. nelsoni was also present in C. virginica at the low salinity Great Wicomico River site in September 1998 (4%), and at high salinity sites in October 1998 (4-8%) and May 1999 (4%). Intensity of//, nelsoni infec- tions was light except for heavy infections found in oysters sampled from medium (1/132) and high salinity sites (1/157). No H. nelsoni was found in C. ariakensis. Other parasites observed in histological sections of C. virginica were the protozoan Haplospo- Comparison Within Between Low salinity Medium salinity High salinity C. virginicii C. virginica C. virginica C. ariakensis C. ariakensis C. ariakensis C. virginica and C. ariakensis 0.116 C. virginica and C. ariakensis 0.01 S C. virginica and C. ariakensis 0.005 Low salinity vs. medium salinity 0.268 Low salinity vs. high salinity 0.931 Medium salinity vs. high salinity 0.440 Low salinity vs. medium salinity 0.034 Low salinity vs. high salinity 0.018 Medium salinity vs. high salinity 0.280 low medium Salinity regime high Figure growth ensis. 6. Interaction between oyster species and salinity on mean rate. Means of 2 sites (± SO). • = C. virginica. ■ = C. ariak- 226 Calvo et al. LOW SALINITY REGIME 100 -, Aug Sep MEDIUM SALINITY REGIME _Ns_BL May Aug Oct May Aug Sep 100 ^ 80 g 60 HIGH SALINITY REGIME 1 1 May Aug 1998 May Aug 1999 _^_NsiH_^ Sep Figure 7. Mean pre>alence of P. mariiius b> salinity regime (N = 2 sites, + SDl, in samples of 25 oysters, from May 19yS to September 1999. Open bars = C. virginica. Solid bars = C. ariakeiisis. ** = Sig- nificant at a = 0.01. * = Not significant at a = 0.05. NS = Not sampled because no C. virginica remained. ridium cosrale (SSO), present at high salinity sites, the tieniatode Bucephalus sp. and a chlamydia-like organism. None of these or other parasites were observed in C. ariakeiisis. Condition At low salinity sites, mean condition index in C. virginica and C. ariakeiisis were, 3.6 and 6.6. respectively, and means were not significantly different (p = 0.121). Similarly, there were no sig- nificant differences (p = 0.121) in mean body weights between species. At medium and high salinity, comparisons between spe- cies were not possible because no live C. virginica remained at those sites at the end of the experiment. For C. ariakeiisis. mean condition index at low. medium and high salinity, respectively, were 6.6, 5.3 and 9.7 and means were not significantly different (p = 0.276). Similarly, there were no significant differences (p > 0. 102) between mean body weights among salinity regimes. The percentage of shell weight relative to whole oyster weight in C. virginica (62%) was similar to that in C. ariakensis at low. me- dium, or high salinity, respectively, .'>9. 61 and 65%. Polydora At low salinity sites, mean prevalence of Polydora spp. was 100% in both oyster species, and there was no significant differ- ence in mean weighted prevalence between oyster species (p = 0. 121 ). At medium and high salinity, comparisons between species were not possible because at the end of the experiment there were no live C. virginica at those sites. For C. ariakensis. there was a trend of decreasing prevalence with increasing salinity. Mean prevalence in C. ariakensis at low. medium and high salinity sites were, respectively, 100. 62 and 12%. However, ranked mean prevalence and weighted prevalence were not significantly differ- ent (p = 0.156) among salinity regimes. Ploidy The baseline sample revealed that prior to deployment 94% of the C. ariakensis in the lot were triploids. Individual certification assured that triploids were exclusively deployed in the field. Dur- ing the course of the study, there were 62 individuals in which combinations of diploid and triploid cells (mosaics) were detected out of 1 164 oysters examined (5.3%). The proportion of mosaics ranged from 0.0 to 13.8% depending on time and site (Table 3). The effect of time on mean percentage of mosaics was significant (p = 0.002). Mean percentage of mosaics increased significantly between the first sampling time and each subsequent sampling time (p < 0.007). After the initial sampling time, the percentage of mosaics did not differ significantly among the remaining sampling times (p > 0.498). Salinity regime had no effect on mean percent- age of mosaics (p = 0.128). Examination of 39 mosaic individuals revealed that 10 were females. 23 were males, one was hermaph- roditic, and fi\e were undifferentiated. DISCUSSION Drought conditions and below normal Chesapeake Bay stream flow starting in fall 1998 resulted in increased salinity and epi- zootics of both H. nelsoni and P. mariiuis during 1999 (Ragone Calvo & Burreson 1999). High disease pressure prevailing in the region was associated with severe infections and high mortality in r. virginica, though not in C. ariakensis. After the first summer of disease exposure, more than 50% of C. virginica had died and prevalence of P. iiiarinus at medium and high salinity sites was 100%. A year later when all C. virginica at medium and high salinity sites were dead, cumulative mortality at low salinity sites was 81% and prevalence of P. marinus was 100% with se- vere infections present. Maximum prevalence of H. nelsoni in C. virginica was 25% whereas no H. nelsoni was detected in C. ariakensis. Presence of H. nelsoni and intensification of P. mariiuis infections in C. virginica at the low salinity Great Wicomico site was undoubtedly favored by drought conditions resulting in salinity greater than 15%' starting in fall 1998 and continuing into spring and summer 1999. Persistence of salinity greater than 15% during summer and fall is conducive to devel- opment of lethal P. marinus infections (Burreson and Ragone Calvo 1996). In comparison, maximum P. marinus prevalence in C. ariakensis reached 84%, but infections never exceeded light intensity and mortality remained low (13-16%). With the caveat that this study spanned only 15 months, C. ariakensis appears highly tolerant of the dominant parasitic diseases affecting Chesa- peake Bay oysters. A limitation of this study was that conditions were not identical for both species before the beginning of the experiment. Since C. ariakensis was quarantined for their first two years in land-based systems with limited infiow of raw York River water, long-term exposure to disease agents may have been reduced in relation to Field Study of C. ariakensis and C.virginica 227 TABLE 2. Prevalence and intensity of P. marimis in C. virginica and in ('. ariakensis by salinity regime, site and date during 1998 (A) and 1999 (Bl. A. Site Date C. virginica C. ariakensis Salinity Prevalencet L* M* H* Prevalence! I* M* H* L CNRV 8/12/98 20% (5/25) 3 T 0 ()<;; ((j/25) 0 0 0 9/3D/98 96% (24/25) 18 2 4 12'/; (3/25) 3 0 0 GWRV 8/4/98 88% (22/25) 21 0 1 24% (6/25) 6 0 0 9/30/98 100% (25/25) 12 4 9 28% (7/25) 7 0 0 M WOCK 8/3/98 100% (25/25) 7 5 13 84% (21/25) 21 0 0 9/30/98 100% (24/24) 7 7 10 68% (17/25) 17 0 0 YKRV 8/3/98 100% (25/25) 16 3 6 8% (2/25) 2 0 0 9/29/98 100% (25/25) 7 11 7 52% (13/25) 13 0 0 H BUBY 8/6/98 100% (25/25) 20 1 4 44't (11/25) 11 0 0 10/7/98 80% (20/25) 13 6 1 825%o). Site codes: CNRV = Coan River. GWRV = Great Wicomico River. WOCK = Woodas Creek. YKRV = York River. BUBY = Burton Bay. BOBY = Bogues Bay. t = In parenthesis number of oysters infected/number of oysters examined. * = Number of oysters with, respectively, light, moderate, and heavy infections. NS = No C. virginica remaining for sampling. that of C. virginica which was initially maintained in the water at a nearby location. Absence of P. niariniis in the baseline sample of C. virginica collected in May 1998 does not preclude the possi- bility of sub-clinical infections. Given the low sensitivity of the standard diagnostic assay for detecting P. marinus infections dur- ing spring (Bushek et ai. 1994, Burreson and Ragone Calvo 1996), it is likely that acquisition of infections occurred the prior summer but that overwintering infections were not detected the following spring. Suminoe oysters tested in this study had comparable survival at all salinity regimes and similar growth rate at medium and high salinity regimes, in agreement with the wide salinity tolerance described for C. ariakensis in its native range (Guo et al. 1 999). By the end of the experiment, when oysters were three years old, mean shell height of C. ariakensis at low, medium, and high salinity regimes was respectively 96, 125 and 140 mm. By comparison, in Zhanjiang Bay (annual salinity range = 7-30%) average shell height of three-year-old Suminoe oysters is 100 mm (Cai et al. 1992). In contrast to C. gigas in previous studies (e.g., Handley & Berquist 1997, Calvo et al. 1999), C ariakensis was not found to be adversely affected by Polythna spp. in this study. Mud worms were present in both C. ariakensis and C. virginica. but infesta- tions were not severe and did not appear to affect growth or sur- vival of either species. However, since the conditions that result in severe infestations are not clearly understood, we cannot dismiss the possibility that C. ariakensis might be susceptible to such infestations. Results of the present investigation suggest that C. ariakensis is more adapted to Chesapeake Bay conditions than C. gigas. In a 228 Calvo et al. TABLE 3. Percentage of genetic mosaics by salinity regime, site and time. Site 1998 1999 Total by site Salinity July .August May .lune July .August Total by salinity L CNRV 0.0% 2.8% 8.6% 8.6% 9.7% 8.0% 6.1% 5.0% (0/35) (1/35) (3/35) (3/35) (3/31) (2/25) (12/196) (20/395) GWRV 2.8% 2.8% 2.8% 5.7% 5.9% 4.0% 4.0% (1/35) (1/35) (1/35) (2/35) (2/34) (1/25) (8/199) M EARV 0.0% 2.8% 11.4% S.6% 13.8% 1 2.0% 7.7% 7.2% (0/35) (1/35) (4/35) (3/35) (4/29) (3/25) (15/194) (28/391) YKRV 0.0% 11.4% 11.4% 8.6% 2.9% 4.3% 6.6% (0/35) (4/35) (4/35) (3/35) (1/34) (1/23) (13/197) H BUSY 0.0% 8.8% 2.8% 5.7%' 0.0% 8.0% 4.1% 3.7% (0/35) (3/34) (1/35) (2/35) (0/33) (2/25) (8/197) (14/379) BOBY 0.0% 0.0% 2.8% 5.7% 6.2% 8.0% 3.3% (0/35) (0/35) (1/35) (2/35) (1/16) (2/25) (6/181) Total hy month 0.5% 4.8% 6.7% 7.1% 6.2% 7.4% (1/210) (10/209) (14/210) (15/210) (11/177) (11/148) Overall total = = 5.3% (62/1164) Salinity codes: L = low (<15'?(1. M = medium (15-25%<). H = high (>25'^?( EARV = East River, YKRV = York River, BUBY = Burton Bay, BOBY = oysters examined. Site codes: CNRV = Coan River. GWRV = Great Wicomico River. Bogues Bay. In parenthesis number of mosaics divided by number of comparative study with C. virgiiiica aniJ C. gigus (Calvo et al. 1999), C. gigas exhibited high cumulative mortality (63%) at low salinity sites and growth rate was not greater than that of C. vir- ginica within Chesapeake Bay. In contrast, C. ariakciisis tested in this study had less than 16% cumulative mortality and greater growth rate than C. virgiiiica within medium salinity sites in Chesapeake Bay. In grow-out trials with C. gigas and C. ariakensis using various culture methods at high salinity locutions on the West Coast of USA. growth rate was the same for both oyster species. For example, C. gigas and C. ariakensis juvenWes (mean shell height = 10 mm) planted directly on the bottom in Puget Sound, Washington, or suspended from floating rafts, in Yaquina Bay. Oregon, increased to 9(1-100 tnm after I year of deployment during 1990 to 1991 (Langdon & Robinson 1996). A direct com- parison between C. gigas and C. ariakensis is not available for the East Coast of USA. However, in this study and in Calvo et al. (1999). both species experienced significantly higher growth rate than the conesponding C. virginica control oysters at high salinity sites on the Atlantic Coast of Virginia. The choice of oyster strain must be considered in the interpre- tation of results. C. virginica offspring frotii wild Mobjack Bay brood stock were employed in this study because they are a stan- dard stock for aquaculture and because they were the only stock available with similar age, size and disease status as C. ariakensis. It is likely that a strain of C. virginica selected for disease resis- tance would tnore favorably compare to C. ariakensis. For ex- atnple. DEBY, a strain of C. virgifiica that has been selectively bred by VIMS against P. inarimis and H. nelsoni for four genera- tions (Ragone Calvo et al. 1997), exhibited similar survival to that of C. ariakensis and higher survival, higher growth rate and lower disease susceptibility than that of unselected C. virginica from Mobjack Bay deployed at a site in the Great Wicomico River from June 1998 to May 1999 (G. Calvo unpublished data). Similarly. CROSBreed strains (Debrosse and Allen 1996) have been selected for dual resistance to both pathogens in the Mid-Atlantic. Use of disease resistant C. virginica strains could provide a tnore relev ant comparison of native and non-native oyster performance for aqua- culture. To some extent, the fact that C. ariakensis were triploids may have allowed them to grow faster than diploid C. virginica. In general, reduced gametogenesis in triploids corresponds with im- proved somatic growth (Barber & Mann 1991 ). However, based on studies with C. virginica ai\d C. gigas, ploidy effects alone are unlikely to account for reduced disease susceptibility and in- creased survival of C. ariakensis. Barber and Mann (1991) found that cohorts of diploid and triploid C. virginica had the same prevalence (96%) and similar intensity of P. marimis after 17 months of deployment at the York River. VA. Meyers et al. (1991) reported equal cumulative mortality for diploid and triploid C. virginica (100%) and lower cumulative mortality for diploid (25%) than triploid C. gigas (34%), after one year of challenge with P. inariniis. In summary, during the course of the study C. ariakensis per- formed better than C. virginica in Chesapeake Bay and the Atlantic Coast of Virginia. Wide salinity tolerance combined with low disease susceptibility was associated with high survival and high growth rate in C. ariakensis. In contrast, high disease susceptibility was associated with low survival and low growth rate in C. vir- ginica. As previously discussed for C. gigas (Calvo et al. 1999). a decision on whether C. ariakensis is, or is not, an appropriate species for use in these environments must include other factors beyond the scope of these field investigations. For exatnple, inter- national organizations have recommended that competent local authorities consider the ecological consequences of species intro- ductions by evaluating the following: ( I ) the possibility that non- indigenous species may carry pests and pathogens into the new environment; (2) the potential relationship of the exotic species with native .species; and (3) the potential range for establishment of the exotic species in the new environnient. A cautious introduction of triploid C. ariakensis for aquaculture purposes following Inter- national Council for the Exploration of the Seas guidelines, as it is Field Study of C. ariakensis and C.virginica 229 currenliy being considered in Virginia, would minimize risks as- sociated witli factors enumerated above. Use of individually cer- tified triploids in a closely monitored research setting allowed us to conduct the present study with minimum risks of unintentionally introducing reproductively capable C. ariuki'iisis into the waters of Virginia. At the present time, however, there is no precedent for using triploids to control oyster populations at a commercial scale. In practice, implementation of such a strategy would require sig- nificant efforts directed at inonitoring the reproductive status of deployed stocks over time, and at designing and enforcing strict quarantine regulations to avoid undesired release of reproductively capable stocks. ACKNOWLEDGMENTS We would like to thank Rita Crockett. Paige Ross and Francis O'Beirn for assistance in the field. Juanita Walker and Rita Crock- ett conducted disease diagnoses. Gregg DeBrosse and staff at Rut- gers University produced triploid oysters. Ploidy analyses were conducted by Aimee Howe and Whitney Chandler. Wanda Cohen and Kay Stubbfield assisted with preparation of the figures. We would also like to acknowledge members of the oyster industry including Odus Cockrell. Lake Cowart Jr.. Ken Kurkowski. Tommy Mason. John Register and John Vigliotta, for general sup- port and access to field sites. VIMS contribution No. 2.'^65. LITERATURE CITED Ahmed. M.. .S. Barkati & Sanuullah. 1987. Spatfall of oysters in the Ghuro- Phiui salt water creek system near Karachi (Pakistani. Pukisinii J. Zoology l9(3):245-252. Allen. S. K.. Jr. 1983. Flow cytometry: Assaying e.xperimental polyploid fish and shellfish. Aqmicidmre 33:317-328. Allen. S. K.. Jr.. S. L. Downing & K. K. Chew. 1989. Hulcheiy manual fai linnhaing triploid oysters. Seattle. WA: University of Washington Press. 27 pp. Barber. B. J. & R. Mann. 1994. Growth and mortality of eastern oysters, Crussoslrea virginica (Gmelin. 1791 ). and Pacific oysters, Crassostrea gigas (Thunberg. 1793) under challenge from the parasite. Perkiiistis marimis. J. of Shellfish Res. 13( 1):109-1 14. Blake, J. A. & J. W. Evans. 1972. Polydora and related genera as borers in mollusk shells and other calcareous substrates. The Veliger l.'>(3):235- 249. Breese. W. P. & R. E. Malouf. 1977. Hatchery rearing techniques for the oyster Crassostrea rivularis (Gould). Aquaciiltnre 12:123-126. Burreson, E. M. & L. M. Ragone Calvo. 1996. Epizootiology oi Perkinsus inarimis disease of oysters in Chesapeake Bay, with emphasis on data since 1985. J. Shellfish Res. 15(l):17-34. Burreson. E. M., M. E.. Robinson & A. Villalba. 1988. A comparison of paraffin histology and hemolyniph analysis for the diagnosis of Hap- losporidnim nelsiini (MS.X) in Crassostrea virginiea (Gmelin). / Shell- fish Re. s. 7(1): 19-23. Bushek, D.. S. E., Ford & S. K. Allen. Jr. 1994. Evaluation of methods using Ray's fluid thioglycollate medium for diagnosis of Perkinsus marimis infection in the eastern oyster. Crassostrea virginica. Annual Review of Fish Diseases 4:201-217. Cai. Y.. C. Deng & Z. Liu. 1992. Studies on the ecology of Crassostrea rivularis (Gould) in Zhanjiang Bay. Tropic Oceanology ll(3):37-44. Calvo. G. W.. M. W. Luckenbach. S. K. Allen. Jr. & E. M. Burreson. 1999. A comparative field study of Crassostrea gigas (Thunberg 1793) and Crassostrea virginica (Gmelin 1791) in relation to salinity in Virginia. J. of Shellfish Res. 18(2):465-+73. Carriker, M. R. & P. M. Gaffney. 1996. A catalogue of selected species of living oysters (Ostracea) of the world. In: A. Eble, V. S. Kennedy. & R. L E. Newell, editors. The eastern oyster Crassostrea virginica. Maryland Sea Grant Publication, pp. 2-18 Cochennec, N., T. Renault. P. Boudry. B. Chollet & A. Gerard. 1998. Bonamia-Wke parasite found in the Suminoe oyster Crassostrea rivu- laris reared in France. Diseases of .Aquatic Organisms 34:193-197. DeBrosse. G. A. & S. K. Allen. Jr. 1996. Cooperative Regional Oyster Breeding (CROSBreed) project. / Shellfish Res. 15(2):5l4-.'il.'i. Downing. S. L. & S. K. Allen. 1987. Induced triploidy in the pacific oyster, Cras.w.strea gigas: optimal treatments with cytochalasin B depend on temperature. AquacuUure 61:1-1.'). Guo. X.. S. E. Ford & F. Zhang. 1999. Molluscan aquaculture in China. / Shellfish Res. 18(1):9-18. Handley, S. J. & P. R. Bergquist. 1997. Spionid polychaete infestations of intertidal Pacific oysters Crassostrea gigas (Thunberg), Mahurangi Harbour, northern New Zealand. Aquaculture 153:191-205. Krebs, C. G. 1972. Ecology: The experimental analysis of distribution and abundance. New York: Harper & Row. 694 pp. Langdon C. J. & A. M. R25°C, high salinity (>17 ppt), and/or low initial cell densities. Alteration of any of these factors will result in changes in P. marinus growth rates. Saunders et al. (1993) and Ford et al. ( 1999) found that cell densities of greater than 10"* to 5 x 10"* per g wet wt oyster"' significantly reduced cell division rate, suggesting that a density dependent feedback mechanism is operative. Decreases in tem- perature and salinity of in vitro culture conditions also resulted in slower P. marinus growth rates (Gauthier and Vasta 1995, Dungan and Hamilton 1995, La Peyre 1996). Likewise, investigations of P. nuirinus activity in nature and in in vivo laboratory experiments have demonstrated the pronounced effect of temperature and sa- linity on P. marinus infection progression in oysters. Field obser- Loss via host (jeath Density dependent negative feedback Growth Parasites / I o Mortality peraturej TransmissioD / Figure 1. Schematic diiij;r;ini of the /'. marinas model. Population Dynamics of Perkinsus marinus 233 TABLE 1. Model forcing functions, state variables, flows, feedbacks and parameters. Description Symbol Default values and/or unit Forcing variables Temperature Salinity State variables Parasites Flows Parasite growth Parasite transmission Parasite mortality Parasite loss via host death Feedbacks Parasite density dependent self-limitation of P,,„ft) Parameters Parasite growth threshold density Parasite maximum density Parasite division rate Parasite transmission rate as function of salinity Parasite mortality rate as function of salinity Parasite loss via host death Initial condition of state variable Parasite (WS Calibration) Parasite (WS Validation) Parasite ( HH ) Parasite (DWSl T(t) S(t> P(0 PoC) Pj15 = 1000 cells d"' at 15 > S> 12 = 500 cells d" at 12 > S > 9 = 100 cells d"' at 9 > S > 6 = 50 cells d"' at S < 6 = 0 cells d"' atS > 12 = 0.0175 d"' at 12 a S a 8 = 0.02 d"' at 8 >S a 3 = 0.0175 d"' at S < 3 = 0.006 d"' at P(t)> 1 00.000 = 0.50 d"' 801 cells (g wet wt)"' 3927 cells (g wet wt)"' 537 cells (g wet wti"' 135 cells (2 wet v\t)"' vations indicate thai parasite proliferation occurs at temperatures above 20°C (Andrews and Ray 198S). This observation has been supported by laboratory experiments in which infection progres- sion was followed in artificially infected oysters held at 10. 15. 20. and 25°C (Chu and La Peyre 1993). Chu and La Peyre (1993) reported that only those oysters held at 20 and 25°C had moderate to heavy infections at the termination of the 7-wk experiment. Similarly, infection progression in oysters maintained in the labo- ratory at 25°C for a 6-wk period was delayed at 12 ppt and ceased at salinities below 9 ppt (Ragone and Burreson 1993). In the model, parasite cell growth, the increase in parasite den- sity as a result of cell division, is dependent on salinity, tempera- ture, and parasite density. The model equation for cell growth was derived as: ■■ P * \i.d * {\-({P-Ap)/[Gp-Ap)] (2) parasite growth, P = parasite density (cells (g wet ■- parasite division rate (d"'), Ap = threshold parasite where P^^- wtf^h/jd density (cells (g wet wt)~'), and Gp = maximum parasite density (cells (g wet wt)"'). Since P. marinus growth has been demonstrated to be density dependent (Saunders et al. 1993, La Peyre 1996) a nonlinear den- sity dependent negative feedback mechanism was incorporated into the equation for cell growth. Maximum density (Gp) was set at 2,000.000 parasites (g wet wt)"'. This value was based on maximum P. marinus densities observed in oysters sampled from Wreck Shoal. James River. Virginia during a study conducted from 1993 to 1994 (Oliver et al. 1998). The maximum density represents the density at which space and or nutrients are so lim- ited that parasite division no longer occurs. The threshold density iAp). the density above which parasite growth is inhibited by intra specific competition for space and or nutrients was set at 10.000 (g wet wt)"'. This value was also inferred from field data and rep- resents an approximation of cell abundance in an oyster having a moderate intensity infection. Parasite division rate (prf) is dependent on temperature and salinity. Field observations indicate that infections progress at tem- peratures exceeding 20°C (Burreson and Ragone Calvo 1996). On this basis, parasite division only occurs at temperatures exceeding 20°C. Temperature control follows a standard exponential form using a Q,,, of 2.0 to calculate cell division at temperatures greater than 20°. Initially the model was run using parasite division rates that were taken from the literature; however, initial model simulations resulted in unnatural population dynamics. Hence an alternative estimation of P. marinus division rate was derived. As an alterna- tive, site specific division rates were derived from field data and laboratory studies. Since 1987, the Virginia Institute of Marine Science has been conducting an intensive survey program to moni- tor P. marinus prevalence and intensity at three oyster bars located in the upper the James River, Virginia. Oysters (/( = 25 ) have been sampled monthly from Wreck Shoal (WS). Honsehead Bar (HH). 234 Ragone Calvo et al. and Deep water Shoal (DWS) (Burreson and Ragone Calvo 1996) (Fig. 2). These oyster bars are located along a salinity gradient with average salinities for the years 1987 to 1994 of 14 ppt at WS. 9 ppt at HH, and 7 ppt at DWS (Burreson and Ragone Calvo 1996). Sampled oysters were examined for P. inarinits using a senii-quantitative assay described by Ray (1954). Using this method infections are categorized according to intensity and as- signed a value from 1-7 for light to heavy intensities, respectively. The weighted prevalence (average infection intensity) can then be calculated yielding a single value that describes the average para- site cell density (Mackin 1962). For the purpose of calculating weighted prevalence for the VIMS data, infection intensities were assigned ranks as follows: 2 for light intensity. 4 for moderate intensity and 6 for heavy intensity. More recently a quantitative assay was developed for determining total parasite density in whole oysters (Choi et al. 1989). Whole oyster parasite densities were found to significantly correlate with Ray assay tissue ranks (Choi et al. 1989. Bushek et al. 1994). The regression equation presented by Bushek et al. (1994) (y = 0.176 - 0.463x + Q205\-. where y = log 10 P. inarinits cells (g wet tissue weight)"' and x = infection rank) was used to convert VIMS oyster disease survey data from weighted prevalence to average parasite density (Fig. 3). Cell division rates were then determined using the calculated para- site densities from each station for the months May to June and June to July 1990 to 1993. During these months temperature is rising above 20''C and cell densities are generally low. so the rates determined are appropriate estimates for an initial division rate. Di\ision rates were calculated using the natural log formula for population growth. The average division rates were respectively 0.009. 0.014. and 0.042 divisions d~' at DWS. HH. and WS. Average salinity at the sites for the same period was 4.6 ppt at DWS. 7.1 ppt at HH. and 11.7 ppt at WS. Using data from a laboratory study presented by Ragone and Burreson (1993). the rate of division at 20 ppt was estimated to be 0.06 d"'. A signifi- cant correlation was observed between salinity and the estimated division rates (P = 0.013. r" = 0.973). and the resultant regression equation (y = -0.009 -I- 0.004x, where x = salinity) was used as a first approximation to incorporate salinity control on fjd into the 37° 20' 37° GO' 76° 40' 76° 30' Figure 2. James River, Virginia oyster disease monitoring sampling sites. Population Dynamics of Perkinsus marinus 235 Figure i. Conversion of P. marimis weighted prevalence (squares) to cells (g wet tissue wll"' (circles). Converted IMMO to 1993 data is shown lor three James River, Virginia oyster bars — Wreck Shoal (topi. Horsehead (middle) and Deepwater Shoal (bottom). Conversions were calculated using equation presented by Bushek et al. 1994. model. Sensitivity testing lead to refinement of this relationship and the final equation for parasite division was derived as: |jL12 ppt and for 3-8 ppt. At <3 ppt, host physiological processes are assumed to be inactive, so parasite mortality rate was set extremely low. In the model parasite mortality occurs only at temperatures <18°C and >3°C. Between 18 and 3°C. temperature control of parasite mor- tality follows a standard exponential decay fonn using a QIO of 2.0. Base rates were derived through sensitivity analysis of values within the range of those observed in the field. The equation for parasite mortality was derived as: : P * \xm (4) where P = parasite density (cells (g wet wt)"') and^;» = specific parasite mortality rate (d"'). Such that: |j.m (t) = )JL/M,„,f (5) where /jm{t) = specific rate of parasite mortality; |ja?7,„, = parasite mortality rate (d"') as a function of salinity where: 236 Ragone Calvo et al. |jLm,„, = 0.01 75 at S > 1 2 and 8 > 3 ppt. = 0.02 at 12 >Ss 8 ppt and rt = 0.003at5< 3 ppt; « = QIO conversion (0.0693 1 ), S = salinity (ppt) and T = temperature (°C). Perkinsiis mariniis Transmission Although it is well documented that transmission of P. marinus is direct from oyster to oyster, the natural dynamics of transmis- sion are poorly understood. The prevailing conceptual model is that transmission occurs during periods of high oyster mortality as infective P. marinus cells are disseminated upon death and decom- position of infected oysters (Andrews 1988). Flow cytometric techniques have been developed that allow quantification of dis- seminated P. inarinns-like cells in the water column (Roberson et al. 1993). This technique was employed to systematically examine the seasonality of P. marinus infection acquisition in oysters in relation to water column abundance of P. marinus-\\ke cells, oyster mortality, and temperature (Ragone Calvo et al. 1995). Distinct peaks of all three parameters occurred during the month of August, following maximal summer temperatures (Fig. 4) (Ragone Calvo et al. 1995). Water column P. marimis-Wke cell abundance, infec- tion acquisition, and oyster mortality decreased from summer maximums as temperatures decreased in September and October and remained at "wintertime" low levels from October through the termination of the study in March. These results support the cur- rently accepted hypothesis that infective stages of P. marinus originate primarily from dying oysters and are most abundant in August. Simulations were run both with and without transmission of parasites. When transmission is incorporated in the model, it oc- curs as a single event occurring once each year on Julian day 2 1 S. This timing corresponds with the early August transmission event observed by Ragone Calvo et al. (1995). In the model, the number of parasite cells transmitted is a function of salinity such that: r2S00 i^'s.,1 (6) where //?,„, = the specific rate of parasite transmission (d ) as a function of salinity and has default values of: 1000 cells at 5 > 15 ppt, 500 cells at 15 >S> 12 ppt. 100 cells at I2>5>9. 50 cells at 9 > S > 6 ppt and 0 cells at S < 6 ppt. Salinity control of transmission is assumed to occur for two reasons: ( 1 ) infective cells are likely diluted by river discharge and (2) P. marinus infection levels and associated oyster deaths are limited in lower salinity areas. Clearly, this is an oversimplifica- tion of a process that is quite complex, but for lack of information it is a first approximation for model development and simulation analysis. Parasite Loss via Host Death It is assumed that the average P. marinus within-host abun- dance in any oyster population decrea.ses as a result of the death of Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar □ Prevalence ■ Temperature n n 500 5 30 -25 20 I (D 15 g. 10-5 n Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Figure 4. Results of /'. marinus transmission study conducted by Ragone Calvo et al. (1995). Perl 100.000 cells(g wet wt)"' and 0.00 at P(I) < 100.000 cells(g wet wt)~'. En vironmenlal Parameters The model is driven by two en\ ironmental factors: temperature and salinity. The water temperature data series for both the model development data set (1990 to 1993) and the validation data set (1994 to 1999) were derived from a continuous monitor at the VIMS pier in the lower York River. Measurements were recorded at 6-min intervals; however, only daily means are utilized in the model. The salinity time series for Deepwater Shoal. Horsehead Rock, and Wreck Shoal were compiled from three sources — biweekly measurements taken by the State Water Control Board, weekly measurements (May through October only) taken in con- junction with the VIMS shellstring survey, and monthly measure- ments taken in conjunction with the VIMS oyster disease moni- toring program. Monthly means for each of the three stations were POPL'LATION Dl'NAMICS OF PERK/NSCS MARINUS 237 determined from the pooled data sets tor incorporation into the model. Model Calihratinn and Validation As described above, since 1987. the Virginia Institute of Ma- rine Science has been conducting an intensive survey program to monitor P. mariiius prevalence and intensity at three oyster bars located along a salinity gradient in the upper the James River. Virginia. Separate portions of this data set were utilized for model calibration and validation. Data collected from Wreck Shoal dur- ing 1990 to 1993 was utilized for model calibration. The model was subsequently validated using 1994 to 1999 time series for Wreck Shoal, Horsehead Rock, and Deepwater Shoal. Since the monthly determinations of P. marinus prevalence and intensity were determined using Ray's fluid thioglycollate tissue assays (Ray 1954), which is only semi-quantitative, it was first necessary to convert the data to a quantitative format. This was accomplished using a conversion equation presented by Bushek et al. ( 1994) as described above. The conversion yielded average within-host abundances of P. mariniis cells per gram wet tissue weight. Initial conditions for both model development and validation simulations were set to correspond to the actual P. marinus abundances that were observed in the first month of the time series (Table 1 ). //; vivo parasite abundance was tracked at a daily time step. Statistical Analysis The significance of the association between predicted and ac- tual P. iiiaiiinis abundance was examined by linear regression using Abacus Concepts, Statview (Abacus Concepts, Inc., Berke- ley, CA, 1992). Additionally, sensitivity analysis was conducted to examine the effect of variation of model parameters {/nil. /jin. /jl. and jjl} and input variables (S and T) on model performance. The rate parameters and input variables were systematically varied by ± 2^^ and sensitivity effects were assessed by calculating the residual, or root, mean square deviation (RMS) (Buzzelli et al. 1999) between the nominal model and the test cases, such that: RMS -- Vil*^'-^''' where/', = model nominal (predicted) value, (5, = sensitivity test case value, and n = number of daily time steps in the six year simulation in = 2190). Differences between comparisons were normalized by calcu- lating percent errors where the average RMS for the -i-20% and -20'7f test conditions is divided by the mean of O,. The model is considered to be sensitive to parameters that yield 9c errors greater than 10%. RESULTS Model Calibration The simulated within-host abundance of P. marinus using a lower York River water temperature time series and Wreck Shoal. James River salinity time series for the years 1990 to 1993 sig- nificantly correlated with actual observed abundance (P < 0.0001. r^= 0.761) (Fig. 5). The simulation generated a dynamic steady state in which parasite abundance exhibited a distinct seasonal periodicity with maximum abundance occurring in mid-August through mid-September following peak summer temperatures. Figure 5. Simulation results for model development-time series, Wreck Shoal, James River. Mrjjlnla 1940 to 1993, showing tempera- ture and salinity time series (topi and predicted versus actual observed P. marinus abundances (bottom). Maximum cell division rates ranged from 0.09 to 0.11 divisions d"'. These rates resulted in generation times from 6-7 days. Para- site abundance declined as temperature decreased during the fall and winter months, and minimums were observed from late April to mid-May. Deviations of the predicted abundance from the actual abundance were most evident in year one and year three at times of minimum and maximum abundance. In each year of the time series maximum P. marinus abundance exceeded 100.000 cells (g wet wt)"' and hence decreases in abundance associated with host death were observed. The occurrence of parasite loss due to host death appears as very short term oscillations in the simulation: population mean abundance is depressed due to parasite loss and then immediately increases as a result of parasite growth. Hence, the duration of the short-term oscillations yields information on the relative extent of parasite-associated host mortality from year to year. The 1990 to 1993 environmental time series was quite variable. Annual salinity minimums and maximums ranged from 0.5 to 9.4 ppt and 15.7 to 19.1 ppt, respectivels. while annual temperature minimums and maximums ranged from 4.2-6.9°C and 26.4- 28.0°C, respectively (Fig. 5). Model Validation The model was validated using a lower York River temperature time series and site-specific salinity time series for Deepwater Shoal. Horsehead Bar, and Wreck Shoal, James River. Virginia for the years 1994 to 1999. The environmental time series included a broad range of conditions with unusually warm years (1997, 1998) and cold years (1996). as well as wet ( 1996. 1998) and dry years (1995. 1999). Annual teinperature minimums ranged from 2.7 to 7.7°C. and annual temperature maximums ranged from 26.7 to 28.4°C (Figs. 6-8). Annual salinity minimums ranged from 2.0 to 238 Ragone Calvo et al. 1995 1996 1997 Year Figure 6. Validation model simulation results for Wreck Siioal, James River, \ Irginia 1^94 to 1999 showing temperature and salinity time scries (topi and predicted versus actual observed P. mariiius abun- dances (bottom). Figure 8. Validation model simulation results for Deepwater Shoal, James Ri\er. \ irginia 1994 to 1999 showing temperature and salinity time series (top) and predicted versus actual observed P. marinus abundances (bottom). 12.5 ppt at WS, 0 to 5.5 ppt at HH, and 0 to 4.5 at DWS, while annual maximums in salinity ranged from 14.1 to 20.5 at WS, 7.3 to 16.0 at HH. and 6.2 to 13.7 at DWS (Figs. 6-8). Predicted P. iiiariiui^ abundance significantly correlated with /- p/ Predicted Actual S6- -6 §5^ A -5 %A- A ■A A / -4 O) / \ A 7 \ / \ / "3- / \ / \ / \ / \ / -3 S ; /, \ / \ A"^^ / ^~-^^ \ / o^- ^^^■: V J V ■^ vV -V -■2 ° 1- -1 0- -0 1995 1996 1997 Year 1998 1999 Figure 7. Validation model simulation results for Horsehead .Shoal. James River, Virginia 1994 to 1999 showing temperature and salinity time series (top) and predicted versus actual observed P. marinus abundances (bottom). actual observed abundance for all three sites; however, the strength of the correlation decreased in an upriver direction (P < 0.0001. r" = 0.724 at WS. 0.598 at HH. and 0.450 at DWS) (Figs. 6-8). As observed with the 1990 to 1993 time series simulations, simula- tions for all three oyster populations generated a dynamic steady state in which parasite abundance exhibited a distinct seasonal periodicity. Parasite abundance exponentially increased during the summer months reaching an annual peak in abundance between mid-September and mid-October. As temperature declined below \8°C during the fall, parasite abundance decreased. This decrease in parasite number continued through the winter and early spring reaching an annual minimum between late April and mid-May. Both the magnitude and timing of annual maximum parasite abun- dance varied both with respect to year and location. Annual abun- dance maximums were higher and generally occurred earlier at WS (Fig. 6) than at HH (Fig. 7) and at HH than at DWS (Fig. 8). The predicted annual periodicity of P. marinus abundance, par- ticularly for WS, closely followed actual observed infection inten- sity and prevalence patterns in upper James River oyster popula- tions. Perkinsiis marinus abundance in the upper James River is typically highest during the months of September and October and lowest in May. Parasite abundance decreases in an upriver direc- tion from WS to DWS and abundance peaks later in the upper river locations. For WS predicted annual averages, maximums and minimums in parasite abundance were always the same order of magnitude as actual observed abundance; however, at HH and DWS differences greater than I order of magnitude were some- times observed. Specifically, at HH. 1998 predicted maximum abundance exceeded actual observed abundance by 1 .03 orders of magnitude, and at DWS 1995 predicted minimum abundance were 1.38 orders of magnitude lower than actual observed values, and 1998 predicted abundance maximums exceeded actual abundance by I order of magnitude. PoPL'LATiON Dynamics of Perkinsus marinus 239 §6- Predicted Actual -6 -5 S4- A i\ f\ 1 -4 1 \ \/ vy -3 S2- J vy r2 2i- -1 en ° n 1999 Figure 9. Model simulation results of predicted and actual observed P. marinus abundance for Wreck Shoal (top), Horsehead Rock (middle), and DeepHater Shoal (bottom! for the period 1994 to 1999 for condi- tion in which annual transmission input is eliminated from model. In the model P. marinus proliferation occurred only at teni- peratures greater than or equal to 20°C at salinity greater than 0 ppt. On average, these conditions were met. and cell division oc- curred 1 39 days per year. The number of days in which the parasite was active decreased in an upriver direction and inter annual vari- ability was relatively high with the annual nuniber of proliferation days ranging from 135-154 days at WS. 135-150 days at HH, and 90-150 days at DWS. Overall predicted mean, minimum and maximum cell division rates (d~') were respectively: 0.061 (± 0.017 SD). 0.023. and 0.104 at WS; 0.034 (± 0.017 SD), 0.0006, 0.073 at HH; and 0.026 (± 0.014 SD). 0.0001. 0.063 at DWS. The predicted division rates correspond to generation tinies ranging from 6 to 34 days. Parasite death as incorporated into the model was modulated by both temperature and salinity. On average, parasite regression oc- curred 196 days per year. Predicted parasite mortality rates ranged from 0.006 to 0.057 d"'. Average predicted parasite mortality rates (d~') were respectively 0.031. 0.023. and 0.021 at WS. HH. and DWS. Overall parasite regression rates were higher at WS than at HH and DWS; however, there were exceptions as both HH and DWS had higher parasite mortality rates than WS in 1995 and 1999. Parasite loss via host cell death was incorporated into the model such that the state variable, parasite abundance, was reduced by 50% at parasite densities greater than 100,000 (g wet wt)"'. At WS, parasite loss was an important regulator of parasite abundance and is evident in the simulations as high frequency short-term high oscillations in parasite abundance occurring coincidentally with summer abundance maximums. The periodicity of parasite cell loss due to host death at WS varied from year to year, being most apparent in 1999 and absent in 1996. Parasite burdens did not escalate to lethal levels at HH and DWS: hence loss via host death did not impact parasite abundance at these sites. Parasite transmission was incorporated into the model as a single annual pulse of parasite cells into the host. The magnitude of the pulse is dependent on ambient salinity at the time. The number of cells transmitted ranged from 50 at salinity between 6 and 9 ppt to 1000 at salinity greater than 15 ppt. In the simulations transmission occurred in all years at all sites and varied from 50 to 100 at DWS, 50 to 500 at HH, and 500 to 1000 at WS. In order to examine the relative importance of annual transmis- sion events in maintaining P. niariiiiis epizootics, the transmission component was eliminated from the model and simulations were run without the occurrence of an annual transmission event. The modified model produced simulations for WS and HH in which parasite abundance remained in a dynamic steady state (Fig. 9). Predicted P. marinus abundance for WS and HH correlated with actual observed abundance {P < 0.001. r-= 0.709 at WS. r- = 0.444 at HH); however, the correlation was not as strong as that produced by the original model. In contrast, P. marinus abundance for DWS did not maintain a dynamic steady state and parasite abundance gradually decreased over time with parasite extinction occurring in year 6 (Fig. 9). Sensitivity Analysis Sensitivity analysis was conducted in order to evaluate how relative changes in model paraineters would effect model perfor- mance. The sensitivity of the model to key rate parameters varied with sampling location. For Wreck Shoal, the model was sensitive TABLE 2. Sensitivity analysis results for simulations for Wreck Shoal. Horsehead Rock, and Deepwaler Shoal 1994-1999. Root mean square deviation (RMS) values are shown for ■f20% variation for each parameter. Average Site Parameter +20% -20% RMS % Error Wreck Shoal ful (),IS 0.23 0.21 5.7 jjin 0.42 0.48 0.45 13.3* M' 0.06 0.10 0.08 2.4 t^l 0.08 0.16 0.12 3.4 S 0.19 0.42 0.30 9.2 T 0.65 1,50 1.07 43.0* Horsehead fd 0.57 0.58 0.57 21.1* tim 0.55 0.76 0.66 22.8* /J' 0.05 0.05 0.05 1.8 fl 0.(X) 0,00 0.00 0.00 S 0.76 0.80 0.78 29.9* T 1.16 1.48 1.32 64.5* Deepwater Shoal pd 0.28 0.35 0.32 15.9* fjni 0.32 0.46 0.39 18.8* A" 0,07 0,08 0.07 3.7 A-/ 0.00 0.00 0.00 0.0 S 0.47 0.98 0.73 49.1* T 0.88 1.64 1.26 59.3* ■ Sensitive parameters. 240 Ragone Calvo et al. to only one rate parameter, parasite mortality, iiin. A 20% cliange in jdm resulted in a 13% error {Table 2). When inodel performance was assessed using the time series for Horsehead Rock and Deep- water Shoal, the inodel was sensitive to both parasite division rate and mortality rate and percent errors ranged from 16 to 23% (Table 2). The model was also very sensitive to temperature and salinity. Percent errors associated with a 20% change in temperature were 43% for Wreck Shoal. 66% at Horsehead Rock, and 59% at Deep- water Shoal. Salinity had a lesser effect than temperature. For Wreck Shoal, the model was not sensitive to 20% change in sa- linity, but for Horsehead Rock and Deepwater Shoal percent errors were relatively high, 30% and 49% respectively. DISCUSSION The model as currently constructed is simple but provides re- alistic population simulations for P. marinus within its host C. virginica. The dynamics of the populations described by these simulations reflects the importance of temperature and salinity in regulating P. marinus population dynamics. The annual cycle of parasite abundance exhibited a strong seasonal periodicity reflect- ing the controlling influence of temperature. As observed in nature (Andrews 19SK. Burreson and Ragone Calvo 1996). simulated parasite abundance peaked approximately 2-3 months after annual maximums in temperature were observed. Under favorable salinity conditions parasite growth occurred at temperatures greater than 20°C, and parasite death occurred at temperatures less than 1 8°C. Hence, abundance increased from late spring until early fall and decreased from late fall to early spring. The magnitude of parasite abundance maximums and mini- mums was modulated by salinity. Perkinsits marinus abundance progressively decreased with decreasing salinity. This pattern, which is the result of slower growth rates at low salinities com- pared to high salinities, is also observed in natural populations. Perkinsus marinus prevalence and infection intensity in oysters sampled along a salinity gradient in the upper James River. Vir- ginia were correlated with salinity (Burreson and Ragone Calvo 1996). Summer and fall P. marinus infection intensities in oysters from the lower-most salinity areas were always light in intensity while moderate and heavy intensity infections were common in oysters collected from higher salinity areas (Burreson and Ragone Calvo 1996). Similarly, the abundance of overwintering parasites was higher in higher salinity areas than in lower salinity areas (Burreson and Ragone Calvo 1996). Correlations between predicted parasite abundance and actual observed abundance were surprisingly strong, indicating that the main factors responsible for the regulation of P. marinus-C. vir- ginica population dynamics in the upper James River. Virginia have been incorporated in the model. However, there remains room for improvement and some discussion regarding the devia- tion of predicted abundance from actual observed abundance. The strength of the correlation between the predicted and observed data decreased with decreasing salinity. This deviation may be in part due to two reasons. First, actual observed abundance data were calculated from a database that relied on a semi-quantitative diag- nostic method. It is likely that some en'or was associated with the conversion of the semi-quantitative data to a quantitative format. The numerical ranks assigned (2. 4, and 6) to convert infection intensity scores (light, moderate, and heavy) to weighted pre\a- lence were restrictive in setting lower and upper limits to the data. Much of the deviation observed in the DWS simulation is likely a result of this conversion. The model consistently generated lower minimum parasite abundance than that observed in the actual data; however, by assigning all infections that were categorized as light a ranking value of 2. the observed data minimums were most likely somewhat inflated. Second, there may be differences in parasite susceptibility between the three oyster populations. Wreck Shoal oysters, located farther down river, have likely experienced much stronger selection pressure from the parasite than Horsehead Rock and Deepwater Shoal oysters. That selection pressure may have affected parasite division and mortality rates, hence, parasite rates that were calibrated based on Wreck Shoal data may not be the best fit for the stations located farther upriver. The results of sensitivity analysis indicated that several factors are important in determining model performance. The model was sensitive to salinity and temperature as well as to both parasite division and mortality rates. Interestingly, the degree of the mod- el's sensitivity to these parameters varied with oyster bar. A twenty-percent change in the parameters had a lesser effect at Wreck Shoal than at Horsehead Rock and Deepwater Shoal. Since Horsehead Rock and Deepwater Shoal are located farther upri\er than Wreck Shoal they more frequently experience salinity de- clines, which can be unfavorable for both parasite and host. It is likely that environmental thresholds affecting parasite and host activities are more frequently approached in the upriver locations. Estuarine environments experience large salinity and temperature fluctuations in relatively short time periods. In developing the model we derived our principal P. );i(()7/)».v-salinity relationships from a record of monthly field data that did not allow us to char- acterize the effect of shorter-term environmental fluctuations. Fur- thermore, it is likely that there are lag periods associated with initiation of parasite and host activities following the occurrence of a threshold condition. As presently constructed, lag times are not considered in the model. More detailed information about the en- vironmental control of parasite division and mortality rates would improve model performance. This information would best be pro- vided by experimental studies under controlled conditions. Perhaps one of the most interesting results of the simulations was the fact that at WS and HH. a steady-state parasite abundance could be generated without the input of annual transmission events. Although parasite extinction ultimately occurred at DWS under conditions of no transmission, extinction took a period of nearly 6 years. These simulations suggest that a single transmis- sion event may be sufficient for P. marinus to become enzootic within specific year classes of oysters in the moderate to high salinity areas, while periodic transmission events are required for the parasite to persist within specific year classes in low salinity areas. In the moderate-high salinity areas, parasite abundance would ultimately decrease as the pool of infected oysters is diluted by the presence of new recruits and the parasite would eventually become extinct when all infected hosts die. Perkinsus marinus became established in the upper James River oyster seed bed areas as a consequence of several consecutive drought years in the mid 1980s. Since that time the parasite has continued to be present in these historically disease-free areas. If the model simulations are correct, this would suggest that more than one transmission event has occurred in these low salinity areas. This clearly presents a very interesting area of focus for future research that could lead to the development of new management strategies. The current con- ceptual model of P. marinus transmission dynamics maintains that transmission occurs as infective cells are disseminated from dead hosts. Since there is little disease-associated mortality of oysters in Population Dynamics of Perkinsi^s makinus 241 these i(n\ s;ilinity ureas, transmitted eells must be originating from live oysters or from downriver sourees. Intensive harvesting of oysters from these downriver source areas, on a rotational basis, may be a means of preventing the recurrence of P. nuiiiiiiis trans- mission into the uppermost James River seed areas. The outcomes of simulations generated by the model presented here are quite different from those presented by Powell et al. ( 1996). Powell et al ( 1996) developed a more complex model for F. imiiiiuis population dynamics coupled with oyster population dynamics in the Gulf of Mexico. Interestingly, their simulations were unable to generate an epizootic simply with changes in tem- perature and salinity. The outcomes of their simulations suggest that oyster populations must first be stressed by some other mecha- nisms before high temperature and high salinity conditions can facilitate epizootic oyster mortalities. They point to food availabil- ity and recruitment failure as two such mechanisms. Until the Chesapeake Bay model presented here is further developed to include host population dynamics the comparison with the Gulf t)f Mexico model remains problematic. However, it is interesting to note that results of simulations from both models indicate that once an epizootic is triggered it is very difficult to eliminate. As currently constructed, this model provides realistic popula- tion simulations for P. inaiiiuis populations in the Chesapeake Bay. Development of the model has provided an excellent format for synthesizing data and for focusing attention on gaps in knowl- edge and research needs. Future development and evaluation of the model will yield a better assessment of thresholds for population maintenance and a better under.standing of conditions that may lead to the termination of /-•. nniriiuis epizootics and parasite ex- tinction. ACKNOWLEDGMENTS Thanks to Bill Seufzer and Jessica Morgan for assistance with Stella programming. VIMS contribution 2.^88. LITERATURE CITED Anderson. R. M. & R. M. May. 1978. Regulation and stability of host parasite interactions I. Regulatory processes. / Aniiii. Ecol. 47:219- 247. Anderson. R. S.. E. M. Burreson & K. T. Paynter. 1995. Defense responses of hemocytes withdrawn from Crassostrea virginicii infected with Per- ktn\ii.\ nicinmis. ./. Iiwerwhr. Pathol. 66:82— 89. Andrews. J. D. 1954. Notes on fungus parasites of bivalve molluscs in Chesapeake Bay. Proc. Natl. Shellfish Assoc. 45:157-16.V Andrews. J. D. 1988. Epizootiology of the disease caused by the oyster pathogen Perkinstis nwriiius and its effect on the oyster industry. Am. Fish. Soc. Spec. Piihl. 18:47-63. Andrews. J. D. & S. M. Ray. 1988. Management strategies to control the disease caused by Perkinsits mtiriiius. .Am. Fish. S<>c. Sfwc. Piihl. 18: 257-264. Burreson, E. M. & L. M. Ragone Calvo. 1996. Epizootiology of Perkinsiis marimis disease of oysters in Chesapeake Bay. with emphasis on data since 1985. / Shellfish Res. 15(1): 17-34. Bushek, D.. S. E. Ford & S. K. Allen. 1994. Evaluation of methods using Ray's fluid thioglycollate medium for diagnosis of Perkinsiis mariniis infection in the eastern oyster. Crassostrea virginica. Aiiiiii. Rev. Fish Dis. 4:201-217. Buzzelli. C. P., R. L. Wetzel & M. B. Meyers. 1999. A linked physical and biological framework to assess biogeochemical dynamics in a shallow estuarine ecosystem. Estuarine. Coastal and Shelf Sci. 49:829-851. Choi. K. S., E. A. Wilson, D. H. Lewis, E. N. Powell & S. M. Ray. 1989. The energetic cost of Perkinsus mariniis parasitism in oysters: Quan- tification of the thioglycollate method. J. Shellfi.ih Res. 8:125-131. Chu, F. L. & K. H. Greene. 1989. Effect of temperature and salinity on in vitro culture of the oyster pathogen Perkinsus marimis ( Apicomplexa; Perkinsea). J. Invertebr. Pathol. 53: 260-268. Chu. F. E. & J. F. La Peyre. 1993. Perkinsus marimis susceptibility and defense-related activities in eastern oysters Crassostrea yirginica: tem- perature effects. Dis. Aquat. Org. 16:223-234. Chu. F. L. & A. K. Volety. 1997. Disease processes of the parasite Perk- insiis marimis in eastern oysters Crassostrea virginica: Minimum dose I'or infection intitiation, and interaction of temperature, salinity, and miective cell dose. Dis. Aquat. Org. 28:61-68. Dungan. C. F. & R. M. Hamilton. 1995. Use of a tetrazolium-based cell proliferation assay to measure effects of in vitro conditions on Perk- insus marimis (Apicomplexa) proliferation. J. Eukarxul. Microhinl. 42(4):379-388. Ford. .S. E. 1996. Range extension by the oyster parasite Perkinsus nuirinus into the northeastern United States: Response to climate change. J. Shellfish Res. 15(l):45-56. Ford. S. E.. A. Schotthoefer & C. Spruck. 1999. hi vivo dynamics of the microparasite Perkinsus marinus during progression and regression of infections in eastern oysters. J. Parasitol. 85(2):27.3-282. Gauthier. J. D. & G. R. Vasta. 1995. In vitro culture of the eastern oyster parasite Perkinsus marinus: Optimization of the methodology. J. In- vertebr. Pathol. 66:156-168. Grenfell, B. T. & A. P. Dohson. 1995. Ecology of Infectious Diseases in Natural Populations. New York: Cambridge University Press. La Peyre. J. F. 1996. Propagation and in vitro studies oi Perkitms marimis. J. Shellfish Res. 15( 1):89-101. Mackin. J. G. 1962. Oyster disease caused by DennocysliJium marimiin and other microorganisms in Louisiana. Publication of the Institute of Marine Science. University of Te.\as 7: 1 32-229. Mackin, J. G.. H. M. Owen & A. Collier. 1950. Preliminary note on the occurrence of a new protistan parasite, Dermocystidium inarinum n. sp. in Crassostrea virginica (Gmelin). Science 111:328-329. 01i\er. L. M.. W. S. Fisher. S. E. Ford, L. M. Ragone Calvo, E. M. Burreson. E. B. Sutton & J. Gandy. 1998. Perkinsus marinus tissue distribution and seasonal variation in oysters Crassostrea virginica from Florida. Virginia and New York. Dis. Aquat. Org. 34:51-61. Powell. E. N.. J. M. Klinck & E. E. Hofmann. 1996. Modeling diseased oyster populations. II. Triggering mechanisms for Perkinsus marinus epizootics./ Shellfish Res. 15( 1 ): 141-165. Quick, J. A. & J. G. Mackin. 1971. Oyster parasitism by Lubyrinihoiny.xa marina in Florida. Florida Department of Natural Resources Profes- sional Paper Scries 13:1 -55. Ragone, L. M. & E, M. Burreson. 1993. Effect of salinity on infection progression and pathogenicity ot Perkinsus marinus in the eastern oys- ter. Crassostrea virginica (Gmelin). J. Shellfish Res. 12(l):l-7. Ragone Calvo. L. M.. E. M. Burreson. C. F. Dungan & B. S. Roberson. 1995. Field and laboratory study of the processes and dynamics of Perkinsus marinus infection in the eastern oyster. Crassostrea vir- ginica. Completion Report NOAA. NMFS Oyster Disease Research Program. Gloucester Point. VA: Virginia Institute of Marine Science. 24 pp. Ray. S. M. 1954. Experimental studies on the transmission and pathoge- nicity of Dermocystidium manniim. a fungus disease of oysters. J. Parasitol. 40:235. Roberson. B. S.. T. Li & C. F. Dungan. 1^93. Flow cytometric enumeration and isolation of imniunotluorescent Perkinsus marinus cells from es- tuarine waters. / Shellfish Res. 12:138. Rohde. K. 1982. Ecology of Parasites. New York: University of Queens- land Press. Saunders, G. L.. E. N. Powell & D. H. Lewis. 1993. A determination of in vivo growth rates for Perkinsus marinus. a parasite of Crassostrea virginica. J. Shellfish Res. 12(2):229-240. Journal oj Shclljiyh Kescunh. Viil. 20. No. I. 24.V265. 2001. A BIOCHEMICALLY BASED MODEL OF THE GROWTH AND DEVELOPMENT OF CRASSOSTREA GIGAS LARVAE ELEANOR A. BOCHENEK,' JOHN M. KLINCK,' ERIC N. POWELL,' AND EILEEN E. HOFMANN- 'Haskin Shellfish Research Laboratory. Ritlgers University; 695^ Miller Ave.. Port Norris, New Jersey US349: -Center for Coastal Physical Oceanography. Crittenton Hall. Old Dominion Universit}; Norfolk. Virginia 23529 ABSTRACT A biochemically based model was developed to simulale the grow th. development and metamorphosis of larvae of the Pacific oyster, Crassosrrea gigas. The model is unique in that ( 1 ) it defines larvae in terms of their protein, neutral lipid, polar lipid, carbohydrate, and ash content; (2) it tracks weight separately from length to follow larval condition index; and (3) it includes genetic variation in growth efficiency and egg quality to better simulate cohort population dynamics. The model includes parameterizations for larval filtration, ingestion, and respiration, which determine growth rate, and processes controlling larval mortality and metamor- phosis. The initial biochemical content of the larva is determined by the composition of the egg. Changes in the initial ratios of protein, carbohvdrate. neutral lipid, and polar lipid occur in response to the biochemical composition of available food as the larva grows. Modeling the process of metamorphosis requires a series of size-based and biochemically based triggers: ( I ) larvae become potentially competent to metamorphose at 275 |a.m. following a decrease in filtration rate at 2.50 (im; (2) larvae become competent to metamor- phose when a daily decline in neutral lipid of 25'7f or more occurs; and (3) larvae metamorphose successfully if neutral lipid stores exceed polar lipid stores. Although based on simple biochemistry, the model succeeds in simulating such basic characteristics of C. gigas larval development and metamorphosis as larval life span and size structure at metamorphosis and the influence of egg and food quality and food quantity on survival. These results suggest that simple biochemical constructs may encompass the biochemical transitions most prominent in determining cohort success. Simulations of larval development show that for the smallest larvae, assimilation does not provide adequate resources to explain observed growth, although measured filtration rates would indicate otherwise. Egg lipid stores are needed to sustain the larva. The simulations also identify egg sizes in the range 37-73 [xm to be viable, very similar to observations. Egg sizes outside this range are predicted to be nonviable due to lipid deficiencies in early larval life. Similarly, simulations identify upper and lower genetic limits on growth efficiency beyond which larvae cannot acquire sufficient neutral lipid stores to successfully metamorphose. As food supply declines, animals with high growth efficiencies are selected in the simulation. Low-protein food diets are predicted to increase larval survival. High-protein diets result in insufficient carbohydrate and neutral lipid to cover metabolic and storage needs. Thus, the influence of growth efficiency is nonrandomly distributed across egg size and respiration rate and the influence seems to be mediated in part by food quantity and to a larger measure by food quality. KEY WORDS: Cnusoslrea gigas. larvae. Pacific oyster, growth, development model, biochemical, food INTRODUCTION For benthic species that have a planktonic larval phase (if their life history, survivability of the larvae is the key deterniinunt of recruitment to the adult population. Thus, much research has gone into identifying factors affecting the growth, development, meta- morphosis, and settlement of larvae, especially for species with commercial value, such as the Pacific oyster, Crassostreu gigas. Larvae of C. gigas undergo growth and development that is typical of bivalve larvae, with progression through D-shaped. umbo, and pediveliger stages (Fig. 1 1. The fraction of develop- mental time spent in each stage is variable and the rate at which the larvae pi-ogress through each is affected by local temperature, sa- linity, and food conditions (Helm & Millican 1977, Malouf & Breese 1977, Nascimento 1980, Gerdes 1983a, His & Maurer 1988. Nell & Holliday 1988; His et al. 1989, Pauley et al. 1988, Arakawa 1990. His & .Seaman 1992. Robinson 1992. Thompson & Harrison 1992, Laing 1993, Thompson et al. 1996). The range of fractional development times reported for the small umbo to pediveliger stages (Fig. 1 ) is a reflection of these effects on larval growth. Once spawned, the ultimate fate of C. gigas larvae is deter- mined by the interaction of a number of factors. The first is the initial biochemical composition of the egg released by the adult, in particular the initial egg lipid content. Studies have shown corre- lations between egg size, lipid content, and bivalve larval devel- opment (Helm et al. 1973, Gallager et al, 1986, Gallager & Mann 1986a, Lee & Heffeman 1991). An implication of these studies is that initial egg lipid content is a determinant of larval survivorship and success at metainorphosis. A second factor is the ability of the larvae to grow and develop so that total time spent in the plankton is minimized, thereby reducing exposure to mortality from predation. Because spawning pulses by adult populations, in nature, occur at widely variable times throughout the spawning season, larvae experience a widely varying set of environmental conditions. The timing of environ- iTiental conditions relative to a particular phase of larval life can greatly affect the total time needed for larvae to complete devel- opment (Dekshenieks et al. 1993). The concept of a critical period immediately post-hatch during which many planktonic larvae are particularly susceptible to low food supply is now well established (Gushing & Dickson 1976, Anger et al. 1981, Taggart & Leggett 1987), and the available data for bivalve larvae support the sensi- tivity of larvae to periods of low food supply (Gallager et al. 1986, His and Seaman 1992, Laing 1995). Studies also suggest that food quality, not just food quantity, is a critical issue (Thompson et al. 1994, 1996), Insufficient dietary lipid, for example, significantly limits larval growth and survival (Helm et al. 1973. Gallager et al. 1986). A third factor affecting larval survival is the ability to acquire sufficient internal energy stores to successfully complete metamor- phosis and set. Studies have shown that larvae can reach the size 243 244 BOCHENEK ET AL. 350 300 250- -§■ 200 S 150 100 50 _ f - pedlvellger/ - large umbo y/ / footed / r / umbo/ / small / / umbo/ / Y D-shaped /embryo - 1 1 1 1 1 0.2 0.4 0.6 0.8 Fractional Development Time 1.0 Figure 1. Schematic of the de\el()pniental stages of Crassiislrea gigas lar>ae as a function of fraction of dexelopnient time. The two curves hraclvet the range of sizes for the different stages, as reported in Ar- akawa (1990). needed for metamorphosis but be unable to complete this step (Robinson 19')2. Haws et al. 1993. Laing 1993). The possibility that a shoit-terni deprivation of food early in larval life may reduce metamorphosis success can be inferred from a variety of studies that emphasize the critical importance of adequate food throughout larval life. The inci'easing emphasis on the importance of an adequate diet, mcluding quantity and quality, in determining a larval cohort's success and the recognition that adequate energy resources are needed for successful metamorphosis suggest the need to incorpo- rate biocheinical transfers into models of larval growth and devel- opment. The existing lar\al models determine growth tYom inges- tion rate, decremented by the energy losses due to respiration and incomplete digestion (e.g., Carlotti & Sciandra 1989, Dekshenieks etal. 1993, 1996). These types of models, however, cannot be used to examine issues of food quality, nor can they be used to simulate the energy-reserve hypothesis underlying metamorphosis success. To investigate the influences of food quality and quantity on growth, development, and successful metamorphosis of C. gigas larvae, a biochemically based model was developed that includes explicit parameteri/ations for the metabolism of protein, carbohy- drate, polar lipid, and neutral lipid within the standard parametei- izations of energy flow via ingestion, assimilation, and respiration. The following section provides a description of the C. gigcis larval model and the pai'ameterizations used in this model. This is followed by simulations that illustrate the effect of initial egg size, food quality, food quantity, and environmental conditions on larval growth, survival, and success at metamorphosis. The discussion section places these simulations within the context of the current understanding of the effect of environmental conditions and food quality on larval growth, survival, and metamorphosis. MODEL DESCRIPTION Model Structure and Governing Equation The change in length for an individual larva over time is given by: It' aL (1) whei'e L is larval length in jxm and a is the rate at which the larva grows and has units of day"'. Larval growth rate (a) is based on formulations that allow differential metabolism of the protein, car- bohydrate, and lipid content of the food ingested by the larva. Thus, net production is expressed as the difference between as- similated ingestion (AI) and respii^alion [R): NP, Al, - R, (2) where / represents the four basic biochemical components included in the model: protein, polar lipid, neutral lipid, and carbohydrate. An increase in larval length occurs when the sum of the four, Y.t= i NP. is positive: when larval condition index is maximal for a given size: and when the restrictions imposed by certain biochemical ratios are simultaneously met. Thus, excess net pioduction, ENP. is the basic quantity responsible for larval growth. The specification of ENP. which determines a, is based on filtration, the metabolism of carbohydrates, polar lipids, neutral lipids, and proteins within the larva, and the conversion of the metabolized food into structural components of the larva versus the conversion into storage material. A basic assumption in this model is that the formation of structural components determines the increase in larval length. Material converted into storage com- ponents (i.e., neutral lipids) does not result in increased larval length. The conversions and parameterizations used to model these processes are described in the following sections. The C. gigas laival model given by Eq. ( 1 ) was solved numeri- cally using a third-order Adains-Bashforth scheme (Canuto et al. 1988) with a time step of 0.1 day. Model Parameterizations Observations from field and laboratory studies were used to derive the relationships that describe the processes affecting net production of C. gigas larvae. The basic units used in the model are grams, joules, and (jtm. and these are not necessarily always consistent with the units used for measurements. Thus, the first step in developing the lar\ al model was to obtain conversions that allow the model calculations and output to be consistent with measurements and to be compared with observations. Also, many of the larval processes vary in amplitude or even form with larval size, requiring that lelationships used to describe these be size dependent. Biochemical Model of C. gigas Larvae 245 Lcngth-I()-I)r\ Tissue Weight Conversion Numerous field and laboratory data sets exist that can be used to derive a length-to-dry tissue weight relationship for C. gigas larvae. However, the reporting of these data sets is quite variable, with some reported in terms of dry tissue weight (Gerdes 1983a) and others reported in terms of whole animal weight (His & Mau- rer 1988. Nascimento 1980. Waldock & Nascimento 1979). Dry tissue weight is the desired unit for deriving this relationship, so the data sets reported as whole animal weight were multiplied by a factor of 0.25. which corrects for the shell being 757c of the animal weight (His & Maurer 1988). The resultant data set (Fig. 2) was used to obtain the length-to-dry tissue weight relationship: at" W (3) where IV is larval weight in ng and L is larval length in (a.m. The coefficients a and h are given in Table I . The correspondence between Eq. ( I ) and the data sets is shown in Figure 2. Length-to-.\sh Weight Conversion His and Maurer ( 1988) provide measurements of the percent of total dry weight as a function of larval size. These values were obtained by summing the percent organic matter represented by larval protein, carbohydrate, and lipids that were determined in- dependently. As a comparison. His and Maurer ( 1988) also deter- mine the percent total dry weight of organic matter by combustion of the whole animal. These data form the basis for developing a relationship to relate larval length to ash weight. Comparison of the organic matter values determined by the sum of the components and those obtained from the heating method differ by about 5%. The difference was assumed to be due to protein, which is more completely measured by the heating method, so this difference was added to the protein value for a given size larva. The percent total dry weight for the largest larval size measured by His and Maurer (1988). 274 p.m, was reduced from 29. 1 % to 26.2% because attempts at using the higher value to produce a relationship between larval length and ash weight gave 2000 1600 G) C ■5, 1200 '5 5 0) 3 (A 250 (xm (7) wheie FR is filtration rate. The coefficients in Eqs. 6 and 7 are given in Table 1 . Initial simulations that used the above relationships for larval Biochemical Model of C. g/gas Larvae 247 Metamorphosis - 50 100 150 200 250 Length (i^im) 300 350 Figure 5. L'pper ciir\ e. nitration rate as a t'unction of Crassostrea gigas larval length calculated from Eqs. 6 and 7 as described in the text. Lower curve: nitration rate after corrections described by Eqs. 8 and 9 (for food concentration of 2 nig 1^ '). tlllratioii rate resulted in growth rates that were too high. Reduc- tion of the growth rate was accomphshed by adding a factor to the fikratioii rate that reduces ingestion efficiency. This factor is of the form: IE-- 7 Food a + p L-L (8) where ingestion efficiency (/£) is a nondimensional quantity that depends on the ambient food concentration (Food) and larval size. Coefficient values are given in Table 1 . The relationship given h\ Eq. 8 results in reduced feeding efficiency for all larval sizes, but with the maximum reduction associated with smaller larvae. The rationale for reducing filtration efficiency is that the vel- lum is a multipurpose organ, so there must be some inefficiency in each of the activities and functions of this organ; otherwise the larvae could only swim at the rate that allows maximum ingestion and dispersal, and escape capabilities would be compromised. Thus, filtration rates measured in a laboratory setting for C. gij^as larvae should be regarded as measures of vellum activity and not as measures of ingestion. In addition, most research has been con- ducted using saturating food concentrations. This would exacer- bate any tendency for more food to be filtered from the water column than could be ingested by the larva. Applying the ingestion efficiency factor (Eq. 8) to Eqs. 6 and 7 resulted in realistic larval growth rates, except for larvae smaller than 80 (Jini. Early in larval life the rapid changes leading to the development of the organs for feeding and digestion should further limit ingestion and/or assimilation efficiency. Thus, the filtration rate for small larvae (FR^) was further reduced by: IE=IE-?>{ 1 + mill 1 L-L, L - L, (9) The above approach is based on the assumption that the vellum in small larvae is very inefficient at capturing food particles and/or that digestion is less efficient in small larvae. Coefficient values and definitions are given in Table I . A relationship between low food supply and increased feeding efficiency due to enlargement of the vellum (.Strathmann et al. 1993) was not included in the model because the influence of vellum enlargement on filtration rate is unknown. The model may underestimate growth rate at low food supply. Temperature and Salinity Effects As with most bivalve larvae, the metabolic processes control- ling growth in C. gigas larvae are affected by temperature and salinity (Lee & Lee 1968, Helm & Millican 1977, Ventilla 1984). His et al. (1989) provide measurements of larval growth rate for salinities between 20%o and 35%c and temperatures between 15°C and 30"C. These measurements start with larvae with a mean shell length of .'^7 ixm, which is prior to the development of the D- shaped stage (Fig. 1 ). and extend through the first seven days of larval growth. The differences in the measured larval growth, for the fed larvae, between day 0 and day 7 were calculated for each temperature and salinity. The resultant values were linearly inter- polated to obtain growth rates at intermediate temperature and salinity values. Normalizing this matrix of growth rates to the growth rate at 25°C and 30%c provides the fractional change in larval growth rate at a given temperature and salinity (Fig. 6). The measured growth rates given in His et al. (1989) were extended to the entire range of temperature encountered by C. gigas larvae by assuming that larval growth rate decreases in a linear fashion to zero between 15°C and 0°C and to zero between 30°C and 35°C. This pattern in growth rate is based on observa- tions of increased larval abnormalities in these temperature ranges (Arakawa 1990). Similarly, for salinity, the growth rate at 207ic was linearly reduced to zero at lO^fr, after which growth rate remained zero. This was based on observations of increased larval abnormalities at low salinities (Arakawa 1990). For salinities above 35%?. larval sirowlh rate at 40%? was assumed to be one-half 0 5 10 15 20 25 Temperature (°C) Figure 6. Fractional change in Crassostrea gigas larval growth rate as a function of temperature and salinity. 248 BOCHENEK ET AL. the larval growth rate at 35%c. and growth rates at intermediate salinities were obtained by linear interpolation. The assumption of reduced growth rate at 40%c is based on measurements (Nell & Holiday 1988) that show C. gigas growth rate at 3%', being one- half the value at 35Vcc. The fractional change in larval growth over the entire range of temperature and salinity was verified by comparing simulated lar- val length with reported larval lengths measured at known tem- peratures and salinities (Helm & Millican 1977. Nell & Holiday 1988. His et al. 1989. Robinson 1992). These comparisons showed excellent agreement between simulated and observed larval length, except for lengths between 150 and 200 ^x.m. At these sizes, mis- matches of 10 to 45 |jim occurred for the higher teinperature and salinity values. Larvae of this size are growing rapidly at salinities above 24'^t and temperatures above 25"C, and hence the growth curves are steep. Therefore, slight mismatches in the reporting of measured length for a given larval age can greatly affect compari- sons with simulated lengths. Food Composition and Assimilation Efficiency The C. gigas larval model allows for differential metabolism of protein, carbohydrate, polar lipid, and neutral lipid. For this to occur, the food ingested by the larva must be expressed in terms of the relative contribution of each of these biochemical constituents. Based upon measurements reported in Utting (1986). Roman ( 1983). and Lee et al. ( 1971 ). the average biochemical composition of marine algae, in terms of ash-free dry weight, was taken to be 3 parts protein, 2.5 parts carbohydrate, 0.6 parts polar lipid, and 0.4 parts neutral lipid. This basic structure defines the food reservoir for the larvae, for most simulations. Handa (1969) provides assimilation efficiencies for plant ma- terial of 1 .0 for protein, 1 .0 for polar and neutral lipids, and 0.2 for carbohydrates. The reduced assimilation efficiency for carbohy- drates arises because 80% of plant carbohydrate is structural or P-linked carbohydrate (e.g.. the refractory portion) that cannot be digested by animals and is therefore not available as food. The available 20% represents labile carbohydrate. Multiplication of these assimilation efficiencies with the corresponding food frac- tion gives an overall assimilation efficiency for C. gigas larvae of about 0.7. which is within the range expected for bixahe lar\ae (estimated from growth efficiency by Jorgensen 1952). Initial simulations showed that the above assimilation efficien- cies resulted in growth rates for lar\ae less than 80 p.m. which were too small. Observations show a drawdown of neutral lipid reserves during early larval life in C. gigas and C. virgiiiica (Gal- lager etal. 1986, Gallager and Mann 1986a, Whyte et al. 1987. His & Maurer 1988). presumably to fill the carbon needs not covered by feeding. In the model, the early life stages of the larva were allowed to use neutral lipid stores to form structural material in the body. This was done by calculating a small larva factor iSLF,) of the form: SLF, = iiuu 0.. \ Ar SL-L SL - L, (10) where / indicates protein, carbohydrate, polar lipid, or neutral lipid. This relationship calculates the proportionate length change for larva smaller than 80 p-m in a given time increment (At), and the neutral lipid reserves are then used in proportion to the carbon requirement needed to sustain the change in length. The maximum neutral lipid that is used, given by X. occurs when larvae are at their initial size. L,,. This amount decreases proportionately as the larva grows and becomes increasingly capable of feeding, and is zero at 80 p.m. The mobilized neutral lipid is then converted into equivalent protein, carbohydrate, and polar lipid using the bio- chemical conversions given previously. This is the only instance in the model where protein is created de novo, rather than being obtained from food. Thus, the assimilated ingestion (AE,) can be expressed as the product of the filtration rate iFR). the ingestion efficiency (/£,). temperature and salinity effects (TSfaclor). food {Food,), the as- similation efficiency iAE,). and the small larvae factor (SLF,) as: At, = FR ■ lE^ ■ TSfactor ■ Food, ■ AE, ■ SLF, (11) Fate of Assimilated Ingestion The assimilated ingestion obtained from Eq. ( 1 1 ) is parameter- ized in terms of protein, neutral lipid, polar lipid, and carbohy- drate, and the fate of each of these biochemical constituents differs within the lar\a (Table 2). Protein assimilated in a given time interval has, as its primary destination, the somatic protein pool. Protein may also be used to cover a respiratory deficit (discussed below) in accordance with the appropriate protein-to-carbohy- drate-to-polar lipid ratio. The carbohydrate needs of the larva are determined by the amount needed to maintain tissue carbohydrate in its proper pro- portion and that needed to co\er the metabolic process of respira- TABLE 2. Destination of assimilated protein, carbohydrate, polar lipid, and neutral lipid in Crassostrea gigas larvae. Food Primary destination Food deficit Food surplus Tissue maintenance constituent in larva response response deficit Early life (<80 fim) Protein somatic P NA NA respiration (P:C:PLj NA Carbohydrate somatic C & respiration (P:C) somatic PL (P:PL) neutral lipid reserve respiration (P:C:PL) NA Polur lipid somatic PL(P:PL) somatic C (P:C) neutral lipid reserve respiration (P:C:PL) NA Neutral lipid neutral lipid reserves somatic C (P:C): NA somatic C (P:C): somatic C; somatic PL; somatic PL (P:PL| somatic PL (P:PL): somatic P; respiration respiration respiration The particular biochemical ratio determining the conversion to individual reservoirs is indicated. Table columns two. three, and four indicate the fate of the food; column five indicates the fate of the tissue. Transfers of food that do not occur in response to deficit or surplus conditions are indicated by NA. Protein, carbohydrate, and polar lipid are indicated by P. C. PL, respectively. Biochemical Model of C. gigas Larvae 249 tion (Table 2). Assimilated carbohydrate is the primary means by which larval respiratory needs are met (Table 2). The required somatic carbohydrate is determined so as to maintain the carbo- hydrate-to-protein ratio (0.01 ). and this amount is debited from the available assimilated carbohydrate and added to the carbohydrate pool. Excess carbohydrate (food surplus response in Table 2) be- comes part of the larval neutral lipid reserve. When tissue imbal- ances occur (e.j;.. insufficient polar lipid to meet the tissue com- positional requirements of the larvae), somatic carbohydrate is used to maintain larval polar lipid in its proper proportion. The primary destination of assimilated polar lipid in the larva is the somatic polar lipid pool in accordance with the protein-to-polar lipid ratio (Table 2). Excess assimilated polar lipid goes to the lar\al neutral lipid pool. When carbohydrate imbalances occur, polar lipid reserves are mobilized to produce somatic carbohydrate in an amount that is consistent with maintaining the protein-to- carbohydrate ratio. Polar lipids are also used to cover tissue main- tenance deficits arising from respiratory demands. The primarv destination of assimilated neutral lipid is the neu- tral lipid pool (Table 2). This internal pool is mobilized to maintain somatic carbohydrate and somatic polar lipid pools in accordance with the appropriate ratios when assimilated protein, carbohydrate, and polar lipid are not present in the proper proportions in the food. The neutral lipid pool can also be used to cover respiratory needs during periods of carbohydrate deficit. This pool also provides a means for small larvae, less than 80 (j,m, to produce somatic car- bohydrate, polar lipid, and protein as well as cover respiratory costs early in larval life. At any point in the development of the larva, the inability to maintain one of the biochemical constituent ratios or the inability to remove a deficit in a biochemical pool results in death of the larva. Respiration Respiration provides the only metabolic loss of assimilated energy in the C. gigas model. Measurements show that respiration lU "^ - ■- -- - Gerdes (1983b)- 25 C ^^ fioegn-ijuiaDerg & Manahan (1995) - 20 C ^ ~ 'l- ^ 10-3 ^ '■a Z c - CM O / 1 /' S 10-4 — / CO — j' / oc I y c - /' X o _ /' / n a y X 8 10-5 =- / a I / / 1 111! INK Ill 1 1 1 1 1 ml 1 II 10-4 Figure 7. Cra tion of larval 10-3 10-1 10-2 Weight (mg) ssostrea gigas lar>al respiration rate measured as a func- weight at two temperatures. rate for C. gigas larvae increases with larval size and with tem- perature (Fig. 7). Laboratory measurements of respiration rate for C. gigas larvae cover a range of larval sizes measured at 25°C (Gerdes 1983b) and 20°C (Hoegh-Guldberg & Manahan 1995) can be described by the relationship: Resp .W" (12) where Resp is given in mL O, consumed ind~' h"' and W is animal dry tissue weight in mg. The base respiration rate, /„, is assumed to reflect genetic variations in metabolic processes that are known to occur for individual C. ^i,'/,i;((i larvae (e.g.. Lannan 1980). Hence, this parameter is specified using a distribution (described in a following section) that is assumed to represent metabolic variabil- ity within the larval population. Other coefficients are defined in Table 1. The respiration rates measured at 25°C (Fig. 7) were used along with the fractional changes in growth rate (Fig. 6) to obtain the full range of temperature and salinity effects on larval respi- ration. Respiration rate was converted to an energy demand using 20.21 J (niL O, consumed)^' to determine the metabolic cost of respiration. Equation (12) provides the metabolic cost of respiration that must be met by the larva. As discussed in the previous section, the assimilated carbohydrate pool provides the first biochemical res- ervoir that is used to meet this demand (Table 2). This pool is converted to equivalent energy units using the conversions given previously, and the needed carbohydrate is removed from the pool. Any excess is added to the neutral lipid pool in an amount that is consistent with the carbohydrate-to-lipid ratio (Table 2). If the assimilated carbon pool is insufficient to meet the cost of respiration, then the remaining deficit is taken from the neutral lipid pool and any remaining deficit is then taken proportionately from the structural components of the larva (Table 2). Periods during which the larva resorts to using structural material to cover metabolic costs result in reduction of larval condition index de- fined in the model as a reduction in the protein-to-ash ratio. Larval Growth Larval growth in a given time interval is based on maintaining the protein-to-ash ratio (Fig. 4. Eq. 5) for a given larval length. Larval growth resulting in an increase in length is assumed to occur when the protein, carbohydrate, and polar lipid pools are in excess of what is needed to maintain the protein-to-ash ratio at a given size. This is the excess net production (ENP) that determines a in Eq. 1. Excess protein is obtained by subtracting from the protein pool the amount that is needed to maintain the ash weight at a given larval length, as is determined by the protein-to-ash ratio. The excess polar lipid and carbohydrate pools are computed from the excess protein pool based on the required structural ratios of these constituents. The excess net production for a given time interval is the sum of the excess protein, polar lipid, and carbohydrate. This gives the excess net production in a given time interval in terms of an increment in larval weight. The weight increment is then used with the length-to-dry tissue weight relationship (Fig. 2) to obtain an incremental increase in length for the increase in weight. During times of protein deficit with respect to ash weight (low condition index), the larva can have a positive net production that increases organic mass and condition index, but produces no ex- cess net production and. hence, no increase in length. 250 BOCHENEK ET AL. Larval Metamorphosis Observations suggest that once C. i>ii;(is larvae reach 275 ^l.m they may initiate metamorphosis and this process may or may not be successful (Ventilla 1984. Kusaki 1991. Laing 1995). Thus, in the model, the larva is assumed to have the potential of becoming competent for metamorphosis at 275 |JLm. Just prior to this point, at a length of 250 |xm. filtration rate declines. Observations show that C. gigas larvae can become competent for metamorphosis over a range of sizes. This implies that the switch between Eqs, 6 and 7 controlling the point where filtration rate changes might contain a size dependency, determined by some metabolic process that is. as yet. unknown. Not having this information, the point at which the larva can become competent for metamoiphosis was fixed at a size a little larger than the size observed by Gerdes (1983a) for the change in filtration rate. Simulations discussed subsequently support this decision. Once the larva reaches 275 \xm, it becomes competent to meta- morphose if it experiences a 25% drop in neutral lipid stores in one day. This is determined by the interrelationship of food supply, filtration rate, and respiration rate. Competency is triggered by a decrease in neutral lipid that, if continued, would impair successful metamorphosis. Once competent, the larva immediately attempts metamorphosis. Successful completion of metamorphosis occurs if the larval neutral lipid pool is greater than the polar lipid pool. This establishes a minimum storage requirement needed to sustain metamorphosis. If this condition is not satisfied, then metamor- phosis is unsuccessful and the larva dies. Biochemically Determined MetaboUc Mortality The simulated larval growth prior to metamorphosis is based on maintaining specific ratios between protein, polar lipid, carbohy- drate, and ash weight. Small variations in these ratios are allowed, consistent with changes that occur in the larva as it grows (cf. Fig. 4). However, large changes are not permissible. The interdepen- dencies of the biochemical ratios results in the protein-to-ash ratio being a good indicator of the biochemical state of the larva. If this ratio is reduced at any time to 70% or less of its needed value, then larval condition index is too low and the larva is assumed to die. This condition is termed starvation in the model. During the initial stages of larval growth, about the first two days, the larva does not filter efficiently (Fig. 5). and hence food ingestion is not usually sufficient to cover metabolic costs. During this period, it is assumed that the larva survives by using its stored neutral lipid supply. However, if during this period the neutral lipid supply approaches zero, the larva is assumed to have reached its metabolic point of no return and dies. Also, inability of the larva to maintain its required protein-to-lipid and protein-to-carbohy- drate ratios results in death. Model Implementation Initial C. gigas Egg Size, Including Genetic Variability The eggs spawned by C. gigas adults have an average size ot 50 jjLm (Quayle 1988. Arakawa 19<-)()). However, using this as the initial condition for the model resulted in mismatches in the initial simulated and observed length-to-weight relationships, which are based on larval size. Thus, simple egg diameter is not the appro- priate metric for use with the length-to-weight and other conver- sions. The discrepancy arises because of the mismatch in volume of a spherical egg and the more ellipsoidal-shaped larva. There- fore, it was necessary to convert initial egg diameter to an equiva- lent larval size. This was done using a diameter-to-length conver- sion factor of 1.096 (Arakawa 1990) that conserves volume in going from a spherical egg to an ellipsoidal-shaped larva. Thus, a 50 p-m egg is equivalent to a 54.8 p.m larva. C. viri;inica egg size is observed to range between 30 and 80 Ijim (Gallager et al. 1986). More limited information is available For C. gigas. but a similar range of egg sizes can be interred (Breese & Malouf 1975). This variation was assumed to represent genetically or environmentally determined variability in the spawning population. Therefore, for each simulation, the initial conditions included a range of egg sizes. To establish the initial biochemical composition of the egg. the larval size immediately post-hatch was used with the length-to-dry tissue weight relationship (Fig. 2) to calculate an initial dry weight, which in turn was used to obtain an initial ash weight value (Fig. 3). The protein component of the egg was then determined by multiplying the ash weight by the protein-to-ash ratio. The egg polar lipid content was determined by multiplying the protein con- tent by the polar lipid-to-protein ratio. The carbohydrate content was taken to be 1% of the dry weight of the modified egg. Egg neutral lipid content was obtained by difference through subtract- ing the protein, polar lipid, and carbohydrate weight from the initial dry weight value. If this calculation resulted in a negative value of neutral lipid, which can occur for very small eggs, the egg was assumed to he non\ iable. Predation and Other Nonmetabolic .Sources of Larval Mortality The larval model provides, as output variables, the total time for larval development, larval size at the end of the simulation, and a description of why the simulation ended. Termination of a simu- lation occurs because of successful metamorphosis, unsuccessful metamorphosis, inappropriate metabolic ratios, and starvation. The simulated larval results are then examined with a submodel that calculates larval survivorship based on the timing of mortality and the larval life span of the survivors for each combination of egg size and respiration rate represented by the genetic variability as- sianed to the cohort. Los.ses to nonmetabolic sources of mortality, such as predation. are evaluated at this point with losses increasing in proportion to the larval life span obtained for each combination of egg size and respiration rate. The resultant simulated distribu- tions of survivorship can then be compared with similar values reported from field and laboratory studies. Predation and other forms of nonmetabolic mortality (EM) are imposed during the larval period using a relationship assumed to be of the form EM(j.k) i.MDU.k) (13) where the daily mortality rate, iii,, is the same as that used for C. virginica larvae (Dekshenieks et al. 1997) and LD is the total time required for a larva with an initial egg size (7) and respiration rate (k) to trigger a mortality event or to successfully metamorphose. The longer the larva takes to develop, the higher the chance of nonmetabolic mortality. (;enetic P'.ffects on Larval Mortality Growth, mortality, and other population processes are appor- tioned based on genetic variability, in which certain combinations of initial egg size and respiration rate are less common in the cohort and in which certain combinations are less viable overall. Biochemical Model of C. gigas Larvae 251 either due to uietahiilic imbalances, metabolic inetYicieneies, or longer lar\;il life spans increasing nonmetabolic mortality. This type of genetically determined outcome. GE. is prescribed with a Gaussian function of the form: GE-. (14) where the Gaussian distributions extend for two standard devia- tiiins (2.sy/,.^,^,. 2,v(y,_,,,,) about a central egg size and respiration rate that are given by £5„ and Resp,,. respectively. Coefficient defini- tions and values are given in Table 1. Equation 14 weights mor- tality or any other population process by a population distribution that is characterized by a certain range of egg sizes and respiration rates. Thus, the surviving larval population represents the com- bined effects of genetics, food composition, and environmental conditions. RESULTS Reference Simiilution The reference simulation was run with near-optimal en\ iron- mental conditions of 25"C. 30'^ff, a food concentration of 2 mg L"'. and food with a protein, polar lipid, neutral lipid, carbohy- drate ratio of 3:0.6:0.4:2.5. Development of C. gigas larvae over the first few days of larval life is primarily sustained by egg neutral lipid stores. The drawdown of neural lipid stores results in a de- crease in the neutral lipid-to-protein ratio (Fig. 8A). The decrease in this ratio is most pronounced for eggs with small initial sizes. For initial egg sizes of 40 to 50 |ji.m. the neutral lipid-to-protein ratio approaches zero. All larvae, independent of initial egg size, reach a maximum ratio value of between 0.15 and 0.165 and a size of >275 |a.ni (Fig. 8B). Larvae arising from larger eggs reach these values earlier, and hence can metamorphose earlier. Once a size of 250 (xni is reached, filtration rate declines and so. too. does the neutral lipid to protein ratio. However, growth continues and most larvae metamorphose at about 300 fjLin. regardless of initial egg size (Fig. 8B). The simulated larval growth is exponential and independent of initial egg size (Fig. 8B). Larvae reach 100 p,m. which corresponds to the small umbo stage (cf. Fig. 1 ). in 4 to 12 days for Initial egg sizes of 70 (j,m and 40 p.m. respectively. The rate of growth ac- celerates after 100 p,ni. The time required for the larvae to reach 300 |j.m ranges from 13 to 19 days for the largest and smallest initial egg sizes, respectively. The development time required for the 50 (j.m egg to reach 300 jxm is 17 days, which agrees with development rates measured at 25°C and iVAt (His et al. 1989). The range of simulated development times is al.so consistent with those reported for C. gigas larvae (Quayle 1988. Arakawa 1990. Laing 1995). The effect of variations in initial egg size and growth efficiency on the fate of the larvae is summarized by the state of the larva at the time it either dies or successfully completes metamorphosis (Fig. 9). Variations in growth efficiency are modeled as variations in respiration rate; however, similar results would be obtained if the variation was in any component of Eq. 2. Initial egg sizes less than 37 p-ni result in nonviable larvae for all respiration rates. Initial egg sizes above 73 (jim result in eggs that do not have sufficient neutral lipid stores after day 2 to continue development at all respiration rates. The large initial size of these eggs results in an iinbalance in neutral lipids that cannot be corrected subse- quently (cf. Fig. 8A). A similar fate occurs for initial egg sizes 0.20 = 0.15 ra q: c a) ^* o a. ■o '5. 0.10- 0.05 0.00 I I I I I 1 I I I I I I I 1 I I I I I I I I I I I I I I I I I A - 70 pm - 60 pm •■■■ 50 pm 40 pm HUU T' T T T T 1 T 1 1 1 [ [ T 1 T 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 - B: \ 300 -_ ? - 3 - '- / /' / 1 ' £ 200 - OI ' / ■ ■••'' ' " c " 0) / ' / / " _l - / '' .•■■'' / : / '' ■•■■' / ■ / /' .•■''' ^ I 100 '^y^.-' / ^ : - ^ - n "1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 i- 0.0 10.0 30.0 40.0 20.0 Time (Days) Figure 8. Simulated time development of the (A) neutral lipid-to- protein ratio and (B) Crassoslrea gigas lenjith obtained for a selection of initial egg sizes from 40 ytm to 70 pni. between 40 and 50 p.ni that yield larvae with high base respiration rates. For these larvae, too much of the neutral lipid pool is re- quired to cover respiratory demand, and this produces a metabolic imbalance from which the larva cannot recover. Above and below the region of initial egg sizes and base respiration rates that pro- duce successful metamorphosis are regions where the larva devel- ops to the point of attempting metamorphosis, but is unable to do so successfully. A drop in neutral lipid triggers metamorphosis in these larvae, but they have insufficient lipid stores to cover the metabolic costs of metamorphosis. At the population level, the fraction of the larvae that survive to complete metamorphosis is dependent on the initial egg size (Fig. lOA). with the distribution of survivorship centered around an initial egg weight of 50 p-m. which is the center of the egg size distribution. Survivorship tapers off toward larger egg sizes, with essentially no survival to metamorphosis at egg sizes greater than 68 (xm. Survivorship at smaller egg sizes decreases abruptly with essentially no survival at sizes less than 40 p.m. Larval survivor- ship as a function of base respiration rate (Fig. lOBj is maximal at 252 BOCHENEK ET AL. Reference Case 400 350 - 300 250 - 200 150 100 50 60 Initial Egg Size(iim I I Initial egg size too small I I Insufficient neutral lipid at day 2 [^^. I Unsuccessful metamorphosis ^^1 Successful metamorphosis Figure 9. Sinuilated time development of Ciassostrea gigas larvae for a range of initial egg sizes and base respiration rates. The simulation used the reference case environmental conditions of 25 C, Wr<. and a food concentration of 2 mg L' with a protein, polar lipid, neutral lipid, and carbohydrate ratio of 3:0.6:0.4:2.5, respectively. rates around 1.05 KJ day"'. Respiration rates above this result in an abrupt decrease in survival, with no survival at rates above 1.254 KJ day"'. Larval survival at respiration rates below 1.05 KJ day"' slowly decreases and is zero at rates below 0.628 KJ day" . Most larvae reach sizes between .^00 and 325 ixin before meta- morphosis (Fig. IOC). Detailed verification of the reference simulation is difficult because data on larval biochemical composition, as it varies with eoa size, development time, and environmental conditions, are meager. General trends in larval success as measured by survivor- ship. lar%'ai size, and larval life span are much better known. Four such trends are observed in the simulation. ( 1 ) Adequate neutral lipid stores are a prerequisite of high survival during the critical period a few days post-hatch and at metamorphosis. The reference simulation demonstrates both effects (Figs. 8A. lOB). (2) Typi- cally, egg size ranges from 40-60 p.m. The reference simulation identifies this range as optimal for C. gigas. based on changes in biochemical composition dictated by larval energetics (Fig. lOA). (3) Successful metamorphosis occurs for larvae of 295-340 (iiii in the reference simulation. This is a frequently observed size range (Ventilla 1984. Kusaki 1991. Laing 1995). (4) Larval life span for the most successful egg sizes varies from about 14 to 17 days, a range that approximates the norm in observation (Ventilla l'-)S4. Arakawa 1990. Laing 1995). Effect of Uirval Filtration The parameterization for larval filtration (Eqs. 6 and 7. Fig. 51 is based on the assumption that the filtration response changes abruptly at 250 p.m. C. gigas larvae are observed to metamorphose at a range of lengths. The change in filtration rate is an important 0.9 1.1 1.3 Respiration Rate (KJ d"'') 0.50 re 0.25 0.00 250 275 300 350 375 400 325 Length (um) Figure 10. Simulated Crassostrea gigas larval survival as a function of (Al initial egg size, (B) base respiration rate, and (C) larval length at metamorphosis, using the reference case environmental conditions given in Figure 9. contributor to the mechanisms by which the model determines the onset and success of metamorphosis. Therefore, the larval length at which the change in filtration response occurs was varied. Varying the onset of a decline in filtration rate does not impact the range of viable egg sizes because this range is determined by events that predate this point in the larva's life span (Figs. I lA. 1 IB). However, if filtration rate changes at 270 |xm. larvae spend a longer time at a high filtration rate. These animals metainorphose successfully over a wider range of base respiration rates than those that support successful metamorphosis when filtration rate begins to decline at 250 |JLm (Fig. 128). Also, the overall population survivorship is higher; 78.8% of the cohort versus 62.4% in the reference simulation (Figs. 9. 10). and the range of lengths at which larvae metamorphose is wider (Fig. 12C). Moving the larval length at which the filtration rate response changes to a smaller size. 230 |xm. results in a limited range of base respiration rates at which successful metamorphosis can oc- cur (Figs. IIB. 13B). Many of the animals survive to attempt metamorphosis, but are unsuccessful at completing the process. The smaller size at which the filtration rate is reduced results in the larvae not storing enough neutral lipid to cover the metabolic needs associated with metamorphosis. Larval length at metamor- phosis (Fig. 13C) is now limited to a small range of sizes. Total population survivorship drops from 62.4% in the reter- Biochemical Model of C. gigas Larvae 253 Filtration Break 270 iim A ' 1 1 1 1 / ■ 1 . ■ 1 !l 1.5 - / - 1.3 - / - 1.1 - / - 0.9 - - 0.7 - _ 0.5 1,1 II 1 — 1 — Filtration Break 230 urn 1.5 1.3 a. 0.9 0.7 0.5 30 40 50 60 Initial Egg Size (Llm) 70 80 I I Initial egg size too small I I Insufficient neutral lipid at day 2 I I Unsuccessful metamorptiosis ^^1 Successful metamorpfiosis Figure 11. Simulated fate of Crassostiea gigas larvae for a range of Initial egg sizes and base respiration rates in which the change in larval nitration rate response is at (A) 27(1 pui and (Bl 230 pm. The simula- tion used the reference case environmental conditions given in Figure 9. ence case to 8.1% with a size trigger of 21(1 \xm rather than 2.'il) (xm. Cohort sur\ ivorship declines dramatically at trigger sizes be- low 250 fxm (Table 3) and asymptotes rapidly above 250 p.m. At trigger sizes above 250 fj.m. some larvae do not metamorphose until reaching sizes much larger than normally observed (e.g.. 375^00 (Xin. Fig. I2C). however. Accordingly, larval growth, as modeled, requires a filtration rate trigger near 250 (xm to obtain observed levels of cohort survivorship and size at metamorphosis. This approximates the size where filtration rate declines are ob- served to take place (Gerdes 1983a). Effect of Diet Information on the effects of diet on growth and survival of C. gigas larvae indicate that high-lipid and low-protein diets are usu- 0.50 ' T-| 1 1 1 1 1 1 I 1 1 1 1 1 1 1 A - c no external mortality o external mortality u. H 0.25 - ~ > . > D cn ■ ^ — --^ ~~~^^~^-^c^ . 1 1 i(^''i ' ' J — ■— i—i — L_ 'T-^^^-'-- ■ ■ 1 30 40 50 60 Initial Egg Size (^m) 70 80 0.50 15 0.25 0.00 0.50 re 0.25 Respiration Rate (KJ d' ' ) 0.00 -1 — I — i — |— rn — TT — I — I — i — r-n r I I J 1 J I L — 1 — 1 k 250 275 300 325 350 Length (tim) 375 400 Figure 12. .Simulated Crassoslrea gigii\ lar>al survival as a function of (.\) initial egg size, (B) base respiration rate, and (Cl larval length at metamorphosis obtained when the change in larval Filtration rate re- sponse occurs at 270 pm. The simulation used the reference case en- vironmental conditions given in Figure 9. ally efficacious (Utfing 1986, Thompson & Harrison 1992. Thompson et al. 1994. 1996). These trends are reproduced by the model. A larval diet that is lacking in neutral lipid but contains the conect ratios of protein, polar lipid, and carbohydiate (3;().6:0:2.5) results in animals that are unable to successfully complete meta- morphosis at all ranges of initial egg size and base respiration rate (Fig. 14). Lack of neutral lipids results in more of the protein and carbohydrate pools being used to cover the demands of respiration and growth and, so, lipid stores are insufficient to sustain meta- morphosis. In addition, the range of egg sizes and base respiration rates in which the neutral lipid store at day 2 is insufficient for further development is greatly expanded relative to what is ob- tained for a diet containing neutral lipid (Fig. 14 versus Fig. 9). A diet in which the protein content is SO'/r higher relative to the standard diet (4:0.6:0.4:2.5) produces a similar result in that no combination of initial egg size or respiration rate results in suc- cessful metamorphosis (Fig. 15A). Larvae either have insufficient neutral lipid reserves at day 2 to continue development or attempt metamorphosis but fail to complete the process. A high-protein diet requires that more of the neutral lipid stores be used to cover tissue structural needs, and hence less lipid is stored for later use in metamorphosis. 254 BOCHENEK ET AL. 0.50 ™ 0.25 0.00 - mortality, metabolic sources only - mortality, including non-metabolic sources 400 No Neutral Lipid Diet 0.50- 75 0.25 0.00 0.50 a 0.25 0.00 0.9 1.1 Length (|im) Figure 13. Simulated Crassoslrea gigas larval survival as a functitin of (Al initial egg size, (Bl base respiration rate, and (C'l larval length at metamorphosis obtained when the change in larval nitration rate re- sponse occurs at 230 urn. The simulation used the reference case en- vironmental conditions given in Figure 9. A diet low in protein (2:0.6:0.4:2.5) is beneficial to the larva (Fig. I5B). Ingestion of low-protein food extends the range of respiration rates that resuU in successful metamorphosis (Fig. 15B versus Fig. 9). More polar lipid and carbohydrate is available to cover metabolic costs and to increase neutral lipid stores. This increases metamorphosis success. Simulations indicate that a low- protein diet increases overall population survivorship under hatch- ery conditions where nonmetabolic causes of mortality are mini- mized (88.8% from 62.4% in the reference simulation), but de- creases survivorship under field conditions, from 6.19^^ in the reference simulation to 4.7%. due to slower growth and longer planktonic times increasing losses to predation. TABLE 3. Total population survivorship for simulated cohorts of Crassoslrea gigas larvae in which the larval size triggering a change in filtration rate was varied from 210 pm to 290 (ini. Trigger size Cohort survivorship 210 8.1% 230 17.2% 250 62.4% .270 78.8% 290 78.8% 50 60 Initial Egg Size (um) [ I Initial egg size loo small I I Insufficient neutral lipid al day 2 I I Unsuccessful melamorphosis P^^^ Protein:asfi ratio outside acceplable range Figure 14. Simulated fate of Crassoslrea gigas larvae for a range of initial egg sizes and respiration rates resulting when the larval diet includes no neutral lipid. The simulation used the reference case en- vironmental conditions given in Figure 9, except that the food had a protein, polar lipid, neutral lipid, and carbohydrate ratio of 3:0.6:0: Effect on Metamorphosis The cue that is assumed to initiate metamorphosis is a 25% drop in the neutral lipid reserves of the larva over a span of one day. This level of decline was chosen as a metamorphosis trigger by comparing the simulated size and biochemical composition of metamorphosing larvae to measured values. The condition and size of larvae at metamorphosis is quite variable, and the metamorpho- sis trigger value used in the model represents one that simulates the average of these conditions. Requiring only a 10% drop in neutral lipids in one day to trigger metamorphosis results in essentially all combinations of initial egg size and base respiration rate producing larvae that successfully undergo metamorphosis (Fig. 16 versus Fig. 9). The lower trigger value allows larvae to have a greater energy store at the time metamorphosis is attempted, and hence the probability of success is increased. At the population level, a wide range of egg sizes and base respiration rates results in survival (Figs. 17A & I7B1. but larval length at metamorphosis is shifted to smaller larvae (Fig. 17C). The population mode in this case extends from 285 10.111 to about 305 p.m. a size range predominately comprising sizes smaller than typically observed and smaller than the 295 p.m to about 335 |xni size range observed in the reference simulation (Fig. 9C). Increasing the trigger to a 40% reduction in neutral lipids in one day (not shown) results in attempted, but failed, metamorphosis at all initial egg sizes and base respiration rates. Thus, the value of 25% produces observed success rates for meta- morphosis at a range of observed egg sizes not achieved by higher or lower trigger values. MetaiTiorphosis success is determined by the ratio of neutral BiocHtMicAL Model of C. gigas Larvae 255 High Protein Diet A 1 1 1 1 1 y 1 1 1.5 - - 1.3 1.1 0.9 - / - - / - 0.7 - / - 0.5 ' III 1 1 1 1 1 Metamorphosis Trigger 10% Low Protein Diet B - ■ 1.5 - ^^^^^^H 1 - 1.3 : ^^^H 1 - 0.9 : ^^^H 1 - ^^^^^^^^^^^Bf - . . ~ 0.7 ^^^^^H^^^^^^^lr .J - 0.5 III 1 1 II 1 30 40 50 60 Initial Egg Size (urn) 80 I I Initial egg size too small I I Insufticjent neutral lipid at day 2 I j Unsuccessful metamorphosis ^^1 Successful metamorphosis Fiiiurt' 15. Simulated fate of Crassoslrca gigas lar\ae for a range of initial v^^ .sizes and respiration rates resulting when the larval diet is (.\) high in protein and (B) lo« in protein. The simulation used the reference case environmental conditions given in Kigure t, except that the food had a protein, polar lipid, neutral lipid, and carbohydrate ratios of 4:0.6:(l.4:2.5 and 2.5:0.6:(l.4:2.5. lipid to polar lipid in the animal at the time that metanwrphosis is triggered. In the reference case (Fig. 9). successful metamorphosis occurs if the content of neutral lipid exceeds the content of polar lipid. Allowing metamorphosis to be successful when the larval neutral lipid reserves are greater than 80'vf of the polar lipid con- tent produces successful metamorphosis over a large range of ini- tial egg sizes and respiration rates (Fig. 18A|. At the population level, larval survival is enhanced over a wide range of initial egg sizes and base respiration rates (Fig. 19A. 19B). The length at metamorphosis also extends over a wide range (Fig. 19C). wider than normally observed, but the mode of the population is still around 308 jxm. as seen in the reference simulation (cf. Fig. IOC). If metamorphosis is successful only when the neutral lipid pool is greater than 1 lO^c of the polar lipid pool, the range of initial egg 400 350 300 250 200 150 100 50 60 Initial Egg Size (urn) I I Initial egg size too small I I Insufficient neutral lipid at day 2 [ I Unsuccessful metamorphosis ^^1 Successful metamorphosis Figure 16. Simulated fate of Crassostrea gigas larvae for a range of initial egg sizes and base respiration rates when the trigger for meta- morphosis is set to be a 10% drop in neutral lipid stores in one day. The simulation used the reference case environmental conditions given in Figure 9. sizes and respiration rates at which successful metamorphosis oc- curs is greatly reduced (Fig. 18B). The effect of egg size on larval survivorship is only somewhat modified (Fig. 20A); however, the range of base respiration rates that result in successful metamor- phosis is greatly reduced (Fig. 20B). as is the range of larval lengths (Fig. 20C). Overall, trigger values above 100%. a 1:1 neutral lipid-to-polar lipid ratio, generate rates of successful meta- morphosis that are too low and larval sizes at metamorphosis that are too large (Table 4). Trigger values below 1009f at first seem to be defensible; however, survivorships are unusually high. Hence, model conditions chosen for the reference simulation include, as the minimally required neutral lipid stores for success at metamor- phosis, a quantity greater than the polar lipid content of the larva (>1:1). Recall that the polar lipid content is constrained to a pre- determined ratio with protein and structural carbohydrate, and so the same results would have occurred had neutral lipid been com- pared to any structural constituent. Temperature and Salinity Effects The temperature and salinity used for the pre\ ious simulations. 25"C and SO'Xt. is near optimal (or the growth and development of C. gigas larvae (cf. Fig. 6). The optimal range of these environ- mental variables is narrow, and therefore relatively small changes in these conditions have the potential of causing large changes in larval growth and development. Population survivorship is non- zero over a relatively narrow temperature range and a somewhat wider salinity range (Table 5). A 5^C reduction in temperature, from 25°C to 20°C for example, results in no combination of initial egg size or base respiration rate that produces successful meta- morphosis (Fig. 21A). A reduction in salinity to 207cc extends the 256 BOCHENEK ET AL. 0.50 ^ 0.25 3 cn 0.00 — 1 — I — I — I— -T 1 1 1 p - mortality, metabolic sources only -mortality, including non-metabolic sources 30 40 50 60 Initial Egg Size (^m) 70 80 0.00 250 275 300 325 Length (pm) 350 375 400 Figure 17. Simulated Crassostrea gigas larval survival as a function of (A) initial egg size, (Bl respiration rate, and (C) larval length at meta- morphosis obtained when metamorphosis occurs in response to a 10% drop in neutral lipid stores in one dav. The simulation used the ref- erence case environmental conditions given in Figure 9. range of initial egg weights and base respiration rates at which the larva exceeds its neutral lipid constraint at day 2 of development (Fig. 21B). Half as many animals exposed to this salinity success- fully complete metamorphosis as in the reference case. Effect of I ariations in Food Resources Environmental food concentration can vary over a wide range. Food concentration in excess of 2 mg L"' used for the reference simulation (cf. Fig. 9) only extends somewhat the range of initial egg weights and base respiration rates that result in successful metamorphosis because 2 mg L"' food is a saturating food con- centration. How ever, a 50% reduction in food concentration, from 2 mg L~' to 1 mg L"'. significantly narrows the range of base respiration rates that result in successful metamorphosis (Fig. 22A|. The strong coupling between respiration rate and food avail- ability is not surprising since respiration is the primary metabolic loss that the larva must cover through ingestion. This result is further substantiated when looking at population level survival trends (Table 6). Survivorship declines rapidly at food concentra- tions below 2 ing L"' and reaches zero at food concentrations of about 0.5 mg L"'. Certain environmental conditions inay spare a decrease in food 400 350 300 Neutral Lipid > 80% Polar Lipid 200 150 400 Neutral Lipid > 110% Polar Lipid B 350 - / 300 - / 250 - -^ /^_y^_^ 200 - ^J 150 - III 11 II 30 40 50 60 Initial Egg Size (urn) 70 I Inrtial egg size too small I ] Insufficient neutral lipid at day 2 j I Unsuccessful metamorphosis m Successful metamorphosis Figure 18. Simulated fate of Crassostrea gigas larvae for a range of initial egg sizes and respiration rates v\hen (he metamorphosis trigger occurs when {A) neutral lipids are greater than 80% of the polar lipid stores and IB) neutral lipids are greater than 1 10% of the polar lipid stores. The simulation used the reference case environmental condi- tions given in Figure 9. supply by permitting an increase in filtration rate. Increasing tem- perature in the previous simulation to 30°C. for example, increases survival by permitting some larvae with high respiratory demands to survive (Fig. 22B). Total population survival increases from 27.2% to 37.6% at the higher temperature. Planktonic values of food of 0.5 mg L"' or less are not unusual. Consequently, larvae may experience times of starvation due to significantly reduced food concentrations. A simulation in which food is available at a concentration of 2 mg L~' for days 1 to 4 of larval development and unavailable after day 5 shows that no combination of initial egg size and base respiration rate results in successful metamorphosis (Fig. 23A). Death results from poor condition in some cases, but more frequently from metabolic im- balances between principal biochemical constituents. Simulated Biochemical Model of C. gigas Larvae 257 0.50 ■S 0.25 — I — [ — I 1 1 — I — i — r 1 1 1 i ^ I I 1 i r- - mortality, metabolic sources only -mortality, including non-metabolic sources 0.5 0.7 0.9 1.1 Respiration Rate (KJ d 1.3 1i 0.50 B 0.25 - 0.00 250 400 325 Length (urn) Figure 19. Simulated Crassoslrea gigas larval survival as a function of (.A) initial egg size, (B) base respiration rate, and (C) larval length at metamorphosis obtained when metamorphosis occurs when neutral lipids are greater than 80 9f of polar lipid. The simulation used the reference case environmental conditions given in Figure 9. 0.50 ■S 0.25 0.00 ■ mortality, metabolic sources only - mortality, including non-metabolic sources 30 50 60 Initial Egg Size (Mm) 80 0.50 75 0.25 0.00 0.50 ra 0.25 Respiration Rate (KJ d" ) -1 — I — I — I — I — I — i — I — I I I I I I r 0.00 '-^ 250 275 300 375 400 325 Length (um) Figure 20. Simulated Crassostrea gigas larval survival as a function of (Al initial egg-size. (B) base respiration rale, and (Cl larval length at metamorphosis obtained when metamorphosis occurs when neutral lipids are greater than UO'^r of polar lipid. The simulation used the reference case environmental conditions given in Figure 9. survival times were similar to Uiose observed tor starved larvae (His & Seaman 1992. Lainy 1993). Survival times of simulated larvae starved after day 4 were two to eight days, with most dying within two days. Larvae starved after day 9 survived two to ten days, with most dying in the first four days. Thirty-three percent of larvae starved after day 14 completed metamorphosis. These were larvae that came from relatively large eggs with low base respira- tion rates (Fig. 2.^B). Such larvae were much closer to metamor- phosis on day 14 when starvation began than other larvae and were able to complete metamorphosis using their energy stores. This last simulation is particularly interesting because a nar- rower range in egg size produced successful larvae. In other simu- lations, varying base respiration rate was much more significant in determining survival than varying egg size. The simulation sup- ports the conclusion of Gallager et al. (1986) and Gallager and Mann (1986b) that egg quality is important in minimizing losses due to low food supply during larval life. Genetics of Egg Size and Respiration Kate As discussed previously, it was necessary to model larval co- horts characterized by a range of egg sizes and growth efficiencies to obtain the observed ranges in larval life span, size al metamor- phosis, and survivorship seen in experimental studies of C. gigas larvae. One important choice, then, was the mean of the frequency distribution chosen for these two variable characters. The average egg size was taken to be 50 p-ni in the reference simulation. With a few exceptions, varying the conditions of the simulation did not vary the range of viable egg sizes to a great degree (Figs. 9-22). The range of viable egg sizes is dictated by the most basic constraints imposed by biochemical composition at birth and by environment (e.g.. food supply) rather than by genet- ics. Not surprisingly, changing the initial egg distribution so that it is centered on a 60 p.m egg, rather than a 50 p-m egg, results in survivorship that is skewed toward larger egg sizes (Fig. 24A versus 9A) because most eggs of 40 to 70 p,m in size are viable TABLE 4. Total population survivorship for simulated cohorts of Crassostrea gigas larvae in which successful metamorphosis occurs when neutral lipid reserves exceed by a given fraction the polar lipid content. Neutral lipid-to-polar lipid ratio Cohort survivorship 0.8:1 80.5'7r 1.0:1 62.4% 1.1:1 26.5'7r 1.2:1 0.0% 258 BOCHENEK ET AL. TABLE 5. Total population sur\ ivorship for simulated cohort.s of Crassostrea gigas Ian at' exposed continuously to different temperatures ( C) and salinities (%c). Temperature 15 20 25 30 35 Cohort survivorshi|i O.O'r 0.0"f 62.4% 63.3% 0.0% Salinity 15 20 25 30 35 Cohort survivorship 0.0% 33.1% 63.3% 62.4% 31.8% (Fig. 9), and moving the average size higher simply increases the total number of eggs spawned in this higher size range. Respiratory effects on survivorship (Fig. 24B) are not aUered by a change to mean egg size. However, the larval length at the time of metamor- 20 C, 30 0/00 A 1 • 1 II II 1.5 - J - T3 1.3 _ / . -5 / Rate - / - Resprration o - / - 07 - y - 0.5 III 1 1 1 1 1 25 C, 20 'oo B III 1 1 1 1 1.5 - / - T3 1.3 _ r-' . -5 2C / Rate - / - c .0 re - /A mm - c/l ■^ 0.7 - ^■P-- :. : 0.5 - 1 . 1 1 r , r 1 30 40 50 60 Initial Egg Size (^m) I Initial egg size too small I I Insufficient neutral lipid at day 2 I . . I Unsuccessful metamorphosis 1^1 Successful metamorphosis Figure 21. Simulated fate of Crassostrea gigas larvae for a range of initial egg sizes and base respiration rates at (.\1 2(1 C, MVit and (Bl 25C 2(1';.. Food concentration was 2 mg ['' with a protein, polar lipid, neutral lipid, and carbohydrate ratio of 3:U.6:().4:2.5. Food 1mg 1' 1,25 C A 1.5 1 1 1 TJ 1.3 / y Rate - ^ - c o - J - 1 0.9 /~v \r ~^^ - iB - 0.7 - - - - 0.5 1 i 1 i 1 1 1 Food 1mg |- 1,30 C B 1 1 1 1 II 1 1 1 1 350 - / - 300 - /^ - 250 - /V_^^ \ A /■ - 200 ^-^ v \J - !. - 150 1 1 1 1 1 1 i 1 1 30 40 50 60 70 80 Initial Egg Size (um) Initial egg size too small I I Insufficient neutral lipid at day 2 I I Unsuccessful metamorphosis ^^1 Successful metamorphosis Figure 22. Sinudated fate of Crassostrea gigas lar>ae for a range of initial egg sizes and base respiration rates and for a food concentration of 1 mg L"' at (A) 25°C and (Bl 30C. The simulation used environ- mental conditions of iO"i( and a food composition with a protein, polar lipid, neutral lipid, and carbohydrate ratio of 3:11.6:11.4:2.5. phosis is somewhat larger (Fig. 24C). Total survivorship declines somewhat from 64% to 60% because more of the cohort that is spawned falls into egg sizes above 70 |xm. Overall, however, the simulation shows that a moderate change in mean egg size does not malerially change the outcome of the simulation. As the simu- TABLE 6. Total population survivorship for simulated cohorts of Crassostrea gigas larvae exposed ccmtinuously to different concentrations of food img F '). Food Cohort survivorship 0.5 0.0% 1.0 27.2% 2.0 62.4%. 4.0 64.4%. Biochemical Model of C. gigas Larvae Food Variation A 0.50 259 Initial Egg Size (n gm) I I Initial egg size too small | | | | Polar lipidprotein ratio not satisfied I I Insufficient neutral lipid at day 2 j I Unsuccessful metamorphosis ^^H Successful metamorphosis Carbohydrate protein ratio not satisfied I Protein ash ratio outside acceptable range Figure 23. Siniululed fate of Crassostrea gigas larMie fur a range of initial egg sizes and base respiration rates and I A ) a food concentration of 2 nig~' available for days 1 to 4 and zero afterward and (B) a food concentration of 2 mg~' available for days 1 to 14 and zero afterward. The simulation used environmental conditions of 25C, MVir and a food composition with a protein, polar lipid, neutral lipid, and carbo- hydrate ratio of .1:().6:(l.4:2.5. lation depicted in Figure 23B sliows. this outcome may not be repeated in cases of nonsaturating food supply. Average base respiration rate for larvae was set at 1 .03 KJ day"'. Increasing the mean base respiration rate by 20% to 1.25 KJ day"' results in larval survival skewed toward lower respiration rates (Fig. 25B) and smaller egg sizes (Fig. 25A), with a rapid decrease in survival above respiration rates of 1.05 KJ day"' and egg sizes above 60 |jLm. Larval survival peaks at a metamorphosis length of 30ii iJim. which is similar to that obtained in the reference simulation (Fig. 9C). However, larval survivorship decreases rap- idly for larger larvae (Fig. 25C). Total survivorship declines dra- matically from 64% to 339f . Higher base respiratiim rates penalize larger larvae because insufficient excess neutral lipid can be ob- tained under the food supply provided to cover tissue maintenance and provide the requisite energy stores for metamorphosis. Thus, the model is considerably more sensitive to moderate changes in base respiration rate than in mean egg size. Conversely, the influ- ■5 0-25 > '> CO 0.00 30 0.50 - mortality, fitetabolic sources only -mortality, including non-metabolic sources 40 50 60 Initial Egg Size (i^im) 70 80 n 0.25 0.00 "T — I — I — I — r 0.5 0.7 0.9 1.1 1.3 1.5 Respiration Rate (KJ d'^) 0.00 250 275 300 325 Length (^m) 350 375 400 Figure 24. Simulated Crassostrea gigas larval survival as a function of (.\1 initial egg size, (B) base respiration rate, and (C'l larval length at metamorphosis for an initial egg size distribution that is centered at 60 pm. The simulation used the reference case environmental conditions given in Figure 9. ence of environment on the range of viable base respiration rates is much greater than on the range of viable egg sizes and. so. a change in mean respiration race might easily produce an alternate result under different environmental conditions. DISCUSSION Perspective The model described here is unique in that it seeks to recreate many of the growth and mortality phenomena observed in C. gigas larvae from basic biochemical principals. The approach was dic- tated by a desire to model the influence of food quality and short- term changes in food supply on larval growth and survival and the influence of egg size and composition on ultmiate success at meta- morphosis. Although biochemically based, the model contains only the crudest biochemical constructions. The larva is modeled as a four- constituent organism composed of protein, carbohydrate, neutral lipid, and polar lipid. Each constituent and the transitions between constituents are modeled using the simplest of flow schemes. So, for example, lipid and carbohydrate are interchangeable, carbohy- drate covers respiratory demand when in sufficient supply, and 260 BOCHENEK ET AL. 0.50 H 0.25 0.00 -mortality, metabolic sources only mortality, including non-metabolic sources 30 40 50 60 Initial Egg Size (|jm) 70 80 0.50 H 0.25 0.00"" 0.50 0.7 0.9 1.1 1.3 1.5 Respiration Rate (KJ d"^ ) TO 0.25 0.00 250 275 350 375 400 300 325 Length (^im) Figure 25. Simulated Crassoslrea gigas larval survival as a function of (A) initial egg size, (Bl base respiration rate, and IC) larval length at metamorphosis for a base respiration rate distribution that is centered at 1.254 KJ day"'. The simulation used the reference case environ- mental conditions given in Figure 9. neutral lipid is the primary storage component. Assimilated protein is used only to create tissue protein, and this necessitates the for- mation of a certain amount of structural carbohydrate and polar lipid. Failure to supply these other components in the amounts required by protein assimilation results in structural imbalances and eventually death. More complex biochemical transformations are excluded. So. for example, although many amino acids can be synthesized, protein, in the model, comes only from protein in- gested as food, with the one exception early in larval life when neutral lipid is used to sustain growth in the first few days after birth. Carbohydrate and protein, though potentially used as energy reserves for metamorphosis or to sustain the larva during periods of negative scope for growth, are only used as the last resort and, then, only in proportion to their fractional contribution to structural tissues. Although based on a simplistic biochemistry, the model suc- ceeds in simulating some of the basic observations of C. gigas eggs and larvae, suggesting that simple biochemical constructs can be successful and may encompass the biochemical transitions most prominent in determining cohort success. Verifications of many of the details in the simulations cannot be accomplished because of limited information on the details of larval biochemical composi- tion as it is influenced by environmental and genetic factors. Nev- ertheless, the model succeeds in simulating known population- level characteristics that permit verification at this higher-order level of integration. Necessities of Model Constnietion A number of special model constructs were required to obtain verifiable simulations. These included resolving a mismatch be- tween egg ash content and earliest larva ash content, the conver- sion of egg size to earliest larval size, the addition of genetic variation, and the need to modify filtration rate, especially in early stage larvae. Birth and Condition Condition is tracked in the model using a protein-to-ash ratio. Tracking condition was necessitated by the desire to model periods of low food supply, including, in the extreme, periods of starva- tion. Particularly once the shell is formed, larval size does not change during periods of restricted food supply, but larval tissue weight does (Laing 1993). The protein-to-ash ratio for eggs is high and does not fit the larval pattern. Hence, initializing the model with the protein-to-ash ratio of the egg consistently failed to produce verifiable simula- tions. Presumably, as the egg develops and hatches, ash is added from inorganic solutes in the surrounding water and the protein- to-ash ratio drops. The model was initialized with the protein-to- ash ratio of newly hatched larvae to circumvent this problem. Reconciling the mismatch between the protein-to-ash ratio of eggs and newly hatched larvae will require additional experimental studies. Length at Birth Eggs are more or less spherical. Larvae, even newly hatched, are not. The model tracks length independently of weight, a ne- cessity imposed by the wealth of data for verification provided in terms of length and the need to follow condition. Initializing the model with egg ""length"" (diameter) fails because the increase in length during egg development is not representative of growth, but simply a result of tissue reorganization. Consequently, a growth model cannot account for this process. We used a conversion from egg diameter to earliest larval length to circumvent this problem. Genetic Variation One of the key observations recorded in the literature is the success rate at metamorphosis and the size of metamorphosing larvae. Considerable variability exists depending upon the condi- tions of larval culture, and egg quality provides a sufficiently strong signal that variations in egg quality should influence suc- cess at metamorphosis in simulated spawns. Considerable variabil- ity also exists within cohorts. Such variability cannot exist if all larvae in the cohort are equivalently affected by environmental conditions. Consequently, it was necessary to add some variation between larvae to the model. The observations of Gallager and Mann ( 1986a) and Gallager et al. (1986) provide a basis for describing a range of egg sizes with a simple Gaussian function to define the frequency of a given egg size in a cohort. This range in egg sizes also produced a range in egg qualities in that larger eggs were relatively more lipid rich. The resulting simulations showed an improved fit to observation in Biochemical Model of C. gigas Larvae 261 that a range of larval sizes and success rales at iiictaniorpliosis were obtained. However, the range in predicted lar\al si/e and success at metamorphosis was still too constrained in comparison to obser- vation. The obvious next option was to include a range of growth efficiencies. Genetic variation in growth efficiency is well de- scribed and may accrue from any number of processes including variations in respiration rate, protein turnover, assimilation effi- ciency, or feeding efficiency (e.g.. Garton 1984. Koehn and Hil- bish 1987. Garton & Berg 1989. Koehn & Bayne 1989. Garton & Haag 1991). In the model, respiration rate and filtration rate con- trol growth efficiency and. although the two processes are some- what differently affected by temperature and salinity, inserting variation in either effectively generates simulated larvae with a range of growth efficiencies. Genetic variation in growth effi- ciency was inserted as a range in base respiration rates using a simple Gaussian construction. This addition produced the range in outcomes at metamorphosis expected from observation. Simulations were run to examine the influence of varying the mean of the Gaussian distribution describing egg size and respi- ration rate. Simulations did not change markedly with a variation in egg size because the range of viable egg sizes was tightly constrained, as discussed later. The model was more sensitive to variations in the mean base respiration rate. Here, however, simu- lations showed that little leeway existed for varying the central tendency of base respiration rate because substantial changes in cohort survival occuiTed with relatively small changes in central tendency. These results are compared, for the most part, to observations taken under saturating conditions of food and near-optimal envi- ronmental conditions. Optimal base respiration rates, to a large extent, and egg sizes, particularly under limiting food supply, sug- gest that changes in the range of egg size and base respiration rate might be adaptive in certain cases that might routinely exist under tleld conditions. One might expect variations in respiration rate ( = growth efficiency) to be the most adaptive. Filtration Rate Measured filtration rates always provided growth rates higher than observed. The mismatch was largest for smallest larvae. In these animals, observed reductions in neutral lipid clearly indi- cated that assimilation does not provide adequate resources to explain observed growth, although measured filtration rates would indicate otherwise. It seems likely that filtration rate and ingestion rate are not equivalent in larvae or that assimilation efficiency is size-dependent. In the model, a size dependency on ingestion rate or assimila- tion efficiency is effectively equivalent, so no attempt was made to distinguish between the two. One might reasonably conclude that both feeding and digestion should be affected by larval develop- ment processes, particularly during early larval life, and that this might lower the amount of energy realized at a given filtration rate. One might also conclude that filtration efficiency in part is a func- tion of the larva's use of the vellum to maintain its position in the water column as well as to feed and. so. particularly under condi- tions of saturating food supply where most filtration rates are measured, a tendency to filter more material than can be ingested should exist. Regardless of the cause, to lower growth rales from levels predicted from observed filtration rates, we imposed a size- dependent penalty on ingestion that was largest for the smallest larvae. The mismatch between observed growth rates and simu- lated growth rates from measured filtration rates, however, points to an area of early lar\ al biology that warrants further study. Metamorphosis The simple biochemical construction of the model required a simple explanation for the metabolic basis for triggering metamor- phosis. A full explanation of how endogenous and exogenous fac- tors control metamorphosis (e.g.. Coon & Bonar 1986. Fitt et al. 1990, Berias & Widdows 1995) does not exist. Accordingly, the approach used was derived within the limitations imposed by the four-pool biochemical construct of the model and five observa- tions in the literature that directly related to it. ( 1 ) Filtration rate drops in larvae of about 230 [xni and larger, probably due to changes in either behavior or the beginnings of tissue reorganiza- tion that must presage metamorphosis. The former option would be sufficient. Older larvae spend more time near the bottom (e.g.. Dekshenieks et al.. 1997) and. thus, may spend less time filtering, a fact that would be interpreted in experiment as a decline in filtration rate. The reduction in feeding rate should ultimately re- duce larval scope for growth, and this should have consequences concerning the decision to metamorphose. (2) Smallest size at metamorphosis is about 275 (j,m. This size should be somewhat larger than the size triggering the decline in filtration rate. (3) Lipid stores decline at metamorphosis. This could be a consequence of a decline in scope for growth as well as a consequence of the energy needed to reorganize tissue. (4) Literature information suggests that larvae require a certain amount of stored energy to metamor- phose successfully. Although any kind of tissue constituent might provide this energy, the reliance of larvae on neutral lipid as the primary energy store suggests that the proportion of neutral lipid is a good measure of energy available for metamorphosis. (5) The quantity of lipid present in the egg influences larval survival. Thus some information on the status of neutral lipid reserves should pertain to the decision to undergo metamorphosis. The process of metamorphosis was modeled using these five observations to generate explicit triggers for certain steps in the process, as follows. (I) Larvae were assumed to become poten- tially competent to metamorphose at 275 ptm, following a decrease in filtration rate at 250 jjim. Simulations showed that the filtration rate decline could not be set at 2M) |xm or 275 ixm. The minimum size for metamorphosis, in most simulations, did not fall below 285 (Jim. so that the 275 iJ-m size limit was rarely invoked. That is. invoking a change in filtration rate at 250 p.m normally resulted in larvae metamorphosing at sizes above 275 |xm. (2) Larvae were assumed to become competent to metamorphose when a daily decline in neutral lipid of a certain level occurred. Our assumption was that larvae might be expected to continue to grow and store lipid as long as a sufficiently positive scope for growth was present and this would enhance success, but that a decline in neutral lipid would reduce success. Accordingly, metamorphosis should occur when scope for growth dropped significantly below zero. The range of observed sizes at metamorphosis suggests that some pro- cess of this sort does occur. Although the decline in filtration rate at 250 jjim predestined larvae to eventually reach the trigger point defined by a significant neutral lipid decline, food quantity and quality and biochemical composition can permit growth much in excess of 250 p.m before scope for growth drops to substantially negative values. Typically, in the model, metamorphosis occurred at sizes of 300-330 |j,m, as observed in culture. (3) Larvae were 262 BOCHENEK ET AL. assumed to metamoqjhose successfully if neutral lipid supplies were adequate. Adequacy was judged as a ratio between energy stores and structural components. We cannot evaluate how accurately the modeled mechanism for metamoiphosis approaches reality, not having available an ad- equate understanding of the biochemistry of the process. However, the simulations reveal some interesting trends. The choice of 250 (Jim as the point where filtration rate declines is based on obser- vation, but the model also indicates that this trigger is tightly constrained to this size. Neither 230 [x.m nor 270 |xm sizes offered verifiable results. The choice of a 25% daily decline in neutral lipid triggering competency is also tightly constrained. Values of 10% and 40% did not provide results equivalent to observations. Both larval size distributions and success rates at metamorphosis varied from observations. The choice of a >1:1 ratio of neutral lipid to polar lipid is also tightly constrained. Values of 0.8: 1 and 1.1:1 produce unrealistic size distributions and success rates at meta- morphosis. Verification of this construction for modeling metamorphosis was directed at evaluating performance in simulating four impor- tant phenomenon: ( 1 ) variations in egg quality significantly influ- enced success at metamorphosis. (2) variations in food quality and quantity significantly influenced success at metamoiphosis. (3 1 larval life span as predicted was well within the range of obser- vations, and (4) larval size structure at metamorphosis was well within the range of observation. Obtaining these four results re- quires a reasonably accurate rendition of growth and survi\al at the biochemical resolution of the model. This suggests that the ap- proach to modeling metamorphosis must reflect, in some signifi- cant way. the process as it actually proceeds in the larva. Conseqiienccs of Model Coiistnictioii Larval success is determined by intrinsic and extrinsic factors. Intrinsic factors include egg size and quality and genetic makeup. Extrinsic factors include temperature, salinity, food quality, and food quantity. Implications of Egg Size Oyster eggs are about 50 |jLm in diameter, with a size range typically of 40-60 |j.m. The model identifies viable egg sizes in the range 37-73 |j.m, very similar to observations. Egg sizes outside this range are predicted to be nonviable due to lipid imbalances in early larval life. Very likely, the lower limit of 37 p,m represents a packaging problem. Egg size is simply too small to provide adequate resources for the structural changes required in forming the first larval stage. In the model, the required structural tissue ratios cannot be achieved and still provide any neutral lipid re- serves. In effect, the larva is never born. The upper limit of about 73 |j,m yields a larva that has insufficient neutral lipid reserves to cover metabolic needs immediately post-hatch. During this time, feeding is inefficient, and some of the larva's carbon needs for growth and tissue maintenance must be met by using neutral lipid reserves. The larger larvae, coming from eggs >73 |jim in diameter. essentially starve to death before they can become competent filter feeders. This may provide one explanation for the small size of most planktotrophic eggs. Presumably, the upper limit on egg size could be extended by increasing neutral lipid reserves; however, the bet hedging mode of life (e.g.. Steams 1976) would limit the amount of energy invested in any one embryo. The trade-off between additional energy ex- penditure and increased success at metamorphosis is clearly indi- cated in Figures 11. 15, and 23. Larger eggs yield successful larvae over a much larger range of respiration rates and environmental conditions than do smaller eggs. Larger eggs yield larvae that reach metamorphosis faster (shorter planktonic time), thus mini- mizing loss to predation and the chance of reduced survival from transient reductions in food supply. Thus, the simulations suggest that the a\erage egg size of 50 [j,ni minimizes the chance of re- productive failure, wliich increases rapidly at smaller egg sizes, while still permitting the spawning of a large number of eggs. As an example, increasing average egg size to 60 |xm reduces total egg output by 31% at a given total energy expenditure. An equiva- lent increase in larval success is not achieved in our simulations. The model also indicates, however, that transient reductions in food supply during larval life may increase the success rate for large eggs relative to small eggs. In this circumstance, the extra energy required to produce large eggs may be better repaid. Whether an increase in fitness is adaptively advantageous requires a better understanding of food supply under field conditions and how this influences larval survival. Respiration (=Gro\vth Efficiencyl Effects In the model, varying respiratory rate is equivalent to varying growth efficiency. Larvae with high growth efficiency have low respiration rates. The model identified an upper and lower limit to growth efficiency under defined environmental conditions. The upper limit varies widely depending upon environmental condi- tions, whereas the lower limit is relali\ely fixed. Simulations show that the upper limit on base respiration rate (e.g., -1 kJ day"' in Fig. 9) is determined by the point at which larvae cannot acquire sufficient neutral lipid stores to successfully metamorphose. Smaller eggs are also less viable because insufficient neutral lipid can be stored to cover larval needs over a few days post-hatch. Interestingly, very tow respiration rates also normally result in unsuccessful larvae. These animals put too much assimilated car- bon into somatic structural tissue and so have insufficient neutral lipid reserves. We are unaware of experimental data upon which to verify this last result. Condition and Mortality Many models do not explicitly follow length and weight inde- pendently (e.g., Powell et al. 1992, Dekshenieks et al. 1993). In bivalves, tracking condition permits observation of larval perfor- mance during periods of low food supply. This requires tracking length and weight independently such that not all increases in weight result in changes in length and such that no decreases in weight result in decreases in length. The performance of the model was evaluated under conditions of food deprivation by simulating the process of starvation. Although the mechanisms of death under these conditions are described in the model, death occurs due to a variety of biochemical imbalances, depending upon the initial sta- tus of the larva. Whether such a degree of complexity actually exists requires more information on the changes in larval bio- chemical composition under conditions of low food supply. How- ever, the higher-level effects that integrate biochemical processes Biochemical Model of C. gigas Larvae 263 were simulated by the model, including a decrease in weight (con- dition), a drop in neutral lipid content, a nonlinear time-dependent increase in mortality, and the still-sLicccssful metamorphosis of older hir\ae. In addition to inadequate food suppl\ . larvae can die if food of inadequate composition is ingested. Thus, rigorous criteria were set for biochemical compositions not allowed in viable larvae. Food having inadequate lipid or being too protein-rich resulted in mortality, even if the quantity of food remained high. These pa- raiiieterizations describing mortality under such conditions are es- sentially aJ hoc constructs (literature observations not being avail- able), but they did produce cohort mortality rates that appeared to be realistic. Effect of Diet Most experimental studies on C. gigcis larvae have used food supplies of >2 mg L '. This level of food saturates feeding and. in fact, raising food quantity from 2 mg L"' to 4 mg L~' in the model has little influence on simulated larval success. However, as in Crassosirea virginica (Dekshenieks et al. 199.^). food quantities below 1 ing L"' dramatically restrict larval growth and survival. As food supply declines, animals with high growth efficiencies are selected for in the iriodel. At high food content, larger eggs with lower growth efficiencies also survive to metamorphosis. With rare exceptions, small eggs with low growth efficiencies never do. Thus, the influence of growth efficiency is nonrandomly distrib- uted across egg size, and the influence seems to be mediated in part by food quantity and to a larger measure by food quality. The influence of food content on C. gigas larval growth and survivorship has received considerable attention (e.g.. Wilson 1978. Waldock & Nascimento 1979. Helm & Laing 1987). Al- though not all studies agree, low-protein diets and high-lipid diets often show improved growth and survivorship. The simulations show the positive effect of a low-protein diet on larval growth and survivorship. With this diet, a relatively larger portion of ingested energy is allocated to energy stores that in turn sustain the larva through metamorphosis. With a high-protein diet, larvae grow too fast and fail to store enough energy to sustain them through meta- morphosis. The destination of protein within the larva is limited in terms of building tissue and covering metabolic needs (Table 2) if insufficient carbohydrate is ingested. Any transfer of excess amino acid into other tissue components is not permitted. Potentially, this allocation of ingested protein is too simplistic, although the simu- lations do provide some insight into the value of a low-protein diet. Simulations with no neutral lipid gave similar results in terms of larval survivorship and growth. Thus, the relative amounts of protein and neutral lipid in larval food are important determinants of growth and sur\ i\al. A number of studies have identified specific components of the lipid pool as important dietary constituents (e.g.. Thompson et al. 1994, 1996). The model could be expanded to track inore complex biochemical pools such as polyunsaturated fatty acids (PUFAs) or sterols. The fact that the model achieves realistic simulations over a relatively wide range of environmental and dietary conditions indicates that the approach used to model larval biochemistry, including the subsuming of a diversity of lipid compounds into two pools, polar and neutral, is sufficient to provide realistic simula- tions of larval growth, metamorphosis and survival. Temperature and Salinity C gigas is known to be relatively stenotopic for the genus. Temperature and salinity conditions describing optimal growth circumscribe a narrow range. The model reproduces this behavior. In this contribution, most simulations were run under optimal con- ditions of 25°C and 30%c. Lower temperatures result in insufficient neutral lipid storage and metamorphosis because feeding rate is low. Temperatures above .30'C result in biochemical imbalances due to high respiratory demand. Low salinity also results in insuf- ficient food ingestion to meet the demands of metamorphosis. Again, data to verify the accuracy of predicted cause and effect on biochemical composition are not available. Growth rate is a complex function dependent upon the balances of ingestion and respiration. The ability of positive environmental conditions to offset a reduction in food supply and vice versa depends upon the relative scaling of their affects on respiration and ingestion. Thus, increased temperature can "spare" a reduction in food, permitting the same growth rate, if the influence of tempera- ture on ingestive processes scales with a larger exponent than the influence of temperature on respiration. The importance of differ- ential scaling in the energy balance of bivalve molluscs and other animals is well known (e.g.. Newell et al. 1977, Powell et al. 1992. Brown et al. 1993). Given the sensitivity of growth and survival to decreases in food supply, the fact that a decrease in food supply often occurs during summer months in C. gigas habitat (Kobayashi et al. 1997. Hyun et al.. in press), when an increase in temperature is likely to be of significance in increasing ingestion rate, suggests that the differential scaling of ingestive processes and respiration is likely of significance for the reproductive success of the species. CONCLUSIONS A model that simulates the growth, development, and meta- morphosis of Crassosirea gigas larvae has been developed. The model is the first of its kind in that it ( 1 ) tracks length separately from weight so that changes in condition can be followed and (2) predicts growth from the ingestion and transformation of bio- chemical constituents, thus permitting the simulation of the effects of changes in food quality. Food quality and feeding rate are important constraints in larval culture, so the model might be used to optimize culture conditions for C. gigas larvae as well as to investigate the influence of critical periods of food supply in larval development in the field. Of particular importance is the investi- gation of "teleconnections" during larval life in which events oc- curring at one point in larval life have consequences at another, temporally distant, point. The model has a crude depiction of the biochemistry of C. gigas larvae. However, the model works well even with this limited biochemistry and indicates that the formu- lation of sophisticated biochemicallv based models offers the promise of substantially improving the population modeling of marine larvae. ACKNOWLEDGMENTS Computer resources and facilities were provided by the Center for Coastal Physical Oceanography at Old Dominion University. We also acknowledge sabbatical funding to Eleanor Bochenek provided by Rutgers University. We akso acknowledge the support of Sea Grant, including the Oyster Disease Research Program, for support of the Rutgers/ODU shellfish modeling group. 264 BOCHENEK ET AL. LITERATURE CITED Anger, K.. R. R. Dawirs. V. Anger & J. D. Costlow. 1981. Effects of early starvation periods on zoeal development of brachyuran crabs. Bml. Bull. (Woods Hole) 161:199-212. Arakawa. K. Y. 1990. Natural spat collecting in the Pacific oyster Cras- sostrea gigos Thunberg. Mar. Bebav. Physiol. 17:95-128. Beiras. R. & J. Widdovvs. 1995. Induction of metamorphosis in larvae of the oyster Crassostrea gigos using neuroaclive compounds. Mtu: Biol. (Bert) 123:327-334. Breese, W. P. & R. E. Malouf. 1975. Hatchery manual for the Pacific oyster. Sea Grant College Program ORESU-H-75-002. Oregon State University. 23 pp. Brown. J. H.. P. A. Marquet & M. L. Taper. 1993. Evolution of body size: Consequences of an energetic definition of fitness. Am. Nat. 142:573- 584. Canuto. C. M. Y. Hussami, A. Quarteroni &. T. A. Zang. 1988. Spectral methods in fluid dynamics. New York: Springer- Verlag. Carlotti. F. & A. Sciandra. 1989. Population dynamics model otEiiieipiiui aculifrons (Copepoda: Harpaclicoida) coupling individual growth and larval development. Mar. Ecol. Prog. Ser. 56:225-242. Coon. S. L. & D. B. Bonar. 1986. Norepinephrine and dopamine content of larvae and spat of the Pacific oyster. Crassosrrea gigas. Biol. Bull. (Woods Hole) 171:632-639. Crisp. E.J. 1971. Energy flow measurements. In: N. A. Holme & A. D. Mclntyre, editors. Methods for study of marine benthos. I.B.P. Hand- book No. 16. Oxford: Blackwell Scientific Publications, pp. 197-297. Cushing. D. H. & R. R. Dickson. 1976. The biological response in the sea to climatic changes. Adv. Mar. Biol. 14:1-122. Dekshenieks. M. M.. E. E. Hofmann. J. M. Klinck & E. N. Powell. 1996. Modeling the vertical distribution of oyster larvae in response to en- vironmental conditions. Mar. Ecol. Prog. Ser. 136:97-110. Dekshenieks. M. M.. E. E. Hofmann, J. M. Klinck & E. N. Powell. 1997. A modeling study of the effects of size- and depth-dependent predaiion on larval survival. J. Plankton Res. 19:1583-1598. Dekshenieks, M. M., E. E. Hofmann & E. N. Powell. 1993. Environmental effects on the growth and development of Eastern oyster. Crassosrrea virginica (Gnielin. 1791). larvae: A modeling study. / Shellfish Res. 12:241-254. Pitt. W. K.. S. L. Coon. M. Walch. R. M. Werner. R. R. Colwell & D. B. Bonar. 1990. Settlement behavior and metamorphosis of oyster larvae (Cras.sostrea gigas) in response to bacterial supematants. Mar. Biol. (Berl.) 106:389-394. Gallager. S. M. & R. Mann. 1986a. Growth and survival of larvae of Mercenaria mercenaria (L.) and Crassostrea virginica (Gmelin) rela- tive to broodstock conditioning and lipid content of eggs. Aquaciilnne 56:105-121. Gallager. S. M. & R. Mann. 1986b. Individual variability in lipid content of bivalve larvae quantified histochemically by absorption photometry. J. Plankton Res. 8:927-937. Gallager. S. M.. R. Mann & G. C. Sasaki. 1986. Lipid as an index of growth and viability in three species of bivalve larvae. Ai/uaciiltiire 56:81-103. Garton. D. W. 1984. Relationship between multiple locus heterozygosity and physiological energetics of growth in the estuarine gastropod Thais haenuistoma. Physiol. Zool. 57:530-543. Garton. D. W. & D. J. Berg. 1989. Genetic variation at the LAP locus and ammonia excretion following salinity transfer in an estuarine snail. Comp. Biochem. Physiol. .4 Comp. Physiol. 92:71-74. Garton. D. W. & W. K. Haag. 1991. Heterozygosity, shell length and me- tabolism in the European mussel. Dreissena polymorpha. from a re- cently established population in Lake Erie, Comp. Biochem. Phxsiol. A Comp. Phy.siol. 99:45^8. Gerdes. D. 1983a. The Pacific Oyster Crassostrea gigas. Part 1. Feeding behaviour of larvae and adults. Aipiacnhiire 31:195-219. Gerdes. D. 1983b. The Pacific Oyster Crassostrea gigas. Pan II. Oxygen consumption of larvae and adults. Aqiuiculture 31:221-231. Handa. N. 1969. Carbohydrate metabolism in the marine diatom. Skel- ehmema costatmn. Mar. Biol. (Berl.) 4:208-214. Haws. M. C. L. DiMichele & S. C. Hand. 1993. Biochemical changes and mortality during metamorphosis of the Eastern oyster. Crassostrea vir- ginica, and the Pacific oyster. Crassostrea gigas. Mol. Mar. Biol. Bio- teclmol. 2:207-217. Helm. M. M.. D. L. Holland & R. R. Stephenson. 1973. The effect of supplementary algal feeding of a hatchery breeding stock of Oslrea cdiilis L. on larval vigour. / Mar. Biol. Assoc. U.K. 53:673-684. Helm. M. M. & I. Laing. 1987, Preliminary observations on the nutritional \ alue of "Tahiti IsochiysLs" to bivalve larvae. Aqiiacidture 62:28 1-288. Helm. M. M. & P. F. Millican. 1977. Experiments in the hatchery rearing of Pacific oyster larvae (Crassostrea gigas Thunberg). .\(piacidlioe 1 1:1-12. His. E. & D. Maurer. 1988. Shell growth and gross biochemical compo- sition of oyster larvae (Crassostrea gigas) in the tleld. Ai/iKuidture 69:185-194. His. E.. R. Robert & A. Dinet. 1989. Combined effects of temperature and salinity on fed and starved larvae of the Mediterranean mussel Mytiliis galloprovincialis and the Japanese oyster Crassostrea gigas. Mar. Biol. (Berl.) 100:455-463. His, E. & M. N. L. Seaman. 1992. Effects of temporary star\ation on the survival, and on subsequent feeding and growth, of oyster (Crassostrea gigas) larvae. Mar Biol. 1 14:277-279. Hoegh-Guldberg. O. & D. T. Manahan. 1995. Coulometric measurement of oxygen consumption during development of marine invertebrate em- bryos and larvae. J. E.Kp. Biol. 198:19-30. Hyun. K.-H., I.-C. Pang. J. M. Klinck. K.-S. Choi. J.-B. Lee. E. N. Powell. E. E. Hofmann. E. A. Bochenek. in press. The effect of food compo- sition on Pacific oyster Crassostrea gigas (Thunberg) growth in Korea: a modeling study. Aquaciilture. Jorgensen. C. B. 1952. Efficiency of growth in Mytilns ediilis and two gastropod veligers. Nature (bond.) 170:714. Kobayashi. M.. E. E. Hofmann. E. N. Powell. J. M. Klinck & K. Kusaka. 1 997. A population dynamics model for the Japanese oyster. Crassos- trea gigas. Aqiiaculture 149:285-321. Koehn. R. K. & B. L. Bayne. 1989. Towards a physiological and genetical understanding of the energetics of the stress response. Biol. J. Linn. Soc. 37:157-171. Koehn. T. K. & T. J. Hilbish. 1987. The adaptive importance of genetic variation. Am. Sci. 75:134-141. Kusaki. Y. 1991. Oyster culture in Japan and adjacent countries: Crassos- trea gigas (Thunberg). In: W. Menzel. editor. Estuarine and marine bivalve mollusk culture. Boca Raton, FL: CRC Press. Inc. pp. 227-243. Laing. I. 1995. Effect of food supply on oyster spatfall. Aquaciilture 131: 3 1 5-324. Lannan. J. E. 1 98(^1. Broodstock management of Crassosliea gigas I. Ge- netic and environmental variation in survi\al in the larval rearing sys- tem. Aquaculture 21:32.V336. Lee. C. K. & J. J. Lee, 1968, The effect of some factors on the mortality of trochophora of oyster, Crassostrea gigas. Bull Korean Fish. Soc. 1 : 45-49. Lee, R. F. & P. B. Heffeman. 1991. Lipids and proteins in eggs of Eastern oysters (Crassostrea virginica (Gmelin, 1791)) and northern quahogs (Mercenaria mercenaria (Linnaeus. 1758)). J. Shellfish Res. 10:203- 206. Lee. R. F.. J. C. Nevenzel & G.-A. Paffenhofer. 1971. Importance of wax esters and other lipids in the marine food chain: Phytoplankton and copepods. Mar. Biol. (Berl.) 9:99-108. Mann. R. & S. M. Gallager. 1985. Physiological and biochemical energet- ics of larvae of Teredo navalis L. and Bankia gouldi (Bartsch) (Bi- valvia: Teredinidae). / E.xp. Mar. Biol. Ecol. 85:211-228. Malouf R. E. & W. P. Breese. 1977. Food consumption and growth of larvae of the Pacific oyster. Crassostrea gigas (Thunberg). in a con- stant flow rearing system. Proc. Natl. Shellfish Assoc. 67:7-16. Biochemical Model of C. gigas Larvae 265 Nascimento. I. A. 1980. Growth of the Uirvae of CmxMisnvd ainiis Thun- berg. fed with different algal species at high cell concentrations. J. Cons. Int. Explor. Men 39:134-139. Nell. J. A. & J. E. Holiday. 1988. Effects of salinit> on the growth and survival of Sydney Rock Oyster (Saccostrea commercialis) and Pacific oyster (Cnissoslrea gigas) larvae and spat. Aquacutnire 68:39^14. Newell. R. C. L. G. Johnson & L. H. Kofoed. 1977. Adju.stment of the components of energy balance in response to temperature change in Osiica ediilis. Oecoiogia (Berl.) 30:97-1 10. Pauley. G. B.. B. van der Raay & D. Troutt. 1988. Species profiles: Life histories and environmental requirements of coastal fishes and inver- tebrates (Pacific Northwest) Pacific oyster. U.S. Dept. Interior. Fish Wildl. Serv., Biol. Rpt. 82(11.85). TR EL-82-4. Powell, E. N., E. E. Hofmann, J. M. Klinck & S. M. Ray. 1992. Modeling oyster populations L A commentary on filtration rate. Is faster always better? J. Shellfish Res. 11:387-398. Quayle, D. B. 1988. Pacific oyster culture in British Columbia. Can. Bull. Fish. Aqiial. Sci. 218:1-229. Robinson. .A. 1992. Dietary supplements for reproductive conditioning of Crassosueu gigas kumamolo iThunberg). I. Effects on gonadal devel- opment, quality of ova and larvae through metamorphosis. J. Shellfish Res. 11:437-141. Roman. M. R. 1983. Nitrogenous nutrition of marine invertebrates. In: E. J. Carpenter & D. G. Capone. editors. Nitrogen in the Marine Environ- ment. New York: Academic Press, pp. 347-383. Stearns. S. C. 1976. Life-history tactics: .\ review of the ideas. Quart. Rev. Biol. 51:3-47. Strathmann. R. R.. L. Fenaux. A. T. Sewell & M. F. Strathmann. 1993. Abundance of food affects relative size of larval and postlarval stnic- tures of a molluscan veliser. Biol. Bull. {\Voo,ls Hole) 185:232-239. Taggart. C. T. & W. C. Leggetl. 1987. Short-term mortality in post- emergent larval capelin Mallotus villosus. I. Analysis of multiple in situ estimates. Mar. Ecol. Prog. Ser. 41:205-217. Thompson, P. A., M.-X. Guo & P. J. Harrison. 1996. Nutritional value of diets that vary in fatty acid composition for larval Pacific oysters iCras- sostrea gigas). Aquaeutture 143:379-391. Thompson. P. A. & P. J. Harrison. 1992. Effects of monospecific algal diets of varying biochemical composition on the growth and survival of Pacific oyster {Crassosireu gigas) larvae. Mar. Biol. {Berl.) 113:645- 654. Thompson. P. A.. D. J. S. Montagnes. B. A. Shaw & P. J. Harrison. 1994. The influence of three algal filtrates on the grazing rate of larval oysters iCrussosirea gigas). determined by fluorescent microspheres. Aqua- culture 119:237-247. Trytek. R. E. & W. V. Allen. 1980. Synthesis of essential amino acids by bacterial symbionts in the gills of the shipworm Bankia setacea (Tryon). Comp. Binchem. Physiol. A Camp. Physiol. 67:419^27. Utting. S. D. 1986. A preliminary study on growth of Crassostrea gigas larvae and spat in relation to dietary protein, .\quaculture 56:123-138. Ventilla. R. F. 1984. Recent developments in the Japanese oyster culture industry. Adv. Mar. Biol. 21:1-57. Waldock. M. J. & L A. Nascimento. 1979. The triacylglycerol composition of Crassostrea gigas larvae fed on different algal diets. Mar. Biol. Lett. 1:77-86. Whyte. J. N. C. N. Bourne & C. A. Hodg.son. 1987, Assessment of bio- chemical composition and energy reserves in larvae of the scallop Patinopeeten yessoensis. J. E.xp. Mar. Biol. Ecol. 113:113-124. Wilson. J. H. 1978. The food value of Phaeodactylum tricornutuin Bohlin to the larvae of Oslrea edulis L. and Crassostrea gigas Thunberg. Aquacullure 13:313-323. Jdiiivul iij SluilJLsh Research. Vol. 20. No. I. Ibl-^lli, 2001. BEHAVIORAL RESPONSES OF VELIGER LARVAE OF CRASSOSTREA GIGAS TO LEACHATE FROM WOOD TREATED WITH COPPER-CHROME-ARSENIC (CCA): A POTENTIAL BIOASSAY OF SUBLETHAL ENVIRONMENTAL EFFECTS OF CONTAMINANTS ANGELIKA PRAEL.' SIMON M. CRAGG.' * AND SUZANNE M. HENDERSON" Institute of Marine Sciences. School of Biolofiical Sciences, University of Portsinoiilh, Ferry Road, Portsnioiitli P04 9LY. United Kin,i>doni: -Forest Products Researcli Centre, Bucl = 0.033). Larvae in the cuvettes continued to be active for at least a day after initiation of the experiments, except in the case of 7-day-old larvae exposed to CCA-oxide leachate. which became totally in- active after 24h. The significantly higher activity of 3-day-old larvae was no longer evident three hours after the larvae were introduced to the leachate. While the 2- and 7-day larvae in seawater without leachate had similar mean upward swimming velocities (0.32 and 0.31 mtn.s"' respectively), three-day larvae were much slower (0.22 tnm.s"') (Fig. 2). There was no significant difference between the velocities of 2-day-old larvae in seawater alone and in seawater with various leachate types, but in the case of 3- and 7-day larvae the presence of leachate resulted in significantly higher swimming velocities (/> < 0.05) (Table 3). Three-day-old larvae subject to CCA leachate swam 1 .4 times faster than those in leachate from untreated wood and 2.7 times faster than those in hatchery water. Seven-day larvae subject to CCA leachate. swam on average twice as fast as those in uncontaminated seawater. Significant cuvette to cuvette varia- tions in mean swimming velocity occurred, accounting for 33%, 1 1% and 12% of variance in the case of 2-. 3- and 7-day-old larvae respectively. DISCUSSION The ,\'aliire of Leachates The analytical data from the leachates indicate that the rate of leaching of copper and chromium from treated wood generally decreased with time. This is consistent with findings of other stud- ies of leaching into seawater from CCA-treated wood (Albuquer- que et al. 1996). Studies to date suggest that the rate of leaching declines exponentially and that this is true for all three elements (Brooks 1996, Cragg et al. 2001 ). Factors such as the retention of preservative, the surface area of wood exposed to the water, the proportion of exposed wood surface that consists of cross-cut wood cells and the volume of water used all affect the concentra- tion of metals in the leachate (Cragg et al. 2001). In this study, the use of small leaching volumes, of a high proportion of cross-cut surfaces and of temperatures at the upper end of the range likely to occur in the environment, will have lead to concentrations of leached elements exceeding the worst case scenario proposed by Brooks (1996). The chromium concentrations measured in the hatchery water (0.17 ppm), greatly exceed the expected chromium concentration of uncontaminated seawater of 0.003 ppm (Brooks 1993). The elevated chromium levels may indicate the presence of stainless steel somewhere in the hatchery water system. The experimental design did not permit the isolation of this effect on the larvae from other factors. Lan'al Responses to Leachate Control larvae swam at velocities somewhat lower than those reported for other oyster veligers, 0.37 to 3.10 mm s"' depending on age and conditions (Mann & Rainer 1990). This inay reflect the small size of the experimental chamber, but Mann and Rainer demonstrated that only at oxygen tensions less than 20% of satu- ration were rates of swimming significantly lower. Toxic effects of high concentration leachates were evident in the decreasing activity of the larvae in the initial experiment and in 7-day-old larvae in the main experimental series. The initially higher swimming activity by 3- and 7-day-old larvae in the pres- ence of CCA leachates might be due to a stimulatory effect of Cu on cell physiology, a direct physiological response, rather than a behavioral response mediated by the nervous system. Evidence that such effects may occur is provided by the increased the rate of Behavioral Responses of Bivalve Veligers to Contaminants 2-day larvae 271 60 50 40 30 - 20 10 0 60 50 40 30 20 10 -I 0 D no leachate Q untreated wood B CCA-oxide m CCA-salt rfn _.*_ % activity 60 - 50 40 30 20 10 - 0 2-6ay larvae 60 50 40 30 20 10 0 60 50 40 30 20 10 0 1 7-day larvae 60 50 40 30 20 10 0 X i 1st day leachate 2nd day leachate Figure 1. Activity of 2-, 3- and 7-da.v-old Crassostrea gigas veliger larvae in cuvettes containing seawater witli or without leachate from wood bh)cks. Leachate obtained after the first or second day of immersion of the blocks. The blocks were with or without treatment by CCA salt or oxide solutions. Numbers of larvae ascending plus numbers descending past a reference line on the cuvettes are expressed as mean + SE of % of the total number of active larvae. cell cleavage and larval growth observed by Hoare et al. (1995) in Mylihis edulis larvae exposed to 8 ppb of copper. However, it is more likely that the increased swimming velocity and the greater overall levels of up and down movement represent kinesis-type responses to inimical conditions. Kinesis responses may also be responsible for the increases in activity and changes in vertical distribution of larvae in response to contaminants observed by Thompson ct al. (1997). His et al. (2000) and Van den Hurk ( 1994). Changes in velocity, activity and vertical distribution may be due to the same response. Increased swimming velocity would result in larvae reaching the surface more rapidly. A larger number of larvae at the water surface is liable to increase the rate of collision between larvae. Larvae that collide tend to perform the fright response, which involves retracting the velum, closing the shell and sinking (La Barbera 1974; Cragg 1980). then resume swimming. By increased swimming activity, a larva will sample water quality over a greater veilical range and, due to vertical differences in current velocity, also horizontal range. The chance of being moved away from the polluting source is thus increased (Dek- 272 Prael et al. velocity (mm s"^) 0.8 0.7 0.6 O.S 0.4 0.3 0.2 0.1 0.0 0.8 0,7 0.6 05 04 03 02 0.1 0.0 0.8 1 0.7 ■ 0.6 05 04 0.3 0.2 0-1 0.0 n no leachate ED untreated wood B CCA-oxide SCCA-salt 2-day larvae 3-day larvae 7-day larvae Figure 2. Vertical velocity (mean + SE) of 2-. 3- and 7-day-old Cras- sostrea gigas larvae in cuvettes containing one of four seawater types: seawater alone; seaviater with leachate from untreated «ood; seavvater Hith leachate from vvood treated with an oxide formulation of tX .\; seawater v>ith leachate from wood treated with a salt formulation of CCA. Leachate obtained after the first day immersion of the blocks. Significantly different means indicated by different letters (Tukey- Kramer. p < 0.(15). concentration leachates. observed in the initial experiments, re- sembles the responses reported by Gruffydd (1976) of veligers of the scallop ChUiinys islaiulica exposed to layers of low salinity seawater. and may represent a generalized response to inimical conditions. The avoidance is achieved by the fright response and is probably initiated by stimulation of a sense organ. Pnteiitial for a Behavioral Response Bioassay The behavioral responses observed in these experiments can be readily measured with simple laboratory equipment and thus have the potential to be used as a bioassay. Behavioral responses are of particular value as bioassays. since they occur at pollutant levels too low to produce clear toxic responses. The results of these experiments highlight the need to include controls for the effect of wood and even for source of water in the assay design. The con- tinued activity of larvae over the day after the initiation of experi- ments, indicates that the cuvettes provided adequate conditions for the larvae and were thus suitable for a bioassay. By using image acquisition and image analysis software, the increased efficiency of data capture would enable a greater level of replication than in this .study, increasing the precision of the assay. This bioassay could be used to evaluate modifications to treat- ment procedures designed to minimize in-service leaching. Among.st the modifications worth testing would be the use of a post-treatment, pre-installation leaching in water intended for mak- ing up further treatment solutions. Our limited data and more extensive measurements reviewed by Cragg et al. (2001 ) indicate that rates of leaching tend to decline rapidly. By using pre- installation leaching, the highest emissions of the elements of CCA would be retained for future treatment rather than released to the environment. Modifications to the post-treatment fixation regime may also improve leach resistance, as has been shown by Lebow (1997). Furthermore, Archer et al. (1994) showed that the formu- lation of CCA used might affect leaching rates, a conclusion supported by the differences in leachate from the salt and oxide formulations used in this study. Minimization of leaching is par- ticularly important where wood is used in aquaculture. such as in buchot culturing of mussels. In view of the concerns raised by Weis and Weis ( 1996) regarding the use of CCA-tieated wood in places where water is almost stagnant, bioassay evaluation of treatment methods may assist in minimizing impacts of this es- sential protection for a useful and renewable construction ma- terial. shenieks et al. 1996). Further information about the nature of the response would be obtained if larvae were kept in a deeper ves.sel, pennitting investigation of whether larvae tended to sink further before resuming swimming. The fact that the .^-day-old larvae swam faster in the leachate of an untreated block than in seawater alone suggests thai the larvae can detect soluble chemicals derived from the wood itself. Possible candidates include soluble sugars, pectins or the stilbene. pino- sylvin, a biocidal compound deposited during heartwood forma- tion in Piniis sylve.stri.s (Eaton & Hale 199.^). The lack of a detectable response in the two-day-old larvae may reflect their developmental status. The ability to close the shell appears some time after initial shell formation (La Baibera 1974; Cragg 1980) and the nerve network that innervates the ciliated cells of the velum (Cragg 1989) may not have become functional. The apparent avoidance behavior of larvae encountering high ACKNOWLEDGMENTS This work was undertaken with the financial support of the Marine Science and Technology program (MAST-III) of the Eu- ropean Union, under contract MAS2-CT94-0100. AP was funded by the European Union Social Fund. Her work was undertaken while studying for an MSc in Shellfish Biology at the School of Ocean Sciences, University of Wales, Bangor. We thank: B. Plun- kett and M. Wells at the School of Pharmacy and Biomedical Sciences. University of Portsmouth, for analyses of the leachates; P. English of Emu Environmental for assistance with conducting fertilizations; Seasalter Hatcheries (Cumbria) and P. Goulletquer of IFREMER for providing lar\ ae; Brian Matthews of the Forest Products Reseaich Centre. Buckinghamshire Chilterns University College for treating the test blocks; A. R. Beaumont and R. A. Eaton for constructive criticism of the manuscript. Behavioral Responses of Bivalve Veligers to Contaminants 273 LITERATURE CITED Albuquerque R. M.. S. M Cragg & J. D. Icely 1996. Leaching from CCA- treated wood submerged in seawater: effects of high loadings, and a comparison between laborator\ and marine conditions. International Research Group on Wood Presersation Document No. lRG/WP/y6- 500:3, 17pp. Anderson. D. G.. J. A. Cornfield & G. R. Williams. 1941. Waterbased fixed preservatives. In: R. Thompson, editor. The chemistry of wood pres- ervation. Cambridge: The Royal Society of Chemistry, pp. 101-116. Anon. 1996. Acute oyster embryo-larval development test guideline. In: Direct toxicity assessment (DTA) methods and guidelines. Environ- ment Agency/Scotland and Northern Ireland Forum for Environmental Research (Sniffer). Archer, K., A. Preston. C. Chittenden & D. Page. 1994. Depletion of wood preservatives after four year' exposure in Mt. Maunganui Harbour. New Zealand: International Research Group on Wood Preservation Doc. No. IRGAVP/94-50036. Brooks. K. M. 1993. CCA treated wood in the aquatic environment: Lit- erature review and assessment of the environmental risks associated with the use of treated wood products in the aquatic en\ironment. Hickson Corporation. 34 pp. Brooks. K. M. 1996. Evaluating the en\ironmental risks associated with the use of chromated copper arsenate-treated wood products in aquatic environments. Esnaiiies 19:296-305. Brown. C. J.. R. M. Albuquerque. S. M. Cragg & R. A. Eaton. 2000. Ef- fects of CCA (copper-chrome-arsenic) preservative treatment on the settlement of calcareous fouling organisms on wood at a range of European sites. BiofouUng 15:151-164. Chang. Y.. S. L. Coon. M. Walch & R. Weiner. 1996. Effects of Hypomo- ihi.s PM- 1 biofilms on the toxicity of copper and zinc to Crassoslrea gigas and Cni.woxlreci virginiai lar\al set. ,/. Shi'lllhli Ri's. 15:589- 595. Chapman. P. M. & E. R. Long. 1983. The use of bioassays as part of a comprehensive approach to marine pollution assessment. Mar. Poll. Bull. 14:81-84. Cragg. S. M. 1980. Swimming behaviour of the larvae Peclcn maximus (L.) (Bivalvia). / Mar. Biol. Assoc. U.K. 60:551-564. Cragg. S. M. 1989. The ciliated rim of the velum in larvae of Pccicii maximus (Bivalvia). / Moll. Snul. 55:497-508. Cragg. S. M. 1995. Impact on non-t;irget organisms of anti-marine wood- borer treatments. In: Marine Sciences and Technologies: 2nd MAST days and Euromar market. Vol. II. Luxembourg: Office for Official Publications of the European Communities, pp. 1227-1236. Cragg. S. M.. C. J. Brown. R. A. Albuquerque & R. A. Eaton. 2001. Rate of emission from CCA-treated wood in the marine environment: mea- surement, modeling and requirements for further research. In: Proc. 5"^ Wood Preservation Symposium: Environment and Wood Preservation. International Research Group on Wood Preservation Doc. No. IRGAVP 01-50166. Dekshenieks. M. M.. E. Hofmann, J. M. Klinck & E. N. Powell. 1996. Modelling the vertical distribution of oyster larvae in response to en- vironmental conditions. Mar. Ecol. Prog. Ser. 136:97-110. Eaton. R. A. & M. D. C. Hale. 1993. Wood: decay, pests and protection. New York: Chapman and Hall. 546 pp. Gruffydd, LI D. 1976. The development of the larva of Chimin s islandica in the plankton and its salinity tolerance in the laboratory (Lamelli- branchia. Pectinidae). Asuirtc 8:60-66. Hayes. C. P. M. T. Curran. & M.J. Hynes. 1994. Preservative leaching from softwoods submerged in Irish coastal waters measured by atomic absorption spectrophotometry. Hiil-forscluiiig 48:463-473. Hidu. H. & H. Haskin. 1978. Swimming speeds of oyster larvae Cras.sos- trea rirginica in different salinities and temperatures. Estuaries 1:252- 255. His, E.. R. Beiras & M. N. L. Seaman. 2000. The assessment of marine pollution - bioassays with bivalve embryos and larvae. Adv. Mar. Biol. 37:3-17. Hoare. K.. A. R. Beaumont & J. Davenport. 1995. Variation among popu- lations in the resistance of Mylilus edulis embryos to copper — adaptation to pollution. Mar. Ecol. Prog. Ser. 120:155-161. La Barbera. M. 1974. Calcification of the first larval shell of Tridacna squamosa (Tridacnidae: Bivalvia). Mar. Biol. 24:233-238. Lebow. S. 1997. Effects of fixation conditions and seawater on leaching of wood treated with marine retention CCA. International Research Group on Wood Preservation Doc. No. IRGAVP/97-50088. Lebow. S. T.. D. O. Foster & P. K. Lebow, 1999. Release of copper, chro- mium and arsenic from treated southern pine exposed in seawater and freshwater. For. Prod. J. 49:80-89. Mann. R. & J. Rainer. 1990. Effect of decreasing oxygen tension on swim- ming rate of Crassostrea virginica (Gmelin. 1791) larvae. J. Shellfish Res. 9:323-327, Mann. R.. B. M. Campos & M. W. Luckenbach. 1991. Swimming rate and responses of larvae of three inactrid bivalves to salinity discontinuities. Mar Ecol. Prog. Ser 68:249-256. Thompson. I. S.. R. Seed. C. A. Richardson, L. Hui & G. Walker. 1997. Effects of low level chlorination on the recruitment, behaviour and shell growth of Mylilus edulis Linnaeus in power station cooling water. Scieinia Marina 61:77-85, Tremblay, M. J. & M. M. Sinclair. 1988. The vertical and horizontal dis- tribution of sea scallop (Placopecten magelkmicus) larvae in the Bay of Fundy in 1984-1985. J. Northwestern Atlanl. Fish. Soc. 8:43-53. Van den Hurk, P. 1994. Effects of natural sediment properties on test results in bioassays with oyster larvae (Crassostrea gigas) on sediment elutriates. J. Aquat. Ecosyst. Health 3:185-191. Weis. J. S. & P. Weis, 1992. Con.struction materials in estuaries: reduction in the epibiotic community on chromated copper arsenate (CCA) treated wood. Mar. Eco. Prog. Ser. 83:45-53. Weis. J. S. & P. Weis. 1993. Trophic transfer of contaminants from or- ganisms living by chromated-copper-arsenate (CCA)-treated wood to their predators. / E.xp. Mar Biol. Ecol. 168:25-34. Weis. J. S. & P. Weis 1996. The effects of using wood treated with chro- mated copper arsenate in shallow-water environments: a review. Estu- aries 19:306-310. Weis. P.. J. S. Weis & E. Lores. 1993. Uptake of metals from chromated- copper-arsenate (CCAl-treated lumber by epibiota. Mar. Poll. Bull. 26:428-430. Wendt. P. H.. R. F. Van Dolah, M. Y. Bobo, T. D. Mathews & M. V. Levisen. 1996. Wood preservative leachates from docks in an esluarine environment. .Arch. Env. Contain. To.xicol. 31:2-1-37.. Joiirihil of Shellfish Rcscaixh. Vol. 20. No. 1. 275-2S1, 2UUI. STABI.E ISOTOPE PROFILES OF SERRIPES GROENLANDICUS SHELLS. II. OCCURRENCE IN ALASKAN COASTAL WATER IN SOUTH ST. LAWRENCE ISLAND, NORTHERN BERING SEA BOO-KEUN KHIM Polar Scii'iHcs Lahoratorx. Korea Ocean Research and Development Institute. P.O. Bo.x 29. Ansan 425-600. Korea ABSTRACT Three bivalve shells (.Sc;r//;t'.v f-menlandicus. Greenland cockle), collected alive in the south of St. Lawrence Island, have been analyzed for the stable oxygen and carbon isotopic composition using a sequential method to trace the variation of hydrographic conditions (i.e., temperature, salinity, and oxygen isotope) of ambient seawater during the life-spans of these organisms. The 6'"0 profiles of three shells exhibit a cyclic pattern interpreted as seasonal variation of Bering Shelf Water, which is the dominant water mass at the collection site. Discernible and distinct light 8"*0 peaks occurred in the summer of 1991, suggestmg unusual hydrographic conditions. These are attributed to the migration of Alaskan Coastal Water into the south of St. Lawrence Island during the summer. The changed oxygen isotopic compositions of seawater between 1990 and 1993 in the study area supports the idea that the flow of Alaskan Coastal Water, intensified by the enhanced freshwater discharges from the Alaskan coast, can be migrated westward into the south of St. Lawrence Island. The 8"C profiles also show apparent cyclicity. reflecting the variation of available carbon isotopes in relation to photosynthesis and oxidation of organic matter. A seasonal bloom of phytoplankton and consequent oxidation of organic matter is the possible cause of the noticeably light 6"C peak in 1992, not the influx of terrestrial carbon or metabolically derived cai'bon. The warm Alaskan Coastal Water, characterized by low oxygen isotopic composition, is clearly recorded in the shell isotope profiles, which provides a record of the hydrographic variation of the ambient seawater. KEY WORDS: Serripes groenhmduus. bivalve, stable isotope, water mass INTRODUCTION Little attention has been paid to the stable isotopic composition of shallow marine biogenic carbonates, mainly because the ex- treme spatial and temporal changes of hydrographic conditions make their interpretation of compositions difficult. In contrast, exceptions are the stable isotope studies on the growth rate and longevity of mollusk shells to obtain the information on population dynamics (e,g., Tanabe 1988. Jones and Quitmyer 1996). Shells of marine bivalves, gastropods, corals, and brachiopods in the near- shore and continental-shelf environment have been the object of a number of stable isotope studies because their skeleton chemistry has the potential to rect)rd the ambient seawater conditions (sum- marized in Rhoads and Lutz 1980). In particular, many investiga- tions have explored the relationship between the stable oxygen and carbon isotopic compositions of bivalve shell carbonates and en- vironmental information using microsampling to highlight high- resolution temporal changes of ambient seawater during the lifespans of living or fossil organisms (Krantz et al. 1987, Krantz 1990. Comu et al. 1993. Weidman et al. 1994, Jones and Allmon 1995. Klein et al. 1996. Beniis and Geary 1996, Bice et al. 1996, Kirby et al. 1998, Khim et al. 2000, Khim 2001 ). The hydrographic conditions in the shallow marine environ- ment of temperate and sub-arctic regimes vary seasonally. In win- ter, the water column is vertically homogeneous due to mixing resulting from the convective cooling and severe wind. On the other hand, in summer, incoming freshwater discharge and heating of the surface water result in the vertical stratification of water masses in these shallow waters. For example, in the northern Ber- ing Sea (Fig. 1), stratification develops particularly during the summer after the sea ice melts (Coachman et al. I97.'>, Muench et al, 1981, Coachman and Shigaev 1992, Schumacher and Kendall 1995). The migration of water mass is of great importance because the transport and dispersal of pollutants is controlled by the coastal current in the shallow marine environment. In addition, the result- ant mixing of water masses plays a potential role in influencing primary productivity by reintroducing the limited nutrients (Walsh et al. 1989). In this paper, three living bivalve (Serripes groenlaiulicus) shells were collected in the northern Bering Sea by dredging in June 1993 during the Helix 171 research cruise. From the patterns of stable oxygen- and carbon-isotope profiles, constructed using the high-resolution sequential sampling, interesting hydrographic events can be identified distinctly. HYDROGRAPHY AND SAMPLING SITE The northern Bering Sea is characterized by one of the largest continental-shelf seas in the world ocean (Fig. la). This sea. which is the connection between the North Pacific and the Arctic Ocean. is deepening slightly toward the Bering Strait. In the northern Bering Sea lie two straits, the Anadyr and Shpanberg straits, on cither side of St. Lawrence Island (Fig. lb). Northward flows through the Bering Strait (Fig. la; Coachman et al. 1975. Walsh et al. 1989) result from sea-level sloping toward the Arctic Ocean (Stigebrandt 1984). Three water masses (Anadyr Water. Bering Shelf Water, and Alaskan Coastal Water) defined by their temperature-salinity pro- files have been observed in the northern Bering shelf (Fig. lb; Coachman et al. 1975). The water masses are also characterized by the unique 8"'0„ (i.e.. S'^'O,, vs. V-SMOW) values (Greb- meier et al. 1 990, Cooper et al, 1 997). Their results showed that the separation of Alaskan Coastal Water from Bering Shelf Water was defined by &"^0„ value of about -2.0 '/it. These water masses are oriented in an east-west direction across the shelf The most saline and coldest Anadyr Water occupies the western part of Anadyr Strait, whereas the least saline and warmest Alaskan Coastal Water occurs in the eastern part of the Shpanberg Strait. Between these lies the Bering Shelf Water of intermediate temperature and salin- ity. The relatively strong hydrographic boundary between Anadyr Water and Bering Shelf Water is maintained during the year, but the corresponding front between Bering Shelf Water and Alaskan 275 276 Khim Figure 1. (a) The schematic flow pattern in the northern Bering Sea (after Schumacher and Kendall 1995). SLI: St Lawrence Island, (b) The bathymetry of northern Bering Shelf and location of sampling site {H\ 171 -059). The three principal water masses (AW: Anadyr Water, BSW: Bering Shelf Water. ACW: Alaskan Coastal Water) are aligned in the east-west direction across the shelf. BS: Bering Strait, AS: Anadyr Strait, SS: Shpanberg Strait. Coastal Water is usually indistincl and gradational and the lateral mixing is quite possible (Coachman and Shigaev 1992). Enhanced freshwater runoff to intensify the Alaskan Coastal Water, as well as mixing process from wind and tidal shear, works the water mass to extend its way across a front into the shelf area. The warmer Alaskan Coastal Water (summertime bottom- water temperature: about 6°C) is characterized by a mixture of freshwater runoff with relatively saline coastal waters (Coachman 1986. Schumacher and Kendall 199.^). The relatively low salinities (generally <3I.8 psu. practical salinity unit) near the Shpanberg Strait are reinforced annually by the addition of Yukon River water. This marked seasonal diffei^ence in freshwater runoff causes the intensification of Alaskan Coastal Water during the warm sea- son. Alaskan Coastal Water also shows the inter-annual variations of salinities which are the same magnitude as those of the seasonal cycle (about 0.5 psu: Coachman and Shigaev 1992). The vertical profiles of tempeiature and salinity measured at the station HX171-059 (61"46.0'N. 17ri6.0'W. Fig. lb) where the bivalve shells were collected are shown in Figure 2. The water column formed a distinct two-layered stratification at the time of shell collection. High-temperature (~6.0°C) and less saline (-,30.7 psu) waters form the upper layer whereas low-temperature (-0.4°C) and more saline (-31.6 psu) waters form the lower layer. A sharp thermocline lies a little deeper than the halocline. both of which maintain the water-column stability. The high temperature and low salinity of the upper layer is probably due to effects of solar radiation, and freshwater runoff and sea-ice melts, although the collection time (late June) was close to the beginning of sum- mer when it is fairly drier. Approximately -2.0 %c (V-SMOW) of bottom water at the depth of 45 m is measured, which seems to be quite low. on the basis of the 8"*0„ definition of Bering Shelf Water (Grebnieier et al. 1990). MATERIALS AND METHODS Three bivalves. Seniles groenlaiicliciis (Greenland cockle), were collected alive near the hydrocasting station H.\17l-059 (45 m deep) south of St. Lawrence Island, northern Bering Sea by dredge in late June. 1993 (Fig. lb). This infaunal and suspended feeder is cosmopolitan in arctic and boreal regions from the sub- tidal zone to about 100 m. The shell of this species is composed entirely of aragonite and grows up to 10 cm as a ten-year-old adult (Andrews 1972). These features and previous research (Khim et al. 200 1 . submitted) suggest that S. groenlandicus would be useful for assessing seasonal and intei'-annual hydrographic variability of the Alaskan Coastal Water from shell isotope profiles. The growth patterns of bivalve shells with their repeated for- mations of annual increments provide information determining age Stable Isotope Profiles And Alaskan Coastal Water 277 Temperature (°C) 80 40 50 5'°0 , (V-SMOW) seawater ' ' = -2.0°/oo(at45m) 30.5 31.0 31.5 32.0 Salinity (psu) Figure 2. ViTtical pr(>file,s of temperature and salinity measured at the station of H\l7l-(l5y. The two-layered stratified waler-colunin is elearly maintained by the strong thermocline and haloeline that lie between Id and 2(1 m. Ihe 5 "'0„ of bottom water (water depth of 45 m) is about -2.0 %c (V-SMOW) (Cooper unpubl.). and growth rates of marine bivalves (Rhoads and Lut/, 1980. Jones and Quilmyer 1996). The bivalve S. groenlandicus produces an- nual growth marks which are correlated to its isotopic record (Khirii et al. 2001). Figure 3 shows the growth rate history of analyzed shells using growth marks. Since the shell specimens were collected alive in 1993, calendar years may be related to the sampling intervals by counting growth bands backward from the shell margin. Consecutive sampling for isotopic determinations of shells is schematically shown in Figure 4. No sampling was done near the umbo (younger ages) because of the thin outer shell layer and near the shell margin (older years) because of the very slow growth rate. Preparation technique for stable isotopic analyses was taken from Krantz et al. (1987). The specimens were soaked in a y/c sodium hypochlorite solution and rinsed in distilled deionized wa- ter to remove the organic matter from the shell exterior. One valve of the mollusk specimen was radically cut along the axis of maxi- mum growth. A serial shell-sampling technique produced a set of individual carbonate powders (approximately 2 mg), which were collected in ontogenetic sequence from the outer shell layer with a 0.3-mm dental drill bur. Care was taken to avoid the penetration of the inner nacreous layer since these boreal bivalves have a thin outer layer. Isotope measurements were carried out at the Univer- sity of Maine. The number of samples obtained from the speci- mens (HX 1 7 1 A. HX 1 7 1 B, and HX 1 7 1 C) are twenty, twenty-three, and twenty, respectively (Fig. 4). Stable oxygen and carbon isotopic ratios were determined us- HX17I-059A HX171-059B HX171-059C 012345678 Growth Year Figure 3. Growth rate of S. groenlandicus estimated by measuring shell height, based on the annual increments. The three shells show their different growth rate in spite of li>ing in adjacent habitats. ing an on-line, automatic carbonate preparation device attached to a VG Isogas PRISM isotope ratio mass spectrometer. The evolved CO, gas was measured after reaction in lOO*?!^ phosphoric acid at 90°C. All isotope values are reported in the standard 8 notation of per mil (%c) unit as the ratio of "*0/"'0 or '''C/'-C in the sample relative to a V-PDB reference (Craig 1957. Coplen 1994). Ana- lytical precision based on NBS-19 carbonate powders yield better than ±0.10<7fc for S'^'O and ±{).{)T/,, for 8"C. RESULTS AND DISCUSSION Stable oxygen- and carbon-isotope profiles of three shell speci- mens are illustrated in Figure 3. The vertical axis is inverted to show the low 8"*0 values toward the top, and the horizontal axis shows the ontogenetic growth of shell from left to right. In general, based on the thermodynamic behaviors of the oxygen isotopes in the skeletal formation (Urey 1947. Epstein et al. 1953. Grossman and Ku 1986). the low S'^O values of the oxygen-isotope profile reflect the high seawater temperature during summer whereas the high values represent the low temperature during winter. Isotope values for sampled years of each shell are summarized in Table 1 . All the oxygen-isotope profiles are clearly cyclic and seasonal, but their patterns are different among the shells in the same year (Fig. 5). It may be due to the different growth rate, as judged from the width of annual bands (Fig. 3). The specimen HXI71A shows a fairly low amplitude of seasonal 8''"*0 variation (Fig. 5a). ranging from 0.45 to \.22%c (Table I). The calculated seasonal tempera- ture variations through simple temperature fractionation of oxygen isotopes (-0.23/°C. O'Neil et al. 1969) are about 2 to 5°C. which corresponds to the temperature of the Bering Shelf Water (Coach- man et al. 1975). The shell HX171B has a small 8"'0 variation (about 0.8 %f) in 1990 similar to those of HX17IA (Fig. 5b). but has a prominent signal of the lowest h'^O value in the summer of 1991. the amplitude of which is up to about 2.2 %c (Table 1). For the same years, the pattern and amplitude of the profile of speci- men HX171C are similar to those of HX171B with an amplitude 278 Khim l>0 u CO (a)HX171-059A(«=20) #1 h- #20 2^ 7K "K Z Z A A 86 87 88 89 90 91 92 93 (b)HX171-059B(/7=2i) #21 #43 7K z^ r 88 89 90 (c)HX171-059C(»=20) 33 4j 1 r 91 #44 92 93 #63 56 6f 7^ — 7^ — z^ — z^ — r 86 87 88 89 90 91 92 93 1^ 10 20 30 40 50 60 70 80 Shell Height (umbo to margin in mm) Figure 4. Sampling sclieinc for nbtaining carbonate ponders Ijv drilling airing the shell growth. The identification of annual bands along with the backward counting from margin provides the matching year for the individual isotope profile. of l.6%c (Fig. 5c). Thus, the most remarkable feature of the oxy- gen-isotope profile.s is the event in the summer of 1991 that marks the lightest &''''0 value obtained. A similar pattern in the oxygen- isotope profiles inteipreted due to changes of hydrographic prop- erties has been reported in mid- Atlantic coastal shells (Krantz et al. 1987). The lack of a distinct peak in the summer of 1991 for specimen HX171A is problematic. However, the most plausible explanation is that this shell lived far from the area where the two specimens (HX171B and HX171C) were obtained because the trawling area for collection was too large. In terms of the seasonality of the 8"*0 profiles, the bivalves can record the seasonal temperature variations during their life-spans through the isotopic fractionation of shell carbonate, under the assumption that the oxygen isotope values of ambient seawater are uniform. A variety of previous investigations substantiated that the h'^O variations are controlled primarily by seasonal changes of seawater temperatures (Cornu et al. 1993, Bemis and Geary 1996. Khim et al. 2000), although the bivalve shell imperfectly records the lowest temperature due to either the growth cessation or slow growth rate during the winter (Krantz et al. 1987, Weidman et al. 1994). The b'^O variation of HX171B in 1991 (about 2.2 7cc) corresponds to a seasonal temperature change of seawater of ap- proximately 9°C, based on temperature fractionation (O'Neil et al., 1969). This seasonal temperature range is larger than expected when compared to the present-day hydrographic conditions occu- pied by the Bering Shelf Water (Coachman et al. 1975. Coachman and .Shigaev 1992) as well as to the seasonal variations expected from HX171A (Fig. i). Thus, at a first approximation, such sea- sonal temperature change in 1991 calculated from the amplitude of oxygen-isotope profile for HX171B is unlikely to occur. The large part of variation may be due to another factor to affect oxygen isotopic composition of shell carbonate when organisms lived. An alternative plausible and potential variable to infiuence the oxygen isotopic composition of shell carbonates is the oxygen isotopic composition of seawater (S'^'O^) itself (Epstein et al. 195,3, Grossman and Ku 1986). In general, the 8"*0,,, values are linearly correlated with the salinity (Craig and Gordon 1965); the high-salinity water mass is characterized by the high S^'O.^ value and vice versa. In the northern Bering Shelf region, the Alaskan Coastal Water has lower 5"*0„, values than the Bering Shelf Water (Grebmeler et al. 1990, Cooper el al. 1997. Cooper unpubl.). Fig- ure 6 illustrates the temporal variation of bottom water 8"*0„ values in the southern part of St. Lawrence Island. In June. 1990. the S'^'O^ variations demonstrate that the water masses are aligned in east-west direction and 8'**0,^, values increase gradually west- ward (Fig. 6a, Grebmeier et al. 1990). Where the shells were collected in June. 1993, the increasing 8"*0„ trend can be simi- larly observed in the south of St. Lawrence Island (Fig. 6b, Cooper unpubl.). However, the absolute 8'**0„ values in 1993 became much lower by as much as approximately 0.5 %c than those in 1990. Such low 8'**0„ values in 1993 are attributed to the migra- tion of Alaskan Coastal Water into this area, resulting in the drop of 8'**0„ value. Therefore, although 8"*0„ data in 1991 are un- available in the present study, the distinct pattern of oxygen- isotope profiles demonstrates that the lightest 8'**0 peak occurred in 1991 is presumably caused to much extent by the low S'^'O.^ value of Alaskan Coastal Water. Krantz et al. (1987) also showed that short-term low-salinity events in the mid-Atlantic coastal area can be identified from the pattern of oxygen-isotope profiles. The carbon-isotope profiles of live shells also show periodic patterns similar to oxygen-isotope profiles (Fig. 5). In general, the high 8''C values are observed dunng mostly the entire season, whereas the low 8"C values occur along with the low 8'*'0 values that refiect the warm season. The 8"C profile of specimen Stable Isotope Profiles And Alaskan Coastal Water 279 o - "Jo d 30 40 50 60 7 -4 3 H '90 '91 ■92 '93 f 2 - A 1 - / -^ 1 . J15''C 0 - *w •-^ W^ 1 - o / Tr/\ 2 - (b) J'hd =%ocP y^W'^o 1 1 40 50 60 70 30 40 50 60 Shell Height (mm) Figure 5. Stable oxygen- and carbon-isotope profiles of living bivalve .S. armiiltinttiais. Ia( HX171A. (b) HX171B. (cl H\171C. Note the low S'"()peaks in IWl of H\171B and HX171C and the hm 5'*C values that occurred in \W2 in HX171B and HX171C. HX171A is cyclical with almost consistent amplitudes over the sampled interval (Fig. 5a). The mean h'^C and range of variation are about -1.18 7ic and 1.70 7cc. respectively (Table 1 ). The mean 8''C values of HX171B and HX171C are -0.33 and -1.24 ';',f. respectively, the latter of which is close to that of HX171A. The enriched 8''C values of HXI71B may be mainly due to the dif- ferent 8"C of seawater dissolved inorganic carbon (DlC) as a background level, although these shells were dredged in a same locality. The 8"C patterns of HX171B and HX171C are similar in spite of different growth rates, but clearly distinguishable from that of HX171A (Fig. 5). In 1992, the light 8"C peak is observed distinctly in both HX 1 7 1 B and HX 1 7 1 C. as similar event of oxy- gen-isotope profile in 1991. The approximate amplitude is up to 2.8 "cc (Table I). In addition, the shapes of oxygen- and carbon- isotope profiles of these two shells are alike in both 1992 and 1993. Several probable variables affecting the carbon isotopic com- position of shell carbonates are as follows: 8"C of the DIC res- ervoir associated with phytoplankton productivity and reoxidation of organic matter, effects of temperature on '-'C fractionation, and seasonal physiological effects related to metabolic activity (Ru- binson and Clayton 1969. Emrich et al. 1970. Tanaka et al. 1986. Rosenberg and Hughes 1991. Wefer and Berger 1991, Romanek et al. 1992, Klein et al. 1996). A primary factor is the available "C atoms in ambient seawater DIC (Mook and Vogel 1968), but other variables can be important. In particular, seasonal variations of the 8"C values of the shell isotope profiles documented in the previ- ous studies have been explained as a result of modification in the level of 8'-'C of seawater DIC (Krantz et al. 1987). Due to the phytoplankton role in fractionating carbon isotopes during photo- synthesis producing organic matter, the heavy carbon atoms are easily incorporated into the shell carbonate. On the other hand, isotopic fractionation through the decomposition of organic matter causes the light carbon isotopes to be assimilated in the shell carbonate (Kroopnick 1980). The low 8'''C values of HX171A in summer may reflect sea- sonal phytoplankton productivity and decomposition of organic matter at the bottom. In the northern Bering Sea, the seasonal bloom of phytoplankton occurs during the warm months when the sea ice is retreating and the influx of nutrient-rich water is increas- ing (Hansen et al. 1993, Springer and McRoy 1993). In the Alas- kan Coastal Water, the blooming duration may not be apparently long enough to produce a distinguishing signal in the '"C atoms (Walsh et al. 1989, Coachman and Shigaev 1992). Alternatively, there is a possibility that the light 8"C peaks may reflect the inclusion of metabolic light carbon formed by growing much faster during warm, favorable conditions (Rosenberg and Hughes 1991, Wefer and Berger 1991). In spite of lacking evidence to demon- TABLE 1. Summary of minimum, maximum and variation of stable oxygen and carbon isotope judged by the individual year over the sampled intervals. Year (,'"0 8"C Specimen s'-o^. 8"'0„.. Ab'XQ " ^min 8"C^,, A6"C HX17IA 1989 1.47 2.64 1.22 -2.27 -11 .-s s 1 69 IWO 1.84 2.29 0.45 -2.46 -0.81 1.65 1991 1.35 2.11 0.76 -2.40 -1.28 1.12 HXniB 1990 1.37 2.18 0.81 -1.43 0.13 1.56 1991 -0.04 2.16 2.20 -0.49 0.38 0.88 1992 0.66 I.9S 1.32 -2.71 0.06 2.77 HX 1 7 1 C 1990 1.62 2.24 0.62 -1.70 -0.59 1.11 1991 0.01 1.66 1.65 -1.80 -0.82 0.98 1992 0.77 1.22 0.45 -3.09 -086 2.23 Note: itallL' numerals are incomplete variation. 280 Khim N-j>4 Russia^3~^ (a) June, 1990 64°N _ Gulf of Anadvr -0.6 . Si Lawrence ^y — \ Island - -0.8 V_x ^ 62°N - -0.8 -1.0 60°N -1.5 St. Matthew Island 1 1 1 -2.0 -2.5 Bering Sea 1 1 1 Nunivak Island^r-^ ^^^'""^ (b)June, 1993 64°N Gulf of n^ ^' Lawrence Anadyr /^->-_/'~N Island 62°N -2.0 -1.5 60"N St, Matthew — Island Bering Sea 1 1 I Nunivak < 1 74 W 1 72' W 1 70 W I68'W 1 74 W 1 72 W 1 70' W 168"W Figure 6. The spatial distribution of S'^O,, values in the south of .St. Lawrence Island, la) .June IWM Ifroni (Jrebnieier et al. 199(0. (b) .June 1993 (from Cooper unpubl.). The absolute S'^O^^values in 1993 are greater by as niuth as approximately 0.5 'c, compared to those in 1990. Such low S'^O,, values in 1993 niav be due to the migration of .Maskan Coastal Water into this area, resulting in the drop of S'^O,, values. strate that the tenestrial carbon species transported by Alaskan Coastal Water (Naidu et al. 1993). the light terrestrial carbon should be considered to interpret. Terrestrial carbon from river discharge is prominently supplied during the warm season by the intensified Alaskan Coastal Water. If the low 6"*0 peaks observed during the summer of 1991 result from the influx of Alaskan Coastal Water, then the water mass maintains plenty of "'O atoms along with low salinity caused by increasing river runoff. How- ever, the 8"C profile in that year does not show the coiTesponding low peaks. Thus, the addition of light teirestrial carbon during the wanii season is unlikely to cause the low peak. The distinct low 5"C peak of 1992 is informative because the S'^O profile has no equivalent peak in the same year (Fig. 3). As explained above, such light peak is not attributable to light carbon of tenestrial origin transported by riverine discharge. The &''C minimum in 1992 can be identified in both HX171B and HX171C. but not in HX171A. It is difficult to judge which pi'ocess causes this discernible low S'^'C peak. The plausible mechanism resulting in this S'V peak is the abrupt change of 5"C of seawater DIC by the oxidation of organic matter genei'ated from phyloplankton blooms. A stratified water-column, a short duration of primary production on the surface layer, and a fairly shallow depth are all conditions likely to change the 8"C of bottom water. Thus, the light carbon is assimilated easily in the formation of shell carbon- ate. ACKNOWLEDGMENTS This study was done during a stay at the University of Dela- ware. I thank Dr. Lee W. Cooper (Oak Ridge National Laboratory) for supplying the live shell specimens and unpublished seawater oxygen isotope data and Dr. David E. Krantz (now US Geological Survey) for discussing the isotope results. Dr. Doug Introne (Uni- versity of Maine) is appreciated for his analyses of the stable isotopic composition of shell carbonate powders. I would also like to thank Dr. Sandra E. Shumway (editor). Dr. Louise Purton- Hilderbrand (Trinity College), and an anonymous reviewer for their critical and constructive comments and suggestions. This work was carried out through the support of the National Science Foundation (DPP-93()()684 and DPP-9423663) and Ministry of Maritime Affairs and Fisheries (PN00062). LITERATURE CITED Andrews, J. T. 1972. Recent and fossil growth rates of marine bivalves, Canadian. Arctic, and Late Quaternary Arctic marine environments. Palaeuj^eogr. Palaeocliniatol. Palaeoecol. 11:157-176. Bemis. B. E. & D. H. Geary. 1996. The usefulness of bivalve stable isotope profiles as environmental indicators: Data from the Eastern Pacific Ocean and Southern Caribbean Sea. Palaios 1 1 :328-339. Bice, K. L.. M. A. Anhur & L. Marincovich. Jr. 1996. Late Paleocene Arctic Ocean shallow-marine temperatures from mollusc stable iso- topes. Paleoceanofiiaphy II: 241-250. Coachman. L. K. 19S6. Circulation, water masses, and fluxes cm the south- eastern Bering Sea shelf. Cont. Shelf Res. 5:23-108. Coachman. L. K. & V. V. Shigaev. 1992. Northern Bering-Chukchi Sea ecosystem: the physical basis. In: P.A. Nagel. editor. Results of the Third Joint US-USSR Bering and Chukchi Seas Expedition (BERPAC) summer 1988. Washington. DC: US Fish and Wildlife Service, pp. 17-27. Coachman, L. K., K. Aagaard & R. B. Tripp. 1975. Bering Strait: the regional oceanography. Seattle. WA: University of Washington Press. 172 pp. Cooper, L. W., T. E. Whitledge, J. M. Grebmeier & T. Weingartner. 1997. The nutrient, salinity, and stable oxygen isotope composition of Bering and Chukchi Seas waters in and near the Bering Strait. J. Geophys. Res. 102:12563-12573. Coplen. T. B. 1994. Reporting of stable hydrogen, carbon, and oxygen isotopic abundances. Pure Api^lied Cliein. 66:273-276. Cornu. S„ J. Patzold. E. Bard. J. Meco & J. Cuerda-Barcelo. 1993. Pale- otemperalure of the last inlerglacial period based on S'^O of Slrombus Stable Isotopk Profiles And Alaskan Coastal Water 281 buhonius from the western Mediterranean. Piilm'(ii;e(if;r. Piildidcliiiui- rol. Pahieoecol. 103:1-20. Craig. H. 1957. Isotopic standards for carbon and oxygen and correction factors for mass-spectronietric analysis of carbon dio.xide. Geochim. Cosmochim. Acta 12:133-149. Craig. H. & L. Gordon. 1965. Deuterium and o\ygen-18 variations in the ocean and in the marine atmosphere. In: E. Tongiorgi. editor. Stable Isotopes in Oceanographic Studies and Paleotemperatures. Pisa. Italy: Consiglio Nazionale Delle Ricerche. pp. 9-130. Enirich, K.. D. H. Ehhalt & J. C. Vogel. 1970. Carbon isotope fractionation during the precipitation of calcium carbonate. Earlli Planet. Sci. Letr. 8:363-371. Epstein. S.. R. Buchsbaum. H. A. Lowenstam & H. C. Urcy. 1953. Revised carbonate-water isotopic temperature .scale. Geiil. Sue. Am. Bull. 64: 1315-1326. Grebmeier. J. M.. L. W. Cooper & M. J. DeNiro. 1990. Oxygen isotopic composition of bottom seawater and tunicate cellulose used as indica- tors of water masses in the northern Bermg and Chukchi Seas, Limmtl. Ocecmogr. 35:1182-1195. Grossman. E. L. & T. Ku. 1986. Oxygen and carbon isotope fractionation in biogenic aragonite: Temperature effect. Chem. Geol. 19:91-132. Hansen. D. A.. T. E. Whitledge & J. J. Goering. 1993. Patterns of nitrate utilization and new production over the Bering-Chukchi shelf. Com. Shelf Res. 13:601-627. Jones. D. S. & W. D. Allmon. 1995. Records of upvvelling. seasonality and growth in stable-isotope profiles of Pliocene mollusk shells from Florida. Lethaiu 28:61-74. Jones. D. S. & I. R. Quitmyer. 1996. Making time with bivalve shells: oxygen isotopes and season of annual increment formation. Pulaios ll:.340-346. Khini. B. K., D. E. Krantz & J. Brigham-Grette. 2001. Stable isotope profiles of last interglacial (Pelukian Transgression) bivalves and pa- leoclimate implications in the Bering Strait region. Quiit. Sci. Rev. 20 Khim. B. K.. D. E. Krantz & L. W. Cooper. Stable isotope profiles of Senipes groenkmdicus shells. I. Seasonal and interannual variations of Alaskan Coastal Water in the Bering and Chukchi Sea. Cont. Shelf Res. Submitted. Khim. B. K.. K. S. Woo & J. G. Je. 2000. Stable isotope profiles of bivalve shells: Seasonal temperature variations, latitudinal temperature gradi- ents and biological carbon cycling along the east coast of Korea. Cont. Shelf Res. 20:843-861. Kirby. M. X.. T. M. Soniat & H. J. Spero. 1998. Stable isotope sclero- chronology of Pleistocene and Recent oyster shells {Crassostren vir- gmica). Palaios 13:560-569. Klein. R. T., K. C. Lohmann & C. W. Thayer. 1996. Sr/Ca and "C/'"C ratios in skeletal calcite of Mytilus tfossulus: Covariation wilh meta- bolic rate, salinity, and carbon isotopic composition of seawater. Geochim. Cosmochim. Acta 60:4207—4221. Krantz. D. E. 1990. Mollusk-isolope record of Plio-Pleistocene marine paleoclimate. U.S. Middle Atlantic Coastal Plain. Palaios 5:317-335. Krantz. D. E.. D. F. Williams & D. S. Jones. 1987. Ecological and pale- oenvironmental information using stable isotope profiles from living and fossil molluscs. Palaeogeogr. Palaeoclimatol. Palaeoecol. 58: 249-266. Kroopnick. P. 1980. The distribution of '^C in the Atlantic Ocean. Earth Planet. Sci. Lett. 49:469-484. Mook. W. G. & J. C. Vogel. 1968. Isotopic equilibrium between shells and their environment. Science 159:874-875. Muench. R. D.. R. B. Tripp & J. D. Cline. 1981. Circulation and hydrog- raphy of Norton Sound. In: The Eastern Bering Sea Shelf: Oceanog- raphy and Resources. Office of Marine Pollution Assessment. NOAA. pp. 77-93. Naidu. A. S.. R. S. Scalan. H. M. Feder. J. J. Goering, M. J. Hameedi. P. L. Parker, E. W. Behrens, M. E. Caughey & S. C. Jewett. 1993. Stable organic carbon isotopes in sediments of the north Bering-south Chuk- chi seas, Alaskan-Soviet Arctic Shelf Cont. Shelf Res. 13:669-691. O'Neil, J. R., R. N. Clayton & T. K. Mayeda. 1969. Oxygen isotope fractionation in di\alent metal carbonate. J. Chem. Phys. 51:5547- 5558. Rhoads, D. C. & R. A. Lutz. 1980. Skeletal growth of aquatic organisms: biological records of environmental change. New York: Plenum Press. 750 pp. Romanek, C. S., E. L. Grossman & J. W. Mores. 1992. Carbon isotope fractionation in synthetic aragonite and calcite: effects of temperature and precipitation rate. Geochim. Cosmochim. Acta 56:419—130. Rosenberg, G. D. & W. W. Hughes. 1991. A metabolic model for the determination of shell composition in the bivalve mollusc. Myiihis edulis. Lethaia 24:83-96. Rubinson. M. & R. N. Clayton. 1969. Carbon- 13 fractionation between aragonite and calcite. Geochim. Cosmochim. Acta 33:997-1002. Schumacher. J. D. & A. W. Kendall. Jr. 1995, An example of fisheries oceanography: walleye pollock in Alaskan waters. AGU U.S. National Report to lUGG (1991-1994). Rev. Geophys. (suppl.):l 153-1 163. Springer. A. M. & C. P. McRoy. 1993. The paradox of pelagic food webs in the northern Bering Sea. 111. Pattern of primary production. Cont. Shelf Res. 13:575-599. Stigebrandt. A. 1984. The North Pacific: a global-scale estuary. / Phys. Oceanogr. 14:464-470. Tanabe. K. 1988. Age and growth rale determinations of an intertidal bivalve. Phacosoma japonicum. using internal shell increments. Le- thaiu 21:231-241. Tanaka, N., M. C. Monaghan & D. M. Rye. 1986. Contribution of meta- bolic carbon to mollusc and barnacle shell carbonate. Nature 320:520- 523. Urey, H. C, 1947. The thermodynamic properties of isotopic substances. / Chem. Soc. 562-581. Walsh. J. J., C. P. McRoy. L. K. Coachman, J. J. Goering, J. J. Nihoul. T. E. Whitledge, T. H. Blackburn, P. L. Parker, C. D. Wirick, P. G. Shuert. J. M. Grebmeier. A. M. Springer. R. D. Tripp, D. A, Hansell. S. Djenidi, E. Deleersnijder, K. Hendriksen, B. A. Lund, P. Andersen. F. E. Miller-Karger & K. Dean. 1989. Carbon and nitrogen cycling within the Bering/Chukchi Seas: source regions for organic matter affecting AOU demands of the Arctic Ocean. Prog. Oceanogr. 22: 277-361. Wefer. G. & W. H. Berger. 1991. Isotope paleontology: Growth and com- position of extant calcareous species. Mar. Geol. 100: 207-248. Weidman. C. R.. G. A. Jones & K. C. Lohmann. 1994. The long-lived mollusc Artica islandica: A new paleoceanographic tool for the recon- struction of bottom temperatures for the continental shelves of the nonhern North Atlantic Ocean. ./. Geophys. Res. 99: 18305-18314. Juiiriuil oj .Slicllji^h Resfiirch. Vol. 20. No. 7, 28.V287, 2001. ASPECTS OF THE GONADAL CYCLE IN THE ANTARCTIC BIVALVE LATERNULA ELLIPTIC A G. BIGATTI.' P. E. PENCHASZADEH,' * AND G. MERCURI^ ' Faciilkul de Ciencias Exactas v Natundes. UBA y Miiseo de Ciencias Naturales-CONICET. Av. Angel GaUardo 470. Buenos Aires, Argentina; 'Instituto Antdrtico Argentina, Cerrito 1248, WW Buenos Aires, Argentina ABSTRACT Vitellogenesis and oocyte growth in LalcrnuUi citipuca. a conimoii hermaphrodite bivalve living in the soft muddy bottoms of the Antarctic continent, are reported. Formation of the gelatinous layer surrounding the oocyte, gonadic develop- ment, accumulation of mature oocytes, and spawning events were studied through histological evidence. Gonads were observed to reach maturity at a size of 49 mm. Vitellogenesis has been found to last seven months, and storage of oocytes before spawning was observed. It has been observed that once sexually mature, the animals remain with sperm cells and oocytes available during the entire year, suggesting that individuals would be prepared to spawn at any moment, probably depending on environmental conditions. KEY WORDS: Latenuila titiptiia. Antarctic clams, vitellogenesis, oocyte growth, first maturation INTRODUCTION Latcnmia cUiptica (King & Broderip 18.^2) is possibly the most cdiiinion bivalve in the Antarctic continent wherever soft muddy bottoms are present (Powell 1965). It buirows deeply into the substrate (Stout & Shabica 1970); densities of 50 individuals per nr with a fresh weight biomass of 2-3 kg/m" have been reported (Hardy 19721. Recently several authors have studied different aspects of the reproduction of L. elliptica (Pearse et al. 1986; Pearse et ai. 1987; Urban & Mercuri 1998; Ansell & Harvey 1997). Several questions remain unanswered, however. This article deals with the vitello- genesis period and oocyte growth within the ovarian portion of the hermaphrodite gonad, the size and age of gonadic maturation, the formation of the gelatinous layer suiTounding the oocyte, the ac- cumulation of mature oocytes, and spawning events, studied from histological evidence. MATERIAL AND METHODS The material was obtained by one of the authors (G.M.) during a joint German-Argentine project (Kloser & Arntz 1995; Urban & Mercuri 1998), at the location of Potter Cove, King George Island, South Shetland Islands. A monthly sampling was conducted at a fixed station at 10 m depth (see Urban & Mercuri 1998, for locality details). Histologic slides were prepared for one complete year of sampling (1993-1994). Gonads were fixed in Bouin's .solution, preserved in 70% alcohol and treated with common inclusion pro- cedures. Sections of 5 microns were stained with Harris" hema- toxylin and eosin (Merck). The ovaries of 30 individuals of stan- dard adult bivalves (of about 65-80 mm in shell length) were *Corresponding author. E-mail addresses: penchas@bg.fcen.uba.ar (P.E.P.); gblgattl(a'bg.fcen.uba.ar (G.B.j studied for each monthly sample. A total number of 1113 oocytes (73-1 17 per monthly sample) with distinct nucleolus were mea- sured, comprising the whole visual field. In a later sample, in February 2000, gonads of 22 individuals of about 27-73 mm were studied with the same procedures in order to determine the size of gonadal development and matura- tion. RESULTS AND DISCUSSION Under the stereoscopic microscope it is possible to observe, in the ovarian portion of mature clams, the mature oocytes and empty spaces related to partial spawning (Figs. 1-3). A gelatinous layer (Figs. 4-9) envelops the mature oocyte. This outer layer becomes visible (3.1 SD 1.3 microns thick) dur- ing vitellogenesis when the oocytes are 90-100 microns in diameter. Maximum thickness (13.6 SD 5.0 microns) of the ge- latinous layer is reached in oocytes of 220 microns in complete diameter (these oocytes have a mean cytoplasmic diameter of 193 microns). According to the histograms of oocyte diameter frequencies, this process of wall formation takes 3-4 months, starting in March-April and ending in September (Figs. 10 and II). Ansell and Harvey (1997) stated that after fertilization this gelatinous envelope condensed to form a strong, sticky, elastic capsule in which further development took place. We consider this gelatinous layer to be a vitelline membrane, formed by the oocyte itself (Huebner& Anderson 1976). The largest oocyte mean diameter we measured, including the gelatinous layer, was 220.7 microns in February 1994 (Fig. 10), but we consider the oocyte to be mature at 171.5 microns external diameter, and the mean mature oocyte is 195.2 microns (including gelatinous layer). 283 284 BiGATTI ET AL. Figures 1-3. Ovarian portion of gonads ol iMliriiiila elliplica (>6(l mm shell lenglhl \ie«ed tlirough stereoscopic microscope, (ll Ripe ovary; (2) Partiallv spent o\ary; (3) Spawned ovary with remnant oocytes. Scale Bar = I2U0 microns. period of vitellogenesis. The period of fa\ orable liglit conditions to phytoplankton growth in Potter Cove appeared to be very limited, approximately 15 months during summer (Schloss et al. 1997). Apart from that, resuspension of benthic material and possible input of terrigenous material constitute the main carbon source in summer time, while resuspended material or secondary bacterial production would account for it during the rest of the year (Schloss et al. 1997). We consider that, as resorption of unspawned oocytes was never observed in any case, there is currently a storage of large mature oocytes in the female follicles, ready to spawn during the entire year. Protected lecitotrophic embryos, nourished by a con- siderable amount of yolk reserves which enable them to avoid a free swimming larvae stage in the first phases of the development, is the reproductive mode for L. clliptica (Pearse et al. 1986; Bosch & Pearse 1988). Urban and Mercuri ( 1 998 ) found that ripe ovaries seem to dominate, with values between eO-SC/r throughout the whole year, and only during the warmest two months (February and March) spent ovaries dominated, with ripe ovaries being reduced to about 25%. The authors suggested that it is most likely that the oocyte development cycle last longer than one year. Our results conclude that the oocytes complete their growth in less than seven months, and are stored until spawn. Once sexually mature, the animals remain with sperm cells and oocytes available during the whole year, suggesting that individuals would be prepared to spawn at any moment, probably depending on cn\ ironmental con- ditions. In a later sample performed during February 2000 in the same study area, microscopic differentiation of both testicles and ovar- ian Follicles was observed at a size of 27 mm. Taking into account growth rates estimated by Brey and Mackense ( 1997) and by Ur- ban and Mercuri (1998). this size corresponds to individuals <2 years old. Maturity in the male portion of the gonad is evidenced by the existence of mature sperm cells at a size of 30 mm shell length (Fig. 6). At a shell length of .'^2 mm. which corresponds to individuals -3 years old. it is possible to observe some oocytes covered by the characteristic gelatinous layer (up to 140 microns in external di- ameter). Ovaric maturity might then be reached when the indi- viduals attain a size of 49 mm shell length, which corresponds to an age of approximately 4 years. It is at this stage that the gelati- nous layer containing oocytes of 170 microns external diameter can be observed. We conclude then that L. elliptica is a simultaneous hermaph- rodite, completely mature at the age of 4 years. Information is still needed about its first stages of development, in order to assess the possibility of early development of any of the gonadic portions relative to the others. ACKNOWLEDGMENTS The modal diameter peaks for bigger oocytes show no variation from January to July (171.5 microns). But modal diameters of the small oocytes show a remarkable constant tendency to grow from January (55.5 microns) to June (142.5 microns); this is the main The authors thank Mariana Lozada for her English correc- tions, and Alfredo Rodriguez Galtero for assistance with the graphics. This research was partially supported by a grant from Fundaciiin ,\ntorchas and Agencia Promocion Cientifica PICT-98- 04321, Argentina. Gonadal Cycle in Laternula elliptic a 285 Figures 4-9. odc'vtt's sh()« another with Scale Bar A Light micrographs of Laleriiii la elliptica gonad. (4) P'cniaie follicles containing both mature and nc" groHing oocytes: (5) Most the characteristic gelatinous layer: (6) Detail of a male follicle full of sperm cells: |7) An o\arian follicle partially spawned and growing oocytes: (8| Ripe o\ar\ with packed mature oocytes. (9) Growing female follicles with both mature and immature oocytes. & C = 500 microns. Scale Bar B = 100 microns. 286 BlGATTI ET AL. ^ 50 1:. 40 o 30 O 20 S" 10 January N=108 iLlUi-^ jIi 5:, 40 o 30 c O 20 S" 10 July II. ill I ■ 1 — 50 ^ ^40 J >> 0 30 . 1 20 01 10 .jH February N=101 ll li 40 August 7^30 N=117 c « 20 ^ 10 "" 0 ...ll II. 50 40 > 30 -^ 50 ^ 40 >. u c O 20 S" 10 30 March N=108 lL_iL April N=90 -~ 50 ^ 40 O 30 c « 20 g" 10 September N=105 ill. ll ~1 50 40 30 20 10 0 November N=100 ..-_i_L- ^- mA^mAJL iL C, 40 0 30 c 01 70 J May N-106 ff 10 "^ 0 I.ILIJIJ..I ^ 50 ^40 o 30 « 20 g" 10 •S: 0 Decemtwr N=114 Ml < 33 83 55.57 84.58 113.57 142 57 171 57 200 57 oocyte diameter (microns) < 33.83 55 57 84 58 113 57 14257 17157 200 57 oocyte diameter (microns) Figure 10. Frequency (%) of oocyte size (microns). N, number of measured oocytes. 200 150 100 SO 0 6 month 10 12 Figure 11. Modal diameter peaks of oocyte size for each sampled month in an annual cycle. Gonadal Cycle in Laternula elupt/ca 287 LITERATURE CITED AnseM. A. D. & R. Harvey. 1997. Protected larval development in the Antarctic bivalve Lalemithi eliipuca (King & Broderip) ( Anomalodes- niata: Lalernulidae). Journal of MoIIkscwi Snulic!, 63:2SS-2S6. Bosch. \. & J. S. Pearse. 1988. Seasonal pelagic development and juvenile recruitment of the bivalve Latemuhi ellipiica in McMurdo sound. Am- arctica American Zoologist 28:89A. Brey. T. & A. Mackense. 1997. Stable isotopes prove shell growth hands in the Antarctic bivalve Lattriiiila cllipiicci to he tbrmed annually. Polar Biology 17:465-468. Hardy, P. 1972. Bioma.ss estimates from shallow-water infaunal commu- nities Lit Signy Island, South Orkney Island. British Antarctic Survey BulUlin .M:43-l()6 Huebner, E. & E. Anderson. 1976. Comparative spiralian oogenesis. Struc- tural aspects: an overview. American Zoologist 16:315-343. Kloser, H. & W. E. Arntz. 1995. RASCALS (Research on Antarctic Shal- low Coastal and Littoral Systems). Polarforsclumg 64( 1 ):27^l. Pearse. .L S.. L Bosch. J. B. Mc Clintock. B. Marinovic & R, Britton. 1986. Cc>ntrasting modes of reproduction b\ common shallovv-water Antarc- tic invertebrates. Antarctic Journal of Zoology of the United Statc\ 19(51:138-139. Pearse. J. S.. I. Bosch, J. B. Mc Clmtock, B. Marinovic & R. Britton. 1987. Contrasting tempos of reproduction by shallow-waters animals in Mc- Murdo Sound. .Antarctica. Antarctii Journal of the United States 21(5):182-184. Powell. A.W.B. 1965. Mollu.sca of Antarctic and Sub Antarctic seas. pp. 333-380. In: P. Van Oye & J. Van Meighem (eds.). Biogeography and ecology in Antarctica. The Hague: Junk. Schloss. L. H. Kloser, G. Ferreyra, A. Curtosi, G. Mercuri & E. Pinola. 1997. Antarctic Communities: species, structure and survival. In: Battaglia. Valencia. & Walton, editors. Cambridge, UK: Cambridge University Press, pp. 135-141. Stout, W. E. & V. C. Shabica. 1970. Marine ecological studies at Palmer Station and vicinity. Antarctic Journal of the United Slates 5(4): 134- 135. Urban, H. J. & G. Mercuri. 1998. Population dynamics of the bivalve Laternula elliptica from Potter Cove, King George Island. South Sh- etland Islands. Antarctic Science 10(2): 153-160. Journal oj Slwllfisli Rfsfarch. Vol. 20, No. 1, 289-295. 2001. FIRST DESCRIPTION AND SURVEY OF THE EGG MASSES OF LOUGO GAHI (D'ORBIGNY, 1835) AND LOUGO SANPAVLENSIS (BRAKONIECKI, 1984) FROM COASTAL WATERS OF PATAGONIA PEDRO J. BARON Ccnrm Nacioiud Patagonico. Conscjo Ncuioital dv Invcstigacioiies Cientificas y Teciiicas. Boulevard Brown s/n. Puerto Madryn (9120). Cluihul. Argentina ABSTRACT A survey was conducled on ihc egg masses of Loligo gain (D'Orhigny. IX.^.')). a target ot" an important squid fishery in the Southwest Atlantic, and L. sanpaulensis (Brakoniecki. 1984). a species exploited as a by-catch species in artisanal fisheries from Brazil to Argentina. The egg masses are described, and several spawning areas are identified. L f>ahi egg masses were composed of a variable number of capsules (6-345) enclosing 40-150 eggs/embryos. L sanpaulensis egg masses consisted of numerous capsules (6-465) containing approximately 240-320 eggs/embryos. The size range of the eggs was 2. 1-2.3 mm for L gain and 1 .2-1 .3 mm for L sanpaulensis from Nuevo Gulf, showing good correspondence with size of the mature oocytes carried in the oviducts by the females of each species at the same location. L galii egg masses were located in San Mati'as. the San Jose and Nuevo gulfs. Cape Tres Puntas, and Beagle Channel, usually attached to hard substrates (e.g.. kelp) or objects laying on the seatloor (e.g.. stones, shells, ropes, or even fishing lines) at depths from 1-15 m. L. sanpaulensis egg masses were found in Nuevo Gulf on sandy and muddy bottoms at depths of 5-15 m. Throughout a survey conducted from March 1996 to February 2000 in Nuevo Gulf, in an area within the distribution range of both species, L. gahi egg masses were found more frequently; L. guhi egg masses were observed every month of the year but February, whereas L. sanpaulensis egg masses were detected only from February to May. KEY WORDS: Loligo gahi, Lnligo sanpaulensis. egg masses, eggs, embryos, spawning grounds. Argentina. Patagonia INTRODUCTION Two loligiiiid squid species inhabit the Atlantic coast of Pat- agonia: Loligo gahi (D'Orbigny, 1835) and L .sanpaulensis (Bra- koniecki. 1984). L gain is a species distributed in the southeastern Pacific from Peru to Tierra del Fuego (Cardoso et al. 1998) and in the southwestern Atlantic from Tierra del Fuego to 36°S (Castel- lanos & Cazzaniga 1979. Vigliano 1985). L sanpaulensis is a species associated with coastal waters, distributed from San Jorge Gulf (Castellanos 1967) to Brazil (20°S) (Roper et al. 1984). Both species are present in the San Mati'as Gulf (Gonzalez 1999) and in the Nuevo and San Jose gulfs (Re et al.. unpublished manuscript). The fishery of L. gahi in the southwestern Atlantic is the second most important loliginid fishery worldwide, reaching average an- nual captures over 56,000 t during 1988-1997 (FAO 1999). L. sanpaulensis constitutes a by-catch species for local fisheries along the Atlantic coast of South America from 23°S (Costa & Femandes 1993) to 43°S (Castellanos & Cazzaniga 1979). The spawning areas of loliginid squids have been studied in some commercially valuable species (Vecchione 1988, Augustyn 1990, Baddyr 1991, Porteiro & Martins 1992, Saner et al. 1992. Sauer et al. 1993). The data on the spawning areas of loliginids in the southwestern Atlantic are scarce (Andriguelto & Haitiiovici 1996. Re et al., unpublished manuscript, Arkhipkin et al.. 2000). Moreover, the location of these areas has been a matter of specu- lation until recently (Chesheva 1990, Hatfield et al. 1990, And- riguetto & Hainiovici 1996). The egg masses of Loligo are typical among those of the cepha- lopods (Arnold & Williams Arnold 1977, Sweeney et al. 1992); however, there are aspects (e.g., the number of capsules per egg mass, the number of eggs per capsule, the egg sizes, and the type of substratum on which the egg masses are attached) that are important to characterize the egg masses to the taxonomic level of species and that have not been reported in L. sanpaulensis and have been studied only recently in L. gahi (Guerra et al. 2001, Arkhipkin et al. 2000). For both species, the existence of one or more laying periods at different locations is still a subject of debate (Hatfield et al. 1990, Hatfield 1991, Costa & Femandes 1993, Andriguetto & Hainiovici 1996). The aims of the present study were ( I ) to find and identify the egg masses of L. gahi and L sanpaulensis. (2) to locate spawning areas of both species along the Atlantic coast of Patagonia. (3) to study the seasonality of both species" spawning in a given area, and (4) to relate seasonality to the local temperature regime to identify critical values of tempera- ture that could restrict or favor spawning. MATERIALS AND METHODS A continuous survey on the egg masses deposited by female loliginid squids in Nuevo Gulf. Argentine Patagonia (Fig. 1), was conducted between March 1996 and June 1998. Four buoyed ropes with weights attached were deployed on the sea bottom at Ameghino Point (March 1996-June 1996) and Kaiser Beach (July 1996-June 1998) (Fig. 1) at depths from 5-15 m (200—1,000 m from the coast). To stimulate Loligo females to lay their eggs on the ropes, artificial egg masses made up of small polyethylene bags filled with polystyrene spheres were attached to the ropes' bottom ends. The ropes were monitored every 2—1 wk depending on weather conditions. Seawater tetnperature, salinity, and pH were registered with a manual electronic sensor (Horiba Inc.. Kyoto, Japan) at every control date. From August 1996 to February 2000, eight other spawning sites were found at Nuevo and San Jose gulfs and Beagle Channel (Fig. 1 ) by scuba diving and by collection of stranded egg masses on the beach. Additionally, two Loligo spawning sites were identified at Cape Tres Puntas and San Mati'as Gulf (Fig. 1 ) from the analysis of egg masses provided by Dr. Alejandro Petovello (Santa Cruz Province Fisheries Department) and Dr. Ratil Gonzalez (Almirante Storni Marine Biology Insti- tute). The capsules from each egg mass were counted, and the aver- age number of eggs/embryos per capsule was estimated from a lO-capsule sample per egg mass or from the capsules available when less than 10. The embryonic stages present in each egg mass 289 290 Baron - I 1 •Lurug'uaV - > cs::--' 35 - a 1 o S :f: 5000 m 2 Mar del Plata y^-'/" -* ^^W/ — >>^^.^^^^^^m f 40-- Las Gnitas-p«_^/»^*|^**^ Puerto \?>'S»B Ja*^ >' Madrvn-«J Oulf .<' j Argentine f Kkvo .' Patagonia J Culf j .-' ATLANTIC OCEAN 45'- Z*^ .ftBl» Jorge ; I / Gulf 1 '•^ „ ^-^ Cap* -'' \ D'e"sl23o^Tr«P-«» \ ^/ r ' ~. ' " - - - W 200^m " "50-- V /'l* .Jl/^~ FalWand (Malvlnas) -^ / '■X/f^/ Islands ^ Tlerra d«l Fu«flo*-' ^wi3^:'K r."^% 55- ^^ V ^-S^^'^, Loligo santxaulBnsis \ *^ Channel '' ' \ Bndges ' Lcligc g^HI LS 65' 60- 55' SO' 1 1 1 1 LW Figure 1. Area of distribution of Loligo galii and L. sanpaulensis in Patagonia and location of the spawning areas identilled in tliis study. Data from V'igliano (1985), Inada et al. (19861, and personal observations. were identified folkiwing Arnold's scale (Arnold 1965). A 100- egg sample per egg mass ( 10 eggs x 10 capsules) was measured in egg masses showing embryonic stages previous or equal to 13 (beginning of the blastoderm expansion; Arnold 1965). On the basis of the egg counts and measurements, the egg masses were classified into different types. To identify these egg inass types to the species level, samples of the mature oocytes were taken from the ovaries and oviducts of 10 L. t^ahi (89-192 mm ML) and 10 L. sanpaulensis (67-157 mm ML) mature females captured by coastal seine and jigging at Nuevo Gulf. For each species, the lengths of 200 oocytes (20 oocytes x 10 females) were measured. Additionally, samples of 100 eggs/mature oocytes were randomly taken in triplicate from each of the egg mass types and from the ovaries and oviducts of the L. gahi and L. sanpaulensis mature females (3x10 oocytes x 10 females) from Nuevo Gulf, dried to constant weight, and weighed. All measurements were done at 25x magnification with a Wild dissecting microscope (Wild, Heer- brugg. Switzerland) equipped with an eyepiece; weights were taken to the nearest 0.01 g using a Mettler PC 440 - Delta Range electronic scale (Mettler Instruments. Zurich. Switzerland). The hatching paralarvae from each egg mass type were characterized on the basis of their mantle lengths and chromatophore arrange- ments. RESULTS Two egg mass types with distinctive eggs/embryos sizes and numbers of eggs per capsule were found in Nuevo Gulf. Each egg mass type, and also the capsules, eggs, and embryos themselves, will be hereafter referred to as type 1 and 2. The type 1 Loligo egg masses were regularly found attached to the ropes at the two lo- cations of the same gulf where they were deployed: Ameghino Point and Kaiser Beach (Fig. I; Table 1 ). They were also located by scuba diving on gravel substrate or on objects such as shells, ropes, or fishing lines at diffeient locations of the same gulf: Pier Piedrabuena (Puerto Madryn). Cuevas Point. Avanzado Hill, and by collection of stranded material on Mimosa Beach (Fig. 1 ; Table 1 ). Type 1 egg masses were also found throughout a wide latitu- dinal range along the Atlantic coast of Patagonia: Las Grutas (San Matias Gulf). Los Pajaros Island (San Jose Gulf). Puerto Deseado (Cape Tres Puntas). and Bridges Islands (Beagle Channel). Type 2 egg masses were found only by scuba diving, anchored in sandy or muddy bottoms of Nuevo Gulf, near the pillars of Pier Almirante Slorni (Puerto Madryn). and in the vicinity of Las Piedras Park, a recreational diving spot located a few hundred meters from the Puerto Madryn"s waterfront (Fig. 1; Table 1). Type I and 2 egg masses are well-defined structures and are made up of several capsules attached to each other and to the substrata at their basal ends by short stalks. Each capsule consists of an inner core of jelly surrounded by a spiral band of jelly that contains a row of eggs and an external covering formed by several layers of translucent material that provides cohesion and protection to the eggs (Fig. 2). The bases of the stalks, composed of the same material as the capsule's core and deprived of any external layer, are entangled, forming a bundle that is attached to the substrate. The number of capsules per egg mass varied from less than 10 to more than 300 for the type 1 egg masses and from less than 10 to more than 400 for the type 2 egg masses (Table 1 ). Some of the egg masses found during the survey consisted of groups of cap- sules, each one containing embryos in a particular stage of devel- opment (multiple egg masses; Table 1 ). This can be attributed to the deposition of egg capsules in a common egg mass by more than one female at various time intervals. The average number of eggs/ embryos per capsule was 69 (h = 340) for the type I and 298 (« = 50) for the type 2 egg masses, but these numbers varied con- siderably between egg masses of the same type (Table 1 ). Both type 1 and 2 eggs were approximately oval in shape, with the animal pole more pointed than the vegetal pole. The average size of the eggs was 2.2 mm (2.1-2.3 mm; ;; = 500) for type 1 egg masses and 1.2 mm (1.2-1.3 mm; n = 200) for type 2 egg masses (Fig. 2). The dry weights of 100 eggs were 0.10 g for each of the three replicates of type 1 eggs and 0.03 g for each of the three replicates of type 2 eggs. Survey of the Egg Masses of L gahi and L. sanpaulensis 291 TABLE 1. Results of the survey on Loligo spp. egg masses in the Atlantic coast of Patagonia. Date of Depth Temperature Salinity Capsules per .\verage eggs Stage of embryonic Record Location collection (ml pH t C) (PPt) egg mass per capsule development Type 1* Las Grutas 3/1 1/84 5 n/a n/a n/a 15 52 13-22 1 -i Puerto Deseado 12/15/95 Stranded n/a n/a n/a n/a 55 28 I 3 Ameghino Point 3/12/96 10 8.1 16.3 34.6 65 46 28 1 4 Ameahino Point 3/12/96 5 8.1 16.3 34.6 6 52 16 1 S Ameghino Point 3/28/96 10 8.08 16.3 34.6 20 80 28 1 h Ameghino Point 4/26/96 10 8.18 14.5 35.0 10 63 29 1 7 Los Pajaros Island 5/5/96 Stranded n/a 15 n/a 153 92 17 1 8 Ameghino Point 5/29/96 15 8.2 12.9 35.0 35 65 29 1 9* Ameghino Point 5/29/96 10 8.2 12.9 35.0 37 94 18-23 1 10 Ameghino Point 5/29/96 10 8.2 12.9 35.0 22 63 12 1 II Kaiser Beach 7/26/96 15 8.21 10.6 35.1 78 61 10 1 12 Pier Piedrabuena 8/13/96 5 8.22 10.2 35.1 38 53 15 1 13* Kaiser Beach 8/29/96 15 8.26 10.6 35.2 52 58 15-17 1 14 Kaiser Beach 9/12/96 10 8.3 11 34.8 49 81 14 1 1? Kaiser Beach 9/12/96 5 8.31 10.8 35.1 123 69 13 1 16 Kaiser Beach 9/12/96 10 8.3 11 34.8 23 63 13 1 17* Kaiser Beach 9/30/96 10 8.29 11.8 35.1 345 63 12-21-30 1 18 Kaiser Beach 10/14/96 10 8.3 13.2 34.6 85 85 19 1 19* Kaiser Beach 10/14/96 10 8.3 13.2 34.6 98 42 13-18 1 20* Kaiser Beach 1 1/27/96 5 8.4 15.4 34.7 69 44 28-29 1 21 Kaiser Beach 11/27/96 5 8.4 15.4 34.7 30 76 10 1 21 Kaiser Beach 12/26/96 5 8.35 14.9 34.6 13 39 0 1 23 Kaiser Beach 1/28/97 10 8.46 IS 34.8 85 73 28 1 24 Kaiser Beach 1/29/97 10 8.46 18 34.8 63 82 13 1 25 Kaiser Beach 3/4/97 10 8.44 19.3 34.2 7 91 8 1 26 Kaiser Beach 3/25/97 5 8.49 16.5 34.8 3 63 29 1 27 Cuevas Pomt 6/23/97 5 8.58 12.8 35.0 58 74 26 1 28 Cuevas Point 6/23/97 5 8.58 12.8 35.0 60 82 19 1 29 Cuevas Point 6/23/97 5 8.58 12.8 35.0 24 88 10 1 30 Mimosa Beach 7/3/97 Stranded n/a n/a n/a 62 55 29 1 31 Kaiser Beach 9/10/97 in n/a n/a n/a 243 48 n/a 1 32 Bridges Islands 6/15/97 12 n/a n/a n/a 51 74 12 1 33* Pier Storni 3/17/98 15 8.5 16.8 34.2 180 309 Various 2 34* Pier Storni 3/17/98 15 8.5 16.8 34.2 465 322 Various 2 35 Pier Storni 3/26/98 15 8.4 16.5 34.0 13 243 13 2 36* Pier Storni 3/26/98 15 8.4 16.5 34.0 82 314 Various 2 37 Avanzado Hill 3/31/98 1 n/a n/a n/a 28 154 11 1 38 Pier Stonii 4/4/98 15 8.22 16.3 34.0 88 n/a 14-22-26 2 39 Pier Storni 4/4/98 15 8.2 16.3 34.0 52 n/a 29 2 40 Pier Storni 5/22/98 15 8.07 14.8 34.0 6 n/a >29 2 41 Kaiser Beach 6/30/98 10 8.25 11.2 34.3 138 60 12 1 42* Bridges Islands 8/17/98 6 n/a 5 n/a 35 63 Various 1 43 Las Piedras Park 2/4/00 5 n/a 17.4 n/a 20 302 21 2 No egg masses were found on the ropes on 8/13/96. 1/18/97, 2/17/97, and 4/23/97. Embryonic stages follows the scale of Arnold (1965). n/a: not available. *Multiple egg mass (groups of capsules with embryos at different stages of development). Besides larger absolute sizes, type 1 embryos show external yolk sacs proportionally larger than those of type 2 embryos at the same stages of development (Fig. 3). Therefore, throughout their development, type I embryos show features closer to those de- scribed by Naef (1928) for the embryos of L. vulgaris (Lamark. 1798), and type 2 embryos show characteristics that resemble those illustrated by Arnold ( 19651 for the embryos of Z,. pealei (Le Sueur, 1821 ). At hatching, type 1 embryos attain mantle lengths of 2.6-3.2 mm. and type 2 embryos attain mantle lengths of 1.4-1.7 mm (Fig. 3). The most common chromatophore arrangements found on the embryos of both types of paralarvae are shown in Figure 4. Chromatophores were red or yellow on the ventral sur- face of both types of paralarvae, brown or yellow on the dorsal surface of type 1 paralarvae, and only yellow on the dorsal surface of the type 2 paralarvae. Orange chromatophores were observed only on the embryos in stages 26-28 of Arnold's scale (Arnold 1965). The cheek patches of the type I hatchlings consisted in four red chromatophores, but three chromatophores on either one or both cheeks were frequently observed. The type 2 hatching paralarvae displayed only two red chromatophores on each cheek patch. The average size of the mature oocytes sampled from the ova- ries and oviducts of 10 Z.. gahi females was 2.2 mm (2.0-2.4 mm; n = 200): the average size of the oocytes sampled from 10 female 292 Baron Figure 2. Aspect of Loligo spp. egg mass types found in tfie Atlantic coast of Patagonia. Left: type 1 egg capsule from Beagle Channel, Center: type 1 Loligo sp. egg capsule from Nuevo C^ulf, Right: type 2 egg capsule from Nuevo Gulf (scale har = 6 nmi). L. scinpaiileiisi.s was 1.2 ( 1.2-1.3 mm; /; = 200). The dry weights of 100 oocytes were 0.10 g for each of the three replicates of L. gain's oocytes and 0.03 g for those of L .scinpauleiisis. Given that L. galii and L sanpaulensis are the only loliginids that inhabit the waters of northeastern Patagonia {Castellanos & Cazzaniga 1979. Inada et al. 19S6). the striking conespondence between the sizes and weights of each egg type and the mature oocytes of either species make it possible to state that the type 1 egg masses belong to L. gain and type 2 egg masses belong to L. sanpaulensis. From the egg masses obtained in Bridges Islands (Fig. 1; Table 1 ). those found in August 1998 showed eggs identical in size to the L. galii eggs from Nuevo Gulf (average 2.2 mm, range 2.1-2.3 mm; n = 200). The egg mass found in the same locality in June 1997 (Table 1) showed eggs somehow bigger than those normally found in the type 1 egg masses (average 2.6 mm. range 2.3-3.0 mm; ;; = 200). Moreover, this was coniposed of two different groups of capsules showing eggs that were 2.3-2.6 mm and 2.6- 3.0 mm. However, the number of eggs per capsule fell into the range reported for the type 1 egg masses (Table 1 ). The dry weight of each of the three lOO-egg replicates taken from this egg mass was 0.20 g. In Nuevo Gulf, L. galii spawned almost continually throughout the year; February was the only month in which egg masses were not found (Table 1). According to historical registers (more than 10 years, hourly records) taken at Nuevo Gulf (unpublished data. Puerto Madryn's Tides Control Station, Hydrographic Ser\ ice. Ar- gentine Navy), the nionthly average of daily niininuun SST reaches a minimum in August (7.3°C). and the monthly average of daily maximum SST reaches a maximum in February (23^C). SST registered in Beagle Channel at the time of collection of type 1 egg masses (August 1998) was 5°C. The.se data show that L. galii can spawn at temperatures as low as 5"C and that spawning could be limited at temperatures close to 23''C. Although the sampling pro- cedure limits the ciuanlitative estimation ou spawning peaks of L. galii in Nuevo Gulf, it is interesting that L gahi egg masses with higher numbers of capsules were registered in September in both 1996 and 1997 (Table 1 ). Type 2 egg masses were found in Nuevo Gulf from March 1998 to May 1998 and in February 2()()0 (Table 1 ). The analysis of the embryonic stages found in these egg inasses "^ ^ > ys Figure 3. Compared aspect of the type 1 and type 2 Loligo fixed embryos. Upper left: type 2 emhryo at stage 30; upper rigth: type 2 embryo at stage 21; lower left: type I embryo at stage 3(1; lower rigth: type 1 embryo al stage 21 (scale bar = 1 mml. ys: yolk sac. Kmbryonic stages following the scale of embryonic de>elopment of .Arnold (1^651. suggests that the egg masses were actually deposited froni March 1998 to April 1998 and in January 2000. These results show that the spawning season of L. sanpaulensis in the study area extends from summer to early fall. During this period, the monthly average of daily niinimum SST reaches a tninimum in April (12'-C). and the monthly average of daily mean SST ranges between 15"C (April) and 17.5°C (February). DISCUSSION It has been observed that some loliginids deposit their egg masses on the ropes and PVC pipes used as traps for octopuses (Porteiro & Martins 1992. Re et al. 1996). Also, artificial egg masses such as those used in this work have been reported to cause a visual stimulus for loliginids to mate and spawn (Arnold & Williams Arnold 1977. Yang et al. 1986. Vecchione 1988). The devices deployed in this study to examine the laying activities of the Loligo species present in Nuevo Gulf are a simplification of those traps. These structures stimulated the spawning of L. gahi but did not show any results with L. sanpaulensis. This seems reasonable, considering that L. gahi deposits its egg masses on hard substrata and L. sanpaulensis spawns on soft bottoms. A distinct selection of substrata for egg laying has also been observed in two sympatric loliginids in the Gulf of Mexico, L. pealei se- lecting hard substrates and L. />/?( (Blainville, 1823) preferring soft Survey of THt Egg Masses of L gahi and L sanpaulensis 293 Figure 4. Chromatophore arrangements on the ventral (left) and dor- sal (rigth) surfaces of the Loligo spp. hatching paralarvae from coastal waters of Patagonia. Top: type I paralarva. hott(mi: type 2 paralarva. Black circles represent red chromatophores, gray circles represent brown chromatophores and empty circles represent yelloH chromato- phores. substrates (Vecchione 1988). Arkhipkin el al. (2000) found egg masses of L gahi attached to no other substrate than kelp and concluded that the spawning sites of the species could be limited to shallow waters. In the present work, egg masses of L. gahi have been found attached not only to kelp but also to other types of hard substrate. The structure of both species' egg masses, several capsules enclosing a row of eggs arranged in a coiled pattern along their longitudinal a.xle. is typical of Loligo species (Arnold & Williams Arnold 1977). The number of capsules per egg mass is variable for L gahi and L sanpaulensis egg masses; this variability is probably related to the size and physiological condition of the mother. This has also been observed in L. vulgaris reynaudii (Lamark. 1798) by Sauer et al. ( 1993). who found concentrations of egg masses more than 3 m in diameter surrounded by smaller ones (1-10 capsules per egg mass), and by Vecchione (1988). who observed Loligo sp. egg masses in the Gulf of Mexico that were made up of 10— tO capsules. However, as it has been observed in the present study, the variability could in part be the result of the deposition of multiple egg masses on a common mass by more than one female. Also, it should be noted that even when rnost egg masses were considered as individual egg masses, it is possible that these were deposited by more than one female within a brief period of time. so they could not be distinguished from each other on the basis of embryo maturity. This has also been suggested by Arkhipkin et al. (2000) who found 6-141 capsules per egg mass in L. gahi. The number of eggs per capsule in the egg masses of i.. gahi is comparable to that reported for other loliginids. such as L. dii- vaiicelii (D'Orbigny. 1848) (125-150 eggs per capsule; Asokan & Kakati 1991), L forbesi (Steenstrup. 1856) (39-52 eggs per cap- sule; Porteiro & Martins 1992). and L. vulgaris reynaudii (148 ± 37 [mean ± SD] eggs per capsule; Sauer et al. 1993). The number of eggs per capsule reported in this study is comparable to that found by Guerra et al. (2001) (50-60 eggs per capsule) from a 12-capsule egg mass of L. gahi found in Reilaca. Chile (32°58'S. 71°32'W). Also, the mean number of eggs per capsule observed in the present work (69 eggs per capsule) is similar to that reported by Arkhipkin et al. (2000) for the same species (71 eggs per capsule). The number of eggs per capsule of L. sanpaulensis is comparable to that for L. plei (200-300 eggs per capsule) reported by Waller and Wicklung (1968). The egg-size range of L. opalescens (Berry. 1911) (2.0-2.5 mm; Fields 1965) shows the greatest resemblance to that of L. gahi. Other species with close egg-size ranges are L vulgaris (2.3-2.7 mm) (Worms 1983. cited in Baeg et al. 1992) and L bleekeri (Keferstein. 1866) (2.6-2.7 mm; Baeg et al. 1992). The size range observed in the eggs of L. gahi from Nuevo Gulf and the egg mass found in Bridges Islands in August 1998 (2.1-2.3 mm) is within the range reported by Arkhipkin et al. (2000) (2.0-2.5 mm) from egg masses found in the Falkland (Malvinas) Islands, and the size range from the eggs found in Bridges Islands in June 1997 (2.3-3.0 mm) is comparable to that reported by Guerra et al. (2001 ) from the eggs found in Reiiaca. Chile (2.6-3.1 mm). On the other hand, the size range of the L. sanpaulensis eggs falls within the range reported for the Northeastern Atlantic species L. pealei (1.1-1.6 mm; Summers 1983. cited in Baeg et al. 1992). The chromatophore arrangements displayed by the hatching paralarvae of L. gahi and L. sanpaulensis were easily distinguish- able from each other. The pattern found in L. gahi closely re- sembles that reported by McConathy et al. (1980) for L. opal- escens. and the pattern found in L. sanpaulensis is similar to that observed by the same authors in Lolliguncula brevis (Blainville, 1823). The pattern of chromatophore arrangements reported by Arkhipkin et al. (2000) for the paralarvae of L. gahi is similar to that found in this study. However, this authors omitted two yellow chromatophores placed on the ventral surface of the head, anterior to the eyes, that were regularly observed in the L. gahi hatchlings from this study, both from Nuevo Gulf and Bridges Islands, and in other loliginids (Naef 1928, McConathy el al. 1980. Baeg et al. 1992. Blackburn et al. 1998). Also. Arkhipkin et al. (2000) incor- rectly report the presence of brown chromatophores on the ventral surface of the hatching paralarvae, which are actually red, and the presence of orange chromatophores. which are actually yellow (orange chromatophores are present only in earlier stages of the embi^onic development). L. gahi shows a year-round spawning season in Nue\o Gulf, which could be limited at the depth range covered in this study (0-15 m) by the maximum SST of the hottest month. Continuous spawning throughout the year has also been reported for L. vul- garis reynaudii (Sauer et al. 1992) from South Africa and L forbesi (Lum Kong et al. 1992) from Irish waters. The observation of two peaks of recruitment of L gahi to the Falkland (Malvinas) Islands fishery has led some authors to consider two major spawn- ing peaks for that area (Patterson 1988. Hatfield et al. 1990). 294 Baron Recently, a third recruitment peak has been identified (Agnew et al. 1998. Hatfield & MuiTay 1999). The hypothesis of a spawning peak in September for L. gahi in Nuevo Gulf, on the basis of this study's preliminary observations on the size of the egg masses, agrees with the observations of Portela et al. (1994). who reported maximum proportions of spawning L, gahi individuals in Septem- ber between 42°S and 49°S during a survey conducted from March to October 1989. On the basis of the results of the egg mass survey in Nuevo Gulf, it can be stated that the spawning activity of L. sanpauloisis near the southern limit of its geographic distribution is restricted to summer. Unpublished data on the reproductive cycle of L. san- paidensis in Northern Patagonia on the size and maturity structure of the adult and juvenile population (Baron & Re. Pers. Comm. ) is consistent with this conclusion. Previous studies on the reproduc- tive cycle of L. sanpaulensis in tlshing areas of central Argentina (Mar del Plata, 38°S) and Brazil (Cabo Frio, 21°S) indicated two spawning peaks per year (Vigliano 1985, Costa & Fernandes 1993). However, an extended spawning period for L. sanpaulensis throughout the year has been considered for southern Brazil in a more recent study (Andriguetto & Haimovici 1996). In Nuevo Gulf, spawning seems to be limited by water temperature below 12°C and facilitated by a monthly average of daily mean tempera- ture in the range of 15-17.5"C. This temperature range is in agree- ment with the observations of Andriguetto and Haimovici (1996) who reported the finding of undetermined Loligo egg masses, presumably of L. sanpaulensis. in Southern Brazil at bottom tem- peratures of I7.5°C-17.8°C. The differences in the sizes and weights of the eggs in the egg mass found in Bridges Islands in July 1997 cannot be fully ex- plained on the basis of the present information. The existence of two groups of capsules enclosing eggs with different size ranges but the same stage of development suggests that this was actually a multiple egg mass deposited by two females within a brief time interval. Given that the number of eggs per capsule of this egg mass is within the range observed for type I egg masses, this could be a L. gahi egg mass adapted to the low temperature conditions of the Beagle Channel. On the other hand, it could be thought of as the egg mass of another Loligo species. Filippova (1969, cited in Castellanos & Cazzaniga 1979) postulated the existence of two different taxonomic entities within L. gahi {L. gahi in waters of the Pacific and L. patagonica in the Atlantic). However. Castellanos and Cazzaniga ( 1979) and Brakoniecki ( 1984) presented evidence that L. patagonica is a junior synonym of L. gahi. The present work shows the wide ranges of latitude and tem- perature suitable for the spawning activity of L gahi along the Patagonian Coast and the narrow range of temperature at which L sanpaulensis can reproduce near the southern limit of its distribu- tion. Further studies on the spawning grounds of L. gahi and L. sanpaulensis must emphasize quantitative estimations to provide predictive data on the recruitment patterns of both species. ACKNOWLEDGMENTS I am grateful to the personnel of the vessel Lago Musters (Pre- fectura Naval Argentina) for their assistance in the survey of egg masses in Nuevo Gulf and to Mr. Hector Monsalve, Dr. Raul Gonzalez, and Dr. Alejandro Petovello. who provided part of the material used in this study. I appreciate the support given by Dr. Maria E. Re and the valuable comments given by Dr. Manuel Haimovici, Dr. Nestor Ciocco, Dr. Jose "Lobo" Orensanz, and two anonimous referees. LITERATURE CITED Agnew. D. J., C. P. Nolan & S. Des Clers. I99S. On the problem of identifying and assessing populations of Falkland Island squid Loligo gahi. In: A. I. L. Payne. M. R. Lipinski. M. R. Clarke. & M. A. C. Roelevelld, editors. Cephalopod biodiversity, ecology and evolution. S. African J. Mar. Sci. 20:59-66. Andriguetto, J. M. & M. Haimovici. 1996. Reproductive cycle of Lolii^o .\ct pealii (Lesueur). Biol. Bull. 128:23-32. Arnold. J. M. & L. D. Williams Arnold. 1977. Cephalopoda: Decapoda, En. Reproduction Mar. hivertchrates 4(5):243-290. Asokan. P. K. & V. S. Kakati. 1991 . Embryonic development and hatching of Loligo diivaucclii Orbigny (Loliginidae. Cephalopoda) in the labo- ratory. Indian J. Fi.sh. 38(4):201-206. Augustyn. C. J. 1990. Biological studies on the chokka squid Loligo vid- gari.s reynaudii (Cephalopoda: Myopsida) on spawning grounds off the south-east coast of South Africa. S. Afr. J. Mar. Sci. 9:1 1-26. Baddyr, M. 1991. Biology of the squid Loligo vulgaris in relation to the artisanal fishing site of Tifnit, Morocco. Bull. Mar. Sci. 49(1-2):661. Baeg, G. H.. Y. Sakurai & K. Shimazaki. 1992. Embryonic stages of Loligo hieekeri Keferstein (Mollusca: Cephalopoda). The Veliger 35(3):234- 241. Blackburn, S.. W. H. H. Sauer & M. R Lipmski. 1998. The embryonic development of the Chokka squid Loliiio vultituis reynaudii d'Orbigny. 1845. The Vc/igcr 41(3):249-258. Brakoniecki, T. F. 1984. A full description of Loligo sanpaulensis. Bra- koniecki, 1984 and a redescription of Loligo gahi D'Orbingny. 1835. two species of squid (Cephalopoda: Myopsida) from the Southwest Atlantic. Bull. Men: Sci. 34(3):435^48. Cardoso, F.. J. Tarazona & C. Paredes. 1998. Aspectos biologicos del calamar patagonico Loligo gahi (Cephalopoda: Loliginidae) en Huar- mey, Peru. Rev. Per Biol. 5(1):9-14. Castellanos. Z. J. & N. J. Cazzaniga. 1979. Aclaraciones acerca de los Loliginidae del Atlantico Sudoccidental (Mollusca: Cephalopoda). Neotropica 25(73):59-68. Chesheva. Z. A. 1990. Biologiya kal'mara Lolifio patagonica Yugo- Zapadnoj Atlantiki. Zool. Zh. 69(2): 126- 129. Costa. P. A. S. & F. C. Fernandes. 1993. Reproductive cycle of Loligo .sanpaulensis (Cephalopoda: Loliginidae) in the Cabo Frio region. Bra- zil. Mar Ecol. Prog. Ser. 101(l-2):91-97. FAO. 1999. Fishery statistics 1997. Vol. 84. Fields. W. G. 1965. The structure, development, food relations, reproduc- tion, and life history of the squid Loligo opalescens Berry. California Department of Fish and Game. Fish. Bull. 131:1-108. Gonzalez. R. A. C. 1999. Biologia y explotacion pesquera del salmon de mar Pseudopercis semifasciata (Cuvier, 1829) (Pinguipedidae) en el Golfo San Mati'as. Patagonia. Argentina. Tesis doctoral. Universidad Nacional del Sur. 135 pp. Guerra, A., F. Rocha, L. F. Buckle & A. F. Gonzalez. 2001. Embryonic .stages of the Patagonian squid Loligo gahi (Mollusca: Cephalopoda). The Veliger 44( I ). Hatfield. E. M. C. 1991. Post recruit growth of the Patagonian squid Loligo i>ahi (d'Orbigny). Bull. Mar Sci. 49:349-361. Hatfield, E. M. C, P. G. Rodhouse & J. Porebski. 1990. Demography and Survey of the Egg Masses of L. gahi and L. sanpaulensis 295 distribution of the Patagonian squid {Ltiliiio t^alii d'Oi"bign\ ) during the austral winter. J. Cons. Cieii. 46(3):306-3l2. Hatfield. E. M C. & A. W. A. Murray. 1999. Objective assessment of maturity in the Patagonian squid Loligo gahi (Cephalopoda: Lolig- inidael froin Falkland Islands waters. ICES J. Mar. Sci. 56(5):746-456. Inada, T., M. Takeda & H. Hatanaka. 1986. Important fishes trawled off Patagonia. I. Nakamura. editor. Japan Marine Fishery Resource Re- search Center. 369 pp. Lum Kong. A., G. J. Pierce & C. Yau. 1992. Timing ol spauning and recruitment in Loligo forbesi (Cephalopoda: l.oliginidac) m Scottish waters. J. Mar. Biol. Ass. U.K. 72(2):3ni-3l I, McConathy, D.A., R.T. Hanlon & R. F. Hixon. 1980. Chromatophore arrangement of hatchling loliginid squids (Cephalopoda, Myopsida). Malacologia 19(2):279-288. Naef, A. 1928. Die Cephalopoden. Embriologie. Fauna Flora Golfs of Neaple. Monograph 35(11). 357 pp. Nesis, K. M. 1987. Cephalopods of the world. L. A. Burgess, editor. THE Publications. 351 pp. Patterson, K. 1988. Life history of the patagonian squid Loligo gain and growth parameters estimates using least squares fits to linear and Von Bertalanffy models. Mar. Ec. Prog. Sen 47( 1 ):65-74. Porteiro, F. M. cS: H. R. Martins. 1992. First finding of natural laid eggs from Loligo forbesi Steenstrup, 1856 (Mollusca: Cephalopoda) in the Azores, Arquipelago Cieiic. Nat. Kif. Earth Sci. 10:1 19-120. Portela, J. M., N. P. Contreras & V. Trujillo Gorbea. 1994. Aspectos reproductivos del calamarete {Loligo gahi). el calamar {lllc.x argeiui- niis) y la merluza {Merliicciiis hiihbsi) en el Atlantico Sudoccidental. Frenle Marilimo 15(a):2l-36. Re, M. E., J. C. Heron & A. Gosztonyi. 1996. First haresiing results of a small cephalopod. Octopus tehuelchus. in the subtidal /one of North Patagonian Gulfs, Argentina. Proc. of the 4"" ICMAM. ,/. Mai Appl. Mulacol. 8(1):119. Re, M. E., A. E. Gosztonyi & I. Eli'as. Loligo sanpaulensis y Loligo gahi (Mollusca: Cephalopoda) en aguas chubutenses y ,su presencia en las captruas con red de costa. Centro Nacional Patagonico - Consejo Na- cional de Investigaciones Cientfficas y Tecnicas, Boulevard Brown s/n, (9120) Puerto Madryn, Chubut, Argentina. Pers. Comm. Roper, C. F. E.. M. J. Sweeney & C. E. Nauen 1984. FAG species cata- logue. Cephalopods of the world. An annotated and illustrated cata- logue of species of interest to fisheries. FAO Fish. Synop., ( 1254) Vol, 3. 277 pp. Sauer, W. H. H.. M. J. Smale & M. R. Lipinski. 1992. The location of spawning grounds, spawning and schooling behaviour of the squid Loligo vulgaris reynaiulii (Cephalopoda: Myopsida) off the Eastern Cape Coast. 5. African Mar. Biol. 1 14(1 ):97-107. Sauer, W. H. H.. C. McCarthy. M. J. Smaile & A. S. Koorts. 1993. An investigation of the egg distribution of the chokka .squid, Loligo vul- garis reytmudii. in Krom Bay. 5. .African Bull. Mar. Sci. 53(3):1066- 1077. Sweeney, M. J.. C. F. E. Roper. K. M. Mangold. M. R. Clarke & S. V, Boletzky. 1992. Larval and Juvenile Cephalopods: A manual for their identification. M. J. Sweeney. C. F. E. Roper. K. M. Mangold, M, R. Clarke & S. V. Boletzky, editors. Washington D.C.: Smithsonian Con- tributions to Zoology No 513. 282 pp. Vecchione. M. 1988. In-situ observations on a large squid spawning bed in the Eastern Gulf of Mexico. Malacologia 29(1 ):135-141. Vigliano, P.H. 1985. Contribucion al conocimiento de la biologia de Loligo brasiliensis Blainville 1823 (Mollusca. Cephalopoda) en aguas argen- tinas. Tesis Doctoral, Facultad de Ciencias Naturales. Universidad Na- cional de La Plata. 183 pp. Waller, R. A. & R. I. Wicklung 1968. Observations from a research sub- mersible— mating and spawning of the squid, Doriteuthis plei. Bio- Science (February):l 10-1 1 1. Yang. W. T., R. F. Hixon, R. F. Turk, P. E. Krejci, M. E. Hulet & R. T. Hanlon. 1986. Growth, behaviour, and sexual maturation of the market squid. Loligo opalescens. cultured through the life cycle. Fish. Bull. 84:771-798. Joiinwl of Shellfish Research. Vol. 20. No. 1. 297-300, 2001. DIGESTION OF CELLULOSE BY STOMACH HOMOGENATES OF GREEN ABALONE (HALIOTIS FULGENS) ALFREDO ENRIQUEZ,' MARIA TERESA VIANA,"* CARLOS VASQUEZ,' AND ARMANDO SHIMADA^ Facultad de Ciencias Marinas, Universidad Aittonoma de Baja California. Ensenada. B.C., Me.xico: 'Institiito de Investigaciones Oceanologicas, Universidad Autonoma de Baja California. Ensenada. B.C., Me.xico: ' Direccion Tecnico-AdministraUva, Direccion General de Zoologicos de la Ciudad de Me.xico. Chapiiltepec, D.F., Mexico; '^Facultad de Estiidios Superiores-Ciiautitkin. Universidad Navionul Autonoma de Mexico. Ajiichitldn, Qro., Me.xico ABSTRACT Three experiments were conducted to study the cellulase activity of stomach homogenates of green abalone. with and without antibiotics added. Forty-eight previously frozen stomachs from wild abalone adults were thawed and individually homog- enized. Four cultured abalone adults maintained under a balanced diet containing cellulose were killed and the fresh stomach samples processed. Cellulase activity was estimated through the production of glucose, incubating the homogenate from each organism with cellulose. To determine the effect of bacteria, each stomach was incubated (at 25°C for 20 h) in a completely randomized design, with and without an antibiotic mixture. For a time series measurement (exp. 2), four homogenates from wild abalone were used; the incubation mixture was followed up taking samples at 0, 1 , 2, 4, 8, 16, 20. 32. 64 and 72 h. The latter samples were also used to measure the enzymatic activity before and after a 72-h incubation, in the absence of antibiotics (exp. 3). Initial bacterial counts in the wild abalone homogenates were 459 colonies per mL. In the presence of the antibiotic mixture, no bactenal growth was observed. Initial cellulase activity was 1.6 Units. Cellulose degradation figures were low, even in the absence of antibiotics, indicating the low dependence of abalone on the nutrients derived from the digestion of the cell walls. Still, the addition of antibiotics depressed the digestion of cellulose (14.5 vs. 10.27f ). suggesting the importance of live bacteria in the digestion process, and that the majority of the degraded cellulose is perhaps acted upon by endogenous enzymes. The time series experiment of samples from wild abalone. with and without antibiotics, showed quadratic effects on cellulose digestion y = 7.55 -i- .29h - .005 h'^ (r^ = .2212), but no significant differences between treatments. The gradual increase in the digestion of cellulose, followed by a plateau is similar to what happens in ruminant digestion. The initial and final enzyme activity without antibiotics showed a significant reduction of 41.1% for the wild abalone whereas in cultured abalone a 50.9% significant reduction was observed. It seems that the cellulase complex enzymes present in the homogenates remain active after 72 h of incubation, even though their activity is lowered by about one half Although the use of antibiotics assure the absence of live bacteria in the incubation mixture, the cellular enzymes possibly freed by the effect of the antibiotics could still be active for up to 72 h after the microbes are no longer viable. Thus the enzymatic activity attributed to the stomach itself could still be confounded. KEY WORDS: green abalone, Halioiis fulgens. cellulose, digestion, bacteria, antibiotics INTRODUCTION in the digestive tract of the abalone (//. midae): however they registered cellulase activity even in gnolobiotic organisms. The information on the nutritional physiology of the abalone. j^^ ^-^ ^f ^^^ p^^^^^, experiments was to study the cellulase spect.cally in relation to the.r capability to digest and metabolize ^^^-^-^^ ^^ ^^^^^^^ homogenates of abalone, with and without polysaccharides such as cellulose, is scarce. Uki et al. (1985) and antibiotics added, using a crystalline form of cellulose as substrate. Uki and Watanabe (1992) established that abalone had a limited capacity to digest cellulose as their growth rate was inversely MATERIALS AND METHODS related to increased levels of dietary cellulo.se (from 0 to 20%). However, it has been observed that abalone grown in banks with Forty-eight stomachs were obtained from wild abalone adults, predominant sea grass, with higher cellulose content, appear to be harvested near the "Emancipacion" Fishery Cooperative in Baja bigger and healthier than those grown in areas were kelp is the California. The stomachs were placed in layers on ice, and frozen predominant vegetation. Monje and Viana (1998) compared two ^t -75°C until used. They were later thawed slowly in temperate purified diets, one containing 19% cellulose and 1% sodium algi- water, their pH and ion strength measured, and then individually nate and a second one in which cellulose was substituted with homogenized. sodium alginate. Although both diets resulted in similar growth Four cultured abalone adults that had been maintained under a rates, the group fed the cellulose-based diet had higher cellulase balanced diet (Table I ) containing 5% cellulose were killed and and alginase activities than those fed the sodium alginate. 'he stomach samples immediately processed as mentioned above. Leighton (1968) stated that the enzymes involved in the diges- Cellulase activity was estimated using the methodology de- tion of complex carbohydrates were secreted in the hepatopan- scribed by Worthington ( 1988). Where an aliquot from the stom- creas. In contrast. Erasmus et al. (1997) suggested that bacteria ^^^ homogenate from each organism was incubated with 300 mg might play a role in the digestive process of the gastropod, since "-cellulose in the presence of a .05 M phosphate buffer solution at several microbial species with cellulase activity could be detected PH -''-''- a"'^ ■^''"C. After 2 h, the reaction was colored with an anthrone-H^SOj solution and the production of glucose measured by spectrophotometer (Genesis 2000) at 625 nm. One unit releases ♦Corresponding author. PC Box 189003-70, Coronado, CA 92178, USA, 0,01 mg glucose per hour from microcrystalline cellulose at 37°C. E-mail: viana@faro.ens.uabc.mx pH 5.5, per mg of stomach homogenate (wet weight). 297 298 Enriquez et al. TABLE I. Ingredient composition of the artificial diet used for the culture abalone. given as g kg"' of dry matter. Ingredients g/kgof dry matter Fish meal' Fish silage- Soy bean meal Com meal' Kelp meal Starch (rice) Cellulose-* Gelatin (50 Blooms) Vitamin mixture^^ Mineral mixture^ Stay C' Choline chloride Methionine Sodium benzoate BHT" 300 50 80 100 100 164 80 60 15 40 2 1 2 2 4 bO'i protein kindly supplied by Productos de Ensenada. Mexico. "^ Made from tuna-fish viscera, given as dry matter basis. ' Maseca. produced in Mexico. ■* Alphacel. a-cellulose. "''As recommended by Uki and Walanabe (1992), kindly supplied by Roche, Mexico. *" Butylated Hydroxytoluene. To determine the effect of the bacteria present in the homoge- nates, each stoinach was incubated (at 25°C for 20 h) in a com- pletely randomized design, with and without an antibiotic mixture as recommended by Erasmus (1996): chloramphenicol (250 |xg/ mLl, cephalosporin (250 |jLg/mL) and ampicillin (600 |ji.g/mL). At the end of the incubation a l-mL sample was taken and after centrifugation (1880 g) the supernatant was used to determine glucose through the procedure indicated previously. The effect of the presence of the antibiotic mixture was assessed using the fol- lowing calculation: Relative digestibility reduction dig. n/a [(dig- dig, n/a) ■ 100 )/ where: (dig. a) is the digestibility with antibiotics and (dig. n/a) the digestibility without antibiotics. The number of bacterial colonies with and without antibiotics was determined before and after incubation of the stomach sample. Thus, a Zobell media was used by blending agar (1.5 g/lOO mL), peptone (.5 g/lOO mL). yeast extract (.1 g/IOOmL) and a-cellulose .2g/100 mL, mixed with filtered seawater to distilled water at a ratio of L3 v/v at pH 5.5. The Petri dishes were incubated at 25°C for up to two days, with and without antibiotics in the same pro- portion as indicated previously. Cellulase activity from the grow- ing bacteria was corroborated using the Congo red procedure (Teather and Wood 1982). For a time series measurement (experiiuent 2). four homoge- nates from wild abalone stomachs were used; the incubation mix- ture was followed up taking samples at 0, I. 2, 4. 8, 16. 20. 32, 64 and 72 h. Glucose was determined as indicated previously and the glucose produced over time was related to the cellulase activity. The latter samples were used to measure the enzymatic activity before and after a 72-h incubation, in the absence of antibiotics (experiment 3). Statistical Analysis A one-way ANOVA test was used to compare the relative digestibility reduction in stomach contents with and without anti- biotics and to compare the initial and final enzyme activity in cultured and wild abalone without antibiotics after 72 h of incu- bation. Figures given in percentage were transformed to their arc- sine sqr (y) to be analyzed. For the time series a quadratic and linear regression model was used to compare the slopes with a variance analysis to estimate the best adjustment according to Zar (1999). The SAS program was used to analyze the data. RESULTS The initial bacterial counts in the wild abalone homogenates were 459 colonies per mL of stomach. In the presence of the antibiotic mixture, no bacterial growth was observed. The initial cellulase activity of wild abalone homogenates was 1.6 Units. The (/) vitro digestibility of cellulose after 20 h incubation was lower when the antibiotic mixture was present (14.5 vs. 10.2%. respec- tively) although the difference was not significant (P = .107; F = 2.666). The relative digestibility reduction was 29.7% (Table 2). The time series experiment of samples from wild abalone, with and without antibiotics (Table 3). showed quadratic effects on the digestion of cellulose y = 7.55 + .29h - .005h- (r" = .2212). but no significant differences between treatments. The estimated equa- tions were y = 8.35 + .31h - .005h- for the treatment without antibiotics and y = 6.76 -i- .29h - .004h- for the treatment with antibiotics; the statistical analysis of these equations showed simi- lar slopes. For the cultured abalone. with and without antibiotics, there was also a quadratic effect, with significant differences (P < .05) between 4 and 16 h. The initial and final enzyme activity without antibiotics showed a reduction of 41.1% {P = .032) for the wild abalone whereas in the cultured abalone a 50.9% (P = .046) reduction was observed (Table 4). DISCUSSION Mammalian cellulose -digesting herbivores can be divided into two groups: those who depend heavily on the resulting metabo- lites, and the ones in which the process only supplements their TABLE 2. Digestion of cellulose in wild abalone stomach homogenates, with or without antibiotics. Parameter Mean (+SDI Total bacteria count/mL of stomach Initial cellulase activity Digestibility without antibiotics, % Digestibility with antibiotics, % Relative digestibility reduction, % Significant differences between treatments 459 (±497) 27 1.6 (±1.7)' 27 14.5 (±13.6)= 38 10.2 (±8.9)= 38 29.7' 38 p = 0.107 ' mg glucose/mg stomach homogenate incubated for 2 hr at 37°C. = mg glucose/mg stomach homogenate incubated for 20 hr at 25°C. ' Relative decrease in digestibility due to the use of the antibiotic mixture. Digestion of Cellulose in Green Abalone 299 TABLE 3. Digestibility of cellulose by wild abalone stomach homogenates as a function of time, with and without antibiotics. .Activity is measured as the increment of glucose production. ,SI) is given in parentheses: n = 4. Hours Antibiotics* 1 2 4 8 16 2(( 32 64 72 Without 4.1 8.8 10.9 14.8 12.3 12.3 11.2 4.0 3.9 (1.1) (2.2) (2..'i) (3.6) (3.8) (3.8) (3.6) (2.6) (2.2) With 4.7 8.1 8.3 II.O 8.9 10.5 12.1 4.3 4.9 (l.Oi (1.4) (1.6) (3.-'i) (1-7) (3.5) (2.9) (2.2) (1.6) * A mixture of chloramphenicol (250 p.g/iTiL). cephalosporin (250 p.g/mL) and ampicillin (600 |j.g/mL) as recommended by Erasmus ( 1996) was used. nutritional budget. Still, both groups are unable to degrade cellu- lose with their own enzymes, relying instead on symbiotic micro- organisms to perform the process. In any case, the digestibility of cellulose, which varies according to a number of factors (Merchen and Bourquin 1994), can be expected to be in the 40-60'^ and the 0-209^ ranges for both animal groups, respectively (Van Soest 19941. Symbiotic bacteria of mammahans" guts digest cellulose by attaching themselves to the substrate and then liberating the re- spective enzymes, which are diffused into the complex structural matrix of cellulose. These enzymes are produced either extracel- lularly or bound in the membrane of microbes. However, when the microbial cells are destroyed, the freed enzymes might still act directly on the cellulose molecules. For example, the ampicillins and cephalosporins used here, act on the bacteria by the activation of enzymes that destroy the cell wall, whereas chloramphenicol acts inhibiting bacterial protein synthesis. In the experiment reported here, cellulose degradation figures were somehow low. even in the absence of antibiotics, which could be an indication of the low dependence of abalone on the nutrients derived from the digestion of the cell wall component. Still, the addition of the antibiotic mixture depressed the disap- pearance of cellulose by around one third, suggesting the impor- tance of live bacteria in the digestion process, and that the majority of the degraded cellulose is perhaps acted upon by endogenous enzymes. Erasmus et al. ( 1997) obtained bacteria-free abalone after im- TABLE 4. Initial and final enzyme activities from wild and cultured abalone stomach homogenates, without antibiotics, after 72 hours of incubation. Enzyme Mean Abalone activity' (±SD) n Wild Initial 1.58 (±1.7) 4 Final 0.65 (±.3) 4 Reduction 41.1% P = 0.032 Cultured Initial 1.65 (±.6) 4 Final 0.84 (±.3) 4 Reduction 50.9% P = 0.046 mg glucose produced/mg stomach homogenate after 2 hr incubation at 37°C. mersing them for 72 h in sea water containing an antibiotic mix- ture. Unfortunately, after several attempts to get gnotobiotic H. fiilgens in our lab, we were not able to repeat their procedure, either by the addition of antibiotics to the water or in the food. In all cases a considerable amount of bacteria appeared in the stom- achs. Moreover, to determine cellulase activity, Erasmus et al. ( 1997) used carboxymethyl cellulose, a chemically modified soluble pre- sentation of cellulose, as substrate to determine cellulase activity. It is known that any form of cellulose that is either acid or alkali- treated has little effect on the overall cellulase system for most of the known microorganisms (Worthington 1999). The enzymatic mechanism whereby certain microorganisms can quite rapidly and completely degrade cellulose is not completely understood. How- ever, according to Worthington ( 1999), two steps involved in the degradation are proposed: first, a pre-hydrolytic step wherein an- hydro-glucose chains are swollen or hydrated, involving an en- zyme designated C,. The second step involves the hydrolytic cleavage of the now susceptible polymers either randomly or end- wise with an enzyme complex termed Cx consisting of exo and endo (3-1,4 glucanases and P-glucosidase (cellobiase) that attack soluble derivatives or cellulose that has been swollen by acid or alkali. Therefore it is recommended that in order to test true cel- lulase activity, the most crystalline form of cellulose available be used. In the present work a-cellulose was used. The effect observed with the homogenates from the wild or- ganisms in the first 32 h of incubation, is similar to what might be expected to occur in the digestive tract of the large mammalian herbivores (Van Soest 1994); that is, a gradual increase in the digestion of cellulose (as the so-called digestion rate in ruminants), followed by a plateau (equivalent to what is known as the extent of the fermentation in ruminants). The low figures for the 64 and 72-h samples could indicate the destruction of the glucose formed from cellulose, due to the long incubation period. The results of experiment 3 could be interpreted in two ways: either that the cellulase enzymes present in the stomach homoge- nates remain active even after 72 h of incubation, although their capability to cleave the substrate is lowered by about one half: or that the final product {i.e. glucose), used as indicative of the cel- lulase activity, was partially spent as in the case of experiment 2. If this is the case, then the activity against carboxymethil cellulose presented by Erasmus et al. (1997) in stomach content after the abalone were treated with antibiotics for 72 hours could not be regarded as indicative of endogenous enzymes participation but to a remnant activity in the stomach. Finally, it is important to stress that although the use of the 300 Enriquez et al. antibiotic mixture assured the absence of live bacteria in the in- cubation process, the cellular enzymes freed by the effect of the antibiotics could still be active even at 72 h after the microbes were no longer viable. Thus the enzymatic activity attributed to the stomach itself could still be confounded. ACKNOWLEDGMENTS The present work was financed by CONACYT. project num- bers 1925PB and G281 19B. We also thank the Emancipacion Fish- ery Cooperative for partially supporting this project. LITERATURE CITED Erasmus, J. H. 1996. The role of enteric bacteria in the abalone, Hatiotis midae. MSc Thesis. University of Cape Town, South Africa, p. 83. Erasmus, J. H., P. A. Cook & V. E. Coyne. 1997. The role of bacteria in the digestion of seaweed by the abalone Hulioiis midae. Aqiiaculitiie 155:377-386. Leighton, D. L. 1968. A comparative study of food selection and nutrition in the abalone, Haliotis rufescens Swainson, and the sea urchin Strongylocentrotus purpuratus (Stimpson). PhD Thesis. University of California. San Diego. Micrnfdm 68-15:685. Merchen, N. R. & L. D. Bourquin. 1994. Processes of digestion and factors influencing digestion of forage-based diets by ruminants. In: G. C. Fahey Jr., editor. Forage Quality, Evaluation, and Utilization. Madison, Wisconsin: Am. Soc. Agron., Inc., Crop Sci. Soc. Am., Inc.. Soil Sci. Soc. Am.. Inc. Monje, H. & M. T. Viana. 1998. The effect of cellulose on the growth and cellulolytic activity of abalone Halioiis fiilgens when used as an ingre- dient in formulated artificial diets. J. Shell. Res. 17:667-672. Teather, R. M. & P. I. Wood. 1982. Use of Congo Red polysacharide. Interactions in enumeration and characterization of cellulolytic bacteria from the bovine rumen. Appl. Em: Microbiol. 43:777-780. Uki. N. & T. Watanabe. 1992. Review of the nutritional requirements of abalone [Haliotis spp. I and development of more efficient artificial diets. In: S. A. Shepherd, M. J. Tegner, and S. A. Guzman del Proo, editors. Abalone of the World: Biology, Fisheries and Culture. Oxford: Fishing News Books, pp. 504-517. Uki, N., A. Kemuyama & T. Watanabe. 1985. Nutritional evaluation of several protein sources in diets for abalone Halions discus hannai. Bull. Jpii. Soc. Sci. Fish. 51:1835-1839. Van Soest, P. J-. 1994. Nutritional Ecology of the Ruminant. 2'"' ed. Corn- stock Publishing Associates. Ithaca. NY: Cornell University Press. Worthington. C. C. 1988. Wonhington Manual. Enzymes. Related Bio- chemicals. Freehold, New Jersey: Worthington Biochemical Corpora- tion. 346 pp. Worthington, C. C. 1999. Cellulase. Worthington-biochem.eom/manual/C/ CEL.html. Jourmil of Shellfish Research. Vol. :0. No. 1. 301-307, 2001. REPRODUCTIVE CHARACTERISTICS OF THE ARCHAEOGASTROPOD MEGATHURA CRENULATA PETER G. BENINGER,' * ROZENN CANNUEL,' JEAN-LOUIS BLIN," SEBASTIEN PIEN," AND OLIVIER RICHARD" ' Labonitoire de Biologic Marine. ISOMer. Facidte des Sciences et Techniques. Universite de Names. 44322 Nantes Cedex, France: 'Syndicat Mixte pour I 'Equipernent du Littoral, Zone conchylicole, 50560 Blainville sur Men France ABSTRACT A histological and histuchemical study was performed on individuals of the archaeogastropod Megalhura crenulata sampled in the field, in order to ascertain the fundamental features of reproductive biology in this species. Basic aspects addressed were gonad and gamete structure, nature of vitelline reserves, and composition of oocyte coat. Stereological counts and oocyte measurements were performed to obtain a quantitative assessinent of the reproductive cycle from June 1999 to June 2000. No simultaneous hermaphrodites were observed. The gonad structure of M. creniilalu consisted of traversing trabeculae from which gametes developed centrifugally. The gonads of both males and females were homogeneous, allowing reliable data to be obtained from a single histological sample of each individual. Mature gametes greatly dominated the profile throughout the study period; coated oocyte diameters were also very stable. These techniques, routinely applied to the study of reproductive cycles, did not allow the identification of spawning preparedness in this species. Vitelline reserves were dominated by non-staining (putatively lipid) vacuoles; no appreciable quantities of glycogen were observed. The oocyte coat was chiefly composed of acid mucopolysaccharides, conferring both mechanical and antimicrobial protection, as well as limiting egg and larval dispersal. KEY WORDS: Megalhura crenulata. Gastropoda, gonads, gametes, reproductive cycle, histochemistry INTRODUCTION Many marine natural products present biological activity in humans, and are used in medical testing or in phaimaceuticals (Munro et al. 1987. Ireland et al. 1989. Kobayashi et al. 1989. Suffness et al. 1989. Faulkner 2000). The archaeogastropod Mega- lluira crenulata is a keyhole limpet native to the California Pacific Coast, with a reported range from Mendocino County (40°N; 124°W) to Isla Asuncio (27°N; 1 15°W). Keyhole limpet hemocya- nin fKLH) of this organism is used in the treatment of certain forms of bladder cancer ( Harris & Markl 1999. 2000). It is thus of considerable medical and economic interest to rear this organism for extraction and purification of the active compound. This goal requires knowledge concerning essential aspects of M. crenidata biology, notably feeding, growth, and reproduction. The sparse studies on reproduction in M. crenulata chiefly concern modes of fertilization, which is external and involves .several hormonal agents (Tyler 1939. Webber 1977). The lack of data on the reproductive biology of M. crenulata may be con- trasted with the relatively abundant information concerning the highly-prized abalones [Haliotis spp). which are commercially im- portant, edible archaeogastropods (Newman 1967. Purchon 1968, Fretter 1984. Hahn 1989). The present study documents the gonad structure, oocyte histochemistry and reproductive cycle of M. crenulata in its natural habitat, with a view toward the long- term objective of cultivation. MATERIAL AND METHODS Specimen Collection and Dissection The animals used in the present study were obtained from June 1999 to June 2000 from the subtidal zone off Long Beach. Cali- fornia (33°45'N, 1 I8°I0'W), The specimens were sent by air to France, where they were dissected and fixed in the SMEL (Syn- *Corresponding author. E-mail: Peter.Beningerfe'isomer.univ-nantes.fr dicat Mixte pour I'Equipement du Littoral) laboratory in Blain- ville. Normandy. Five individuals were dissected each month, ex- cept in October 1999. when no sampling was possible. In order to verify the structural homogeneity of the gonad, the entire organ was removed in certain individuals of certain months, and several regions were examined (Table 1 ): a median region (M). a distal extremity (DE) and a proximal extremity (PE). For samples from February. May. and June 2000. tissue pieces were obtained from labelled L. M and R regions, allowing the distinction of left (L), median (M). and right (R) regions of the gonad. The types of histological sample are summarized in Table I for each individual and each sampling date. Histological Techniques In most cases, entire gonads were fixed in aqueous Bouin's solution. In an effort to improve fixation, small pieces of the gonad were fixed in March and April; however, the resulting sections showed that this provoked extensive leakage of gametes, such that the sections could not be used for males in either month, and for females in March. From February. May. and June 2000. in which the dissections allowed the distinction of three different regions of the gonad, histological samples were directly fixed. After rinsing for a minimum of 10 h and dehydration in an ascending ethanol- Bioclear(B) series, the biopsies were embedded in paraffin and sec- tioned at 7 |jim. Several staining protocols were developed for this study, based on Martoja and Martoja-Pierson (1967), Gabe (1968) and Vacca (1985). as indicated in Table 2. A modified Masson's trichrome protocol using fast green, irioxyhematein. and acid fuschin al- lowed the topological study of the gonads, distinguishing connec- tive tissue, oocyte coat, cytoplasm, nucleus, chromatin, and nucleolus. Periodic acid-Schiff reagents with positive (oyster di- gestive gland) and negative (amylase-digested sections of gonad) controls were used to determine the eventual presence of glycogen in the oocyte cytoplasm, as well as neutral mucopolysaccharides (NMPS) in the oocyte coat. Alcian blue was used to determine the presence of acid mucopolysaccharides (AMPS) in the oocyte coat. 301 302 Beninger et al. TABLE 1. Megalhura crenulata. Summary of histological sample types for males and females examined. Male gonad Female gonad Date Males regions Females regions 06/99 _ _ 1 I - - 2 M,P,D - - 3 I 07/99 I I 1 I 2 M.P.D -) I 3 I - - 08/99 1 M.P.D 1 M.P.D 2 M.P.D T I 3 I - - 09/99 1 I 1 M.P.D 2 I 2 I - - 3 M.P.D 10/99 - - - - 11/99 1 I 1 I 2 I 2 I - - 3 M.P.D 12/99 1 I 1 I 2 I 2 I - - 3 M.P.D 01/00 1 M.P.D 1 I 2 I 2 M.P.D 3 I - - 02/00 1 R.L.M 1 R.L.M 2 R. L.M "> R.L.M - - 3 R.L.M 03/00 - - - - 04/00 - - 1 R.L.M 05/00 1 R.L.M 1 R.L.M 2 R.L.M -) R.L.M 3 R.L.M - - 06/00 - - 1 R.L.M - - 2 I - - 3 R.L.M - - 4 I I: undetermined gonad region; M: median gonad region: P: proximal region of non-oriented gonad (organ previously removed from individual); D: distal region of non-oriented gonad; R: right region of oriented gonad; L: left region of oriented gonad. Stereology Stereological counts were performed in order to quantify the proportions of the different tissue types in the histological sections; variations in the proportions of the different tissue categories thus reflected variations in the different phases of reproductive activity (Weibel et al. 1966. Briarty 1975. Beninger 1987. Morvan & Ansell 1988. Pazos et al. 1996. Mayhew 2000); counts were per- fomied on surfaces of measured area, using a 9 x 9 point matrix on a microscope projector. In males, four tissue categories were identified for stereological purposes: trabecular tissue, developing gametes, mature gametes, and unoccupied space. In females, three tissue categories were identified: coated oocytes, non-coated oocytes, and trabecular tis- sue. Due to the loose nature of the female gonad tissue, it was impossible to determine whether observed unoccupied spaces were real or artefacts of gonad dissection. Stereolological counts were therefore performed only on areas without visible unoccupied space. Nine counts were performed for each individual and region of the gonad, and the means (with 95'7f confidence intervals) of all male and female counts were plotted. Oocyte Diameters Oocyte diameters were measured for 30 coated oocytes for each female and for each gonad region, using a calibrated optical micrometer. In order to standardize the measurements, only oo- cytes in which the nucleolus was visible (approximately the center of the cell) were selected. The evolution of oocyte diameters could then be recorded throughout the year. RESULTS Males Gonad Structure The general structure of the male gonad is presented in Figure 1 area I. The gonad was composed of connective tissue trabeculae, from which arise centrifugally the germ cells, first visible as de- veloping aflagellate gametes, and then the mature flagellate sper- matozoa. The mature spermatozoa occupied the majority of the sectional area (Fig. 1 area I. Fig. 2). Stereology Testicle homogeneity was verified in 1 to 3 individuals from July, August 1999 and January, February and May 2000, using stereological counts. The results shown in Fig. 2 indicate that the male gonad was structurally homogeneous and gametogenetically synchronous. Reliable data for histological study may therefore be obtained from only one histological sample per individual. The male gonad presented a stable histological profile through- out the study period from July 1999 to June 2000. Mature sper- matozoa occupied almost the entire gonad, with mean volume fractions of 0.77 (May 2000) to 0.88 (November 1999). In com- parison, developing gametes occupied a low proportion of the gonad, from 0.095 (November 1999) to 0.16 (July 1999). Trabec- TABLE 2. Summary of stains used, and celullar and molecular components targeted. Stain used Structures and molecules targeted Cellular and tissue constituents Acid fuschin Fast green Trioxyhematein PAS-alcian blue Cytoplasmic granules Reticulate fibres, collagen Nucleic acids Neutral and acid mucopolysaccharides glycogen Cytoplasmic granules Connective tissue Nucleus, nucleolus and chromatin Oocyte coat Possible glvcos;en Reproduction in Megathura 303 Figure 1. U. cniuiUilu nonad. Photomicrographs of paraffin-enihcddtd stclion.s. Area ( 1 ) Histdlo^iial siction of male gonad. Modified Masson's trichrome protocol. T: trabecular tissue; DG: developing gametes; MG: mature gametes; FL: flagella; US: unoccupied space. Area (2) Histo- logical section of female gonad. Modified Masson's trichrome protocol. N: nucleus; T: trabecular tissue; CO: coated oocyte; O: non-coated oocyte; C: coat; Nil: nucleolus: PL: proximal layer of oocyte coat; DL: distal layer of oocyte coat; CY: cytoplasm; CH: heterochromatin; CG: unstained cytoplasmic globule. Area (3, 4) Female gonad. Trioxyhematein and alcian blue stains. DL: distal layer of oocyte coat; PL: proximal layer of oocyte coat; N: nucleus; NU: nucleolus; t'H: heterochnmiatin. Area (5, 6) Female gonad. PAS-alcian blue protocol. N: unstained nucleus; DL: distal layer of oocyte coat; PL: proximal layer of oocyte coat; CG: cytoplasmic globule. 304 Beninger et al. 0.4 SBR DLHM I Trabecular tissue Developing gametes feature gametes Unoccupied space Figure 2. M. crenulala. Volume fractions of tissue categories in right (R), left (L), and median (M) gonad regions of tliree males sampled in May 2000. The 95% confidence intervals are too small to be seen. ular tissue occupied a low and stable proportion of the testicle (approximately 0.03). Unoccupied space was rare (Fig. 3). Females Gonad Structure and Oocyte Histochemistry The structure of the female gonad is shown in Figure 1, area 2. Trabeculae consisted of connective tissue (Fast Green positive) and were often masked by the acid fuschin. Uncoated oocytes adhered to the trabecular tissue and were small (approximately 25 (Jim) their cytoplasm was less intensely stained compared to coated oocytes. Unstained regions of trichrome-stained sections corre- sponded to the oocyte coat, which like the oocytes themselves appeared to be of constant dimension when the sectional plane passed through the nucleus. The cytoplasm of coated oocytes stained intensely with both trioxyhematein and acid fuschin. indi- cating the presence of numerous cytoplasmic globules. Many un- stained globules (poorly visible in photographs due to the great staining heterogeneity of the sections) were present in the cyto- plasm of the coated oocytes (Fig. I areas 2, 5, 6). Of the major biochemical tissue constituents. Masson's trichrome does not stain lipids (which are extracted during section preparation) or AMPS; as the globules were not alcian-blue positive (Fig. 1 areas 3-6). they were very probably lipid in nature. The intensity of the cytoplasmic staining obscured the nucleus of those oocytes for which the sectional plane did not pass through the nucleus. When visible, the large nucleus (approximately 75 (xm) presented dispersed heterochromatin and a single nucleolus (Fig. I area 2). The oocyte coat appeared to be composed of two layers: a high-density proximal layer and a lower-density distal layer (Fig. 1 area 2). Staining with alcian blue confirmed this structure, and identified the principal coat constituent as AMPS (Fig. 1 areas 3-6). Counterstaining with trioxyhematein was extensive in the cytoplasm, indicating the presence of large quantities of nucleic acids, suggesting considerable anabolic activity (Fig. I areas 3. 4). The negative PAS reaction indicated both an absence of appre- ciable quantities of glycogen in the oocytes, and an absence of NMPS in the oocyte coat (Fig. I areas 5. 6). Stereology Ovarian homogeneity was verified in June. August. September. November and December 1999, and in February. April, May and June 2000. The near-identical volume fractions for the three iden- tified ovary regions demonstrated the histological homogeneity of this organ (Fig. 4). These data confirm that the female gonad was structurally homogenous and gametogenetically synchronous. As was the case for the male gonad, representative histological data may thus be obtained from a single histological sample per indi- vidual. Similarly, the female gonad showed a stable tissue profile throughout the sampling period. Coated oocytes represented the great majority of the mean volume fraction (Fig. 5). from 0.83 (June 1999) to 0.94 (November 1999). Uncoated oocytes repre- sented a small mean volume fraction, from 0.01 (November 1999) to 0.04 (June 1999). The mean volume fraction of trabecular tissue was also small: 0.035 (May 2000) to 0.12 (June 1999). Oocyte Diameters Given the homogeneity of the ovary, oocyte diameters were pooled for all females of a given sampling date. Mean diameters varied only slightly, from 125 to 135 |xm without the coat (Fig. 6). 0.8 0.4 ■ Trabecular tissue 0 Mature gametes 13 Developing gametes D Unoccupied space Figure 3. M. crenulala. Evolution of tissue volume fractions in males. July 1999 to June 2000. The 95% confidence seen. Data unavailable in October, March, and April due to sampling difficulties. M intervals are too small to be Reproduction in Megathura 305 0.4 IQI I ZR QL MM Trabecular tissue Coated oocytes Non-coated oocytes Figure 4. M. crcniilata. \ olume fractions of tissue categories in right (R). left (U, and median (M) gonad regions of three females sampled in February 2U00. The 95% confidence intervals are too small to be seen. This difference was not statistically significant (parametric ANOVA. normality and heteroscedasticity \enfied. P < 0.05). DISCUSSION To our knowledge, the results of the present study constitute the first report on the gonad structure, reproductive cycle and oocyte histochemistry in Megathura crenidaia. A much more abundant literature exists for the commercially exploited archaeogastropods of the family Haliotidae. to which frequent reference will be made. Gonad Structure The gonad structure of M. creniitata. with gametes developing centrifugally from traversing trabeculae. resembles the well- known example of the Haliotidae (Newman 1967. Young & De- Martini 1970. Cochard 1980l. In all of these cases, no ciliated evacuating ducts were observed; gametes are presumably expulsed via contractions of the gonad tegument, as is in H. luidac (Newman 1967). No simultaneous hermaphrodites were observed in any of the specimens studied: the dominant possible sexual modes for M. creiuilata are therefore either gonochoric or successive hermaph- rodite. Most prosobranchs are gonochoric. but there are a small number of hermaphroditic species (Fretter & Graham 1964. Fretter 1984). Complete resolution of this question in M. creniitata will require extensive sampling and long-term rearing. Gonad structure was shown to be homogeneous for both male and female M. creniitata. as is also the case for Hatiotis inidae (Newman 1967). This result will facilitate future studies on the gonad of this species. By standardizing the histological sampling zone, it should be possible to reduce even further the residual inter-individual variation. Reproductive Cycle The marked stability of both the male and female M. creniitata gonad histological profile throughout the sampling period, as well as the uniform oocyte size, preclude the use of either criterion in determining gamete maturity, or even the state of spawning readi- ness of the gonad. The stable histological profile raises an inter- esting possibility: spawning readiness may depend on fine-tuning oocyte reserves rather than on synchronizing protracted periods of vitellogenesis. This is in contrast to the situation in the Haliotidae. A bimodal oocyte size distribution was observed in H. midae (Newman 1967): in H. roei. three oocyte maturation stages, char- acterized by different diameters, were observed (Shepherd & Laws 1974). Several oocyte sizes co-exist in H. tiibercutata. with pro- gressive growth from January to mid-July (Cochard 1980). Simi- larly, marked changes in histological profile characterize the re- productive cycle of H. iniitiw males (Newman 1967). Whereas mature spermatozoa were found in H. rtifescens. this appeared to correspond to a lack of variation in gonad indices and therefore year-round dribble spawning (Young & DeMartini 1970). Simi- larly. H. asinina synchronously spawns every two weeks on the Southern Great Barrier Reef (Jebreen et al. 2000). In the present case, the gonad appears ready to spawn at any moment in the reproductive cycle. The true state of gamete maturity must there- fore be ascertained by other means, as suggested below. A crude indicator of spawning readiness could involve deter- mination of a gonosomatic index (DeVlaming et al. 1982): indeed, despite the unifomi histological profile, considerable variations in gonad volume were observed over the sampling period in the few indi\iduals dissected for histological processing. However, such a I Trabecular tissue 0 Coated oocytes D Non-coated oocyles 0.8 0.4 J 99 i.llJ o i JOG w^m^m=^ Figure 5. M. crcniilata. Evolution of tissue Milunie fractions in females. June 1999 to June 2(100. be seen. Data unavailable in October and March due to sampling difficulties. M A M J The 95 ''f contldence intervals are too small to 306 Beninger et al. 160 120 - E 3. 0) o 80 E c CD 0) 40 ji j r*i n Si [il r* J 99 O N D J 00 M M Figure 6. E\ulution of A/, creniilata ciiated-oocyte mean diameters (measured witliout their coats) from June 1999 to June 2000. Vertical bars are 95% confidence intervals. Data unavailable in October and March due to sampling difficulties. teL-liniqiie requires the sacrifice of relatively large numbers of ani- mals and is therefore not feasible in the context of the commercial exploitation of this species reared in captivity for repeated extrac- tion of valuable hemolymph. Oocyte Histochemistry The presence of an oocyte coat, observed in mature oocytes of M. creniilata. is typical of archaeogastropods (Newman 1967. Co- chard 1980). In gastropods with coated oocytes, gamones inter- vene in the modification of the coal to permit fertilization (Fretter & Graham 1964. Webber 1977. Fretter 1984). The results of the present study establish AMPS as a dominant component of the oocyte coat in M. creniilata. AMPS possess several chemical and mechanical properties which may confer important advantages to the oocytes: (i) Due to the high viscosity of AMPS (Beninger & St-Jean 1997. Davies & Hawkins 1998). they provide mechanical protection to the oocytes: (ii) AMPS reduce frictional resistance, thus allowing better water movement (Hoyt 1975. Daniel 1981. Davies & Hawkins 1998) over the egg masses, and hence im- proved gas exchange and metabolic waste removal. The reduced frictional resistance would also reduce the probability of dislodg- ing the egg masses; (iii) AMPS possess anti-microbial properties (Sasikala & Subramoniam 1987. Subranioniam 1991. Beninger & Larocque 1998). potentially conferring protection from opportu- nistic microbes in the egg masses; (iv) the relatively high density of the AMPS coats could act to confer negative buoyancy to the otherwise positively buoyant (due to the high lipid content) oo- cytes, allowing the egg masses to remain on the substrate rather than in the water column. This characteristic is important in spe- cies which limit propagule dispersal: further studies on the repro- ductive biology of A-/, creniilata could address this possibility; (v) AMPS adhere to and agglutinate particles strongly, as shown in the context of bivalve particle processing (Beninger & St-Jean 1997). This property would once again reduce the dispersal of the oo- cytes, which are oviposited as egg masses. The lack of positive PAS staining in the oocyte cytoplasm eliminates the possibility of glycogen as a reserve in the oocytes of this species. The chief oocyte reserve appears to be lipid (visible as clear globules in the trichrome-stained sections), as is the rule in the Mollusca (Gallager & Mann 1986. Lucas et al. 1986. Caers et al. 1999. Lu et al. 1999). Large lipid reserves have been reported in the ovaries of both Halintis and Megatliiini genera (Webber 1977). Although histological examination is a well-established tech- nique for the detailed documentation of reproductive cycles (Web- ber 1977. Beninger 1987, Barber & Blake 1991). the results of the present study show that this approach, while very useful for elu- cidating other aspects of the reproductive biology of M creniilata, cannot be used to follow and pinpoint spawning preparedness in this species. However, the eventual rearing of M. creniilata will require this information: even more desirable would be a non- destructive technique of monitoring the reproductive status of broodstock. Such biological monitoring of broodstock could lead to increased fertilization success, and hence increased production of adults for pharmacological use. Several additional aspects of the reproductive biology of M. creniilata which could be usefully pursued include the dynamics of ganietogenesis (especially the transition from small uncoated oo- cytes to large coated ones), the buildup of vitteline reserves, the characteristics of gamete storage, and the mechanisms of gameto- genetic synchronization. Such information will be most helpful in both the management of wild stocks, and in future aquaculture operations. ACKNOWLEDGMENTS We thank Herr O. Kottwitz and Biosyn Arzneimittel GmbH (Fellbach. Germany) for having made research funding available for this project, as well as F. Hennequart for his initial role as intermediarv. LITERATURE CITED Barber. B. J. & N. Blake. 1991. Reproductive physiology. In: S. E. Sham- way, editor. Scallops: biology, ecology, and at|iiaciiltiire. Amsterdam: Elsevier, pp. 377-428. Beninger, P. G. 1987. A qualitative and quantilati\e study of the repro- ductive cycle of the giant scallop, Placopecteii magellanicus. in the Bay of Fundy (New Brunswick, Canada). Can. J. Zool. 65:495-498. REPROnLICTION IN MEGATHL'RA 307 Beninger. P. G. & R. Larocque. 1998. Gonopod legumenlal glands: a new accessory sex gland in the Brachyura, Mar. Biol. 132:435—444. Beninger, P. G. & S. D. Saint- Jean. 1997. The role of mucus in particle processing by suspension-feeding marine bivalves: unifying principles. Mar. Biol. 129:389-397. Briarty. L. G. 1975. Stereology: methods for quantitative light and electron microscopy. Sci. Prog. 62:1-32. Caers. M.. P. Coutteau. K. Cure, V. Morales. G. Gajardo & P. Sorgeloos. 1999. The Chilean scallop Argopeclen purpuranis (Lamarck). 1. Fatty acid composition and lipid content of six organs. Coiup. Biochem. Physiol. 1236:89-96. Cochard, J,-C. 1980. Recherches sur les facteurs determinant la sexualite et la reproduction chez Haliotis iiiherciilala L. These de Doctorat, Uni- versite de Brest France. 167 pp. Daniel, T. L. 1981. Fish mucus: in .\itii measurements of polymer drag reduction. Biol. Bull. 160:376-382. Davies, M. S. & S. J. Hawkins. 1998. Mucus from marme Molluscs. Aclv. Mar. Biol. 34:1-72. DeVlaming. V.. G. Grossman & F. Chapman. 1982. On the use of gono- somatic index. Comp. Biochem. Physiol. .\ 73A:3l-39. Faulkner. D. J. 2000. Highlights of marine natural products chemistry (1972 - 1999). Nat. Prod. Rep. 17:1-6. Fretter, V. 1984. Prosobranchs. In: A. S. Tompa, N. H. Verdonk & J. A. M. van den Biggelaar, editors. The Mollusca. vol. 7: Reproduction. Or- lando: Academic Press, Inc. pp. 1-45. Fretter, V. & A. Graham. 1964. Reproduction. In: K. M. Wilbur & C. M. Yonge, editors. Physiology of mollusca. vol. I. New York. London: Academic Press, pp. 127-164. Gabe. M. 1968. Techniques histologiques. Paris: Masson & Cie. editors. 1113 pp. Gallager. S. M. & R. Mann. 1986. Growth and survival of larvae of A/cr- cenaria mercenaria (L.) and Crassoslrea virginica (Gmelin) relati\e to broodstock conditioning and lipid content of eggs. .Aqiiaciill. 56:105- 121. Hahn. K. O. 1989. Survey of the commercially important abalone species in the world. In: K. O. Hahn. editor. Handbook of culture of abalone and other marine Gastropods. Boca Raton. FL: CRC Press. Inc. pp. 3-11. Harris, J. R. & J. Mark 1. 1999. Keyhole limpet hemocyanm (KLH): a biomedical review. Micron. 30:597-623. Harris. J. R. & J. Markl. 2000. Keyhole limpet hemocyanin; molecular structure of a potent marine immunoactivator. Eur. Urol. 37(suppl 3):24-33. Hoyt. J. W. 1975. Hydrodynamic drag reduction due to fish sliines. In: T. V. T. Wu, C. J. Brokaw & C. Brennan, editors. Swimming and Hying in nature. New York: Plenum Press, pp. 653-673. Ireland, C. M., T. F. Molinski, D. M. Roll, T. M. Zabriskie, T. C. McKee. J. C. Swersey & M. P. Foster. 1989. Natural products peptides from marine organisms. In: P. J. Scheuer, editor. Bioorganic marine chem- istry, vol. 3. Berlin: Springer-Verlag. pp. 1-46. Jebreen, E. J.. R. T. Counihan. D. R. Fielder & B. M. Degnan. 2000. Synchronous oogenesis during the semilunar spawning cycle of the tropical abalone Haliotis asiiiina. J. Shellfish Res. 19:845-851. Kobayashi. M.. J. Kobayashi & Y. Ohizumi. 1989. Cone shell toxins and the mechanisms of their pharmacological action. In: P. J. Scheuer. editor. Bioorganic marine chemistrs, \ol. 3. Berlin: Springer-Verlag. pp. 71-84. Lu. Y. T,. N. J. Blake & J. J. Torres. 1999. Biochemical utilization during embryogenesis and metamorphosis in the bay scallop. .Argopecten ir- radiaiis concentricus (Say). J. Shellfish Res. 18:425—429. Lucas, A., L. Chebab-Chalabi & D. Aldana Aranda. 1986. Passage de I'endotrophie a I'exotrophie chez les larves de Mytilus edulis. Oceanol. Acta 9:97-103. Martoja, R. & M. Martoja-Pierson. 1967. InitiatLon au\ techniques de I'histologie animale. Paris: Ma.sson & Cie. 345 pp. Mayhew, T. M. 2000. 3D structure from thin sections: applications of stereology. Eur. Microsc. Anal. November: 17-20. Morvan. C. & A. D. Ansell. 1988. Stereological methods applied to re- productive cycle of Tapes rhomboides. Mar. Biol. 97:355-364. Munro, M. H. G.. R. T. Luibrand & J. W. Blunt. 1987. The search for antiviral and anticancer compounds from marine organisms. In: P. J. Scheuer, editor. Bioorganic marine chemistry, vol. 1. Berlin: Springer- Verlag. pp. 93- 176. Newman. G. C. 1967. Reproduction of the south african abalone Haliotis inidae. Inrestl. Rep. Div. Sea Fish. S. Afr. 64:1-24. Pazos. A. J.. G. Roman. C. P. Acosta, M, Abad & J, L. Sanchez. 1996. Stereological studies on the gametogenic cycle of the scallop. Pecien nui-ximus. in suspended culture in Ria de Arousa (Galicia. NW Spain). Acpiaculture 142:119-135. Purchon. R. D. 1977. Reproduction. In: R. D. Purchon. editor. The biology of Mollusca. 2nd ed. Oxford: Pergamon Press, pp. 269-332. Sasikala, S. L. & T. Subramoniam. 1 987. On the occurrence of acid mu- copolysaccharides in the spermatophores of two marine prawns, Pe- naeus indicus (Milne-Edwards) and Megapenaeus monoceros (Fabri- cius) (Crustacea: Macrura). / Exp. Mar. Biol. Ecol. 113:145-153. Shepherd, S. A. & H. M. Laws. 1974. Studies on southern australian abalone (Genus Haliotis). Part II: Reproduction of five species. Aust. J. mar. Freshwul. Res. 25:49-62. Subramoniam, T. 1991. Chemical composition of spermatophores in de- capod crustaceans. In: T. Bauer & J. W, Martin, editors. Crustacean sexual biology. New York: Columbia University Press, pp. 308-321. Suffness. M.. D. J. Newman & K. Snader. 1989. Discovery and develop- ment of antineoplastic agents from natural sources. In: P. J. Scheuer, editor. Bioorganic marine chemistry, vol. 3. Berlin: Springer-Verlag. pp. 131-168. Tyler, A. 1939. Extraction of an egg membrane-lysin from sperm of the giant keyhole limpet (Megathura crenulala). Proc. Natl. Acad. Sci. USA 25:317-323. Vacca, L. L. 1985. Laboratory manual of histochemistry. New York: Raven Press. 578 pp. Webber. H. H. 1977. Gastropoda: Prosobranchia. In: A. C. Giese & J. S. Pearse. editors. Reproduction of marine Invertebrates. Vol IV: Mol- luscs: Gastropods and Cephalopods. New York: Academic Press, pp. 1-97. Weibel, E. R., G. S. Kistler & W. F. Scherle. 1966. Practical stereological methods for morphometric cytology. / Cell. Biol. 30:23-38. Young. J. S. & J. D. DeMartini. 1970. The reproductive cycle, gonadal histology, and gametogenesis of the red abalone. Haliotis rufescens (Swainson). Calif. Fish and Game 56:298-309. Journal nf Slu-llfish Research. Vol. 20. No. 1. 309-315. 2001. MODELING THE GROWTH OF THE CHH^EAN LOCO, CONCHOLEPAS CONCHOLEPAS (BRUGUIERE, 1789) USING A MODIFIED GOMPERTZ-TYPE FUNCTION LUCIANO RODRIGUEZ." * GIOVANNI DANERl/ CRISTIAN TORRES,' MATIAS LEON,' AND LEONARDO BRAVO' ^Escuela dc Pesqiierias y Ciiltivos. Universidad del Mar. Carmen 446. Cerro Placeres. Vcdparaiso. Chile: 'Centra de Clencias y Ecoloiiia Aplivada (CEA). Universidad del Mar. Carmen 446. Cerro Placeres. Valparaiso, Chile ABSTRACT A modified Gompertz-type function was fitted by nonlinear numerical methods to the absolute growth rates(mm/day) of the marine gastropod ConclmleiHis coneholepas. known in Chile as "loco." The data for this study were obtained from laboratory and field experiments as well as from published and unpublished reports of loco growth. The fit of the modified Gompertz-type function was compared to the fit obtained with the commonly used von Bertalanffy function using the Akaike uiformation criteria corrected by sample size bias (AIQ). The modified Gompertz function used in this study was ilL/ilt = (a,^''' )*(e.xp~"-'') + a,, where L is the penstomal length (mm), and «(, . a, a^ and u,. are the parameters of the nnxlel. The biological meaning of the constant a, of the function is equivalent to the asymptotic growth rate for the species. The von Bertalanffy growth rate equation can be expressed as dUdt = - k (L.^ - L). This equation represents a straight line with a negative slope equal to -k and with the >-axis intercept ilL/iir = 0. providing an estimate of L^. The fitted parameters obtained by fitting a modified Gompertz-type of model to the pooled growth rate data were a„ = 2.93O0e-05, a, = 4.2605. a, = 0.201 1. and a< = 0.02852 with a degree of freedom corrected /' = 0.595. For the von Benalanffy function the fitted parameters were k = 5.943e-04 (day"') and L, = 150.6 mm with r = 0.387. The AlCe of the Gompertz fit was -6,086; whereas, the AlCe of the von Bertalanffy fit was -5.61 1 . The smaller AICc value -6.086 indicates that the modified Gompertz function provides a better fit to the loco growth rate data. The form of the projected modified Gompertz-type growth curve differs notably from the traditional von Bertalanffy curve for juvenile growth. The modified Gompertz model does not reach an asymptotic growth level, predicting instead infinite growth. This latter property could easily explain reported historical findings of loco individuals measuring o\er 170 mm in peristomal length, KEY WORDS: Growth rate. Gompertz model, intcrtidal gastropods, aquaculture INTRODUCTION The loco (Coneholepas coneholepas Bruguiere. 1789) is a mu- ricid gastropod endemic to the Southeastern Pacific, found along the coasts of Pert and Chile (Bustos et al. 1986). Because of the great economic importance attained by the loco in the past two decades, its natural stocks have been under intense exploitation pressure. This has prompted the Chilean Government to enact legal measures aimed at protecting this species by regulating the capture effort and by establishing a minimum legal size of extraction ( 100 mm. peristomal length) (Montt et al. 1977). The locos are mainly exploited by the artisanal sector, which adds a social dimension to its economic importance. The factors above have made the loco one of the better-studied marine gastropods in Chile, with most of the research effort directed toward commercial farming of this species. Growth-rate studies of loco juveniles have been carried out in the laboratory (Mendez & Cancino 1992), in suspended lantern nets ( Varela & Perez 1987), and in the field, especially in intertidal rocky shores (Guisado & Castilla 198.^. Bustos el al. 198.5. Bustos et al. 1986. Adierstein 1986). The early studies of age and growth of loco were initially descriptive and were not used to produce mathematical growth models (Tobella 1975. Lozada el al. 1976. Acufia & Stuardo 1979). Later studies were oriented toward sat- isfying the management requirements of the loco fishery. Initially, several authors applied linear functions to describe the growth of loco juveniles (Guisado & Castilla 1983). More recently. Reyes and Moreno (1990) employed linear and exponential models to *Corresponding author Current address: Escuela de Pesquerilas y Culti- vos. Universidad del Mar. carmen 446. Cerro Placeres. Valparaiso. Chile, Tel.: -h56-32-501839; Fax: ■I-56-32-798731. E-maii: lrodngu@udelmar,cl describe the growth of juveniles under 10 mm length from the Mehui'n marine reserve in the X Region of Chile (Fig. I ). How- ever, the function most widely used to model loco growth and to regulate its fishery is the von Bertalanffy (1957) growth equation (Bustos et ((/. 1985. Bustos et al. 1986. Adierstein 1986. Castilla & Jerez 1986. Lepez 1987). The von Bertalanffy growth equation can be expressed as dUdt = - k (L._,_ - L). This differential equation represents a straight line with a negative slope equal to - k and with the y-axis intercept dUdt = 0. providing an estimate of L-^_. Use of the von Bertalanffy equation is limited, because often it does not adequately reflect the growth of immature or juvenile individuals. In addition, in its original form, the von Bertalanffy equation does not include the seasonal oscillation of growth, an important factor if the growth of organisms is to be described on short time scales (Pauly & Gaschutz 1979, Gaschutz et al. 1980, Akamine 1986, Allison 1994, Askew 1995) and specifically for intertidal gastropods, (Ekaratne & Crisp 1984. Santarelli & Gros 1985. Noda 1991, Shepherd et al. 1995). An alternative to the von Bertalanffy function is the Gompertz function, which in a modified version, has been used to model individual growth rates of juvenile and adult abalones (Haliotis asinina) (McNamara & Johnson 1995). pre- and postlarval stages and adult fishes (Smith & Kostlan 1991: Williams & Lowe 1997: Zweiffel & Lasker 1976) and crustaceans (Misra 1957). In this paper, data from growth experiments of loco individuals grown in the laboratory and in suspended lanterns, as well as data obtained fiom a wide variety of literature sources, were pooled together employing meta-analysis techniques. The use of meta- analysis makes it possible to integrate and summarize information coming from a wide variety of sources. As applied to stock as- sessment, meta-analysis involves the compilation of pre-existing 309 310 Rodriguez et al. [ RADA DE QUINTAY Figure 1. Geographic distribution of Concliolepas coiicholipas in Ciiile and location of study area Quintay Bay, V region. datasets to determine tlie \aliies of parameters of models or to develop prior probability distributions of these parameters (Cooper & Hedges 1994, CFSAM 1998). This approach permits a test of the hypothesis that a modified Gompertz equation provides a better model of the loco growth curve than the commonly used von Bertalanffy function. It is hoped that the development of more accurate loco growth models will improve the management of this commercially and socially important marine resource. MATERIALS AND METHODS The data used in this study were taken from growth studies of loco juveniles carried out in laboratory conditions and in sus- pended lanterns anchored in the Quintay Bay (33°ir S; 7142' W). V Region of Chile (Fig. I). Loco juveniles were collected from the intertidal rocky shore. Individuals were marked with circular plastic tags (2-mm diameter) and the peristomal length (Fig. 2) was measured with a precision caliper (± O.OS-mm). The laboratory growth experiments of loco juveniles were car- ried out in two 40-L glass tanks (0.5 x 0.4 x 0.4), using the facilities of the School of Fisheries. Universidad del Mar. Each tank was furnished with independent airflow systems powered by Elite 8t)0 electrical pumps with a 2.10()-cc air/min capacity. Dur- ing the study, the water temperature in the aquarium tanks was 20°C ± 2°C. Growth controls throughout the study were taken on days 19. 34. 3.5, 69. and 90. Loco individuals grown in laboratory and in culture systems were fed ad lihinnn with a monospecific diet, based on live intertidal mus.sels. Peiuniytihis piirpiinitiis (Lamarck 1819), one of the main prey items of the locos in their natural environment (Castilla et al. 1979). The initial peristomal length-size range of the locos grown in laboratory conditions was 24 to 64 mm. The growth experiments in suspended lanterns were carried out in the Quintay Bay during two periods lasting 34 and 140 days, respectively. The loco growth lanterns were especially designed, having a hard plate made of fiberglass that offered an anchor substrate to the locos (Fig. 3). The initial peristomal size range of the loco juveniles grown in lanterns was in the same peristomal size range (24 to 64 mm). The size increments were calculated using the absolute growth rate, mathematically represented by: AL It' (1) where L, and L, are the final and initial peristomal lengths and r^ and r, are the final and initial times corresponding to the afore- mentioned lenszths. The calculated izrowth rates values were fitted Growth Rath of Conchqlepas concholepas 311 ^ where L is peristomal lengtli (mm) and. a,,, ci,. <;, and Oj are the fitting parameters where a, is equivalent to the asymp- totic growth rate for the species. The first term of this expres- sion {a„ L "' ) is analogous to the power allometric equation proposed by Parker and Larkin (1959); whereas, the second (exp ~ "- ' ) is a factor that produces a decreasing growth rate beyond the maximum value of the function. This inflection point is represented by the size of individuals at which sexual inaturity is reached. The von Bertalanffy function is expressed as: dL at (3) where L^ and k are the a.symptotic peristomal length and the cur- vature parameters, respectively. The selection of the best-fit growth cur\ e was made based on information theory (Akaikel974. Burham & Anderson 1998). The specific tool used for model comparison and selection was Akaike's information criteria corrected for sample size bias. AIC^ defined as: I- — Penstoinallength (Pl) — -I Figure 2. Diagram showing a loco {Concholepas concholepas). to the modified Gompertz and von Bertalanffy functions using the Levemberg - Marquardt algorithm fVoni the Table Curve 4.0 pack- age (SPSS 1998). The modified Gompertz function is expressed as: ''^ „ „/ — = ((/uL"')*(exp "- )-Ki3 (2) Figure i. Diagram ol (In suspended lantern net designed lo grow locos in natural conditions. AlCr = n*]oeiMSE) + 2K + 2K{K+ I) in- K- \} (4) where MSE is the mean square error from the model, ii is the number of datapoints. and K is the number of parameters to be estimated. The first term of this expression reflects the lack of fit, the second is a penalty term for complexity or lack of parsimony of the model, and the third term is a correction factor for sample size bias. Lower A/Q values indicate a better model fit (Burham and Anderson 1998). If we know the functional relationship between growth rate and length, an individual's body size as a function of age or time frame can be obtained by reorganizing Eq. 2 and integrating: J-/ ,/Z. u> (a,, Z."') * (exp "''l-hrt, (5) where L, is the final peristomal length (mm) of the individual at age or time t. and L„ is the initial peristomal length (mm) of the individual at the beginning of the experiinent or when r = 0. Equation 5 cannot be solved analytically; however, a solution can be obtained numerically using a numerical integration subroutine available in any specialized mathematical software package. Because it may be convenient to use a simpler and more prac- tical method for growth projections in the farming business, we also use a recursive-type approximation based on a discrete time interval St. which is mathematically represented as: L, = L 2 '"0 L'," ■ e\p "-'-' + a j) * \t (6) where Z,, is the peristomal length in the j-th period, and L,,; is the peristomal length in the (j-J )-lh period. The term between paren- theses is the growth rate expressed in mm/day of an individual in the (j-l )-lh period, which is dependent on length and the discreet increment L, = (/, ,_, -i- IL ,_,), and Si is the timestep interval in days. The timestep utilized was one day, and the unit growth period utilized was monthly. The loco growth data from this study were pooled with pub- lished and unpublished loco growth dalasets (Table 1 ). This al- 312 Rodriguez et al. TABLE 1. Published and unpublished data sources for growth rate of locos used in this study. Data source (author and year) Peristomal size range (mm) Growth rate range or average (mm/day) Locality (region) Rabi & Maravi (1997) Stotz& Perez (1992) Gallardo(1979) Giocochea et al. (1991) Castilla (1983) Lepez (1987) Lepez et al. (1991) Olivares et al. (1990) Varela& Lopez (1989) Lozadaet al. (1976) Rivas (1994) Mora (1994) Tobella (1975) Reyes & Moreno ( 1990) lFOP(1992) Acuiia & Suiardo (1979) Guisado & Castilla (1983) Castilla & Jerez (1986) Mendez & Canciiio (1992) This study 25-88 0.115 SD = 0.079 0.107 SD = 0.050 10-90 0.016-0.250 1.8-30 0.104 <10 0. 1 30 >10 0.103 14-15 (I.IOO 14-15 0.1 10 Not available 0.041-0.071 7,05 0.061 8-20 0.018-0.054 Not available — 54-63 0.073-0.200 63-81 0.100-0.106 63-90 0.100 44-64 0.039-0.071 36-65 0.063-0.073 14-113 0.045-0.300 1.9-10 0.022-0.058 10-145 0.008-0.086 mean = 0.0284 SD = 0.0197 20 0.100 11-49 0.05-0.232 40-110 0.01.3-0.050 5-12 0.046-0.075 24-64 0.015-0.117 mean = 0.070 SD = 0.027 24-64 0.020-0.093 mean = 0.038 SI) = 0.014 lie Port Catarindo Bay (Peru) Tagged in situ Pta. Lagunillas (IV) Modal progression and tagged Mehuin (X) Aquaria laboratory Mehuin (X) Calela Leandro (Vlll) Tagged in .\ilii Mehuin (X) Ramuntcho (VIII) Aquaria laboratory (Vlll) Suspended system .Aquaria laboratory Modal progression Pta. Saliente (IV) Cale(a Leandro (VIII) Talcahuano (Vlll) Suspended system Caleta Lirquen (VIII) Suspended systeiu Caleta Lirquen (VIII) Modal progression Caleta Leandro (VIII) Modal progression Mehuin (X) Tasked in siiit Piinta Corona (X) Modal progression Montemar (V) Modal progression Las Cruces (V) Las Cruces. El Quisco. Quintay (V) Aquaria laboratory Las Cruces & El Quisco (V) Suspended system Quintay Bay (V) Aquaria laboratory lowed a substantial increase in the size distribution of the data fitted to the von Bertalanffy and the modified Gompertz growth function. RESULTS AND DISCUSSION The average growth I'ate of loco individuals in the laboratory (0.038 mm/day and standard deviation 0.014 nim/day). was about one-half the average growth rate for individuals grown in sus- pended lantern nets in the field (0.070 nint/day and standard de- viation of 0.027 mm/day). The growth rate ranges observed in the laboratory and in the field are, however, in ugreement with growth rates reported for other niuricid species by Spight ci id. (1974). The observed differences between growth rates in the laboratory and suspended lanterns probably reflect better growth rate condi- tions in the field. The locos grown in suspended lanterns were stibjected to lower and more variable changes of ambient tempera- ture (range 10°C -20''C) than the locos grown in the laboratory. Higher temperatures have been associated with higher growth rates in locos grown under laboratory conditions (Mendez & Can- cino.1992); therefoi'e, the higher growth rates obtained under colder field conditions would indicate that factors other than tem- perature should be used to explain the observed growth rate dif- ference between laboratory- and lantern-grown locos. This con- clusion must be reached with caution, however, because the results reported by Mendez & Cancino (1992) (Table 1) are for juvenile locos below the 20-mm peristomal size length; whereas, the piesent study included indi\ iduals in the 24 to 64-mm size range. For the same reason, growth rates obtained in the laboratory in this study are not directly comparable to other laboratory studies of TABLE 2. Fitted modified Gompertz and von Bertalanffy parameters for loco growth rales. Standard Function Parameters error /-value AICc Gompertz modified a„ = 2.69e-()5 1 .625e-05 1.654717 -6086.8 a, = 4.094 0.276217 14.82061 a, = 0.183 0.010620 17.25293 a, = 0.0281 0.000777 36.00687 Von Bertalanffy k = 5.943e-()4 2.722d-05 21.82896 -5611.1 L.. = 150.60 3.138003 48.00212 Note. The /-values serve as indicators of the degree of certainty with the parameters are determined. The parameter with the highest /-value, in addition to having the greatest contribution to the fit, will also been de- termined with the greatest level of certainty (SPSS 1998). Growth Rate of Concholepas concholepas 313 50 100 Peristomal length (mm) Figure 4. Gompcrtz curve lit to growth rate of locos. 50 100 Peristomal length (mm) Figure 5. Von Bertalanffy curve fit to growth rate of locos. loco growth reported by other authors: 0.103 - 0.130 mm/day (Goicochea et al. 1991). 0.018-0.058 mm/day (Olivares et at. 1990) and 0.046 - 0.075 mm/day (Mendez & Cancino 1992), mainly because these authors worked with loco individuals below the 20-nini size range. Reyes and Moreno ( 1440) obtained values from 0.022 to 0.043 mm/day for natural loco populations (<20 mm umbal lengths) from the Mehuin marine reserve in the X Region of Chile. This would indicate that environmental pressure can depress the potential juvenile loco growth in the field. The growth rates reported here are in better agreement with results reported by Rabi and Maravi' (1997) for culture experiments in Peru and with growth data reported by Castilla and Jerez ( 1986) for natural loco population in the V region of Chile. Similar growth rates (0.039-0.073 mm/day) were obtained by Mora (1994) and Rivas (1994) in culture experiments in Caleta Lirquen in the VIII region of Chile (Table 1 ). The fitted parameters for the pooled growth rate data obtained during this study and from published and unpublished data for the Gompertz- function (Table 2 ; Fig. 4) were a,, = 2.9300e-05. a , = 4.2605. a. = 0.201 1 . and a, = 0.02852 with a degree of freedom corrected r" = 0.595 and AICc = -6,086. For the von Bertalanffy function (Table 2 ; Fig. 5), the fitted parameters were k = 5.943e- 04 (day-') and L,, = 150.6 with r- = 0.387 and A/Cc = -5611. The lower AIC^ value for the modified Gompertz function indi- cates that this function provides a better fit. Although the corrected r" only accounts for 59.5% of the variability of the data, this is, nonetheless, better than the 38.7% obtained with the von Berta- lanffy model. 300 250 200 ■S 150 100 50 1 1 ^ y^ ^ / ■ii^' -f=^ / ^ o jompertz function /on Bertalanffy function Field Data V region k age- age: ■12-120 month ( IFOP jnpublished data) 0 50 100 150 200 Age (month) Figure 6. Crovvth curve of peristomal length (mm) at age (months), for locos measured in natural enviroment. Gompertz function I- ■ Berlalanflv function (-.-.-). I and von 314 Rodriguez et al. The inflection point of the fitted growth-rate curve or point of maximum growth rate (Fig. 4) represents the individual's change of state from a sexually immature to a sexually mature individual. The energy budget of immature individuals is exclusively destined to growth, but when sexual maturation occurs, individuals redirect energy toward the production of gonads, resulting in a marked reduction in growth rates. This inflection point indicates that locos reach sexual maturity at about 25 to 30-mm peristomal length. This is well below the 50-mm minimum size value reported by Ramorino ( 1979), and it may well reflect a behavioral shift toward earlier time of sexual maturity in the loco population, which, dur- ing the last two decades, has been subject to a more intense level of exploitation. A growth projection of this species was performed as a function of time by using the discrete character expression represented by Eq. 6 and compared to the same projection of the fitted von Ber- talanffy growth curve from this study (Fig. 6). For the sake of comparison, the sizes of individuals under a year reported by Guisado and Castilla (1983). and the mean length between one and ten years (IFOP unpublished data) for the Region V are superim- posed. The growth projection using the modified Gompertz func- tion indicates that the legal size of capture of 100-mm peris- tomal length, should be reached in 66 months. The modified Gompertz-type model does not reach an asymptotic size as pre- dicted by the von Bertalanffy expression. Instead, the projected growth, especially for individuals over a year old. has a resulting curve that is notably similar to the infinite growth pattern de- scribed for some other marine invertebrates (Tanaka 1982. Tanaka 1988. Phillips et al. 1983. 1992, Ariyama 1993. McNamara & Johnson 1995). This could easily explain historical findings of loco individuals measuring over 170 mm in peristomal length. Further- more, the fitted von Bertalanffy curve does not adequately describe the ju\enile loco growth pattern. Our results indicate that the modi- fied Gompertz equation provides a stronger model of the loco growth cur\e than the commonly used \on Bertalanffy function. Based on these results, we propose that the modified Gompertz equation described in this study should be used routinely in the management models of the commercially and socially important loco species. ACKNOWLEDGMENTS The contribution of anonymous reviewers helped to improve this manuscript greatly. We thank the Fondap-Humboldt Research program grant to G. Daneri. LITERATURE CITED Acuna. E. & J. Stuardo. 1979. Una estimacion de cla.ses anuales y crec- imiento relative en muestras de dos poblaciones de Conchotepas coii- cholepas (Bruguiere. 1789). Biiilogia Pesquera Chile 12:131-142. Adierstein, S. 1986. Age and growth of Concholepas concholepas (Bru- guiere. 1789). MS Thesis. University of Washington. Seattle. 179 pp. Akaike. H. 1974. A new look at the statistical model identification. IEEE Trans. Aittom. Contr. AC-I9:716-72.\ Akamine. T. 1986. Expansion of growth curves using a periodic function and BASIC programa and Marquardl's Method. Bull. Japan. Sea Reg. Fis. Res. Lab. 36:77-107. Allison, E. H. 1994. Seasonal growth models for great scallops [Pecien ma.>:imus (L.)] and queen scallops. [.Aeipiipecten opeixularis iL.)] J. Shellfish Res. 13(2): 555-564. Ariyama. H. 1993. Growth of swimming crabs Ponunus (Pominns) triruberciilatiis in Osaka Bay. Nippon Siiisan Gakkaishi. 59(8):!269- 1277. Askew, C. G. 1995. A generalized growth and mortality model for assess- ing the economics of bivalve culture. Research Paper 3. Centre for the Economics and Management of Aquatic Resources. (CEMARE). Uni- versity of Portsmouth. 14 pp. Burham. K. P. & D. R. Anderson. 1998. Model selection and inference: a practical information theoretic approach. New York: Springer. 353 pp. Busies, E., E. Pacheco, A. Zabala, E. Lara, H. Robotham. A. Reyes & A. Augsburger. 1985. Estudio de ed.id y crecimiento del recurso loco {Concholepas concholepas) en la 111 y X region. Intorme IFOP a CORFO (10/85). 42 pp. Bustos. E.. H. Robotham. E. Lara & E. Pacheco. 1986. Edad y crecimiento de Concholepas concholepas y consideraciones a la aplicacion de la ecuacion de von Bertalanffy (Gastropoda -Muricidae). Invesrigacion Pesquera (Chile) 33:33-45. Castilla, J. C, Ch. Guisado & J. C. Cancino. 1979. Aspectos ecoldgicos y conductuales relacionados con la alimentacion de Concholepas conc- holepas (Mollusca: Gastropoda: Muricidae) Biologfa Pesquera. Chile 12:99-114. Castilla. J. C. 1983.Pesqueri'a de moluscos gastropodos en Chile: Conc- holepas concholepas. un caso de estudio. In: P. Arana, editor. Analisis de pesquen'as Chilenas. Escuela de ciencias del mar. Valparaiso: Uni- versidad Catolica de Valparaiso, pp. 37-50. Castilla, J. C. & G. Jerez 1986. Artisanal fishery and development of a database for managing the loco Concholepas concholepas. resource in Chile. In: G. S. Jamieson & N. Bourne, editors. North Pacific Work- shop on Stock Assessment and Management of Invertebrates. Cana- dian Special Publication Fish. Aquat. Sci. 92:133-139. CFSAM. Committee on Fish Stock Assessment Methods 1998. Improvini; Fish Stock Assessments.The National Academy of Sciences, ihttp:// w\vw.nap.edu/openbook/0309057256). Cooper. H. & L. V. Hedges. 1994. The handbook of research synthesis. New York: Russell Sage Foundation. Ekaratne, S. U. K. & D. J. Crisp. 1984. Seasonal growth studies of inter- tidal gastropods from shell micro-growth band measurements, includ- ing a comparison with alternative methods. J. Mar. Biol. Ass. U.K. 64:183-210. Gallardo, C. 1979. El ciclo vital del Muricidae Concholepas concholepas y consideraciones sobre sus primeras fases de vida en el bentos. Bio- logia Pesquera Chile 12:79-89. Gaschutz. G.. D. Pauly & N. David. 1980. A versatile BASIC program for fitting weight and seasonally oscillating length growth data. Int. Coun- cil for the Exploration of the Sea. CM. 1980/D:6. Statistics Committee. 23 pp. Goicochea, O.. V. Aravena & R. Jaramillo. 1991. Curva de crecimiento de juveniles de Concholepas concholepas en condiciones de laboratorio. Resiimenes XI Jornadas de Cienciasdel Mar. 57-58. Guisado, Ch. & J. C. Castilla. 1983. Aspects of the ecology and growth of an intertidal ju\enile population of Concholepas concholepas (Mol- lusca: Gastropoda: Muricidae) at las Cruces. Chile. Mar Biol. 78:99- 103. IFOP. 1992. Diagnostico de las Principales Pesquen'as Nacionales 1992. Pesquen'as Bentiinicas III,-IV-.X Region. Estado Situaciiin y Perspec- tivas del recurso. SGI -IFOP 93/7. 75 pp. Lepez. M. 1. I987.EcoIogia intermareal de Concholepas concholepas (Brugiere. 1789) bajo dos regi'menes de intervencion antrdpica. M.S. thesis. Universidad Austral de Chile. 136 pp. Lepez. I.. O. Aracena. O. Olivares & G. Pena. 1991. Reclutamienlo de Concholepas concholepas en el intermareal de Ramuntcho. Octava region. Resiimenes XI Jornadas de Ciencias del Mar. 43 pp. Lozada. E.. M. T. Lopez & R. Desqueyroux. 1976. Aspectos ecologicos de Growth Rate of Concholepas concholepas 315 poblat-iones chilenas de loco Concholepas concholepas (Bruguiere, I7S9) (Mollusca. Gasteropoda. Muricidae). Bioloi;ia Pesqneia Chile 8: 5-29. McNamara. D. C. & C. R. Johnson. 1995. Growlh of the Ass's Ear abalone {Haliolis asinma Linne) on Heron Reef. Tropic. East. Australia. Mar. Freshwater Res. 46:571-574. Mendez. M. A. & J. M. Cancino. 1992. CreLinucnlo de ejeniplares post- nietamorficos y juveniles de Concholepas concholepas (Bruguiere) en condiciones de laboratorio. Efecto de densidad. dieta. y temperatura. Biolojiia Pesqiiera 21: 21-30. Misra. R. K. 1957. An expression for the growth coefficient a in the law y = b*xa of constant differential growth ratio, expressing the growth relationship between the body size x and the organ size y, in various organic forms. Proc. Nat. Inst. Sci. India, 23 B(l-2):42^7. Montt, I. & Cia. S. A. Abogados. 1997. Compilacion de Regulaciones Pesqueras Extractivas Vigentes. Actualizado al 5 de Febrero de 1997. 5()fi pp. Mora. R. F. 1994. Intluencia de la alinientacion en el crecimienlo del recurso loco Concholepas concholepas (Bruguiere. 1879). niantenido en sistenia suspendido de cultivo. Infornie de Practica para optar al litulo de Tecnologo en Recursos del Mar. Pontificia Universidad Catolica de Chile. Sede Regional Talcahuano. Departamento de Cien- cias del Mar. 45 pp. Noda, T. 1991. Shell growth of the sand snail. Umboniiim costatum (Kiener) in Hakodate Bay. Bull. Fac. Fish. Hokkaido Univ. 42(4): 1 15- 125. Olivares. O. I. Lepez & O. Aracena. 1990. Concholepas concholepas: consumo y crecimiento de juveniles mantenidos en acuarios.Re- siimenes X Jornadas en Ciencias del Mar. 75 pp. Parker, R. R. & P. A. Larkin. 1959. A concept of groulh m fishes. J. Fish. Res. Bd. Can. 16:721-745. Pauly, D. & G. Gaschiitz. 1979. A simple method for fitting oscillating length growth data, with a program for pocket calculators. Int. Council for the Exploration of the Sea. CM. 1979/G:24. Demersal Fish Com- mittee. 24 pp. Phillips. B. F., L, M. Joll. R. L. Sanland & D. Wright. 19S3. Longevity, reproductive condition, and growth of the Western Rock Lobster. Paindirns cyiiniis George, reared in aquaria. Aust. J. Mar. Freshn: Res. 34:419-429. Phillips. B. F.. M. J. Palmer R.Cruz & J. T. Trendall. 1992. Estimating growth of the Spiny Lobters. Paimlirus cygims. P. ari;ns. and P. or- ixatus. Aust. J. Mar. Freshwater Res. 43:1 177-1 188. Rabi, M. & C. Maravi 1997. Growth curves and specific growth rate of Concholepas concholepas (Bruguiere. 1789) (Gastropoda: Muricidae) in culture experiments. Scientia Marina 61 (Supl. 2):49-53 pp. Ramorino. L. 1979. Conocimiento cientiTico actual sobre reproduccicin y desarrollo de Concholepas concholepas. Biolojiia Pesqiiera 1 2:59-70. Reyes. A. & C. Moreno. 1990. Asentamiento y crecimiento de los primeros estadios bentonicos de Conclutlepas concholepas (Mollusca: Muri- cidae) en el intermareal rocoso de Mehuin. Chile. Revista Chilena de Historia Natural 63:157-163. Rivas, H. E. 1994. Cultivo suspendido en bandejas para el recurso loco Concholepas concholepas (Bruguiere. 1879) en Caleta Lirquen. bahia Concepcion. Chile. Informe de Practica para optar al ti'tulo de Tecnol- ogo en Recursos del Mar. Pontificia Universidad Catolica de Chile. Sede Regional Talcahuano. Departamento de Ciencias del Mar. 38 pp. Santarelli. L. & P. Gros. 19X5. Determination de Page et la croissance de Buccinwn nndatiini L. (Gasteropoda: Prosobranchia) a I'aide des iso- topes stables de la coquille et de 1' ornamentation operculaire. Oceano- logicaActa. 8(2):22 1-229. Smith M. K & E. Kostlan. 1991. Estimates of age and growth of Ehu Etelis carhimclus in four regions of the Pacific from density of daily incre- ments in otoliths. Fisher. Bull. US. 89:461-472. Seber. G. A. F. & C. J. Wild. 1989. Nonlinear regression. New York: Wiley-Interscience. 767 pp. Shepherd. S. A. D. Al-Wahaibi & A. R. Al-Azri. 1995. Shell growth checks and growth of the Omani abalone Hidiolis mariac. Mar. Fresh- water Res. 46:575-582. Spight. T. M.. C. Birkeland & A. Lyons. 1974. Life histories of large and small Mure.xes (Prosobranchia: Muricidae). Mar. Biol. 24:229-242. SPSS. 1998. Table Curve 2D Automated Curve Fitting & Equation Dis- covery, version 4.0 for Windows 95. NT & 3.1. Users Manual. Stotz, W. & E. Perez. 1992. Crecimiento y productividad del loco Conc- holepas concholepas (Bruguiere. 1789) como estimador de la capacidad de carga en areas de manejo. Invest. Pesq. 37: 13-22. Tanaka. M. 1982. A new growth curve which expresses infinitive increase- . Pubis. Amakusa Mar. Biol. Lab. (>:\(>1-\11. Tanaka. M. 1988. Eco-physiological meaning of parameters of ALOG growth curve. Pubis. Amakusa Mar. Biol. Udi. 9:103-106. Tobella. G. 1975. Crecimiento de Concholepas concholepas (Bruguiere, 1789) (Moll. Gast. Muricidae), Boletin Sociedad de Biologia de Con- cepcion 49:185-189. Varela. C. E & E. Perez. 1987. Cultivo de Concholepas concholepas B. en sistemas suspendidos y estanques en el CEACIMA-METRI. Re- siimenes VIII Jornadas en Ciencias del Mar, 147 pp. Varela C. E & D. A. Lopez. 1989. Manejo de los reproductores de Con- cholepas concholepas (Bruguiere. 1789) en el disefio de una estrategia de repoblacion. Medio Ambienle. Chile 10( 1 ):3-12. Von Bertalanffy, L. 1957. Quantitative laws in metabolism and growth. Quart. Rev. Biol. 32:217-231. Williams, H. A. & M. K. Lowe. 1997. Growth rates of four Hawaiian deep slope fishes: a comparison of methods for estimating age and growth from otolith microincrement. Can. J. Fish. Aijuat. Sci. 54:126-136. Zweiffel. J. R. & R. Lasker. 1976. Prehatch and posthatch growth of fishes— a general model. Fish. Bull. 74(3): 609-621. Joiirmil iH' Shellfish Research. Vol. 20. No. I. 317-321. 2001. UNCOVERING ECONOMIC BENEFITS OF CHIVITA {MELONGENA MELONGENA LINNAEUS, 1758 AND MELONGENA CORONA BISPINOSA PHIEIPPI, 1844) MICHAEL D. KAPLOWITZ* Michigan Slate University. 31 lA Natural Resoiircci. East Lansing. Michigan 48S24 ABSTRACT West Indian crown conch Meliin)>ena melongena Linnaeus and other crown conch e.g.. Melongenci corona Gmehn ha\e been viewed by most researchers and marine resource beneficiaries as predatory species with little or no value. Focus groups and individual interviews were conducted w ith local residents as part of the design phase for an economic valuation study of mangrove ecosystems in Yucatan. Mexico. They examined how local inhabitants use, perceive of. and understand ecological services associated with their shared mangrove ecosystem. The data revealed that collection of Melongena melongena and Melongena corona bispinosa Philippi (collectively called "chivita") has become an increasingly significant component of these communities' economic activity. It was learned that chivita collection is replacing other marine resource-based subsistence strategies in these communities. These findings place Melongena melongena and Melongena corona hispinosa in a new light. KEY WORDS: Mangroves, qualitative methods, economic value. Mexico INTRODUCTION The West Indian crown conch {Melongeini inchmgeiia Lin- naeus, 1758) is generally thought of as a predatory species that preys on other "valuable" species such as the oyster (Crassostrea virginica. Gmelin) (Villarreal 1989). The published research on the West Indian crown conch and its close relative the Florida crown conch (Melongena corona Gmelin) has centered on the predatory nature and behavior of these species (Bowling 1994. Dalby 1989. Ellison and Farnsworth 1992. Garcia-Cubas 1981. Villarreal 1989). Some of the only reported research on Melongena corona hispinosa. Philippi. 1844 appears as part of conference proceedings several years ago and centers on the reproductive life of the conch (Manzano et al. 1998; Zarate and Arana 1998; Zarate et al. 1998). In Mexico and in this paper, the West Indian crown conch (Melongena melongena L.). Caracol Negro [Black conch] (Melongena corona hispinosa). and the Florida crown conch (Melongena corona) are collectively referred to as ""chivita" (Secretaria de Medio Ambiente Rucursos Naturales y Pesca |SEMARNAP| 1995). While there have been a few passing ref- erences to some economic value associated with chivita. there appear to be no docutnented studies in the literature on economic, ecological, or social benefits associated with these shellfish. In several instances, researchers have made reference to some commercial value associated with chivita. Villarreal (1989) makes the passing remark that in Veracruz, Mexico, "fishermen . . . be- lieve that \Melongena melongena] should be exterminated. On the other hand, since this snail is of economic importance, its capture is considered more attractive." However. Villarreal Chavez does *Tel.: -i-l-.'il7-355-OI01; Fax: -^ 1-5 17-353-8994; E-mail: kaplo\vit@msu. edu 'Focus groups are carefully planned discussions designed to learn about subjects" perceptions on a defined area of interest in a permissive, non- threatening environment. They are conducted by a skilled moderator who follows a discussion guide and involve as few as 2 to as many as 12 informants. "Individual interviews (also called unstructured, exploratory, intensive, in- depth, and depth interviews) are guided conversations whose goal is to elicit from interviewees (also called informants) rich, detailed materials that can be used in qualitative analysis. The interviewer used the same discussion guide as used in focus groups to guide the one-on-one conver- sations. not support, amplify, or explain what is meant by this. Elsewhere. Solfs-Ramfrez ( 1994. 20) includes chi\ itas (M. melongena and M. corona) as commercially exploitable shellfish species in one of that author's tables on the mollusks of the Yucatan Peninsula. Unfortunately. Solfs-Ramfrez does not refer to chivita in the arti- cle's text nor further explain the chivita reference. Most recently, Zetina Zarate et al. (1998). in their introductory remarks on the reproductive cycle of M. corona hispinosa. make otherwise un- supported statements to the effect that the collection of chivita is important and subject to increasing collection pressures. As a re- sult of the absence of information on economic benefits associated with chivita collection, this paper provides some useful informa- tion on the economic significance of the collection of West Indian conch and Black crown conch (chivita) in coastal communities in northern Yucatan. Mexico. Hypotheses Complex environmental and natural resources, such as the Yucatan's mangrove wetlands and coastal resources, represent substantial sources of cultural, intergenerational. environmental, and economic wealth (Aylward and Barbier 1992. Bann 1997, Barbier 1994. Barbier et al. 1997. Carson 1998. Perrings 1995). As part of an effort to identify the range of relevant ecosystem ser- vices for a study of the economic value of mangrove ecosystems in northern Yucatan, Mexico, I undertook a series of qualitative in- quiries. Two qualitative research methods were used — focus groups' and individual interviews.^ This paper examines research questions concerning the economic significance of chivita collec- tion in two coastal fishing conmiunities bordering Chelem Lagoon near Progresso. Mexico. This paper focuses on how local benefi- ciaries perceive, use. and value chivita. First, the hypothesis is explored that chivita collection is a positive economic activity for the communities of Chelem and Chubuma. A second hypoth- esis examined whether there is a significant positive economic benefit associated with chivita collection. It should be noted that findings concerning conllicting management agendas and the rela- tive merit of the use of focus group and individual interview data for resource valuation are presented elsewhere (Kaplowitz and Hoehn 2001. Kaplowitz 1999. Kaplowitz 2000a, Kaplowitz 2000b). This paper briefly describes the research design and locale before discussing the collection method and analysis of data on the economic significance of chivita collection. Finally, the results are 317 Kaplowitz discussed before concluding with some observations and implica- tions of the findings. MATERIALS AND METHODS Researchers in diverse fields of study regularly use qualitative methods such as. focus groups and individual interviews, as com- prehensive research tools and as important components in design- ing and implementing reliable research studies (Chilton and Hutchinson 1999, Krueger 1994. Morgan 1997. Schwarz 1997. Sudman et al. 1996, Weiss 1994). Qualitative methods for analyz- ing and managing shellfisheries have been reported previously (Whitlatch and Osman 1994). Individual in-depth interviews are efficient means for collecting information on beneficiaries" use and understanding of mangrove ecosystems at the local level (Ko- vacs 1999). Furthermore, focus groups and individual interviews have proven to be complementary in identifying conflicting re- source agendas (Kaplowitz 1999). They are also useful for learning from local beneficiaries how they use. perceive, and value envi- ronmental and natural resources (Kaplowitz and Hoehn 2(101. Ka- plowitz 200()a. Mandondo 1997). Studies have also shown that resource beneficiaries' ideas about and use of natural resources may differ greatly from those of scientists and so-called experts (Talawar and Rhoades 1998). Locale and Procedure The communities of Chelem and Chubuma, Mexico are located along a 15-kilometer stretch of coa.stal fringe that borders the Gulf of Mexico on one side and Chelem Lagoon on the other. These communities are. respectively, about 3 and 13 kilometers west of the port city of Progresso. The villages are comprised of families that have traditionally relied upon the natural resources of the region, including the mangrove wetland, for their subsistence and livelihood. The inhabitants of Chelem and Chuburna share similar socio-economic characteristics and have roughly 475 and 215 households respectively (Instituto Nacional de Estadfstica Geograffa e Informatica [INEGI] 1992). Traditionally, these com- munities have relied upon a combination of activities for their subsistence and economic gain. They have been able to survive long periods (November to March) of seasonal bad weather "'nortes" by developing a multiple use and activity strategy of combining fishing in the sea and lagoons, small scale salt extrac- tion, agriculture, and tourism activities (Pare and Fraga 1994). A total of 97 year-round residents from the two communities were interviewed in one of 12 focus groups or 19 individual in- depth interviews. The research design allowed for examination of the collected data across interview type, gender, and community. Research assistants canvassed randomly and selected sections of the target communities at staggered times of day to recruit partici- pants. The focus groups were comprised of between four and seven individuals of the same gender from the same village. No respon- dent or their family members participated in more than one focus group or interview. The focus groups and individual interviews were designed and implemented following the generally accepted practices of Morgan et al (1998), Morgan (1997), Morgan (1996), and Weiss (1994) respectively. A Mexican, professional moderator using a specially prepared discussion guide conducted the focus groups and individual interviews. All focus group and individual interviews were tape-recorded and subsequently transcribed. Data Analysis The data analysis allowed me to ( 1 ) discover themes, (2) con- sider the choice and meanings of words, (3) consider the context(s) of data collection, and (4) consider the consistency of responses (Krueger 1994). The analysis did not produce simple counts of things, but rather "fractured" the data and rearranged it into cat- egories that facilitated understanding the data and comparing it w ithin and between categories (Maxwell 1996. Strauss and Corbin 1990). The 12 focus groups and 19 individual interview transcripts resulted in more than 500 pages of text. An iterative, grounded theory approach (Strauss and Corbin 1990) was used to code the transcripts. First, almost every word of a randomly selected subset of transcripts was coded (open coding). Next a set of thematic or summary codes was developed (axial coding). When no new open codes were necessary to code additional transcripts, all of the study's transcripts were axial coded. The final iteration of coding of the text (selective coding) focused on organizing the data into 36 categories relevant to respondents" resource use. value, under- standing, perception, and control of the ecosystem. The reported research here focuses on local beneficiaries' use, perception, and understanding of chivita collection in Chelem La- goon. The data analysis created and used multiple-response vari- ables to record instances that focus-group discussions and indi- vidual interviews raised for discussion matters pertaining to chivita (e.g., shellfish collection, sale of chivita, market price for chivita). For example, the variable "Lagoon fishing"" captured and docu- mented a wide range of particular "fishing activity"" (e.g., mullet, shrimp, crab, chivita, and other species in the lagoon). Use of such variables accommodated the wide range of the discussion topics as well as allowed the coded transcript data to be subsequently ana- lyzed using statistical software. Furthermore, this approach al- lov\ed the focus group and individual interview transcript data to be transformed into variables to test research hypotheses and gen- erate value estimates. RESULTS The variable, interview type, records the type of interview (e.g., focus group or individual interview) associated with each case of coded data. Other variables capture those fishery services raised by respondents during the focus groups and Individual interviews. Table I illustrates the four primary fishery ser\ices as.sociated by local beneficiaries with the lagoon and that emerged from the focus group and individual interview data. Table 1 also presents some representative references as well as the percentage of focus group and individual interview sessions that raised for discussion each fishery service. As can be seen, participants, in general dis- cussions of the region, the ecosystem, and what they do, talked about "fishing in the lagoon"" quite frequently. Interestingly, par- ticipants discussed chivita collection most frequently and in almost all of the sessions. The other significant "fishing activities"" asso- ciated with the lagoon and raised by participants include the col- lection of crab as bait for use during the limited octopus (Octopus nuna and Octopus vulgaris) fishing season and the increasingly infrequent collection of shrimp in the lagoon. These results are especially noteworthy when one considers that local researchers, regional coastal resource managers, environmentalists, and gov- ernment officials failed to mention chivita collection as a local Uncovering Chivita Benefits 319 TABLE 1. Fishery service variables. Topic variable Example % sessions raisin;; topic Clii\ita Mfluiifteiia iiieloiigena: Meluiijieiici 97 corona bispinosu: used for food and commerce Lagoon fishmg we fish in lagoon; people conic lo Ml fish in wetland; some use their nets day & night at lagoon entrance Crah collected as bait; frozen for use 61 during two month octopus season Shrimp seawater sometimes brings shrimp; 39 when shrimp here, all fish for them; not as many shrimp as in past activity or an ecosystem service in more than 12 interviews con- ducted prior to the focus groups and individual interviews. As the data illustrate, the qualitative research revealed infor- mation about the use and undeistanding of the complex coastal ecosystem as a fishery. The qualitative data revealed the impor- tance of chivita collection to the subsistence of the local inhabit- ants of the Chelem Lagoon region. Focus groups and qualitative interviews can produce opinion and economic data that is consistent with traditional survey re- search data (Reynolds and Johnson 1978. Vaughn et al. 1996). The data were also coded to capture respondents'comments on catch rates and prices associated with the collection of chivita. As Table 2 illustrates, respondents reported a range of chivita collection rates and market prices. Some participants mentioned collection rates per working day ranging from as few as 0.5 kilograms to as many as 4 kilograms. Many respondents spoke about once being able to easily collect more than 10 kilograms per day during good times. During the sessions, respondents also volunteered a range of prices being paid for chivita by brokers or middlemen in the re- gion. As Table 2 shows, the aveiage price per kilogram was about 8 pesos or approximately $1.13. The brokers, in turn, transported and sold the chivita to hotels and restaurants in the tourist trade for use as appetizers and table snacks. It must be pointed out that this area enjoys a potentially lucra- tive octopus fishing season from mid-August through late- November. While there are reports that chivita and other lagoon species are collected during Octopus season, many men reported focusing on the octopus season, if able. This is not true for the women of the area who reported collecting chivita to sustain their families throughout the year, especially while their husbands go to fish in the sea. As Figure I illustrates, the communities of Chelem and Chuburna derive significant economic benefits from their e\- TABLE 2. Chivita collection reports Variable Low High Mean Catch rate (kg/day) Price* (Pesos/kg) 0.5 5 4 10 IS 7.9 ■ Exchange rate at the time about 7 pesos to il US. tractive use of chivita from the Chelein Lagoon. Using the median reported catch rates and the median reported market price, it is estimated that chivita collection represents approximately $230,000 to $350,000 dollars of income to these cominunities. That is. a family relying on the lagoon for chivita collection year- round can generate about $580 dollars annually, or about $390 dollars from chivita if they devote four months completely to octopus fishing. These findings are especially significant when compared to the Mexican minimum wage of 14 pesos ($2) per day paid at factories in the region, when there is available work. DLSCUSSION As the foregoing analysis demonstrates, chivita collection is the most significant economic activity associated with the lagoon fish- ery for the communities of Chelein and Chuburna. While the par- ticipants mentioned the collection of small crabs, it turns out that the unidentified species collected is too small to support human subsistence and is used as bait for the brief octopus (Octopus inaya and Oclopus vulgaris) season in the nearby Gulf of Mexico. The existing economic difficulties facing these communities and Mexico as a whole repeatedly came to light during the focus groups and individual interviews. The sessions were replete with discussions of the difficulty in providing for one's family. Increas- ing commercial fishing pressure by trawlers in the offshore and near-shore fisheries of the Gulf of Mexico is reported to have decimated the once rich coastal fishing resource. Local beneficia- ries have responded by increasing their reliance upon the lagoon and its mangrove ecosystem for subsistence. The focus groups and individual interviews left no doubt that lagoon fishing for chivita is of utmost importance to local people. The collection of chivita from the muddy bottom of the Chelem Lagoon has become the predominant subsistence strategy for the regions' communities. Repeatedly. I was told of mothers placing their small children on planks of wood so that they could attend to them while digging for chivita in the lagoon's muddy bottom. Chivita collection has replaced other inore conventional lagoon fishing and the collection of crabs as the key lagoon fishery ser- vice. When the Gulf of Mexico's fishery was thriving, there were reports that local beneficiaries were using the lagoon only as a place to collect their bait. However, the decline of the near-shore fishery has resulted in changes. While shrimp collection in the lagoon was mentioned in the sessions, it occurs only occasionally. According to participants, the recent construction of a duck habitat restoration dike by Ducks Unlimited and Mexican Navy activities, have caused drastic cur- tailment of the once annual or biannual inundation of shrimp in the lagoon. It is reported that these projects have stopped the circular flow of seawater through the lagoon that resulted after Hurricane Gilbert. Another lagoon-based activity that has been curtailed is salt collection. At one time, individuals in the region could con- struct salt ponds; flood them with seawater, allow the water to evaporate and then collect and sell crystallized sea salt. However, the area's lucrative salt mining business has been defunct for years. This change followed the Hooding and ipso facto enlarging of Chelem Lagoon when the Mexican government dredged and con- structed a safe harbor and naval station in the lagoon in the late 1960s and eariy 1970s (Pare and Fraga. 1994). It seems that virtually every family in the two communities, at one time or another has adopted chivita collection as part of its subsistence survival strategy. Furthermore, it is coinmon for al- 320 Kaplowitz Seasonal Chinita Collection (Excluding Octopus Season) 8 mo. season x 24 working days/mo. x 1 .8 kg/day x 7.9 pesos/kg = 2,733 pesos/household or ± $ 390* per household Year-Round Ciiivita Collection 12 mo. X 24 working days/mo. x 1.8 kg/day x 7.9 pesos/kg = 4,095 pesos/household or ± $ 585* per household per year Aggregate Annual Value to Chelem & Chuburna (+ 600 households) Excluding Octopus Season Year-Round Collection 1,639.800 pesos per year 2,457,000 pesos per year or or ± $ 234,257* per year ± $ 351,000* per year * Exchange rate of 7 pesos per dollar Figure 1. Benefit of chivita cullectiun. most everyone in the area to refer to himself or herself as a "pesca- CONCLUSION dor" (fisherman) despite the fact that many provide for themselves and their families by working in nearby factories or doing con- This study demonstrates that chivita collection is not only a struction work. Individuals when speaking about lagoon fishing benefit to these communities but it is a significant source of their perceive themselves as "fisherpeople", and it was learned repeat- subsistence activity. The data show the value of using indi\ idual edly during the sessions that respondents include chivita collec- interviews and focus groups to learn from local beneficiaries about tion, crab and shrimp collection together with line and net fishing their ecosystem and natural resource use. The finding that the West for other species. What makes this especially interesting, is that Indian crown conch (Melongeiia melongena L.), Caracol Negro researchers from nearby Merida working on coastal zone manage- [Black conch] (Melongena corona hispinosa). and the Florida ment in the region were surprised to learn of the extent to which crown conch (Melongena corona) are significant sources of sub- the respondents relied upon chivita collection. It was their belief sistence and income, place these species in a new economic light, that chivita was a minor component of residents" subsistence strat- For the people of Chelem Lagoon, these species are welcome egy and that near-shore fishing in the gulf was the predominant visitors not predators, occupation in the area. In the words of one respondent: ACKNOWLEDGMENTS We u.sed to make a living tlshing in the sea . . . Now you can't make a profit more than 2 lo 3 months from tlshing in This research paper was made possible, in part, by support from the sea... The same problem is also happening in the la- [^e Inter-American Foundation (lAF) and the Organization of goon, it used to be that you could take all the crab you , . „, . //-xacx »«• u i r-v i^ i ■• ■ a • • . ^ ■ ., , . ,, J ,,,u-, American States (GAS). Michael D. Kaplowitz is an Assistant wanted. Now only the small ones are around . . . While '^ some try to work elsewhere, people sustain their families Professor in the Department of Resource Development at Michi- wiih chivita from the wetland (Transcript 18). gan State University. LITERATURE CITED Aylward. B. & E. B. Barbier. 1992. Valuing environmental tunc- Barhier, E. B. 1994. Valuing environmental functions: Tropical wetlands. tions in developing countries. Biodiversity and Conservation Land Economics 10A55-\7?<. '■^■*~-^*'- Barbier. E. B.. M. Acreman & D. Knowler. 1997. Economic valuation of Bann, C. 1997. An economic analysis of alternative mangrove management wetlands: A guide for policy makers and planners. Cambridge, UK: strategies in Koh Kong Province, Cambodia. Singapore, Singapore: Ramsar Convention Bureau, Department of Environmental Economics Economy and Environment Program for South East Asia. and Management, University of York. 135 pp. Uncovering Chivita Benefits 321 Bowling, C. 1994. Habitat and size of the Florida crown conch [Mehmgena corona GnielinI: Why big snails hang out at bars. J. Exp. Mar. Biol. Ecol. I75(2):18I-I95. Carson. R. T. 1998. Valuation of tropical rainforests: Philosophical and practical issues in the use of contingent valuation. Ecological Econom- ics 24: \ 5-29. Chilton. S. M. & W. G. Hutchinson. 1999. Do focus groups contribute anything to the contingent valuation process? Journal of Economic P,svc/i()/(>,?y 20:465-183. Dalby. J. E. 1989. Predation of ascidians by Meloniiena corona (Neogas- tropoda: Melongenidae) in the northern Gulf of Mexico. Bulletin of Marine Science 45(3):708-7I2. Ellison, A. M. & E. J. Famsworth. 1992. The ecology of Beli/ean man- grove-root fouling communities: Patterns of epidont distribution and abundance, and effects on root growth. Hydrobiologia 247:87-98. Garcia-Cubas, A. 1 98 1 . Moluscos de un sistema lagunat tropical del sur del Golfo de Mexico. Mexico; National Autonomous University of Mexico lUNAM). Report nr 5. 182 pp. Instituto Nacional de Estadfstica Geograffa e Informatica (INEGI). 1992. Yucatan-resultados definitivos: Datos por .\GEB urbana. XI censo gen- eral de poblacion vivienad, 1990. Aguascalientes, Mexico: Instituto Nacional de Estadfstica. Geograffa e Informatica. Kaplowilz. M. & J. Hoehn. 2001. Do focus groups and personal interviews reveal the same information for natural resource valuation? Ecological Economics 36:237-247. Kaplowitz. M. D. 1999. Conflicting wetland agendas in the Yucatan. In- ternational Review of Comparative Public Policy 1 1:141-156. Kaplowitz. M. D. 2000a. Identifying ecosystem services using multiple methods: Lessons from the mangrove wetlands of Yucatan, Mexico. Agriculture and Human Values 17(2):169-179. Kaplowitz M. D. 2000b Statistical analysis of sensiti\e topics in group and individual interviews. Quality and Quantity: International Journal of Methodology 34(4), Kovacs J. M. 1999 Assessing mangrove uses at the local scale. Landscape and Urban Planning 43:201-208. Krueger R. A. 1994 Focus groups: A practical guide for applied research. Thousand Oaks, CA: Sage Publications. Mandondo, A. 1997. Trees and spaces as emotion and norm laden com- ponents of local ecosystems in Nyamaropa communal land, Nyanga District, Zimbabwe. Agriculture and Human Values 14:353-372. Manzano, N., L., Carrillo & D. Aldana Arana, 1998. Irradiation of eggs capsules of crown conch. Melongena corona hispinosa. with non- coherent and non-polarized light from arseniuro de galio emitting di- ods. Proceedings of the Gulf and Caribbean Eisheries Institute 50:594— 611. Maxwell, J. A. 1996. Qualitative research design: An mieractive approach. Thousand Oaks, CA: Sage Publications. Morgan, D. 1997. Focus groups as qualitative research. Thousand Oaks, CA: Sage Publications. Morgan, D. L. 1996. Focus groups. In: J. Hagan, K. S. Cook, editors. Annual review of sociology. Palo Alto, CA: Annual Reviews, pp. 129- 152. Morgan, D. L.. R. A, Krueger. A. U. Scannell & J. A. Kmg. 1998. Focus group kit. Thousand Oaks. CA: Sage Publications. 6 v. p. Pare. L. & J. Fraga. 1994. La costa de Yucatan: Desarrollo y vulnerabilidad ambiental. Mexico: Universidad Nacional Autiinoma de Mexico. Insti- tuto de Investigaciones Sociales. Perrings, C. 1995. The economic value of biodiversity. In: V. H. Heywood. editor. Global biodiversity assessment. Cambridge. UK: Cambridge University Press. Reynolds F, D„ D. K. Johnson.. 1978. Validity of focus-group findings. Advertising Research 18:21-24, Schwarz. N. 1997. Cognition, communication, and survey measurement. In: R. J. Kopp, W. Pommerehne, N. Schwarz. editors. Determining the value of non-marketed goods. Boston: Kluwer-Nijhoff. Secretaria de Medio Ambiente Rucursos Naturales y Pesca (SEMARNAP) 1995, Norma Oficial Mexicana NOM-013-PESC-I994. Diario Oficial de la Federacion CDXCIX{\by.5-l . Solfs-Ramfrez, M, J. 1994. Mollusca de la Peninusia de Yucatan. In: A. Yanez-Arancibia. editor. Recursos faunisiticos del litoral de la Penisula de Yucatan. Campeche (Mexico): Universidad Autonoma de Campeche. pp. 13-32. Strauss. A. & J, Corbin. 1990. Basic of qualitative research: Grounded theory procedures and techniques. Newbury Park, CA: Sage Publica- tions. Sudman, S.. N. B. Bradburn & N. Schwarz. 1996. Thinking about answers: The application of cognitive processes to survey methodology. San Francisco: Jossey-Bass Publishers. Talawar, S. & R. E. Rhoades. 1998. Scientific and local classification and management of soils. Agriculture and Human Values 15:3-14. Vaughn, S., J. S. Schumm & J. Sinagub. 1996. Focus group interviews in education and psychology. Thousand Oaks, CA: Sage Publications. Villarreal, C. G. 1989. Impact of predation by Melongena melongena (L,) on the populations of the oyster Crassostrea virginica (Gmelin) in Laguna de Tampamachoco. Veracruz. Ciencias Marinas I5(2):55-65. Weiss. R. S. 1994. Learning from strangers: The art and method of quali- tative interview studies. New York: The Free Press. Whitlatch. R. B. & R. W. Osman. 1994. A qualitative approach to man- aging shellfish populations: Assessing the relative importance of trophic relationships between species. Journal of Shellfish Research 13(11:229-242. Zetina Zarate. A. I. & D. Aldana Arana. 1998. Juvenile growth of the marine snail Melongena corona bispinosa (Philippi. 1844). under labo- ratory conditions. Proceedings of the Gulf and Caribbean Fisheries Institiae 50:584-593. Zetina Zarate, A. I.. T. Brule Demarest & D. Aldana Arana. 1998. Repro- ducion of the chivita snail Melongena corona bispinosa (Philippi. 18-14) in Chubuma Yucatan. Mexico. Proceedings of the Gulf and Caribbean Fisheries Institute 50:574-583, JuKi-iuil cif Shellfish Research. Vol. 20. No. I, 323-328. 2001. YELLOW AND BROWN SHELL COLOR MORPHS OF CORBICVLA FLVMINEA (BIVALVIA: CORBICULIDAE) FROM SICHUAN PROVINCE, CHINA, ARE TRIPLOIDS AND TETRAPLOIDS ANDONG QIU,' ANJING SHI,' AND AKIRA KOMARU^* ' Sichuan University. Department of Biology. Chengdu 610064 Sichuan. People'.^: Republic of China: 'Mie Universir\: Faculty of Bioresources, Tsu-city. Mie. 514-8507 Japan ABSTRACT Yellow and brown shell color morphs were distinguished in samples of Corhiciila fluminea collected from Anyue County, Sichuan Province. China. Shells of yellow morphs are straw yellow externally and white internally, while those of brown morphs are dark brown and purple, respectively. Karyological and DNA microfluorometric analyses revealed that yellow and brown morphs are triploid and tetraploid. respectively. Both are simultaneous hermaphrodites with non-reductional and biflagellate sperma- tozoa and brood their larvae in the inner demibranchs like diploid C. fluminea in Taiwan and Japan. These reproductive characteristics are similar to triploid Corhkula leanu in Japan, which is self-fertilizing. These results suggest that Corbicula fluminea at different ploidy levels may also reproduce by self-fertilization. KEY WORDS: Corbicula fluminea. triploid. tetraploid. karyotype. DNA microfluorometry INTRODUCTION The fre.shwater bivalves. Corbicula Icivni (Miiiler) and Cor- bicula fli(niinea (Miiller). are hemiaphroditic and brood their lar- vae (Miyazaki 1936. Kraemer and Galloway 1986). In both spe- cies, one individual can reproduce by self-fertilization (Ikematsu and Yamane 1977. Kraemer 1978). Both species show little alloz- yme polymorphism (Hillis and Patton 1981. Sakai et al. 1994). This monogenic condition has been explained as self-fertilization (Mcleod and Sailstad 1980) or gynogenetic development (Oka- moto and Arimoto 1986). C. leana in Japan is triploid and C. flwiunca in Taiwan and Japan is diploid, both species producing non-reductional spemia- tozoa (Okamoto and Arimoto 1986, Komaru et al. 1997). We have also produced cytological evidence of spontaneous androgenelic development in C. leana (Komaru et al. 1998). These similar re- productive characters suggest that diploid C. fluminea may also reproduce by self-fertilization. In this study, we identified two obviously different shell color morphs of C. fluminea and studied their chromosomes, spermato- zoa, and somatic cell DNA to identify genetic differences between and reproductive modes in them. MATERIALS AND METHODS The Samples The Corbicula fluminea (Miiller) samples were collected from a stream in Anyue County. Sichuan Province. People's Republic of China, in September 1997. The two color morphs were collected at the same sampling site. The samples were transferred to Sichuan University. Chengdu. Sichuan. China and the National Research Institute of Aquaculture. Nansei. Mie. Japan, and kept in labora- tory aquaria. Chromosomes The samples were transferred to freshwater containing 0.002% colchicine for 4-5 hours. The gills and gonads were then dissected out and hypotonically treated with distilled water for 30 min and fixed in Camoy's fixative. Small pieces of gill or gonad were *Corresponding author. E-mail: komaru@bio.mie-u.ac.jp minced in 50% acetic acid and isolated cells were placed on slides, air-dried, and stained with 2% Giemsa in phosphate buffer (pH = 7.2) (Okamoto and Arimoto 1986). For chromosome classification and karyotyping, metaphase chromosomes were measured and paired on the basis of relative length and centromeric index. DNA Microfluorometry Spermatozoa and somatic cells from the mantle were isolated in distilled water, placed on slides, air-dried, and fixed with 70% ethanol. The slides were then stained with a DAPl staining solu- tion. Relative DNA content was estimated by microfluorometry after DAPI staining (Komaru et al. 1988). Light Microscopy Gonads and gills were fixed with 10% formaldehyde and pro- cessed for paraffin sectioning. Slides were stained with Mayer's hemalum and eosin. Spermatozoa Measurement Spermatozoa were isolated from the gonads in freshwater and observed by phase contrast microscopy. The lengths of the sperm head and flagella were estimated using an ocular micrometer. RESULTS Shell Color and Size Figure 1 shows the two shell color morphs (yellow and brown) of Corbicula fluminea from Anyue County. Externally, the shell surface of the brown morph is dark brown (Fig. lA), and its inner surface is deep purple along the ventral margin, but rather white from the pallial line to the umbo (Fig. IB). The external surtace of the yellow morph is straw-colored (Fig. lA) while the inner sur- face is white (Fig. IB). This form is also characterized by purple flashes along the anterior and posterior lateral teeth. Shell lengths (SL). heights (SH). and widths (SW) were not significantly different between the yellow and brown morphs. but the three ratios of SW/SL. SW/SH. and SW/ (SL-i-SH-t-SW) are significantly larger in the yellow than in the brown morph. whereas the ratio of SH/SL in both morphs is similar (Table 1 ). 323 324 QlU ET AL. Figure 1. External (.4) and internal (B) views of the shells of yelloH deftl and brown (right I inorphs of Corbicula fluminea collected from .^nyue County, PRC. Chromosomes and Relative DNA Content of Somatic Cells Chromosome numbers of the yellow morph were 54 in the gills (Fig. 2A) and 18 in the gonads {Fig. 2B). and 72 in the gills of the brown morph (Fig. 2C). Eight well-spread mitotic metaphases of the yellow morph and seven of the brown were measured and karyotyped. Figure 3 shows the karyotype taken from one meta- phase of the yellow morph containing 1 8 triplets arranged on the basis of shared relative lengths and centromeric indices. It consists of 2 metacentrics. 13 submetacentrics {including 3 submetacen- trics-metacentrics and 6 submetacentrics-subtelocentrics), and 3 subtelocentrics {Table 2). Figure 4 shows the karyotype of the brown morph. It contains 18 quadraplets. consisting of 2 metacen- trics. 13 subcentrics (including 3 submetacentrics-metacentrics and 6 submetacentrics-subtelocentrics). and 3 subtelocentrics {includ- ing 2 subtelocentrics-submetacentrics) (Table 3). The meiotic chromosomes of the yellow morph have triplet-like chromosomes and an incomplete synapsis of homologous chromosomes was ob- served (Fig. 2B). Meiotic chromosomes could not be observed, however, in any studied individuals of the brown morph. The relative DNA content of the yellow morph was almost the same as that of the triploid C leana collected from Mie Prefecture. Japan, whereas that of the brown morph was about 1.3 times that of triploid C. leana (Table 4). DNA Content of Sperm and Somatic Cells As shown by DNA microfluorometry of the spermatozoa and somatic cells (Table 5), the relative DNA content of the former was almost identical to that of the latter. Seven yellow morph and eight brown morph individuals all produced spermatozoa with a DNA content similar to that of their own somatic cells. Spermatozoa Measurements The sperm heads of both C. fluminea morphs are slightly curved and elongated and both have two flagella. The length of the brown morph spermatozoa is significantly longer than that of the yellow. The sperm head size was 20.3 ± 1.0 |xm (n = 10) for the yellow morph and 24.0 ± 0.5 ixm (n = 101 for the brown moiph. Flagella length was 46.2 ± 1.5 ixm {n = 10) in the yellow morph and 49.5 ± 1.9 p-m (n = 10) in the brown one. Gonad and iMrval Incubation Gonad and gill sections of 14 yellow and 12 brown morphs were examined. All individuals were simultaneous hermaphro- dites. Three of the 14 yellow morph individuals and four of the 12 brown morph were brooding larvae in their inner demibranchs. TABLE 1. A comparison of shell dimension ratios between yellow and brown morphs of Corbicula fluminea. .SH/SL SW/SE SW/SH SWASL + SW + SH) Mean ± SD Mean ± SD Mean ± SD Mean ± SD Y morph B morph 0.822 ±0.022 0.847 + 0.021 0.567 ±0.016 0.478 ± 0.020 0.690 ±0.016 0.564 ± 0.024 0.238 ± 0.007 0.206 ± 0.006 Y: yellow morph of C. fliimiuea B: brown morph of C. fluminea Yellow and Brown Shell Color Morphs 325 Figure 2. Mitotic and meiotic chromosomes of yellow (A, B) and brown (C) morphs of C. fluminea. Scale bar = 5 jim. DISCUSSION C. flmninea from Japan and Taiwan have been reported to be hermaphroditic and diploid with In = 36 in the gills and 18 bivalents in the gonads (Kornaru et al. 1997). In the present study, yellow and brown color morphs of C. fluminea were discovered in Anyue County. Sichuan Province. China, where they are sympa- tric. According to our karyological and DNA content analyses of these two morphs of C. fluminea. the yellow one is triploid with 3« = 54 in the gill tissues, which can be classified into 18 triplets on the basis of their relative lengths and centromeric indices, and m 2 3 4 AAH 5 AAA 6 UA AAA 7 8 AA# 9 10 ill 11 12 13 14 15 16 17 A^A 18 Figure 3. Karyotypes taken from yellow (3h = 54) morph of C flu- minea metaphases. Scale bar = 5 pm. v\ilh 18 trivalents in the gonads. The brown morph is tetraploid with -Xn = 72 in the gill tissues, which can be classified into 18 tetraplets. A comparison of karyotypes from triploid and tetraploid C. fluminea shows that they have a similar formula containing 2 metacentrics. 13 submetacentrics, and 3 subtelocentrics. This im- plies that they may have the same origin. However, there are also some differences between the karyotypes, so that the No.2 chro- mosomes of both the triploid and tetraploid had submetacentrics of different relative lengths. These differences indicate that they may have different genetic properties. The shell color of C. fluminea is highly polymorphic (Morton 1987, Chen et al. 1995). Britton and Morton ( 1986) also identified two different color morphs of C. fluminea in USA with signifi- cantly different mean shell dimension ratios, and suggested that such polymorphism in shell color and shell size were the result of environmental induction. Morton (1987) identified the same two morphs in Hong Kong and Tsoi et al. (1991) showed there was no genetic differentiation between the two morphs by allozyme analy- sis. However, triploid and tetraploid C. fluminea with different shell colors and sizes in Anyue County are possibly sympatric, which indicate that these distinct differences in the two shell color morphs may be mainly derived from their different genetic con- stitutions. Physiological studies should be implemented to deter- mine the fitness of both morphs and the significance of the poly- ploidy. Triploid and tetraploid C. fluminea are hermaphrodites and brood their larvae in the inner demibranchs. The meiotic chromo- somes comprise 18 trivalents with an incomplete synapsis of ho- mologous chromosomes in the gonads of triploid C. fluminea. These could not be observed in tetraploid C. fluminea. These re- sults imply that the triploid C. fluminea cannot carry out normal meiosis as in the triploid C. ieana. and tetraploid C. fluminea fail to achieve meiosis. DNA microfluorometry of spermatozoa and somatic cells of the triploid and tetraploid C. fluminea showed that they all produce non-reductional and biflagellate spermatozoa similar to C. Ieana in Japan and C. fluminea in Taiwan and Japan (Komaru et al. 1997). These results indicate that these polyploids can reproduce in ways different from the normal sexual method. It has been reported that triploid C. Ieana in Japan reproduces by androgenesis which Komaru et al. ( 1998) cytologically confirmed. In androgenesis. all maternal chromosomes were expelled as polar bodies at the first division after self-fertilizaton. which results in the forming of only one male pronucleus from the spermatozoa but no female pronucleus. Therefore, only the chromosomes deriving from the male pronucleus formed the mitotic chromosomes for the first zygotic cleavage. Diploid (Komaru et al. 1997). triploid, and tetraploid C. fluminea have the same reproductive characters as C. Ieana. i.e., hermaphroditic, brooding, and the production of non- reductional spermatozoa. It is possible that C. fluminea with dif- ferent ploidies may also reproduce by androgenesis and be fertil- ized by autosperm as in C. Ieana. In the animal kingdom, polyploids are rather rare, as compared with plants. Animals reproducing by parthenogenesis are some- times polyploid (Hughes 1989). Three different ploidy levels have been known in Corbicula up to now. How did they evolve from an ancestral species? We have shown that diploid and triploid Cor- hicula produce non-reductional spermatozoa (Komaru et al. 1997). These data suggest that diploid androgenetic individuals may evolve from a diploid ancestral population with normal meiosis. After androgenesis arose in an ancestral population, mutations for 326 QlU ET AL. TABLE 2. Measurements and classincation of metaphase chromosomes taken from the yellow morph of Corbiciila fluminea. C h rnm n^oinp Relative length Arm ratio Centromeric index \^ 111 (71IIU9UI1IC pair no. Mean SD Mean SD Mean SD Classification 1 10.87 0.23 1.385 0.126 42.05 2.28 m ~i 4.66 0.25 1.339 0.023 42.82 0.44 m 3 7.89 0.42 1.904 0.478 34.97 5.96 sm-m 4 5.55 0.38 1.899 0.200 36.06 1.15 sm-m 5 3.93 0.20 1.724 0.027 36.84 0.33 sm-m 6 4.72 0.23 2.117 0.077 32.45 0.42 sm 7 4.73 0.12 2.187 0.087 31.48 0.86 sm 8 4.61 0.11 2.098 0.046 32.28 0.47 sm 9 4.10 0.16 2.402 0.163 29.58 1.43 sm 10 7.41 0.66 2.494 0.611 28.77 4.87 sm-st 11 6.29 0.71 2.413 0.340 28.85 1.99 sm-st 12 5.74 0.18 3.014 0.038 25.15 0.36 sm-st 13 5.61 0.28 2.781 0.041 27.19 0.75 sm-st 14 4.93 0.09 2.690 0.056 27.13 0.26 sm-sl 15 4.43 0.11 2.902 0.093 25.99 0.43 sm-st 16 5.15 0.14 3.488 0.087 22.54 0.54 St 17 4.76 0.23 3.536 0.064 22.12 0.47 St 18 4.46 0.13 3.578 0.042 22.37 0.38 St polyploidy occuired. Lokki { 1976) pointed out that in parthenoge- netic animals, polyploidy would help to maintain functional loci when mutant alleles are accumulating. This could explain polyp- loidy in Corbiciila. ll is likely that the triploids and tetraploids in Corbiciila have arisen from diploids. To increase the ploidy, we assume two possible androgenetic events in Corbiciila: either (1), production of spermatozoa with increased ploidy levels derived from meiotic failure during spermatogenesis or (2), incomplete polar body formation, i.e.. one or two maternal chromosome sets were not extruded and participate in pronucleus fomiation. The taxonomy of Corbiciila has problems (Morton 1986) be- cause the genus contains species and populations that can repro- uni^ AAXA «A)I* AAAA /VAAA iiAA* AiHAf^ iiXA 10 A/^A^^ 13 11 14 AAAA 12 15 ♦AAA A^A#^ 16 17 18 Figure 4. Karyotypes taken from brown (4n = 72) morph of C. fluminea metaphases. Scale bar = 5 jim. Yellow and Brown Shell Color Morphs 327 TABLE 3. Measurements and classification of metaphase chromosomes talien from the brown morph of Corbicula fliiminea. Chromosome pair no. Relative length Arm ratio Ccntromeric index Mean SD Mean SD Mean SD Classification 1 9.96 0.22 1.268 0.069 44.09 1.29 m 2 4.87 0.04 1.796 0.060 37.08 0.48 m 3 6.33 0.08 1.806 0.029 35.59 0.39 sm-m 4 5.07 0.11 1.728 0.041 36.67 0.4.'; sm-m 5 4.18 0.06 1.923 0.059 34.25 (1.71 sm-m 6 4.88 0.07 2.198 0.034 31.79 0.56 sm 7 4.97 0.03 2.268 0.025 30.58 0.23 sm 8 4.80 0.05 1.975 0.023 33.74 0.18 sm 9 4.68 0.03 1.960 0.017 33.75 0.19 sm 10 6.95 0.11 2.487 0.021 28.68 0.34 sm-st 11 6.42 0.16 2.597 0.092 27.86 0.47 sm-st 12 5.71 0.10 2.689 0.044 27.24 0.32 sm-st 13 5.02 0.08 2.818 0.083 26.21 0.27 sm-st 14 4.48 0.06 2.734 0.055 26.83 0.54 sm-st 15 4.07 0.12 2.742 0.066 26.84 0.48 sm-st 16 7.22 0.09 3.049 0.048 24.78 0.56 st-sm 17 5.90 0.07 3.042 0.036 24.78 0.44 st-sm 18 4.22 0.05 4.152 0.029 19.51 0.41 St duce by self-fertilization and androgenesis. The taxonomy of Cor- bicula should, thus, be determined on the basis of not only shell morphology but also genetic and ecological information. The key to Corbicula taxonomy is how species and populations reproduce TABLE 4. Ploidy estimation by DNA microfluorometry of yellow (Yl-10) and brown morphs (Bl-11 1 of C. fliiminea collected from Anyue County. No. Mean (S.D.) of sample" Mean (S.D.) of 3n standard"" Ratio' Plody Y-1 376.76(11.08) 369.95(11.08) 1.01 Y-2 341.78(15.06) 0.92 Y-3 381.26(11.56) 1.03 Y-4 352.75(12.01) 377.85(14.97) 0.93 Y-5 378.23 (9.24) 1.00 Y-6 315.68(14.10) 344.65 (5.48) 0.92 Y-7 310.63(12.50) 0.90 Y-8 337.37(18.56) 0.98 Y-9 347.40(12.00) 355.60(13.90) 0.98 Y-10 350.45(10.10) 0.99 B-1 478.37(11.96) 353.60 (9.50) 1.35 B-2 488.20(11.83) 1.38 B-3 497.37 (22.49) 356.62(14.93) 1.37 B-4 484.89(10.92) 1.36 B-5 459.39(22.62) 1.29 B-6 4.54.56(15.31) 344.07(12.60) 1.32 B-7 468.14(12.34) 1.36 B-8 484.95(12.09) 375.08(10.90) 1.29 B-9 478.75(20.00) 1.28 B-10 477.95(11%) .352.09(8.03) 1.36 B-ll 484.12(11.83) 1..37 3n 3n 3n 3n 3n 3n 3n 3n 3n 3n 4n 4n 4n 4n 4n 4n 4n 4n 4n 4n 4n ■' Mean of Relative DNA content (tluoroscence intensity) of at least 20 cells at each sample. "" Triploid Corbicula leana collected from Mie Prefecture, Japan. " Ratio = Mean of sample/Mean of 3n standard. and how they evolved from an ancestral diploid species. Further studies on the genetic differences and reproductive modes of both morphs of Corbicula fluminea are required to identify the evolu- tionary significance of androgenesis and polyploidy. Interestingly, there are a few brown individuals with almost the same DNA content as yellow ones, such as B-2 and B-8 in Table 5. It is unknown whether they belong to another polymorphic type of C. fluminea. but it is worth studying further. Maybe the genetic polymorphism of C. fluminea is much more intricate than what we know until now. TABLE S. Mean and standard deviation of relative DNA content of spermatozoa and gill cells in yellow (Yl-7) and brown morphs (Bl-8) of C, fluminea collected from Anyue County, People's Republic of China by microfluorometry. Shell color morph DNA content Sperm Somatic cell Y-1 257.48 ± 12.58 (n = 40) 262.50 ± 20.65 (n = 33) Y-2 293.71 ±27.43 (n = 49) 314.41 ± 23.76 (n = 53) Y-3 276.81 ± 10.68 (n = 48) 299.23+ 16.05 (n = 52) Y-4 257.48 ± 12.58 (n = 33) 262.50 ± 20.05 (n = 40) Y-5 286.63 ± 15.08 (n = 60) 284.48 ± 23.08 (n = 60) Y-6 293.71 ±27.43 (n = 49) 314.41 ±23.76 (n = 49) Y-7 257.48 ± 12.58 (n = 33) 262.50 ± 20.05 (n = 40) B-1 405.43 ±25.01 (n = 58) 443.38 ± 31.99 (n = 58) B-2 316.23 ± 16.90 (n = 47) 299.00 ± 19.22 (n = 42) B-3 333.31 ±23.61 (n = 55) 357.62 ± 35.00 (n = 45) B-4 358.38 ± 13.13 (n = 50) 362.23 ±16.93 (n = 50) B-5 372.36 ±23.71 (n = 52) 405.94 ± 24.36 (n = 51) B-6 405.43 ±25.01 (n = 58) 443.38 + 31.99 (n = 59) B-7 358.93 + 24..56 (n = 66) 349.24 ± 20.93 (n = 65) B-8 247.00 ± 17.65 (n = 31) 249.44 ±21.81 (n = 34) The numbers in parenthesis are the number of cells estimated. 328 QlU ET AL. LITERATURE CITED Britton, J. C. & B. Morton. 1986. Polymorpliism in Corbicula flwmnea (Bivalvia: Corbiculoidea) from North America. Malacol. Rev. 19(1/2): 1-44. Chen, T. C. K. L. Liao & W. L. Wu. 1995. Anatomy on Corhwula fluminea (Bivalvia:Corbiculidae). Bull. Malac. ROC 19:9-19. Hillis. D. M. & J. C. Fatten. 1981. Morphological and electrophoretic evidence for two species of Corhuiila (Bivalvia: Corbiculidae) in North America. Amer. Midi. Nat. 108:74-80. Hughes, R. N. 1989. A functional biology of clonal animals. London: Chapman and Hall. 311 pp. Ikematsu, W. & S. Yaraane. 1977. Ecological studies of CorhiciiUi leana PRIME-3. Bull. Jap. Soc. Fish. 43:1139-1146. Komaru. A.. Y. Uchimura, H. leyama & K. T. Wada. 1988. The detection of induced triploid scallop. Chlamys nobilis. by DNA microtluorom- etry with DAPI staining. Aquaculture 69:201-209. Komaru. A., K. Konishi. I. Nakayama, T. Kobayashi, H. Sakai & K. Kawamura. 1997. Hermaphroditic freshwater clams in the genus Cor- bicula produce non-reductional spermatozoa with somatic DNA con- tent. Biol. Bull. 193:320-323. Komaru, A., T. Kawagishi & K. Konishi. 1998. Cytological evidence of spontaneous androgenesis in the freshwater clam Corbicula leana. De- velopment. Gene and Evolution 208:46-50. Kraemer, L. R. 1978. Corbicula fluminea (Bivalvia: Sphaeriacea): the functional morphology of its hermaphroditism. Bull. Anier. Malac. 40- 49. Kraemer, L. R. & M. L. Galloway. 1986. Larval development of Corbicula fluminea (Miiller): an appraisal of its heterochrony. Amer. Malac. Bull. 4:61-79. Lokki, J. 1976. Genetic polyinorphism and evolution in parthenogenetic animals 8:Heterozygosity in relation to polyploidy. Hereditas 83:65-72. Mcleod, M. J. & D. J. Sailstad. 1980. An electrophoretic study of Cor- bicula fluminea (Bivalvia: Corbiculacea) in the Catawba River. Bull. Amer. Malac. 17-19. Miyazaki, 1. 1936. On the development of bivalves belonging to the genus. Corbicula. Bull. Jap. Soc. Fish. 5:249-254. Morton, B. 1986. Corbicula in Asia-an updated synthesis. Amer. Malac. Bull. 2:113-124. Morton, B. 1987. Polymorphism in Corbicula fluminea (Bivalvia: Corbi- culidae) from Hong Kong. Malac. Rev. 20:105-127. Okamoto, A. & B. Arimoto. 1986. Chromosomes of Corbicula japonica. C. sandai and C iCorbiculal leana (Bivalvia: Corbiculidae). Venus 45:194-202. Sakai, H., K. Kamiyama, S. R. Jeon & M. Amio. 1994. Genetic rela- tionship among three species of freshwater bivalves genus Cor- bicula (Corbiculidae) in Japan. Nippon Suisan Gakkaishi 60:606- 610. Tsoi, S. C. M., S. C. Lee, W. L. Wu & B. Morton. 1991. Genetic variation of Corbicula fluminea (Bivalvia: Corbiculidae) from Hong Kong. Ma- lac. Rev. 24:25-34. Joiirmil of Shcllt'ish Research. Vol. 20. No. I. 329-335. 2001. LACK OF SURFACE-ASSOCIATED MICROORGANISMS IN A MIXED SPECIES COMMUNITY OF FRESHWATER UNIONIDAEt S. J. NICHOLS/ * J. ALLEN,' G. WALKER," M. YOKOYAMA,' AND D. GARLING' '(7.5. Geological Survey. 1451 Green Rd. Ann Arbor. Michigan 4S105: 'Eastern Michigan Universiry, Ypsilanti, Michigan 48197: ^Michigan Slate Univer.'iity, East Lansing. Michigan 4HS24 ABSTRACT To determine whether iinionids contain surface-attached endosymbiotic bacteria, cihates, or I'ungi. we tt.sed scanning electron micro.scopy to examine the epithehal surface of various organs within the digestive system and mantle cavity of temperate river and lake unionids on a seasonal basis. We also cultured material removed from the lumen of these same organs and from the mantle cavity to detect cellobiose-. cellulose- and chitin- degrading microbes. No true endosymbiotic fauna were observed attached to the surface of the digestive or mantle tissues of any species of unionid. Microbial growth on cellulose or chitin bacteriological media varied by season and habitat, but not by unionid species or source of the isolate. Lake unionids did not contain microbes capable of digesting cellulose or chitin, whereas unionids from the river site did in March and August, but not in December. Since these cultured cellulose- and chitin-degrading bacteria were never found attached to any unionid tissues, they appeared to be transient forms, not true endosymbionts. Microbes capable of digesting cellobiose were found in every unionid collected in all seasons and habitats, but again, no microbes were directly observed attached to unionid tissues. If unionids, like most other invertebrates, lack digestive enzymes re(|uired to initiate pnmary bond breakage, then the lack of cellulolytic and chitinolytic endosymbionts would affect the ability to utilize cellulo.se or chitin foods. Thus, in captivity dry feeds based on com. soybeans, or nauplii should be pre-digested to ensure maximum absorption of nutrients by unionids. The lack of cellulolytic or chitinolytic endosymbionts should not affect relocation success, though the seasonal role of transient microbes in unionid nutrition requires further investigation, KEY WORDS: cut microtlora, endemic microbes, Unionidae INTRODUCTION Freshwater unionids in North Airierica are being extirptited at alarming rates due to factors such as habitat degradation and com- petition from the exotic zebra mussel (Dreissena polyinorpha Pal- las) (Williams et al, 1993), At this time, conservation efforts for adult unionids include relocation into new habitats and intensive aquaculture, but mortality rates have been high with most adults surviving less than three years (Cope & Waller 1993; Lellis and Johnson 1998; Gatenby et al, 1999). Although these high mortality rates are undoubtedly due to a host of different factors, one prob- lem area has been identification of dietary requirements. Little information exists on dietary preferences and required nutrients necessary for formulating a captive diet, and evaluating food re- sources in new habitats or refugia. The few dietary studies in natural habitats have shown that unionids ingest a wide variety of potential food items including algae, detritus, fungus, rotifers, and zooplankton, with detritus as the dominant component in both the mantle cavity and gut lumen (Lefevre & Curtis 1910; Jiffry 1984; McMahon 1991; Nichols & Garling 2000). The ingestion of large amounts of detritus implies that detritus is actively selected and thus may play a significant dietary role. This role is not easy to characterize as detritus in aquatic systems represents a complex matrix often containing cellulose frotii aquatic and terrestrial sources, chitin from fungus, rotifers, and zooplankton. and which has in turn been colonized by bacterial, fungal, and protistan fauna. Thus, ingested detritus could function tnerely as a substrate for the preferred food item, the associated epiorganisms. or could function as a direct source of nutrients obtained from cellulose or chitin degradation. One complicating factor is that cellulose and chitin are highly complex polysaccha- rides that require specific enzymes to initiate primary bond break- *Corresponding author. Tel: -1-1-734-214-721 S; E-niail: SJerrine_ Nichols@USGS.gov tThis article is Contribution 1 142 of the USGS Great Lakes Science Cen- ter, age. Such primary enzymes are rarely produced endogenously in aquatic invertebrates, and are recorded for very few marine bi- valves, although cellulases in general have been detected in both marine and freshwater mollusks. In unionids, the concentration of cellulases present has actually proven to be a successful technique for assessing the health status of the animal (Haag et al, 1993; Farris et al, 1994). However, these cellulase concentrations were obtained from total animal bioassays. from unionids freshly re- moved from the field, and were not further identified to type. It has not been determined if primary cellulases were present, and if so, whether they were produced endogenously by the unionid. or ex- ogenously by the microbial community associated with ingested food items. Furthermore, chitinases of any type have not yet been reported for either freshwater or marine bivalves. Most aquatic animals that feed directly on detrital cellulose access the necessary primary cellulases through a symbiotic rela- tionship with sotne type of bacteria/microbe, and contain recog- nizable endosymbiotic fauna somewhere in their digestive tracts (black fly larvae. Taylor et al. 1995; crane fly larvae, Klug & Kotarski 1980; mullet. Mountfort & Rhodes 1991), Microbial and endosymbiont/nutritional relationships among bivalves are far more complex than those reporled for other aquatic invertebrates and often vary by species as well as by location. True endosym- biotic relationships range from the consistent obligate communi- ties of microbes directly buried within gill tissues of deep-sea hydrothermal vent bivalves (Wood & Kelly 1989). to spirochaetes merely attached to the outside epithelial layer of digestive tissue or to the crystalline style (Bernard 1970; Conway & Capuzzo 1989: Prieur et al. 1990). Geographical variability is common, with en- dosymbionts existing inside some species at some locations, but not in others (Bernard 1970). while some marine bivalves never exhibit endosymbiotic relationships (Garland et al. 1982). Studies on freshwater bivalve- microbial relationships are very limited (Siariiper et al. 1997) and have focused mainly on bacterial com- munities found in the gut lutnen or passing through the intestinal tract. As in marine bivalves (Prieur et al. 1990; Harris et al. 1998), 329 330 Nichols et al. /a 1 ,%i '.jTv; •• i^fSS; Figure 1. Scanning electron micrograpli of unionid digestive tissue examined for the presence of endos) mbionts. a = Cross-section of the stomach (si, intestine (i(. and digestive gland tubules (dtl; b = digesti\e gland tubules and lumen ill with food particles: c = surface of crystalline style (cs) showing mucous sheet and food bolus Ibl; (dl ciliary tuft found In the intestinal tract: (el and (fl mucous sheet and food material (diatoms and detritus) on the surface of the intestinal tract. Scale bars = 1mm lal: KHIjim (b and cl: 10 pm Id and fl: Ipm (el. bacterial communities passing through the gut of freshwater union- ids usually differ in species composition from those found in the surrounding water column, retlecting selective predation and sub- sequent enhancement of bacteria species not directly consumed. However, these are not necessarily residential, or endosymbiotic. populations. The objective of our study was to determine if certain unionid species contained endosymbiotic microflora that might aid in the digestion of dietary materials such as cellulose and chitin, and if present, did this flora vary by location, season, or species of unionid. MATERIALS AND METHODS Study Sites and Unionids The unionids used in this survey were collected from three sites in southeast Michigan, U.S.A.: the Huron River, Four Mile Lake, and Vineyard Lake. The Huron River is a regulated stream, con- sisting of a series of impoundments connected by free-flowing stretches of river. Our study area focused on the middle section of the river, where the water is shallow (1 m deep), with an average water velocity of 0.5 m/sec and heavy canopy cover dominated by deciduous trees (Livingston Co., 42°25'41"N. 83°54'40"W). A to- tal of 17 unionid species occur in this river. Four were selected for analysis.' Lximpsilis veiilricosa Barnes, L. siliqiioidea Barnes, Pty- chobranchus fasciolaris Rafinesque, and Pyganodon grandis Say. Four Mile Lake is a 25 hectare lake once used as a marl mining site (Washtenaw Co., 42''20'16"N, 83°58'14"W). Except for the marl pit. the lake is shallow. <2 m deep, with no current flow, and a soft substrate covered with Cham spp. Canopy cover is lacking, al- though some wetland vegetation occurs at the edges of the lake. Only one species, P. grandis. regularly occurs in this lake. Vine- yard Lake is only 12 hectares is size, is 6.5 m deep, and is sur- Gut Microflora of Freshwater Unionidae 331 Figure 2. Scanning electron micrograph of fungal hjpliae growing on the mantle tissue of a single specimen of siliqiinidea collected from the Huron Ri\er in August. Scale (till found Lamps His bar = 1pm. rounded by subdivisions (Jackson Co.. 42°04'57"N. 84°12'35"W). Only one species. ElUptio dikitata Rafinesque, was observed in this lake. Examinalidii for Surface-Attached Microbes Sampling tor unionids occun'ed in March. August, and Decem- ber, and 4 individuals of each species were collected from the Huron River (four species). Four Mile Lake (one species), and Vineyard Lake (one species) (total it = 24 unionid specimens for each month for all three sampling locations). Unionids were placed in damp mesh bags in a cooler, and transported to the laboratory. The shell e.xterior of each unionid was then scrubbed with a brush to remove external Ooru. rinsed in distilled water, followed by a 10% acetic acid rinse, a second distilled water rinse, and then the shell was pried open. The mantle cavity was rinsed for 3-5 minutes with distilled water and all the rinse water and flushed material was saved. The foot, gills, labial palps, siphons from the mantle cavity, stomach, digestive gland, style sac. style, and fore-mid- hind sections of the intestinal tract were excised and processed for scanning electron microscopy (SEM). Distilled water used to rinse and flush contents from the lumen of each part of the digestive tract with the contents from each organ, as well as the material rinsed from the mantle cavity, were saved separately for culture on bacteriological media. For scanning electron microscopy (SEM), the separate organs and tissues removed from each unionid were immersed in 10% neutral buffered formalin (primary fixation). After the tissue firmed, serial sections of approximately 1 mm were taken using a scalpel. The tissue was then isolated if needed, trimmed, and washed three times for 20 minutes each in cacodylate-HCl buffer at a pH of 7.2. Secondary fixation involved immersion in \% cacodylatc-HCl buffered osmium tetroxide for one hour. Samples were then dehydrated in a graded alcohol series: 509} (15 min), 65% (15 min), 75% (15 min), 85% (15 min), 95% (15 min), 100% (3 changes 15 min each). Tissues were then critical point dried in CO, transition fluid, mounted on 2.5 cm stubs, sputter coated with gold, and examined under a scanning electron microscope. Microbial Growth on tiacltriological Media The contents and rinse water were removed from seven areas within each unionid: mantle cavity, stomach, digestive gland, style sac. fore-, mid-, and hind gut. These samples from each of the four individual animals of a particular species collected at each locality were pooled and then used to inoculate three standard enrichment broths, cellulose (carboxymethylcellulose), chitin. or cellobiose. The same amount of inoculant (about 0.10 ccl. was removed from each well-stirred pooled sample and placed into a test tube con- taining cellulose, chitin, or cellobiose with two duplicate sets made (7 internal sources of the inoculant x 2 replicates x 3 media types = 42 total samples per sampling date). This yielded a total of 168 samples for all species in the Huron River each sampling date, and 42 for Vineyard and Four Mile lakes since each had one unionid species present. The duplicate samples for each species and site were randomly split into two groups (3 types of culture media from 7 tissue locations listed above) and then placed in anaerobic or aerobic conditions at room temperature (24°C) using techniques described in Bryant and Burkey ( 1953. substitute chitin for cellobiose). Bryant (1972). and Hungate (1950). Anaerobic conditions were maintained under a gas phase of 100% CO, with- out agitation. Culture tubes were examined for the presence of fermentation end products (gas production) and media degradation (clouding of media and breakdown of larger chitin particles) after 12. 24. 48. 72, 96. and 120 hours. RESULTS Examination for Surface-Attached Microbes Examination of tissues using the scanning electron microscope showed no attached bacteria, fungi, microciliates. or other types of Figure 3. Scanning electron micrograph of iiiiionid iiilcstinal tract showing ciliated enterocvles and the loose junctions hetween the cells. Note, no obvious mass of endosymbiotic microflora can be observed. Scale bar = 10 pin. 332 Nichols et al. microbes found on the epithelial surface of various mantle and digestive organs (foot, gills, labial palps, stomach, ciliary tufts, crystalline style and style sac, digestive gland, and fore-mid-hind intestinal tract) except in one animal (Fig. la-d). All epithelial layers of these tissues were covered in March and August with a heavy layer of mucus. Food particles, including bacteria, were often seen trapped in this mucus layer (Fig. le and f). but no attached bacteria or other microbes were associated with this mu- cus layer or below it on the surface of the epithelium itself. The one exception was a fungal mat growing on the mantle tissue, at the base of the excurrent siphon, collected from one L. siliqiioidea in August (Fig. 2). Even though no obvious attached microbes were seen, every unionid examined, from all sites, contained what appear to be attached round structures, possibly spores, in their intestinal tracts. These structures were always found below the mucus layer and between the enterocytes comprising the intestinal wall (Figs. 3 and 4). They were large, about 1 |jl in diameter, and were attached by a stalk to the unionid tissue (Fig. 4d and f). Though consistently seen, attempts to rear these spores (if that is what they were) in isolation, for further elucidation, failed. Microbial Growth on Bacteriological Media Microbial growth varied by locality, media type, and season but not by unionid species, replicate, organ source, or the presence or absence of oxygen. All ( 100% of the 336 samples) of the samples from the Huron River population collected in March and August Figure 4. .Scanninjj fitclron micrographs of the unidentltli'd allacht'd spore-likf structures found In the Intestinal tract of various species of freshwater unionids collected from several sites in southeast Michigan: (a) I'yganodon grandis. Huron River; (b) iMinpsilis siliqiioidea. Huron River; (c) iMinpsilis ventricosa. Huron River: Id) Ptychobranchus fasciolaris, Huron River; (e) Elliptio dilatata, \ inevard Lake; and (f) Pyganodon grandis. Four Mile Lake. Scale bars = 1pm (a-e). Gi'T Microflora of Freshwater Unionidae 333 contained cellulose-, chitin-, and cellobiose-degiading bacteria, re- gardless of unionid species, organ from which the inoculant was obtained, or whether the sample was grown aerobically or anaero- hicallv (Tabic 1 ). In all 3,^6 samples, bacterial growth was noted within 1 2 h of inoculation based on gas production and clouding of the media. The bacterial communities developing in the cellulose and chitin media were similar in that all were motile, were fer- menters. and facultative anaerobes (Fig. 5). but they differed in size. The rod-shaped bacteria growing in the chitin were longer (avg. 3.96 |j.m) than those in the cellulose media (avg. 2.26 ixni). The bacterial community that grew in the cellobiose was domi- nated by non-motile cocci, which were also fermenters and facul- tative anaerobes. Rod-shaped bacteria were present, but even smaller in length (avg. 1.95 ta.m) than those found in the other types of iriedia. However, in the December samples, none of the 1 12 samples inoculated in either cellulose or chitin showed any growth up to 120 hours. In contrast, the December cellobiose samples ()! = 56) did show a lOC/c response to this media within 12 h, regardless of unionid species, organ source, or whether the sample was grown aerobically or anaerobically. The lake samples, P. grcmdis from Four Mile and E. dilamia from Vineyard, differed from Huron River unionids in that no microbial growth occurred in either cellulose or chitin media re- gardless of season, or organ source (total of 168 .samples for both lakes) (Table 1). However, like the Huron River animals, these lake unionids consistently showed activity in 100% of the 84 cel- lobiose samples, regardless of season, lake, or organ source. As with the Huron River unionid cellobiose samples, the bacterial community growing on the cellobio.se media was dominated by non-motile cocci that were fermenters and facultative anaerobes. In size and appearance they were identical to those observed in the Huron Ri\er unionid samples. DISCUSSION None of the unionids samples examined contained true en- demic microflora. In order to be considered a true endosymbiont, TABLE 1. Growth of niicrnorganisms removed from the lumen of mantle cavity (mantle) and digestive Iraet orjjans (gut) on cellulose, cellobiose, and chitin bacteriological media, under aerobic and anaerobic (in parenthesis) conditions. Season Species Cellulose Chitin Cellobiose March and Aug. Dec. Lumpsilis mantle +( + ) mantle M + ) mantle +(+) vemiicosa Huron River gut H+) gut +( + ) gut +(+) Lumpsilis mantle -)-(-H) mantle -K-I-) mantle +(+) siliqoiiidea Huriin River gut +( + ) gut M + ) gut +(+) Ptxchohrancl}tis mantle M + ) mantle +( + ) mantle +(+) fcisciolaris Huron River gut +( + ) gut +(-)-) gut +(+) P\i^anod(in mantle M + ) mantle +( + ) mantle +(+) grandis Huron River gut +( + ) gut -K+) gut +(+) Pygcmodon mantle 0(0) mantle 0(0) mantle +(+) gnmdis Four Mile Lake gut 0(0) gut 0(0) gut +(+) EUiptio dUatata mantle 0(0) mantle 0(0) mantle +(+) Vineyard Lake gut 0(0) gut 0(0) gut +(+) Lumpsilis mantle 0(0) mantle 0(0) mantle +(+) venlricosa Huron River gut 0(0) gut 0(0) gut +(+) Lampsilis mantle 0(0) mantle 0(0) mantle +(+) silicjuoidea Huron River gut 0(0) gut 0(0) gut +(+) Ptyclwbrunchus mantle 0(0) mantle 0(0) mantle +(+) fasciolaris Huron River gut 0(0) gut 0(0) gut +(+) Pyganodon mantle 0(0) mantle 0(0) mantle +(+) grandis Huron River gut 0(0) gut 0(0) gut +(-H) EUiptio dilauita mantle 0(0) mantle ()(()! mantle +( + ) Vineyard Lake gut 0(0) gut 0(0) gut -)-(-l-) Pyganodon mantle 0(0) mantle 0(0) mantle +( + ) grandis Four Mile Lake gLU 0(0) gut 0(0) gut -l-(-l-) Response code; -i- indicates bacterial growth and media degradation occurred within 12 hours of culture; 0 indicates no hactenal growth seen from 12-96 hours after inoculation. Sample size differs and is discussed in the methods section of text. 334 Nichols et al. Figure 5. Scanning electron micrograph ol the liacterial community cultured Irniii Ihc maiilic ca\il\ and t;iil cmilent of freshwater Unionidae. (a) shows the bacterial community that developed in cellulose media, while (b) shows the community that developed in the chitin media. Scale bar = I fim. microbial fauna must demonstrate certain characteristics which include: at least part of the endosymbiont community must be physically attached to the host's body tissues to ensure retention inside the animal; endosymbiont densities should be high OlO'"/ g), and; the endosymbiont must be consistently found inside the host regardless of season or location (Yokoyama and Johnson 1993; Taylor et al. 1995). These criteria were not met. One prob- lematic question remains regarding the unidentified spore-like structures attached or associated with e\ery unionid intestinal tract we examined. Our inability to culture or isolate enough of these structures for further identification prevents us from labeling them as endosymbionts or to hypothesize their relationship with the unionid host. While in appearance and size these structures re- semble the fruiting bodies, or zoospores, produced by fungus or a bacterium, a hyphal vegetative phase was only found once and in the mantle, not gut. The one exception, the fungal hyphae, attached to a specimen of L. siliquoidea (Fig. 2) and lacked spores or fruiting bodies. Even if these attached structures are spores, their limited numbers are not consistent with endosymbiont community criteria, and likely represent remnants of transient fauna. Fungi are a common component of the planktonic material drifting in the water of both Four Mile Lake and the Huron River (Nichols and Garling 2000). Occasional microbial spores are not uncommon in the intestinal tracts of some marine bivalves (Kueh and Chan 1985). The presence of copious amounts of mucus in the mantle ca\ ity and gut lumen of unionids (Fig. le) may mask or prevent the colonization of tissues by endosymbionts. Hyphal structures of fungi could easily be hidden in thick mucus layers. While surface- associated microsymbionts including fungi have been identified in many animals including fish (Mountfort and Rhodes 1491) and ruminants (Lowe et al. 1987; Sijtsma and Tan 1993) using the same techniques we used in this study, none of these vertebrates produce and utilize mucus in food handling and processing as do unionids. However, during December, when concurrent studies in the Huron River indicated unionids were not feeding (Nichols and Garling 2000). very little mucus was present inside the animals, and yet no attached microtlora could be observed. A true endo- symbiotic community would still have remnant, attached fauna, even when feeding was not occuning. Garland et al. (1982) have hypothesized that in oysters heavy mucus production prevents en- dosymbiont attachment on tissues. Such mucus hindrance might also limit the development of surface-attached endosymbiotic communities in unionids as well. The variability of the results of attempted culture of microor- ganisms on cellulose or chitin media support our conclusion that unionids lack true endosymbiotic microbes capable of digesting these substrates. Microbial growth occun-ed on cellulose and chitin during periods that concurrent studies in the Huron River indicated the unionids were feeding (late February to late November. Nichols and Garling 2000). Similarly, microbial growth on these culture media did not occur during December when the unionids in this river were not feeding. If endosymbionts were present, we should have been able to culture them in December samples. The differences in microbial response to culture media, in combination with the fact that no microbes were found attached inside the unionids, indicate that cellulolytic and chitinolytic microbial com- munities represent transient fauna and not endosymbionts. The total lack of microbial growth in cellulose and chitin media by the Four Mile and Vineyard lake samples is surprising. Likely, this may correspond to significant differences in lake versus river mi- crobial communities. The ability to culture a microbial community on the cellobiose media from all unionids, regardless of species, season, or habitat, is certainly a trait associated with endemic microbial fauna. How- ever, since no microbes were found attached to any unionid inter- nal tissues, we have concluded that these starch-degrading mi- crobes were transient fauna and not endosymbionts. While the spore-like structures found attached between the enterocytes of all unionid intestinal tracts could hypothetically be related to this cellobiose community, the lack of further identification of these structures prevents making this direct association. Assuming primary cellulases and chitina.ses are not endog- enously produced, the lack of cellulolytic or chitinolytic endosym- bionts in unionids means that initial bond breakage must rely on transient microbial fauna. A reliance on such transient microbes could influence captive maintenance success if dry diets based on cellulose (corn/soybean) or chitin (nauplii) are fed to unionids. Efficient utilization of these feeds by these bivalves would require prior microbial predigestion in order to provide the necessary initial bond breakage to break these complex polysac- charides. Unionids appear unK|ue among aquatic invertebrates in that they consume detritus and yet lack microbial endosymbionts that would aid in digesting complex polysaccharides. While this may imply that detritus merely serves as a convenient substrate for preferred food items, i.e., attached epimicroorganisms, further work is needed to fully identify digestive enzymes produced by Gut Microflora of Freshwater Unionidae 335 unionids. Dependence of unionids on microbes appears complex and combines elements of selective predation, subsequent en- hancement of non-prey species, and a reliance on transient fauna. Our observations do not support the endosymbiont model charac- teristic of well-known \erIebrate/microbe associations. The role of transient microbial fauna, particular!) the importance of different microbial species assemblages, may very well be key factors in- fluencing unionid survival in relocation and aquaculture efforts. ACKNOWLEDGMENTS We recognize and thank Dr. Sue Hengemuehle for her invalu- able assistance in culluring microbes and sample preparation. LITERATURE CITED BL'rnard. F. 1470, Occurrence of the spirochaete genus Crixtispini in west- ern Canadian marine bivalves. Veliger 13(l):33-.^6. Bryant. M. 1972. Commentary on the Hungate technique for culture of anaerobic bacteria. Am. J. Clin. Niitr. 2.'i: 1324-1 328. Bryant. M. & L. Burkey. 1953. Cultural methods and some characteristics of some of the more numerous groups of bacteria in the bovine rumen. J. Daily Sd. 36:20.S-217. Conway. N. & J. Capuzzo. 1989. The role of endosymhioni bacteria in the nutrition of Solemya veluiir. Evidence from a stable isotope analysis of endosymbionts and host. Limnol. Oceanogr. 34(1 ):249-25.'i. Cope. W. & D. Waller. 1995. Evaluation of freshwater mussel relocation as a conservation and management strategy. Rc.e"/ /^nc/s Res. Manag. I1(2):I47-I56. Farris. J.. J. Grudzien, S. Belanger. D. Cherry & J. Cairns. Jr. IW4. Mol- luscan cellulolytic activity responses to zinc exposure in laboratory and field stream comparisons. Hydroliiologiu 287:161-178. Garland, C. G. Nash & T. McMeekin. 1982. Ab.sence of surface-associated microorganisms in adult oysters (Cm.ssostrea gigas). App. Environ. Microbiol. 44(5 ):l 205-121 1. Gatenby. C. B.Parker. D. Smith, K. Duncan. & R. Neves. 1999. Use of pond refugia for holding salvaged unionid mussels. Abstract. In: The First Symposium of the Freshwater MolUisk Conservation Society. March 17-19. 1999. Chattanooga TN. Haag. W., D. Berg. D. Carton & J. Farris. 1993. Reduced survival and fitness in native bivahes in response to fouling hy the introduced zebra mussel {Dreissena pnlymnrplun in western Lake Erie. Can. J. Fish. Aqital. Sci. 50:1.3-19. Harris. J.. C. Burke & G. Maguire. 1998. Characterization of the digestive tract of greenlip abalone. Haliotis laevigata Donovan. II. Microenvi- ronment and bacterial flora. / Shellfish Res. 17(4):989-994. Hungate. R. 1950. The anaerobic mesophilic cellulolytic bacteria. Bacte- rial. Rev. 14:1-49. Jiffry, F. 1984. Loss of freshwater shellfish and some ecological impacts after water drawdown in Lake Sebasticook. Maine. MS Thesis. Orono, ME: University of Maine at Orono. Kueh, C. & K. Chan. 1985. Bacteria in bivalve shellfish with special reference to the oyster. / Appled Bacreriol. 59:41-47, Klug. M. & S. Kotarski. 1980. Bacteria associated with ihc gut tract ot larval stages of the aquatic cranefly Tipiila alnlonnnalis (Diptera:Tipul- idae). Appl. Environ. Microbiol. 40:408-416. Leilis. W. & C. Johnson. 1998. Experiences with captive maintenance of Atlantic slope unionids in small research systems. Abstract published in Freshwater Mussel Symposium: Conservation. Captive Care, & Propa- gation. Columbus OH. March 1998. pp. 25. Lefevre, G. & W. Curtis. 1910. Studies on the reproduction and artificial propagation of fresh-water mussels. Bull. U.S. Bur. Fish. 30:105-201. Lowe. S.. G. Griffith. A. Milne. M. Theodorou & A. Trinci. 1987. The life cycle and growth kinetics of an anaerobic rumen fungus. J. Gen. Mi- crobiol. f33:1815-1827. McMahon. R. 1991. Mollusea: Bivalvia. In: J. H. Thorp and A. P. Covich, editors. Ecology and classification of North American freshwater in- vertebrates. San Diego. CA: Academic Press, pp. 315-399. Mountfort. D. & L. Rhodes. 1991. Anaerobic growth and fermentation characteristics of Paecilomyces lilacinus isolated from mullet gut. Appl. Exp. Microbiol. 57(7): 1963-1968. Nichols, S. & D. Gariing. 2000. Food web dynamics and trophic level interactions in a multispecies community of freshwater Unionidae. Can. J. Zoology 78:871-882. Prieur, D., J. Mevel, G. Nicolas, A. Plusquellec & M.Vigneulle. 1990. Interactions between bivalve molluscs and bacteria in the marine en- vironment. Oceanogr. Mar. Biol. Annu. Rev. 28:277-352. Sijtsma, L. & B. Tan. 1993. Degradation and utilization of grass cell walls by anaerobic fungi isolated from yak, llama, and sheep, A?i/m, Feed Sci. Technol. 44:221-236. Starliper, C. R. Villella. P. Morrison & i. Mathias. 1997. Sampling the bacterial tloraof freshwater mussels. USGS Biological Information and Technology Note. No. 97-007. Taylor. M., S. Moss & M. Ladle. 1995. Scanning electron microscopy of the digestive tract of larval Sinmlium ornalnin Meigen (Complex) (Diptera; Simuliidae) and its associated microbial flora. Can. J. Zool. 73:1640-1646. Yokoyama. M. & K. Johnson. 1993. Microbiology of the rumen and in- testine. In: D. Church, editor. The ruminant animal: digestive physiol- ogy and nutrition. Prospect Heights, IL: Waveland Press. 564 pp. Wood, A. & D. Kelly. 1989. Methylotrophic and autotrophic bacteria isolated from lucinid and thyasirid bivalves containing symbiotic bac- teria in their gills. / Mar Biol. Assoc. UK 69:165-179. Williams. J., M. Warren, Jr., K. Cummings, J. Harris & R. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Am. Fish. Soc. l8(9):6-22. Jiiiinuil «/ Shclljlyh Resi-anh. Vol. 20, No. 1, 337-346, 2001. METHODS FOR MASS REARING STAGES I-IV LARVAE OF THE AMERICAN LOBSTER, HOMARUS AMERICANVS H. MILNE EDWARDS, 1837, IN STATIC SYSTEMS BRIAN F. BEAL' * AND SAMUEL R. CHAPMAN" ^ University of Maine at Mucliids. 9 O'Brien Avenue. Macliias. Maine 04654: 'WaldolMno Shad Hcilchery. Walclolwro. Maine 04572 ABSTRACT We conducted a series of five lahoratoiy experiments (7-18 days in duration) to test the interactive effects of stocking density, aeration rates, and food types on survival of American lobster ^Homunis americainis) larvae through their first three planktonic stages (I-III) to the postlarval stage (IV). Experimental units and culture protocols were designed to replicate a 1:100 scaled-down version of equipment used in association with a fishermen-sponsored, stock enhancement lobster hatchery located in Cutler. Maine. The first four trials revealed that extremely high rates of aeration (ca. 240 niL air sec"') were necessary to distribute larvae and food sufficiently to reduce cannibalistic encounters: however, the best survival from stage I-IV (at stocking densities of 7-26 L"' fed ad lihinim with enriched Arremia) was only 24%, The final experiment (stocking density = 20 L"') yielded a mean survival rate (± 95% CI) of 75.8 ± 10.2% (range = 62.7'7o to 90.1%: n = 6). One important difference between the last and first four experiments was how stage I larvae were managed prior to their culture. In the first four trials, unfed larvae were collected from a relatively small (46 cm X 30 cm X 20 cm), screened capture basket located near the discharge pipe of a broodstock holding tank at the hatchery where they may have resided for > 12 hr. Lar\ae used in the final laboratory experiment were collected directly from the broodstock tank within 30 min after being liberated from the mother's swimmerets. Larvae, at relatively high densities within the screened box, likely had many more cannibalistic encounters prior to their culture than those collected directly from the broodstock tank and, therefore, suffered high rates of mortality during the first four laboratory trials. Mass rearing methods for larval American lobsters developed in conjunction with these laboratory experiments were used successfully by staff at the Cutler Marine Hatchery from 1988 to 1992. During this period, survival from stages I-IV averaged 44%, and approximately 875,000 stage IV animals were released to the wild. These culture methods have withstood the test of time as a private lobster hatcheiy in Maine adopted our protocols in 1993, and they continue to be in use. Further, the general techniques described here have been used since 1994 to culture European lobsters (Homarus gammarus) at a commercial lobster hatchery in the southeast of Ireland. KEY WORDS: lobster, Hiwhini.s amcricamis. culture, static systems, stock enhancement, Anemia, nncroalgae, stage IV larvae INTRODUCTION Mass production of juvenile lobsters {Hoitianis americanus H. Milne Edv\ards) for stock enhancement programs began as early as IS72 in the United States (Rathbun 1886). Lobster "parks," natural tidal impoundments, were used in Maine and Massachusetts as spawner sanctuaries where egg-bearing females were held until releasing their stage I larvae. The earliest lobster hatchery was established in 1883 at Woods Hole. Massachusetts. Here, eggs were detached from females, placed into hatching jars, and stage I lobsters released to the wild (Nicosia and Lavalli 1999). In time, enhancement of lobster stocks using early planktonic individuals was abandoned in favor of methods to rear larvae for release at their earliest benthic stage (reviewed in Nicosia & Lavalli 1999). Perhaps the most successful long-term stock enhancement ef- fort in the United States with respect to juvenile production oc- curred at the Massachusetts State Lobster Hatchery and Research Station on Martha's Vineyard from 1931 to 1997. Culture proto- cols and equipment were developed that became the standard for production of massive numbers of stage IV individuals (Hughes 1972, Hughes et al, 1974). In particular, a 40-L round-bottom rearing vessel, or kreisel. was invented thai, with sufficient flow rates (ca. 10 L min"'), provided enough turbulence to distribute homogeneously both larvae and food, and minimized cannibalistic encounters (Hughes et al. 1974), Stocking these kreisels with 2,000 stage I larvae of W. americanus and supplying them with live brine shrimp [Anemia .mlina L.) during the ten to fourteen days in culture reportedly resulted in survival rates to stage IV between 75%-85% (Hughes et al. 1974, Serfling et al. 1974a, Schuur et al. *Corresponding author: E-mail: bbeal@acad.uniin, niainc.edu 1976, Van Olst et al. 1980, Aiken & Waddy 1989, Waddy and Aiken 1998). Improvements on the basic design enabled the cul- ture system to operate efficiently either using recirculating or flow- through seawater (Sertling et al. 1974a, Schuur et al. 1976, Van Olst et al. 1977). During the past thiee decades, the "Hughes pot," or planktokreisel, has been used widely by scientists in the United States and Canada to produce large numbers of juvenile lobsters for enhancement of wild stocks, commercial operations, and labo- ratory and field experiments. In addition, similar larval culture systems have been used in Europe with H. i^ammanis L. (Beard & Wickins 1992. Burton 1992, Browne 1999). In Maine, during the mid-1980's, lobster fishers decided they wanted to fund up to five lobster hatcheries along the coast to produce stage IV individuals for stock enhancement purposes (Plante 1986). At the same time, one of us (S. Chapman) was working on a simple system to rear lobster larvae from stages I-IV in 400-L conical tanks similar to those u.sed to culture bivalve larvae (Castagna 1983). The process involved culturing phy- toplankton to enrich brine shrimp that were then added (after a period of growth) to static culture tanks containing lobster larvae. The work was carried out primarily because Chapman was unable to repeat the success of others using the smaller 40-L kreisels (Hughes et al. 1974) and he was interested in developing culture protocols that could be transferred easily to groups of fishers. In 1986, the first fishermen-sponsored stock enhancement pro- gram for lobsters in Maine was initiated in the eastern coastal community of Cutler, and the Cutler Marine Hatchery (CMH) was created (Nicosia & Lavalli, 1999). From 1986 to 1992, this pro- gram was the only enhancement effort funded by lobster license fees through Maine's Department of Marine Resources, Although we both served as technical advisors to CMH during this time, the 337 338 Beal and Chapman hatchery was managed by a fisherman and elderly housewife who. before that time, had no technical training in any aspect of the biological sciences. During the first months of production at CMH in 1986. survival rates of stage I-IV lobster were very poor (< 10%). Here, we describe a series of five sequential laboratory experiments con- ducted between July and October 1986 that were designed to im- prove on this survival rate. MATERIALS AND METHODS A series of five larval lobster culture experiments (A-E) were conducted at ambient room temperatures at the University of Maine at Machias (UMM) between 15 July and 17 October 1986. Experiments were designed to assist staff at the CMH with culture protocols to maximize lobster larvae survival from stages I to IV. At UMM. it was not possible to replicate the culture vessels em- ployed at Cutler's Hatchery (ca. 4()0-L conical-shaped fiberglass tanks filled to 375 L and aerated from the apex of the cone). Instead, we used 4.73-L plastic, flat-bottomed, round buckets (25.4 cm diameter at the top. tapering to 20.3 cm diameter at the base). A 3.2 mm hole was drilled in the middle of the bottom of each bucket (experimental unit) and a ceramic air stone fitted into each hole. A piece of tubing connected each ceramic stone to a manifold that received a supply of air from a pump ( 1-hp Conde. Westmoor Ltd.. Sherill. NY). All experiments were conducted in 3.78 L (1 gallon) of seawater. which represented approximately 1/100"' of the water volume used at Cutler. Buckets were not covered; there- fore, photoperiod varied from one experiment lo another. An at- tempt was made to detemiine if experimental results could be re- peated at the Cutler Marine Hatchei^ (see Results — Expenment B). Stage I lobster larvae used in Experiments A-D (Table I ) were collected at 0700 hr from a screened wooden box (46 cm x 30 cm X 20 cm) located near the discharge pipe of a broodstock holding tank at CMH. where they may have resided for > 12 hr. Larvae were transferred immediately to moist paper toweling nested within a stainless steel sieve (64 p.). placed inside a styrofoam cooler, and transported to UMM (a process that took between 30 to 45 min). Larvae used in Experiment E were taken directly from the broodstock tank with a small fish net only moments after they had been liberated from the mother's swininierets {sensu Talbot and Helluy. 1995). Filtered (I |ji) seawater used in all experiments was collected at CMH. At UMM. larvae were immediately transferred to a 40-L aquarium from which indi\iduals were removed, counted, and added to experimental units containing a source of food (see be- low). We assumed that animals transferred to experimental units were unaffected by our handling, although we did not conduct a specific handling experiment. We used only larvae that were swimming prior to initiating each experiment. Every 48 hours until the end of an experiment, larvae in this static culture system were poured through a nylon fish net (1 mm aperture) that retained TABLE I. Duration and treatments used in Experiments A in = 4), B-D (h = 3). and E (n = 6). Experiment Duration' Densitv/unit" Food type' Aeration rate^ 15-21 JiiK 2S July-ll August 1 1-25 .August 4-22 September 1-17 October 1(1(1 lUU 150 150 100 100 150 150 200 200 50 50 100 100 150 150 25 25 50 50 100 100 75" 75" LBS LBS LBS LBS LBS LBS LBS LBS LBS LBS Clam FBS Clam FBS Clam FBS Clam + LBS Mackerel + LBS Clam + LBS Mackerel + LBS Clam + LBS Mackerel + LBS LBS LBS Low Moderate Low Moderate Moderate High Moderate High Moderate High High High High High High High High High High High High High High Hieh ' Experiments conducted in 1986. - Number per 3.78 L (1 gallon). ' LBS = Live, enriched Brine Shrimp; FBS = Frozen Adult Brine Shrimp. ■* Low = 20 mL air sec '; Moderate = 80 mL air sec"'; High = 240 niL air sec"'. " Octagonal strips (33 cm x 33 cm) of black plastic mesh (6.4 nim aperture) added to each experimental uni( ''Control; no plastic strips added to experimental units CuLTURiNG Larval Lobsters 339 larvae and any uneaten food. Larvae were returned immediately to a clean (rinsed with 60°C freshwater) bucket containing 3.78 L of seawater (room temperature) and food. Live brine shrimp (Arteinia salina). used as a .source of food in Experiments A. B. D. and E. were reared in 4()-L glass aquaria at UMM. Artemia were fed algae (Tahitian Isochiysis galhana Parke [clone T. iso] and Chaelocerns gracilis Ehrenberg) ad libitum that had been cultured at CMH. Brine shrimp were cultured for four days (total length ca. 2-3 mm) at a temperature of 30^C under constant fluorescent lighting before transferring to e.xperimental units (ca. 40 brine shrimp per mL). In all experiments, there was always an excess number of live brine shrimp in experimental units after 48 hours. This number was estimated by counting a I mL sample from each unit using a Sedgewick-Rafter cell. These excess brine shrimp were returned to the experimental unit from which they had come and new. four day-old brine shrimp were added to bring the density to 40 mL"'. In Experiments A and B (Table 1 ). aeration rate was one of two factors tested. Altogether, three aeration rates were examined (Low = 20 mL air sec"'. Moderate = 80 mL air sec"'. High = 240 mL air sec"' as measured crudely using a water displacement technique). Qualitatively, the low flow rate resulted in a few small air bubbles per second. Water movement was minimal and re- stricted to a small column the width of the air stone (ca. 20 mm) in the center of each bucket. At low aeration, both larvae and brine shiinip were visible; however, lobster larvae appeared to congre- gate on or near the bottom of the buckets. With moderate aeration rates, water motion also was restricted to the center of the buckets, although the diameter of the circulation zone was larger (ca. 80 mm). At this flow rate, larvae were moved in a gentle upward motion and then immediately downward in a vertical circulating current. High rates of aeration resulted in vigorous vertical and lateral movement of larvae throughout the water column. Larval motion was too rapid to observe individuals. At these aeration rates, the water surface within each bucket was moving so rapidly that it appeared to be boiling. Experiments A-D were completely factorial with two fixed factors per trial. Experiments were sequential in that results from a particular trial were used to design the ensuing trial. In Experi- ments C and D, food type was varied. For Experiment C, 6-8 mm pieces of fresh soft-shell clam, Mya urcnaha L., mantles and similar sized frozen adult brine shrimp were used to feed lobster larvae. In Experiment D, 6-8 mm pieces of fresh clam and frozen and thawed muscle tissue from mackerel (Scomber scombnis L.) were added to experimental units containing live, algal-enriched brine shrimp. In Experiment E. octagonal strips (33 cm x 33 cin) composed of extruded, black plastic netting (6.4 mm aperture) were positioned at a 45' angle within six experimental units. Net- ting extended from the bottom of the experimental unit to a few cm above the surface of the seawater. The purpose of the netting was lo increase spatial heterogeneity within the experimental units and provide a surface or substrate for larvae to encounter and. poten- tially, reduce conspecific cannibalistic encounters. .Another six units served as controls (no netting). For Experiments A-D, a two-factor, model I analysis of vari- ance (ANOVA) was performed on the angular-transformed per- cent survival data (Sokal and Rohlf 1995). Orthogonal, a priori contrasts were performed for each source of variation with more than a single degree-of-freedom (see Results). A single-factor, model I ANOVA was performed on the angular-transformed per- cent survival data from Experiment E. In all cases, the transfor- mation served to either normalize the data (Shapiro-Wilks test [SAS 1998]) and/or homogenize variances (Cochran's Test [Winer et al. 19911). RESULTS Experiment A The experiment was terminated on day seven because exces- sive heat inside the laboratory on 21 July (32°C), in addition to high cannibalism rates during the first six days, resulted in the deaths of 98% of the larvae. Only 41 stage III individuals of the initial 500 larvae stocked in the experimental units (Table 1 ) were alive on 2 1 July. There was no cumulative effect on survival to this stage due to density (P = 0.094); however, nearly 9x as many larvae survived the seven-day trial in experimental units receiving moderate (37 animals) versus low (4 animals) rates of aeration (P = 0.021). Experiment B Seawater temperature varied throughout the experiment from 19° to 24°C, which is within the range of natural conditions ex- perienced by H. americcmus larvae (Van Olst et al. 1980). Salinity was constant at 31 ppt. Eight days after the experiment was initi- ated, a mean of at least 88* mortality had occurred in all treat- ments (Table 2). By this date, all surviving larvae (/; = 216) had reached stage III in their development. On 1 1 August, fourteen days after the experiment was initiated, only 124 of 900 (13.8%) larvae remained across all six treatments of which 1 19 (96%) had metamorphosed to stage IV. ANOVA revealed that all main and interaction sources of variation were significant (P < 0.05; Table 3). The effect of aeration rate on mean survival depended on stocking density. A priori contrasts indicated that 7.7% of the larvae reared at 100 unit"' (ca. 26 L"' ) survived from stage I to IV independent of aeration rate. This was nearly twice the average TABLE 2. Percent survival (±95 '^f CI) of larval lobsters hi Experiment B (28 Julj to II August 1986). Density .Aeration Day 6 Dav 8 Day 10 Dav 12 Dav 14 100 \iO 200 Moderate High Moderate High Moderate High 21.5(12.35) 18.0(9.43) 13.3(8.72) 18.6(6.65) 15.5 (6.95) 9.8 (7.56) 9.7(1.47) 12.7 (6.25) 5.6(6.69) 10.4(3.40) 5.5 (5.46) 7.3(I0.(M) 7.3(1.43) 11.3(6.22) 4.9(6.27) 7.8(7.47) 4.7(2.58) 4.8 (4.70) 6.0(4.31) 10.7 (6.25) 2.7 (4.38) 9.2(8.29) 3.9 (4.30) 4.5(1.25) 5.7 (2.87) 9.7 (6.25) 1.9(1.65) 6.9(4.17) 3.2 (2.46) 3.2 (3.99) Density = number per 3.785 L. Aeration (Moderate = 80 mL air sec '; High = 240 mL air sec '). Day refers to days after hatching, n 3. 340 Beal and Chapman TABLE 3. ANOVA on the angular-transformed survival data from Experiment B (1 1 August). Source of variation df Sum of squares Mean square F value Pr > F Aemlion Density (100 vs. [150 & 2001/2) (150 vs. 200) Density x Aeration (100 vs. [150 & 200[/2l X moderate ' (150 vs. 200 X moderate vs. high) Error Total hi>;h 1 b6.59 66.59 17.43 0.(»13 2 106.76 53.38 13.98 0.0007 1 100.28 100.28 26.26 0.0003 1 6.48 6.48 1.70 0.2172 2 .^6.47 18.23 4.77 0.0298 1 0.44 0.44 0.11 0.7412 1 36.03 36.03 9.44 0.0097 12 45.83 3.82 — — 17 2.55.65 — — — A jiriori contrasts appear with density and interaction sources of variation. See Table 2 for aeration rates, n survi\al rate (3.8'7f ) of larvae at the two higher densities (P = 0.0003). High aeration rates significantly improved survival rates of larvae stocked at \5Q unit"' (40 L~') compared with moderate aeration rates, but this same improvement did not occur for larvae initially stocked at 200 unit"' (53 L"'>. Because this experiment had shown that high aeration rates (ca. 240 mL air sec"' ) and a density of 100 L"' may result in 5-10'7f survival, a two-week test of these culture variables was initiated in three 400-L conical tanks at CMH from 1 1 to 25 .August. Culture ves,sels at Cutler were treated similarly to those in the laboratory. We used algal-enriched, live brine shrimp (40 mL"' ) as a source of food and added approximately 10,000 stage I larvae to each vessel, which was aerated vigorously. A regression equation was devel- oped using mass (g) and larval number to estimate number of stage I larvae stocked per conical tank (Fig. la) and number of stage IV larvae (Fig. lb). These equations enabled us to estimate percent sur\ ival on a per-tank basis over the two-week culture period at CMH. Culture conditions at Cutler were very similar to those at UMM. Seawater temperature varied between 20" and 24°C and salinity was 31 ppt during the two-week interval. Survival of stage I to IV larvae in the three conical tanks was: 7.5%, 10.2*7^ and 4.9%. ExperiiiHiil C Previous experiments had demonstrated that mortality between stages I and III was severe (usually > 90%) in both the large conical tanks at CMH and the experimental units at UMM. Be- cause we observed many cannibalistic encounters during the first week, one possible explanation for this might be that larvae pre- ferred a larger food particle than the 2-3 mm brine shrimp. Seawater temperature varied between 21.5° and 24°C and sa- linity fluctuated between 31 and 34 ppt. On 15 August, four days after the experiment was initiated, survivorship summed across all treatments was 24.1% (Table 4). When the experiment was termi- nated on 25 August, mean survivorship over all treatments was 1.5%. ANOVA (Table 5) demonstrated both a density and food type effect. Larvae survived best (2.4%) at the lowest density (50 unit"') and frozen brine shrimp provided a significantly higher survival rate (1.9%>) than chopped soft-shell clam (1%) (P = 0.0307, Table 5). Experimenl D Because the extremely poor survival rates observed in the pre- vious experiment suggested that larger food particles, alone. should be rejected as a source of food for culturing lobster larvae, we examined the interactive effects of diet (large food particles [6-8 mm pieces of clam and mackerel] together with live brine shrimpi and slocking density (25, 50. 100 larvae unit"' I. Seawater temperature varied between 17" and 23°C and salinity between 31 and 35 ppt during the 18-day trial. On 16 September, an experimental unit from the low-density fish treatment was found without water. All larvae inside were dead and this replicate was removed from the experiment. During the final two days of the experiment (20 to 22 September), all four larvae found in one experimental unit from the high-density fish treatment appeared to have become infected with the ciliate VorricelUi spp. No other larvae appeared to have the ciliate. Unlike previous experiments, overall mortality did not exceed 80% until the 16''' day of the experiment (20 September, Table 6). The experiment was terminated on day eighteen when 96% of the individuals had metamorphosed to stage IV (Table 7). Because it is not possible to separate efficiently stage III from stage IV indi- viduals. Table 7 can be used as a guide in developing strategies for releasing early benthic phase H. amerkaims from hatcheries (such as Cutler) that do not have space to rear stage IV -i- animals to larger sizes either individually or communally. For example, on Day 14, 35% of the animals were at stage III while 65% were at stage IV (0.65 x 180 = 117 stage IV individuals). Two days later, 94% of the remaining larvae were stage IV (.94 x 143 = 134). The additional 48 hours yielded 15% more stage IV individuals even though the percent of individuals reaching the fourth stage in- creased over this same time interval by nearly 30% (65% to 94%). This suggests that both overall percent survival and percent of individuals that have attained the first benthic stage should influ- ence decisions when to release larvae to the wild. ANOVA (Table 8) was unable to detect a difference in mean survival between the two food types (clam + live brine shtlmp = 19% vs. mackerel -i- live brine shrimp = 13%. P = 0.084). There was no overall effect due to density; however, the a priori contrast of low (22%) vs. the mean of the two higher densities ( 13%) was significant i.P = 0.048). Although mean survival at the lowest density was better than in previous experiments (Table 6), these rates were much lower than published accounts of lobster larval survival to stage IV. either by other public hatcheries (e.g., the Massachusetts State Lobster Hatchery at Oak Bluffs, Martha's Vineyard. MA. Hughes et al.. 1974) or research laboratories (Serfling et al. 1974; Schuur et al. 1976; Aiken et al. 1981; Eagles et al. 1984). CuLTURiNG Larval Lobsters 341 250 500 750 Number of Stage I Larvae Number of Stage IV Larvae Figure 1. Mass();)-to-number relationships for stage I la) and IV )b) larvae of Hoinanis umericaiiiis. (Stage 1: \ = (1.(1824 + 0.(10633 X, n = 30, r- = 0.988, 157.9 larvae g"': Stage IV: Y = 0.0896 + 0.03821 X, ii = 16, r^ = 0.999, 1 1.2 larvae g"' ). Larvae of both stages were caught using a nylon mesh (I mm aperture! and transferred to a sieve (125 ji). A piece of paper toweling was used to absorb moisture from the bottom of the screening for 30 seconds before measuring mass using a Sarto- rius electronic balance to the nearest 0.01 g. ExpcrimeiU E Seawater temperature varied between 17^ and 22^C and salinity between 30 and 33 ppt. For no apparent reasim. all larvae within one of the buckets containing a strip of the plastic netting died within the first 36 hours. On Day 15, 83'7f of all surviving indi- viduals had reached stage IV. When the experiment was termi- nated 48 hours later, 98% had metamorphosed to stage IV. Dra- matic differences in survivorship occurred between this test and all previous experiments (Fig. 2). By the end of the experiment. 18.1 ± 13.9% (95% CI) (range = 6.7% to 34.7%) of the animals in the experimental units containing plastic strips had survived to stage IV whereas 75,8 ± 10.2% (range = 62.7% to 90.7%) of the larvae attained stage IV successfully in the control units. These survival rates were highly significantly different (P < O.UOOl) Transfer of Techniques from the l.ahonttory lit llw Culler Marine Hatchery In 1986. an estimated total of 20,000 stage IV lobsters were released from CMH. The best survival rates observed during that season, which lasted from mid-.lune to mid-.Sepiember was 12.5% (25 August to 9 September; Beal. pers. obs.). In 1987. using tech- niques learned from Experiment E (above), approximately 93.000 stage IV animals were released from CMH during a similar time interval. Survival rates were as high as 60%'. but averaged 42% for the year (Beal. unpubl. data). From 1988 to 1992 (the last year CMH operated), an average of 175.000 stage IV lobsters per year were released by staff at CMH into coastal waters near Cutler, Maine (Beal et al. 1998). During that time, survival in the 400-L conical tanks varied between 25% and 65% with an average of 44%. DISCUSSION Experimental results and observations involving the initial han- dling of stage I lobster larvae described here led to the adoption of culture techniques that enabled staff at Cutler's Marine Hatchery to become successful in rearing Honuirus aiuericanus larvae through their planktonic stages (I-III) to first benthic stage (IV) for stock enhancement purposes. That program operated from 1986 through 1992 when public funding for lobster hatcheries in Maine ceased. The culture techniques described below have been in use since 1993 by staff at a private facility in Maine (the Southwest Harbor Oceanarium. SHO) who are engaged in stock enhancement efforts in the waters near Bar Harbor, Maine (T. Montague, hatch- ery manager. SHO. pers. com.. 7 luly 2000). In addition, similar culture techniques have been adopted for Homcmis gammarus pro- duction of stage IV animals for stock enhancement efforts in Ire- land (Browne & Mercer 1988). In Maine, wild broodstock were obtained from fishers on an as-needed basis. Maine's Department of Marine Resources issued special permits annually to those who were active in the CMH stock enhancement program. (Aiken and Waddy |1985) reported that in Canada, it was difficult to obtain egg-bearing females from fishers. This led them to develop methods to domesticate brood- stock for year-round seedstock production). In eastern Maine, broodstock with eggs ready to hatch (i.e.. those that display a light bluish hue), were commonly available beginning the second week of June through mid-September from 1986 to 1992 (Beal. pers. obs.). Beginning in 1987 and continuing through the 1992 season, egg-bearing females entering CMH were immediately bathed in a 19f solution of Betadyne (povidone iodide) in filtered (1 (x) sea- water to remove fauna and ectoparisites such as copepods, nema- todes, or Vonicella spp., as well as reduce the probability of bac- terial infections. The bath consisted of immersing the entire ani- mal, except its mouthparts and gills, for 8-10 min. Egg-bearing females were then moved to a static, communal tank ( 1.320 liters) that was completely covered to provide constant darkness. No more than five female lobsters were held in this tank at any time. .Ambient, filtered (1 \y.) seawater in the tank was changed every two days and the tank thoroughly cleaned on a weekly basis. Hatching occurred generally throughout the day. but the hatching period for a particular female lasted no more than five days (see Hughes and Matthiessen 1962). Stage I swimming larvae were taken from the broodstock tank within 30 min of being expelled from the clutch using a small nylon aquarium fish net. Sometimes, large batches of stage 1 larvae were available early in the morning 342 Beal and Chapman TABLE 4. Percent survival (±95'>'f CI) of larval lobsters in Experiment C (11-25 August 1986). Density Diet l)a> 4 Dav 6 Da> 8 Day 10 Day 12 Day 14 50 M. aremiria 30.0(4.97) 13.3(22.39) 5.3 (5.73) 2.0 (0.00) 2.0 (0.00) 2.0 (0.00 Brine shrimp 30.0(30.22) 16.7(28.68) 8.0 (9.941 4.7(2.87) 3.3 (5.74) 2.7 (2.87) 100 M. arenaria 20.0(7.45) 7.0(9.93) 3.0 (4.97) 1.7 (3.79) 1.0 (2.48) 0.7(1.43) Brine shrimp 30.7(19.92) 13.3(10.34) 8.3 (8.69) 4.3 (7.99) 2.0 (2.48) 2.0(2.48) 150 M. tireiniriii 18.0(7.22) 8.0(3.32) 3.3 (2.87) 2.2 (3.45) 0.7 (1.66) 0.4 (0.96) Brme shrimp 15.6 (9.9S) 64(3.82) 2.9 (2.52) 2.2 (0.95) 1.3 (0.00) 0.9 (0.96) Densily = number per 3.78 L. Aeration (High Day refers to days after hatching, n = 3. 240 niL air sec ' ). Diet (6-8 mm pieces): soft-shell clam dV/. un-nuria) or frozen hrine shrimp. (i.e., after CMH staff came to woric ca. 0700); however, these larvae were not added to culture tanks. Instead, they were liberated because results from E,x.periments A-E suggested that a dense assemblage of unfed stage I larvae would not do well under cultt(re conditions. After collecting recently hatched larvae, each netted batch was weighed to provide an estimate of numbers placed into a 400-L culture tank (Fig. la). Over a two-day period, between 7,300 and 10,000 stage I larvae were added to a vigorously aerated (front the apex of the cone) conical tank filled to ca. 375 L (ca. 20-27 larvae L"'). Tanks contained 275 L of seawater ( I8°-20°C) plus 100 L of a I ; I mixture of the flagellate Tahitian Isochrysis galbcma and the diatom Cluietoceros gracilis or C. nuilleri Ehrenberg at a cell density of 2^ x 10'^ cells mL"'. In addition, lobster larvae re- ceived a supply of live, four day-old Arteiiiia (initially 40-50 mL"') that were cultured and enriched at CMH. Approximately 80 niL of dry brine shrimp eggs (Aquarium Products*, Glen Burnie, MD) were first decapsulated (Brugger- nian et al. 1980) and then added to a 400-L conical tank containing only 100 L of filtered (1 (jl> seawater (25°-29°C). After brine shrimp had hatched, the tank was filled with cultured algae (as above). After four days, the tank was drained and brine shrimp (ca. 3 mm) retained on a small sieve, the contents of which were added to one 400 L conical tank with laival lobsters. Lobster larvae remained in the static culture tanks for two days before being transferred to an adjacent, clean tank containing vig- orously aerated seawater, microalgae, and brine shrinip. The trans- fer process lasted no more than two minutes. We devised a dip net that was approximately 40 cm long x 30 cm wide x 20 cm deep made from nylon window screening. To transfer larvae, we would dip once fiom the larval tank and imniediately empty the contents of the net into the clean tank. After six to seven dips, 959c of the larvae had been transferred. Next, we turned off the air and drained the seawater from the larval tank, quickly capturing the remaining larvae on a sieve (500 |j.). This iT)ethod also retained uneaten, live brine shrinip that also were added to the clean tank along with the lobster larvae. This transfer process continued for approxiniately two weeks, or until 85-90'/f of the surviving larvae had reached stage IV, when they were released to the wild. It never was nec- essary to add any antibiotics or other medication to a culture tank. From 1988-1992, survival rates from stage I to IV at CMH aver- aged 449r. These early benthic phase aniinals were released at depths of 10-20 m on or near cobble bottom in along a 170 km stretch of coast from Cutler (44°I4'N, 67°27'W) to Tenant's Har- bor (43°33'N, 69=03'W) Maine. The larval culture techniques employed at CMH were different from methodologies that had been used to that date (Hughes et al. 1974; Serfling et al, 1974a; Schuur et al. 1976; Van Olst et al. 1980) primarily because of work conducted at the University of Maine's Darling Marine Center (DMC, Walpole, ME) from 1983- 1985 by one of us (.S. Chapman). Prior to the mid-l980"s, most culture work on planktonic American lobsters was conducted in 40 L round-bottoni, fiberglass tanks known as kreisels, planktokre- isels, or "Hughes pots," after work published by John Hughes, former director of the Massachusetts State Lobster Hatchery (MSLH; Oak Bluffs, Martha's Vineyard. MA; Hughes et al. 1974) and his colleagues in California (Serfling 1974b; Ford et al. 1975; Van Olsl et al. 1977). Highest estimated survival rates to stage IV per kreisel (35-85%) occuired when stage I larvae were stocked at densities between 500 and 3,000 (12.5-75 L"' ) and fed either live or frozen Anemia throughout the culture period (Hughes et al. 1974), Van Olst et al. ( 1980) reported that. "When food and larvae are uniformly distributed and proper densities of both are )T)ain- tained, larval survival to staae IV will average 60-75% ." Schuur et TABLE 5. ANOVW on the annular-transformed survival data from Experiment C (25 .\ugust). Source of variation df Sum of squares Mean square F value Pr > F Density (50 vs. [100 & 150J/2) (100 VS.150) Food Density x food Error Total 1 1 1 2 12 17 61.28 53,02 8.26 27.88 6.86 55.87 1 5 1 .89 30.64 53.02 8.26 27.88 3.43 4.66 6.58 11,39 1,77 5.99 0.74 0.01 IS 0.0055 0.2075 0.0307 0.4990 See Table 4 for description ot food types CuLTURiNG Larval Lobsters 343 TABLE 6. Percent survival (±95 "^f CI) F Density (25 vs. [50 & 1001/2) (50 vs. 100) Food Density x food EiTor Total 1 1 1 2 11 16 252.02 161.01 91.01 117.46 4.19 358.01 731.68 126.01 161.01 41.01 117.46 2.09 32.54 3.87 4.95 2.80 3.61 0.06 0.0533 0.0480 0.1227 0.0840 0.9379 Refer to Table 6 for specific information about food type, n = 3. pieces of infoirnation we discovered during our experimental trials that subsequently influenced our culture practices was in the tim- ing of collection of the stage I larvae from the broodstock tank. We have been unable to find many specific references in the literature concerning this aspect of culturing lobsters. Aiken and Waddy (1995) state, "'Hatching larvae are screened from the outflow of the maternal female's tank each evening and stocked in specially de- signed larval tanks (planktokreisels) that disperse water in an up- ward rotation that keeps the larvae in uniform suspension, reducing injury and cannibalism." Sastry (1975) and Chang and Conklin (1993) each describe a broodstock holding/larval collection system similar to those used at CMH in 1986. That is. filtered seawater flows unidirectionally into individual compailtnents holding egg- bearing females, which carries stage I lobsters into plexiglass catch baskets (no size given) with 1 mm mesh screening. "Each morning the removable catch baskets are examined for larvae, and. if present, the larvae are rinsed into a larval-rearing system" (Chang and Conklin 1993). Baskets facilitate the ease with which stage I larvae are collected and transferred to kreisels; however, we aban- doned their use at CMH after 1986. We found both experimentally and in routine culture situations that, because of their cannibalistic tendencies, stage I larvae, if confined to small areas at high den- sities without Anemia or other food, either die immediately or become stressed or injured, resulting in high rates of mortality (i.e., > 50%) within the first week of culture (e.g.. Experiments B-D). If capture or collection baskets were emptied often enough (i.e.. before a critical density occurred within each), they would be useful. Instead, we opted to place 3-5 broodstock at a time in a small ( 1 320 L), covered, static, lightly-aerated tank and collected larvae periodically (every 30-60 min) through the day using small, aquarium, fish nets. Stage I larvae hatched during the evening were not added to culture tanks at CMH; instead, they were liberated in a cove adjacent to the hatchery. The culture methods outlined here are not specific to the American lobster. In 1992. a lobster (Hoinanis t>ciiiiiinini.s) stock enhancement effort was initiated in Ireland at the National Uni- versity of Ireland, Galway Shellfish Research Laboratory (SRL) in Cama (53° 1 8' N, 9° 49' W). The system was based on using 40-L kreisels (Hughes et al. 1974) that were connected to a healed seawater reservoir system ( 18°C). Stage I larvae (density = 25-37 L~') were suspended by a continuous tlow of recirculated seawater at 9 L min ' and fed a combination of chopped mysid shrimp {Neomysis spp.), mussel (Mytilus edulis L.), and occasionally supplemented with live Anemia. After 17 days, survival rates were 9.8% (Browne 1999). In March 1993, the culture system was converted to one more closely resembling the one at CMH (Browne 1999: Beal, pers. obs.). Similar mass-to-number relation- ships were discovered for stage I H. gammants larvae (Browne 1999) so that careful stocking estimates could be obtained. After 1 1-14 days in the hatchery. Browne and Mercer (1998) observed an average larval survival rate to stage IV of 47%. The culture techniques used at SRL were transferred to a fish- ermen's cooperative lobster hatchery in southeast Ireland during the mid-1990's. In August 2000, one of us (B. Beal) visited this hatchery (Came. County Wexford. Ireland) where lobsters have been produced for stock enhancement since 1994. The culture methods used are similar to those described by Browne and Mercer ( 1998) and Browne ( 1999). Stage I larvae are collected soon after release by the female and are placed into vigorously aerated sea- water ( 18°-21°C) within 80 L conical tanks (hoppers) at densities (0 > £ CO c o CL c to 100 - 90 - ^I T ^r^^ - \ 80 - i ^ ^^, ,. ^ I ~* 70 - [ 60 - 50 - 40 - \ 30 ■ "^ ^^ ! _^ 20 - "^ ) — •— Control 10 - — o— Plastic netting 13 17 Day Figure 2. Experiment E. Mean percent survival ± 95 '7f CI of Hiimanis americaniis larvae through time for control (•) (// = 6) experimental units and those containing a piece of extruded, plastic netting ( Z ) (6.4 mm aperture, n = 5). Larvae were stocked at 75 unit"' (ca. 20 L ' ). led four dav-old live brine shrimp and received vigorous aeration (i.e., 240 mL air sec"' ) during the entire experimental interval. Experiment was terminated on day 17 after 9S% of larvae had metamorphosed to stage IV. CuLTLiRiNG Larval Lobsters 345 of 1 300-2000/tank (19-25 L"'). Tanks contained two species of cultured phytoplankton (Tahitian Isocliiysis galbana and Cluieio- ceros gracilis) and live, algal-enriched brine shrimp. Larvae are transferred every second day to freshly prepared hoppers. Survival rates of lobster larvae from stage I to IV average 45''( according to hatchery staff (N. Perrella. hatcher\ manager, pers. com.. 24 Au- gust 2000). Culture of H. gaiiiiiianis for stock enhancement also took place in the United Kingdom in Wales and Scotland for nearly two decades beginning in 1973 (Beard et al. 1985. Lee and Wickins 1992. Wickins 1998). Larvae, stocked at 25 L"' and fed twice daily with mysid shrimp {Neoinysis spp.) and supplemented three times per week with newly hatched Artcmia nauplii (Beard and Wickins 1992). were cultured in 100 L modified Hughes pots (planktokreisels) containing 80 L of seawater (16"-19°C). Flow rates of 7.5 to 12 L min"' kept larvae and food suspended for the 1 1-17 days it took to reach stage IV. Survival rates (10-15%) of stage I to IV at the Ministry of Agriculture. Fisheries and Food's (MAFF) Fisheries Laboratory at Conway. North Wales were much lower than those reported from the two culture facilities in the Republic of Ireland (Beard et al. 1985. Beard and Wickins 1992). Similar survival rates were obtained at MAFF's Seafish Marine Fanning Unit in Ardtoe. Scotland by Burton (1992). The culture techniques described here are relatively simple, straightforward, and have withstood the test of time in a variety of hatch-and-release settings both in Maine (Honionis uiiicricauiis) and Ireland (H. gaminuiiis]. Experiments A-D and the first year of production at CMH ( 1986) demonstrated that keeping lobster lar- vae and food well suspended within culture containers is impor- tant, but. without careful management of stage I larvae (i.e., col- lecting for culture within .^0-60 min of release by the female), typical survival rates from stage I to IV will be less than 15%. Our techniques that produced a range of survival rates from 60% to 90% in the laboratory are not intended to replace those described elsewhere (Hughes et al. 1974, Sertling et al. 1974a, Van Olst et al. 1977, Aiken and Waddy 1989, Lee and Wickins 1992). These, too, have been used successfully in a number of disparate locations and production scenarios. Rather, we believe the culture methods de- scribed here provide another option for tlshers. mariculturists, or research scientists to produce large numbers of stage IV clawed lobsters for stock enhancement, commercial enterprise, or experi- mental purposes. ACKNOWLEDGMENTS We wish to thank the coninumity and tlshers of Cutler, Maine and, especially, the Cutler Marine Hatchery Committee for their encouragement and assistance with this project. The manuscript was improved by R. Browne, J. Hinson, and A. Lewis. Funding for culture work at the Darling Marine Center was provided by the University of Maine and the Maine/New Hampshire Sea Grant College Program. Funding for Experiments A-E was provided by the University of Maine at Machias. We thank the Population & Community Ecology class at UMM for assistance on Experiments D and E. The manuscript was prepared while B. Beal was a vis- iting Fulbright Professor at the National Lhiiversity of Ireland, Galway and he thanks the Zoology Department and staff at the Martin Ryan Institute for providing office space and supplies dur- ing his stav. LITERATLI Aiken. D. E.. D.J. Martin. J. D. Meisner& J. B. Sochasky. 19S1. Influence of photoperiod on survival and growth of larval American lobsters [Homariis americaiuis). J. World. Maricult. Soc. 12:225-230. Aiken. D. E. & S. L. Waddy. 1985. Production of seed stock for lobster culture. Aquuciilliire 44:103-114. Aiken. D. E. & S. L. Waddy. 1989. Lobster culture. In: A. D. Boghen, editor. Coldwater aquaculture in Atlantic Canada. Moncton. New Brunswick: Can. Inst. Res. Region. Develop, pp. 77-122. Aiken, D. E. & S. L. Waddy. 1995. Aquaculture. In: J. R, Factor, editor. Biology of the lobster Homanis americiinii.s. New York: Academic Press. Inc. pp. 15.V175. Beal. B. F.. S. R. Chapman. C. Irvine & R. C. Bayer. 1998. Lobster (Homanis amencanm) culture in Maine: A community-based, fisher- men-sponsored public stock enhancement program. In: L. Gendron, editor. Proceedings of a workshop on lobster stock enhancement held in the Magdalen Islands (Quebec) from October 24th to 31st 1997. Can. Ind. Rep. Fish. Aquat. Sci. 244:xi+ 135:47-54. Beard. T. W., P. R. Richards & J. F. Wickins. 1985. The techniques and practicability of year-round production of lobsters, Homurus gwnmanis (L.), in laboratory recirculation systems. Fish. Res. Tech. Repl. No. 79. Ministry Agricult. Fish. Food Director. Fish. Res., Lowestoft, 22 pp. Beard. T. W. & J. F. Wickins. 1992. Techniques for the production of juvenile lobsters {Homarus giiiiimciriis (L.). Fish. Res. Tech. Repl. No. 92, Ministry Agricult. Fish. Food Director. Fish. Res.. Lowestoft. 22 pp. Browne, R. M. 1999. Studies on the European clawed lobster [Homanis giiinnianis Linnaeus 1758). fisheries practices, resource management, cultivation and stock enhancement techniques with observations on the fishery and larval cultivation of spiny lobster Patiniinis rU'phcis (Fab- RE CITED ricius 1787). Ph.D. dissertation. Galway. Ireland: Natl. Univ. Ireland. 500 pp. Browne. R. M. & J. P. Mercer. 1998. The European clawed lobster (Hoinarus gammarus): slock enhancement in the Republic of Ireland. In: Gendron. L. (ed.). Proceedings of a workshop on lobster stock enhancement held in the Magdalen Islands (Quebec) from October 29"' to 3r' 1997. Can. Ind. Rep. Fish. Aquat. Sci. 244:xi+135:33-+l. Bruggerman, E., P. Sorgeloos & P. Vanhaeke. 1980. Improvements in the decapsulation technique of Arlcmiu cysts. In: Persoone. G.. P. Sorgeloos, O. Roels & E. Jaspers, editors. The brine shrimp Anemia vol 3.: Ecology, culturing, use in aquaculture. pp. 261-270. Burton. C. 1992. Techniques of lobster stock enhancement. Seafish Rept. No. 396, Seafish Indust. Author. Mar. Farm. Unit. 51 pp. Castagna, M. 1983. Review of recent bivalve culture methods. J. World Maricult. Soc. 14:567-575. Chang. E. S. & D. E. Conklin. 1993. Larval culture of the American lobster iHomariii americaiiK.s). In: McVey. J. P.. editor. CRC handbook of mariculture. 2nd ed. Boca Raton. FL: CRC Press, pp 489-f95. Eagles, M. D.. D. E. Aiken & S. L. Waddy. 1984. Effect of food quality and feeding schedule on survival, growth, and development of larval American lobsters fed frozen adult brine shrimp. / World Mariciill. Soc 15:142-143. Ford. R. F.. J, C. Van Olst. J. M. Carlberg. W. R. Dorhand & R. L. Johnson. 1975. Beneficial use of thermal effluent in lobsier culture. Proc. .Anmi. Meet. World Maricult. Soc. 6:509-522. Hughes. J. T. 1972. Lobster culture. In: Smith. W. L. & M. H. Chanl, editors. Culture of marine invertebrate animals proceedings. Greenport, N.Y. New York: Plenum Press, pp. 221-227. Huehes. J. T. & G. C. Matthiessen. 1962. Observations on the biologv of 346 Beal and Chapman the Amercian lobster. Hmnanis americaiuis. Liiiinol. Oceiwnf^r. 7:41-1— 421. Hughes. J. T. R. A. Shleser & G. Tchobanoglous. 1974. A rearing tank for lobster larvae and other aquatic species. Prog. Fish-Cull. 36(3);12y- 132. Lee, D.. O'C. & J. F. Wickins. 1992. Crustacean farming. Oxford. UK: Blackwell Scientific Publications. Nicosia. F. & K. Lavalli. 1999. Homarid lobster hatcheries: Their history and role in research, management, and aquaculture. Mar. Fi.sh. Rev. 61:1-57. Plante. J. M. 1986. Inside lobster hatcheries: How do they work? Commci: Fi.sh. News 13(12): 34-35. Rathbun, R. 1886. Notes on lobster culture. Bull. U.S. Comm. Fish Fish. 6: 1 7-32. SAS Institute. Inc. 1988. SAS/STAT user's guide, release 6.03 ed. Cary. NC: SAS Institute. Inc. Sastry. A. N. 1975. An experimental culture-research facility for the American lobster. Homanis uinericaiuis. Proc. 10"' Europ. Mar. Biol. Symp.. vol. 1:419^35. Schuur. A.. W. S. Fisher. J. C. Van 01st. J. M. Carlberg. J. T. Hughes. R. A. Schleser & R. F. Ford. 1976. Hatchery methods for the production of juvenile lobsters (Homanis americamis). Univ. California Sea Grant College Pgm IMR Ref. 76-6. Sea Grant Publ. No. 48. 21 pp. Serfling. S. A.. J.C. Van Gist. & R. F. Ford . 1974a. A recirculating culture system for larvae of the American lobster. Homarus americamis. Aqiia- ciilntre 3:303-309. Serfling. S. A.. J. C. Van Olst & R. F. Ford. 1974b. An aulomalic feeding device and the use of live and frozen Artemia for culturing larval stages of the American lobster, Homanis americamis. Aquacullure 3:311- 314. Sokal. R. R. & F. J. Rohlf. 1995. Biometry 3''' ed. New York: W. H. Freeman and Company. Talbot, P. & S, Helluy. 1995. Reproduction and embryonic development. In: J. R. Factor, editor. Biology of the lobster Homanis americamis. San Diego, CA: Academic Press. Inc. pp. 177-216. Van Olst. J. C, J. M. Carlberg & R. F. Ford. 1977. A description of intensive culture systems for the American lobster, Homanis america- mis. and other cannibalistic crustaceans. Proc. .Aimii. Meel.-Worlil Mariciilt. Soc. 8:271-292. Van Olst, J. C, J. M. Carlberg & J. T. Hughes. 1980. Aquaculture. In: J. S. Cobb, J. S. & B. F. Phillips, editors. The biology and management of lobsters, vol II. New York: Academic Press, Inc. pp. 333-384. Waddy, S. L. 1988. Farming the homarid lobsters: State of the art. Worid Aquuciill. I9(4):63-71. Waddy. S.L. & D. E. Aiken 1998. Lobster {Homanis americamis} culture and resource enhancement: the Canadian experience. In: L. Gendron. editor. Proceedings of a workshop on lobster stock enhancement held in the Magdalen Islands (Quebec) from October 29'" to ?V' 1997. Can. Ind. Rep. Fish. Aquat. Sci. 244:xi-h135:9-1S. Wickins. J. F. 1998. Lobster [Homanis gammanis) stock enhancement investigations in the United Kingdom from 1983 to 1993. In: L. Gend- ron. editor. Proceedings of a workshop on lobster stock enhancement held in the Magdalen Islands (Quebec) from October 29'" to 31"' 1997. Can. Ind. Rep. Fish. Aquat. Sci. 244:xi-Kl35:19-22. Winer. B. J.. D. R. Brown & K. M. Michels. 1991. Statistical principles in experimental design. 3"' ed. New York: McGraw-Hill. Inc. Juunuil uj Slwlljhh Rcifaixli. Vol. 20, No. 1. 347-352, 2UUI. EFFECTS OF DIET ON NEPHROPS NORVEGICUS (L.) LARVAL AND POSTLARVAL DEVELOPMENT, GROWTH, AND ELEMENTAL COMPOSITION GUIOMAR ROTLLANT.' * MIREILLE CHARMANTIER-DAIIRES,^ GUY CHARMANTIER,^ KLAUS ANGER,' AND FRANCISCO SARDA' 'liisiltiil lie Ciencies del Mar. CStC. Passeig Joan dc Burba s/n. E-0HU39 Barcelona. Spain: -Lahoratoire d'Ecophysioloyie dc.s Invertehres. Universite Montpellier 11. F-34095 MontpelUer Cede.x 05. France: Blologisclie Anstalt Helgoland. Meeresstation. Stiftiing Alfred-Wegener-lnsiitul far Polur- iind Meeresforsclning. 2749H Helgoland. Gcrniiuiy ABSTKiCT Few .>,ludies have been conducted on the cuUure ot Nephrop.s imrxvgiciis larvae and none of them have permitted significant numbers of postlarvae to be obtained {.stage IV). The goal of this study was to determine a suitable diet for rearing N. nor\'egicus to juvenile stage V. Survival, molting periodicity, growth, and elemental composition (CHN) of the young postembryonic stages were used as criteria of physiological condition. Four different diets, fresh enriched nauplii of A rremUi (FENA), artificial feed for shrimp (AF), commercial frozen mussels (M), and commercial frozen adult Arwinia (FAA) were tested from hatching to staee V at ambient temperature (14-16°C). Five day old stage I larvae fed with AF, M or FAA showed a similar biomass as freshly hatched larvae, mdicating that the energy derived from food only allowed them to survive but not to gain length. FENA was the only diet that resulted in development to stage V. Larvae fed with FENA pre,sented the shortest time of development, the highest rate of survival, and the highest growth rates in size and biomass. KEY WORDS: A'cplirops. spiny lobster, larva, growth, diet, C,H,N INTRODUCTION The commercial fisheries of the Norway lobster. Nephrop.s mn- vefiiciis (Linnaeus. 1758), are economically important in Europe, and many projects regarding this species have been developed with the support of the European Community (ICES 1999. Sarda 1996). In the Mediterranean Sea, the production and average size of N. norvegiciis have decreased during the past 10 years, while fisheries effort has increased (Franquesa 1996. Lleonart 1995. Sarda 1998a). Growth, reproduction, molt and feeding have recently been studied in the Mediterranean Sea (Sarda 1998bl. Despite the economical importance of this species, information concerning the larval stages is scarce compared to that for adults and to other commercial lobster as the species of the genus Hoinaru.s Weber. 1795, The early postembryonic development consists of three lar- val stages (I, II, III). These were morphologically described by Sars (1884. 1890) from Norwegian waters, by Jorgensen (1925) from the northeast coast of England, and by Santucci (I926a.b,c. 1927) from the Tyrrhenian Sea. The larval phase is separated by a metamorphic molt from the postlarva (stage IV). which presents the morphological characters of the adult. The postlarvae swim for a few days before becorning benthic. There is still a controversy about the terminology of lobster development: some authors sug- gest that stage IV is already a juvenile. Others prefer to call stage IV a postlarva or decapodid and stage V the first juvenile (see Phillips & Sastry 1980. Charmantier et al. 1984). Although larval distribution has been studied in different re- gions of the Atlantic Ocean and the Mediterranean Sea from a fisheries point of view (Brown et al. 1995. Eiriksson 1970. Farmer 1975. Lindley 198.3. Milligan & Nichols 1988. Nichols et al, 1983. Nichols et al, 1987, Thomson et al. 1986). only a few studies have been conducted on requirements for the development of N. nor- vegiciis larvae and postlarvae. Figueiredo and Vilela (1972) and Figueiredo (1971 ) reared N. norvegiciis larvae in the laboratory at *Corresponding author. E-mail: guioCs'icm. csic.es; Fax: -1-34-932217340. different salinities (33-40%r) and temperatures (7-l7°C), In this study, the larvae seemed to be unaffected by salinity in the tested range, while increasing temperature accelerated not only growth, but mortality as well. Larvae reared at 1I-14°C presented the highest survival (25% to stage II, 13% to stage III and 7% to stage IV) and juveniles lived as long as three months reaching 18 mm in total length. Increased temperatures also reduced the intermolt period. 14-15 days at 7-10='C to 5-6 days at 13-17°C. Anemia Leach, 1819 nauplii proved to be insufficient as a diet and better growth was achieved with eggs of Crangon crangon (Linnaeus. 1758) as food. Thompson and Ayers (1989) tested temperatures ranging from 5°C to 15'C finding the same effects as Figueiredo and Vilela (1972) and Figueiredo (1971). The best results were observed at IS'C. where survival was 12% from stage I to stage II, 40% between II and 111. and 8% between III and IV, In addition, the estimated duration of the intermolt period was 6.9 days from stage I to stage II, 6.1 days between 11 and III. and 9,0 days between III and IV. In these studies, small Anemia sp, were pro- vided as a food, occasionally complemented with wild plankton and encapsulated feed. Recently. Dickey-Collas et al. (2000). reared N. noirefiiciis larvae at different temperatures from 8.6°C to I4,7°C with a mixed diet of Tetraselmis siiecica and freshly hatched Anemia nauplii. Mortalities were greater than 50% by the end of stage I at all temperatures tested and only one larvae was raised to stage IV at 1 1.2°C. Stage I lasted from 7 to 18 days and no correlation with the rearing temperature could be made. Sur- vival exceeding 90% to stage III was achieved by feeding the larvae with larger /Irrcm/a juveniles (Anger & Puschel 1986), but this rearing experiment was not continued thereafter. Development time was 8.7 days for stage I. 1 1 .3 days for stage 11. and 1 1 .5 days for stage III, The elemental composition (CHN) of the larvae did not change very much during development. Only freshly hatched larvae showed significantly lower percentages of organic constitu- ents than older larvae. Hence, further information about the culture conditions of N. norvegiciis is critical to successful commercial culture. Aquacul- 347 348 ROTLLANT ET AL. inn J AF sn- 60- 40- 2n- 0- — -»—-«•> . 1 =h ■ 1 ■ 20 40 60 00- ^nm^ p M 80- 60- 40- 7L 20- ^%stff^ 0- .-^ ^^ 20 40 60 Days after hatching 80 Figure 1. PtTcent survival and occurrence of successive stages over lime for lar>al and postlarval de\elopnient of ,\'eplirops n(>ne,i;iciis reared with four different diets: fresii enriched nauplii of Anemia (FENA — after hatching, nauplii were enriched for 24 h with lipids and vitamins using DHA Seico, IN\ E); artificial feed for shrimp. DIBAQ (AF); commercial frozen mussels (M); and commercial frozen adult Arlemia (FAA). 1 to 3 weeks in individual cylinder-conic containers protected by a mesh to avoid escape of freshly hatched larvae. These containers were placed in a 3000 L tank with running sea water at a tem- perature of 15 ± l"C and 12 h light/ 1 2 h dark photoperiod. Females were fed with frozen mussels twice a week. After hatching. 2.169 larvae from a group of 30 females were cultured indi\idually in connected boxes fed with running seawa- ter at ambient temperature ( 15 ± T'C). Four different diets were tested: 1. FENA: fresh enriched nauplii of .A;7(^/);u( AF480 (21.3 mg/g DW HUFA & 0.04 DHA/EPA). After hatching, the nauplii were enriched for 24 h with lipids and vitamins using 300 mg/L of DHA Selco (200 mg/g DW HUFA & 2.96 DHA/ EPA) made by INVE AQUACULTURE NV (Oeverstraat 7. 92()0 Baasrode. Belgium). 2. FAA: commercial frozen adult Anemia produced by Mundo Acuatico. S.L. (C/ Llobregos 9. 08032 Barcelona. Spain). 3. AF: artificial feed for shrimp (feed granules of 0.2 mm containing 12% lipids. 15% carbohydrates. 52% proteins, vitamins, oligo-elements. and 5.500 Kcal) produced by DIBAQ AQUACULTURE (Poligono Industrial "Baix Ebre"" Parcela 114. Camp-redo. 43897 Tortosa. Spain). 4. M: commercial frozen mussels. Food was provided ((/) lihitum twice a day. The different treat- ments were applied homogeneously through the circuit bo.xes in order to ensure that differential mortality was not influenced by the rearing system. Growlh Rale and Molting Frequency Sur\ i\al and molting of larvae and postlarvae in each treatment were recorded dailv. Carapace length (CL) was measured to the nearest 0.1 mm directly from photographs or with a dissecting microscope for larvae fixed in 4% formol. Carapace length of postlarvae was measured with callipers. Estimates of growth pa- rameters were based on the model of Von Bertalanffy (1938). as this curve has been commonly accepted for adults (e.g. Mytilineou and Sarda 1995. Sardit 1985). For fitting the curve. FISHPARM software was used (Prager et al. 1987). The slopes and intercepts of the different growth curves were compared by applying an ANOVA. ture of nephropid lobsters has been reviewed by Van Olst et al. ( 1980). showing that efforts to develop techniques for the culture in controlled environmental systems have been concentrated on species of the genus Hinnnnis. By comparison with N. norvfiiiciis. a large body of information is now available on the culture and physiology of Hnmanis (review in Factor 1995). Generally, poor nutritional quality has been shown to increase the time of development and to decrease survival and growth of lobster larvae (Conklin 1995). The goal of the present study was to improve growth, molting, and nutrition of M iion-egiciis larvae and postlarvae from hatching to stage V. and to evaluate the prospects of a culture to complement the capture fishery. MATERIALS AND METHODS Culture of Lanae and Postlarvae Ovigerous female Norway lobsters were caught off Barcelona harbor by a commercial fishing fleet at 400 m depth from January to March. They were brought to the laboratory and maintained for Measurements of Biomass Larvae from 6 different females were taken immediately after hatching and on day 5 (stage I. intermolt stage C) after feeding with the test diets (FENA. FAA. M, AF). Biomass was measured as wet weight (WW), dry weight (DW). carbon (C), nitrogen (N). and hydrogen (H). The CHN values are presented as absolute (per individual) and relative contents (CHN in % of DW). and as C/N and C/H ratios. Samples were briefly rinsed in distilled water, blotted on fluff- free filter paper, and WW was measured to the nearest 0.01 mg on an Ohaus AP250D-0 balance. Thereafter, the material was stored at -80'^'C. Before the analysis, the samples were freeze-dried over- night in a GT 2 (Leybold-Heraeus) apparatus and weighed again to the nearest 0. 1 p.g on a Mettler UM 3 balance to obtain DW. Larvae were combusted at I020''C in a CHN analyser (Carlo Erba Science. Model 1 108. Version Fisons). The arithmetic means and the standards errors of biomass were calculated from 4 to 8 replicates. The significance of differences Nutrition in Larval Spiny Lobster 349 between mean values of freshly hatched and 5 day old larvae fed different diets were determined using a one-way analysis of vari- ance ( ANOVA) followed by a multiple test at a probability level of p<0.05. RESULTS Molliiiii Frequency Among the four diets tested, only the FENA diet pennitted development to stage V. Commercial frozen adult anemia allowed development to stage IV, while AF- and M-fed larvae reached only stage II (Fig. 1). In addition, only larvae fed with FENA showed a regular pattern of molting and development through the succes- sive stages. Commercial frozen adult Anemia fed larvae showed a higher rate of survival until day 30 as compared to the other diets. Thereafter, however, the larvae fed with FENA had a slightly higher survival. Survival and duration of the developmental stages are given in Table I. Survival rate from stage I to stage II was 21.3% with FENA, 0.9% with AF, 9.4% with M, and 3.4% with FAA. The average time of development from hatching to stage II was 9. 10. 16, and 18 days, respectively, with FENA, AF, M. and FAA. With the two Anemia diets, some larvae reached stage IV; the time to reach stages III and IV was 20 and 32 days with FENA, and .^0 and 60 days with FAA, respectively. Griiwlh Rate {Increments in Body Size) Lar\ ae fed wuh AF and M survived for up to 40 days, with high variability in length (Fig. 1) and in larval size (carapace length, CL|. Larvae and postlarvae fed with FENA or FAA, in contrast, survived for over 70 days (Fig. 1 ). The Von Bertalanffy growth curves (VBGC) could be fitted only in these two groups (Fig. 2). Their slopes were significantly different (P<0.001), showing that the larvae fed with FENA showed a higher rate of growth than those led with FAA. Changes in Biomass Biomass changes (WW, DW, CHN per individual) measured between freshly hatched and 5 day old larvae fed with different diets are documented in Table 2. Changes in the water content (% WW) and in the relative elemental composition (CHN in % of DW; C/N and C/H ratios) are documented in Table 3. Weight and elemental composition (C, N) are shown in Figure 3 for freshly hatched (fh) and 5 day old stage I larvae of one female (no. 23), where sufficient material from all dietary conditions was available. / - CL = l.l.gj-d-e"""' "**-"') • !--•- • .;••. >-'■' • iJ ^^_^ • • ^ 4- • . □ • D . 1 • *• -- • ° u • _-— ^^-''''' ' t- ^— -— S "a ° ° •-'^iiUr-'^^ 2- DO^ '-"'^^' : "°° ° CL = 4099-{l-e*"^**^^°^'f^*^^^^*) 1- , , 0:1 ^ |— o— FAA ] 60 Days after hatching Figure 2, Comparislon of Von Berlantanfly growth curves (VBGC) for Sephrops niirvei'iciis lar\ae and postlarvae reared under t«o dif- ferent nutritional conditions: Iresh enriched nauplii of Arteniia (FENA) and commercial frozen adult Artemia (FAAl. See Figure 1 for further details. This comparison shows that all measures of biomass per individual (WW. DW, C, N; H is not included but shows the same trends) were consistently highest in the FENA group. WW reflected the dietary quality less clearly. Likewise, the percentage CHN figures (in % of DW) were significantly higher in the FENA treatment than in all others. The water content was on average lower in larvae fed with FENA than in the t)ther treatments. In 5 day old FAA-fed larvae, all biomass values were consistently below those of freshly hatched larvae; these losses were statistically significant in the percentage of C. N, and H values (Table 3, Fig. 3). Hence, FENA was the only diet that allowed the larvae to reach a significantly higher biomass within 5 days of hatching. With FAA, M, and AF. the larvae conserved the initial biomass they already had at hatching but did not gain weight. No significant differences in the C/N and C/H ratios were observed between freshly hatched larvae and those fed for 5 days with any of the test diets. In addition to nutritional effects, significant differences were detected between larvae originating from different females (see Tables 2 and 3). DISCUSSION The rate of survival to the juvenile stage V was very low (maximum 0.6%) with all diets. Figueiredo and Vilela (1972) and TABLE 1. Survival and duration of the larval and llrsl postlarval stages of Nephrops norvegicus fed with different diets. Survival Duration Survival Duration Survival Duration Survival Duration Stage I Stage I Stage 11 Stage 11 Stage III Stage 111 Stage I\ Stage IV Diet n (%l (days) (%) (days) (%) (days) (%) (days) FENA 521 21.30 8.81 27.4.^ 10.92 41.93 12.23 53.85 19.71 FAA 859 5.47 17.71 3 1 .4 1 12.09 10.00 6.67 0 - M 352 9.37 15.73 (1 - - - - - AF 437 0.92 10.5* (1 - - - - - FENA: Fresh enriched nauplii of Anemia ( aher hatching, nauplii were enriched for 24 h with lipids and vitamins), FAA; commercial frozen adult Anemia, M: commercial frozen mussels, and AF: artificial feed liir shrimp. * Only two molts observed. 350 ROTLLANT HT AL. TABLE 2. Changes in «et weight (WW), dry weight (D\V), carhon (C), nitrogen (N) and hydrogen (H) in Xeplirops nonegiciis larvae freshly hatched or fed with different diets. Condition ;/ WW X (mg) ±SD DW (mg) c (mg) N (mg( H (mg) Female X ±SD X ±SD X ±SD X ±SD 5 Freshly hatched 4 5.66 0.39 0.656 0,095 0.207 0,0.^5 0,046 0,007 0,03 1 0,006 23 Freshly hatched 5 5.35 0.39 0.508 0.102 0.158 0,031 0,035 0,007 0,023 0,005 23 FENA 8 5,69 0.56 1.001 0.157 0,349 0,063 0,079 0,015 0,056 0,011 FAA 5 5.80 0.43 0.699 0.117 0,220 0,053 0,050 0,011 0.034 0.()(_)9 M 5 5.31 0.33 0.667 0.092 0,192 0,032 0,045 0,009 0,028 0,005 AF 5 5.12 0.29 0.538 0.075 0,150 0,020 0,034 0,006 0,025 0,006 21 FENA T 5.63 0.04 0.83 1 0.123 0,261 0,042 0,061 0,007 0,040 0,007 FAA 5 5.85 0.22 0.536 ().()i)0 0,157 0,032 0,036 0,007 0,023 0,005 M 5 4.65 0.95 0.586 0.059 0,166 0,020 0,040 0,005 0,025 0,003 AF 5 4.70 0.78 0.564 0.087 0,160 0,021 0,038 0,005 0,023 0,003 26 FENA 5 5.59 0.94 0.628 0.182 0,219 0,070 0,05 1 0,016 0,033 0,012 29 M 5 6.36 1.18 0.598 O.IOI 0,173 0,033 0,041 l),()()8 0.026 0.005 I) FAA 8 5.76 0.64 0.551 O.lll 0,167 0,034 0,039 0,008 0,024 0,005 FENA: fresh enriched nauplii of Anemia: FA.'\: commercial frozen adult Anemia: M: commercial frozen mussel; AF: artificial feed for shrimp. Thompson and Ayers ( 1989) found similar rates of survival with a mixed diet (nauplii and adult Anemia, wild plankton, or eggs of Craiigon crangon). However. Anger and Piischel (1986) obtained a survival of over 90% to stage III with living adult Anemia. though they did not monitor survival after this stage (considering stage IV as the first juvenile instar). In the present study, FAA-fed larvae showed a high rate of survival through the tlrst 30 days, but the developmental period in stage I was nearly twice that found with other diets in this study or reported by other authors. Also in another clawed lobster. Hunninis amcricaniis (H. Milne Edwards. 1837). the nutritional value of frozen adult Anemia was found to be less than living adult Anemia (Conklin 1993). The nutritional value of adult Anemia apparently decreases greatly during storage in frozen condition. Figueiredo and Vilela (1972) and Dickey- Collas et al. (2000) showed that early Anemia nauplii were also insufficient for N. noiTegicus larvae, although they were success- fully used as a diet for the larvae of paliiiurid lobsters (Van Olst et al. 1980). The intemtolt duration of larvae in stage I was shorter for that fed with FENA. The duration of larva stages fed with FENA was very close to the duration observed by the other authors at similar temperatures (Anger and PUschel 1986. Figueiredo and Vilela 1972. Thompson and Ayers. 1989). Growth has been studied in N. non-egiciis adults and juveniles, both in natural populations and in captivity (Belchier et al. 1994, Mytilineou & Sarda 1995. Sarda 1985. Tuck et al. 1997, Verdoit et al. 1999). In the present study, the generally applied Von Berta- lanffy (1938) growth curve fit for larvae and postlarvae fed with either FAA or FENA but not for the other treatments. The growth curve obtained with FENA presented a significantly steeper slope, with a CL.^ value close to 14 inm in the fourth larval stage. Figueiredo and Thomas (1967) found that one year old N. iioi- vegiciis individuals reach 13-15 mm carapace length, correspond- TABLE 3. Changes in the percentage of wet weight (% WW), dry weight {% DW), carbon (% C). nitrogen (% N), and hydrogen {% H), and C/N and C/H ratios in Ncphrops iioncgiciis larvae freshlv hatched or fed with different diets. Water C N H Condition H (% WW) (% DW) (% DW) (% DW) C/N C/H Female X ±SD X ±SD X ±SD X ±SD X ±SD X ±SD 5 Freshly hatched 4 88.38 1.63 3 1 ,58 1,14 6,97 0.32 4,74 0.24 4..54 0.19 6.66 0. 1 3 23 Freshly hatched s 90.39 2.37 31,15 1,23 6,99 0.32 4,58 0.24 4.46 0.04 6.81 0.09 23 FENA 8 82.15 3.79 .^4,75 1,41 7,83 0.4 5,54 0..14 4.44 0.14 5.61 0.15 FAA 5 87.99 1.34 31,29 2,23 7,07 0.5 4,78 0,44 4.43 0.11 6..56 0.14 M 5 87.46 1.23 28,7 1,49 6,68 0.5 4,26 0,27 4.3 0,12 6.75 0.11 AF 5 89.52 1.06 27,9 1,05 6,39 0.36 4,68 1,2! 4.37 0, 1 6 6.24 1 .33 21 FENA -) 85.24 2.08 31,35 0,35 7.37 0.26 4.79 (Ill 4,26 0,20 6.55 0.08 FAA 5 90.43 1.37 29,20 1,22 6.72 0.50 4,20 0.23 4,36 0,20 6.95 0.54 M ? 86.87 3.50 28,23 0,94 6.80 0,29 4,22 0.19 4,15 0,05 6.69 0.10 AF 5 87.80 2..^9 28,53 1.05 6.81 0.40 4,15 (1.12 4,11) 0,09 6.87 O.II 26 FENA 5 87.63 3.18 3 1 ,75 1,58 7.38 0,39 4,80 0.38 4,30 0,08 6.62 0.19 29 M 5 90.21 2.80 28.85 1,45 6.81 0,42 4.26 0.26 4.24 0.05 6.78 0.09 9 FAA 8 90.39 2,00 .10,35 1,14 7.04 0,19 4.39 0.18 4.31 0.12 6.91 0.19 FENA: fresh enriched nauplii of Anemia: FAA: commercial frozen adult Anemia: M: commercial frozen mussel; AF: artificial feed for shrimp. Nutrition in Larval Spiny Lobster 351 FENA FAA m FENA FAA M 111 FENA FAA M Diet Diet Diet Figure 3. Body composition of freshly hatched (fli) and 5 day old Nepbnips nonegiciis stage-I larvae, reared in the laboratory under 4 different nutritional conditions: fresh enriched nauplii i<( Arlciiiia (KKNAl; artificial feed for shrimp (AFl: commercial frozen mussels (Ml: and com- mercial frozen adult Anemia (FAA). See Figure I for further details. Wet weight (WW), dry weight IDWI, Carbon (C), and nitrogen (N) content per individual: C, N also in % of DW. ing to 20-25 mm total length. Comparing our growth curves for larvae with those obtained by Sarda (19S5) for juveniles, we can conclude that the average growth rate of larvae and postlarxae is approximately 4 to 5 times faster than in juveniles. Feeding in homarid lobsters begins immediately after hatching and the larvae may not recover from nutritional stress if inadequate food is pixnided during the first stage (Anger et al. 1985. Aiken & Waddy 1995). In the present study, after 5 days of feeding with AF. M. or FAA. stage 1 N. nor\-egicus larvae showed similar absolute and relative CHN values and a similar water content as freshly hatched larvae. In FENA-fed larvae, by contrast, the CHN contents (both per individual and as proportions of DW) were higher and the water content (in ^WW) was lower than at hatch- ing. The CHN values measured by Anger and Piischel (1986) in successfully developing larvae of N. non'egicus al.so increased significantly after hatching. These data indicate that the enei-gy taken from inadequate diets (AF. M. FAA) allowed only for sur- vival but not for gaining length. In addition, carnivorous larval decapods have limited enzymat- ic capabilities during the planktonic development. Since a gastric mill is absent, the survival of lobster larvae with the proteolytic enzyme levels must depend upon the high energy, easily-digestible nature of the zooplankton prey, together with an extended gut residence time. In N. non-cgiciis larvae trypsin-like activity and content per individual increases from stage 1 to stage III (Kumulu and Jones 1997). The CHN results obtained in the present study and the enzymatic works confirm the importance of providing high-quality food immediately after hatching to assure a successful development. In our data, variation in the C/N and C/H ratios did not reflect the developmental stage or differential nutritional quality of the tested diets. This agrees with observations by Anger and Piischel (1986). who found that the absolute biomass increased signifi- cantly during the development of N. non'egicus larvae, but the relative composition (CHN in 9r of ash-free DW) remained fairly constant. Hence, for larvae of this species, the relative composition of CHN seems to be less sensitive as an indicator of the nutritional state than the water content. N. non'egicus is economically a very important species and its culture is far from being accomplished, particularly when com- pared with other homarid lobsters. The data presented demon- strates that larvae fed with FENA showed the best nutritional condition. However, future studies in culture systems, water qual- ity, broodstocks, and feeding are necessary to improve Norway lobster larval development. ACKNOWLEDGMENTS The authors thank Ms. C. Piischel. G. Fuster. C. Garcia and J. Riba for technical support. Dr. L. Recasens for her help in the giowth studies, the Captains and crews of the fishing vessels Maireta III. Francesc i LIuis. and Zorrilla from Barcelona harbor, and Ms. Lluis Boix (DIBAQ AQUACULTURE) for donating commercial artificial food for shrimps. This study was supported by a joint grant from the Catalonian (Spain) and Languedoc- Roussillon (French) Communities and by the "Centre de Referen- cia de Recerca i Desenvolupament en Aqiiicultura de la Generalitat de Catalunya." REFERENCES Aiken. D. E. & S. L. Waddy. 1995. Aquacultiuc, In; J.R. Faclor. editor. Biology of the lobster Himhinis nmcncuniis. San Diego: Academic Press, pp. 153-176. Anger. K.. V. Storch. V. Anger & ,1. M. Capuzzo. 1985. Effects of star- vation on moult cycle and hepalopancreas of stage I lobster [Homanis americainis) larvae. HelgtjUintlcr Mccresiinlers 39:107-116. 352 ROTLLANT ET AL. Anger. K. & C. Puschel. 1986. Growth and exuviation of Norway lobster (Nephrons nnnegiciis) larvae reared in the laboratory. Ophelia 25(3): 157-167. Belchier. M.. P. M. J. Shelton & C. J. Chapman. 1994. The identification and measurement of fluorescent age-pigment abundance in the brain of a crustacean (Nephrops non'egiciis) by confocal microscopy. Comp. Biochem. Physio!. 108B: 157-164. Brown. .1. H.. L. Femand. D. B. Bennett & J. H. Nichols. 1995. A physical retention mechanism for Nephrops non'egiciis. ICES reports. CM I995/K:.M Ref C:l-9. Charniantier. G.. M. Charmantier-Daures & D. E. Aiken. 1984. Neuroen- docrine control of hydromineral regulation in the American lobster Homanis wnericainis H. Milne-Edwards. 1837 (Crustacea, Decapoda). 2. Larval and Postlarval Stages. Gen. Com. Endocrinol. 54:20-34. Conklin. D. E. 1995. Digestive physiology and nutrition. In: J. R. Factor, editor. Biology of the lobster Homanis americanus. San Diego: Aca- demic Press, pp. 441-464. Dickey-Collas. M.. N. McQuaid. M. J. Armstrong. M. Allen & R. P. Briggs. 2000. Temperature-dependent stage durations of Irish Sea Nephrops larvae. J. Planknm Re.s. 22:749-760. Eiriksson. H. 1970. On the breeding cycle and fecundity of the Norway lobster at Southwest Iceland. ICES reports. CM1970/K: 6. 1-3. Factor. J. R. 1995. Biology of the lobster //o/H(I)».'; americanus. San Diego: Academic Press. 528 pp. Farmer. A. S. D. 1975. Synopsis of Biological Data on the Norway Lobster Nephrops NoiTegiciis. FAO Fish Synop 1 12:1-97. Figueiredo. M. J. & H. J Thomas 1967. Nephrops nonvgiciis (Linnaeus. 1758) Leach — A review. Oceunogr. Mar. Biol. .Ann. Rev. 5:371—107. Figueiredo. M. J. 1971. Sobre a cultura de crustaceaos decapodes em laboratorio: Nephrops non'egicus (Lagostim) e Penaeus kcrailninis (Camarao). Bol. Inform. Inst. Biol, marit. 1: 1-17. Figueiredo, M. J. & M. H. Vilela. 1972. On the artificial culture of Neph- rops norvegiciis reared froin the egg. Aqiiaciillure 1:173-180. Franquesa. R. 1996. Quantitative analysis of the relations which condition the northwestern Mediterranean fishing system. CEE contract No DGIV-MED/9.3/022: 150 pp. ICES. 1999. Report of the working group on Nephrops stocks. Copen- hagen. Denmark. 504 pp. Jorgensen. O. M. 1925. The early stages of Nephrops nor\-egicus, from the Northumberland Plankton, together with a note on the post-larval de- velopment of Honicuiis vulgaris. J. Mar. Biol. Assoc. Plymouth 13: 870-879. Kumulu. M. & D. A. Jones. 1997. Digestive protease activity in planktonic crustaceans feeding at different trophic le\els. J. Mar. Biol. ,Ass. U.K. 77:159-165. Lindley, J. A. 1983. Continuous plankton records: Larvae of Nephrops norvegicus (L.) in the northwestern Irish sea. 1971-82. ICES reports. CM 1983/K:12: 1-3. Lleonart. J. 1995. La pesqueri'a de Catalunya y Valencia: descripcibn glob- al y planteaniiento de bases para su seguimiento. CEE contract No. DGXIV- 1989/3: 1634. Milligan. S. P. & J.H. Nichols 1988. Nephrops larvae surveys off the northeast coast of England in \mi. ICES reports. CM 1988/K:5: 1-13. Mytilineou. C. & F. Sarda. 1995. Age and growth of Nephrops norvegicus in the Catalan Sea. using length-frequency analysis. Fish. Res. 23:283- 299. Nichols. J. H., D. B. Bennet, D. J. Symonds & R. Grainger 1983. Estima- tion of the stock size of adult Nephrops from larvae surveys in the western Irish Sea in 1982. ICES reports. CM 1983/K:6, 1-30. Nichols. J. H.. D. B. Bennet, D. J. Symonds & R. Grainger. 1987. Esti- mation of the stock size of adult Nephrops noivegicus (L.) from larvae surveys in the western Irish Sea in 1982. / Nat. His 21:143,3-1450. Phillips. B. F. & A. N. Sastry. 1980. Larval ecology. In: J. S. Cobb & B. F. Phillips, editors. The biology and management of lobsters. New York: Academic Press, pp. 11-57. Prager. M. H.. S. B. Saila & C. W. Recksiek. 1987. FISHFARM: A mi- crocomputer program for parameter estimation of non linear models in fishery science. Norfolk. VA: Old Dominion University. Santucci. R. 1926a. Lo sviluppo posi-embrionale dello Scampo [Nephrops noiyegicus (L. ) del Quarnero. Boll. Mus. Zool. Anat. Comp. Univ. n.a. VI (2):7-10. Santucci. R. 1926b. Lo sviluppo post-embrionalc dell'aragosta iPalimirus vulgaris Latr.) e dello scampo {Nephrops nonegicus (L). Boll. Mus. Zool. Anat. Comp. Univ. n.a. VI (6):31-33. Santucci. R. 1926c. Lo sviluppo e d'ecologia post-embrionali dello "Scampo". [Nephrops non'egicus (L.) nel Tirreno e nei mari nordici. R. Comitato Talassografico llaliano CXXV:3-34. Santucci. R. 1927. Uno stadio di sviluppo non ancora descritto dello "Scampo" (Nephrops iwiregicus (L.)). R. Comitato Talassografico Italiano CXXVin:\~6. Sarda. F. 1985. Estudio de la cdad. ciecimiento y frecuencia dc muda en cautividad J' Q :38° I 39° 40° 41° 42° 43° o Land ALONSHORE-SOUTHWARD V Ocean -T-^ /• ONSHORE SAMBOROMBON' MAR CHIQUITA Mat del Plata QUEQUEN BAHIABLANCAQUEQUEN GRANDg,-^- SALADO lUEN [,57- v^ ,100 Km. \ ABP 5r J F M AM J J A S ON D Month MBP 66° 64° 62° 60° 58° 56° LONGITUDE [W] 54° 52° 50° 48° Figure 1. (Al Geographic position of the study sites: Saniborombon bay, Mar Chiquita coastal lagoon. Quequen Grande river, Quequen Salado river. Line along the coast represents geographic distribution of Lea iirugiiayensis. ABP: Argentinean Biogcographic Province, MBP: Magellanic Biogeographic Province. (Irid points position for the M\\ F data analyzed ( I to 4l. .Solid lines delimit the lour areas according to coastal bearing. (Bl Kkman surface layer field current (dark arrows) forced by «ind field (background arniHsl that gives favorable drift (southward and onshore) contributing to the niegalopac recruitment. Histograms 1 to 4 correspond to monthly occurrence (absolute number of records) of southward and onshore conditions during the time period analyzed. 36°22'S. 56°45'W) of the Bahi'a Saniborombon (100 km long coastal basin within La Plata river estuary. Argentina). (2) the Mar Chiquita coastal lagoon (37°46'S; 300 km south), and (3) Quequen Grande estuary of the river (38°34'S; 200 km further south). All areas are characterized by fiddler crabs {Uca iirugiiayensis) and bun'owing crabs iChasiiuii;iuilhiis gninidata) populations (Spivak et al. 1991. Iribarne and Maitinez 1999). The litloi'al zone extends into an extensive Spanina-dommaXeA saltmarsh (Bortolus and Irib- arne 1999) and the fiddler crab population inhabits the upper open intertidal zone near the Spartina densiflora fringe, with few indi- viduals living at vegetated areas. Information was also obtained froni Quequen Salado river (38"56'S,) a site located 150 Km west of the Quequen Grande river. Biological Characteristics Crab saniples were collected monthly from October 1998 to February 1999. covering the reproductive season (see Spivak et al. 1991). Each time, a randoit) sample of 100 individuals was ob- tained to evaluate size frequency distribution, sex ratio and i-epro- ductive stage. Given that distribution of this species is patchy, to estimate density from each study site we obtained 50 sampling units of 1 m", 10 m apart, covering 500 m along the coast. Within dense patches (clearly noticeable crab beds) also 10 samples were obtained using the same procedure. Both sampling schemes were intended to have a measure of the mean overall density (density along the whole beach) and mean density of the Lk'a beds. ANOVA tests were used to evaluate differences between sites for both sampling strategies (Zar 1999). Given logistic constraints, it was impossible to keep track of settlement and recruitment across the southern boundary. There- foi'e. for each study site, size frequency distribution as well as diverse population and individual characteristics were evaluated to consider possible post-recruitment process that could produce the latitudinal abundance failure. From each sample, number of niales and, o\ igerous and non- ovigerous females were counted. All individuals were nieasured (maximum carapace width, precision 0.02 mm). Comparing crab body conditions from different sites, male and female (non- ovigerous and ovigerous without eggs) dry weight were considered covering homogeneously size classes found at each site. Crabs Fiddler Crab Distribution 355 were dried at 70°C during 48 hours, and weighed (precision 0.001 g). The null hypothesis of no difference in the relationship between dr\ weight and crab size, among categories (i.e., males, non- ovigerous females and ovigerous females) and sites, was evaluated with an ANCOV.'X test (Neter et al. 1990). To compare the amount of energy allocated to reproduction between sites, carbon proportion (as ash free dry weight. AFDW) between the egg masses and the females body (without eggs) was estimated. This proportion was called "energetic allocation". AFDW was obtained by incinerating the samples at 500°C for 4 hours, after having them drying for 48 hours at 70°C. We first evaluated the relationship between "energetic allocation" and fe- male si/e; given the lack of relationship we Just tested the null hvpolhesis of no difference in the mean \alues using Kruskall- Wallis test (Zar 1999). To compare egg production between sites, eggs number were estimated from dry weight of the egg mass (precision 0.0001 g) weighing 500 eggs from each site. Two egg stages were defined; "early", when embryo were undifferentiated and the eggs were completely filled with \itellum. and "late" when embryo is differ- entiated, with eyes and the egg was less than 409^ filled with \ilcllum. The null hypothesis of no difference in the relationship between carapace width and number of eggs among sites was evaluated with an ANCOVA test (Neter et al. 1990). To evaluate differences between number of early and late stage eggs for each site, the average egg mortality during development was estimated. The null hypothesis of no difference was evaluated with an ANCOVA test (Neter et al. 1990). Early stage eggs from at least 50 females ( regardless of cara- pace width) from each site were sampled, measuring their ma,\i- mum length with an ocular micrometer (precision 0.01 mm). The null hypothesis of no difference among sites was evaluated with a Kolmogorov-Stnirnov test (Conover 1980). Morphomelric characteristics were evaluated in order to exam- ine possible phenotypic plasticity. Different body measures were also obtained from males and females from each site (50 each), in order to perform a whole morphometric comparison. From each individual we measured: (1) carapace length (CL), (2) maximum carapace width (CW), (3) minimum carapace width (MCW), (4) distance between eyes (DE) (5) merus length of the third pereiopod of the left side of females and the hypertrophied side of males (Lpl ), (6) width of the fourth segment of the abdomen (AW), and (7) corporal height (CH). For males, we also obtained several measures of the hypertrophied cheliped: (8) merus length (ML), (9) carpus length (CLC). (10) total propodus length (PL), (k) pro- podus width (PW), (11) propodus height (PH), and (12) dactylus length (DL). Individuals of similar sizes were selected from each site to avoid the effect of size. Data were log transformed to comply with statistical analysis assumptions (Hair et al. 1995). Sizes and morphometry of larvae were also analyzed. Five females from two extreme sampling sites (Samborombon and Que- quen) were maintained in aquarium with seawater salinity of 34 at continuous aeration (following Rieger 1996). When eggs hatched, 50 larvae (zoea I) were preserved in 3'> formalin. Then, the fol- lowing measures were obtained: (1) total rostral length (RL), (2) frontal spine length (FSL), (3) dorsal spine total length of the (DSL) and (4) carapace length (CL). All measures were obtained with a precision of 0.02 mm. Then a discriminant analysis (DFA) (Hair et al. 1995) of both larvae and adults, was performed inde- pendently to evaluate the null hypothesis of no differences be- tween sites. Wind Pattern Wind-driven coastal circulation estimated as Ekman (1905) transport was assumed to be responsible for larval transport in coastal areas (McConnaughey et al. 1992). Ocean surface monthly mean winds (Mean Wind Field, MWF) from ERS-1, ERS-2 and ADEOS-1 scatterometers database (IFREMER 1998) were em- ployed in this analysis to infer Ekman transport. The period cov- ered was from August 1 99 1 -February 1998, totaling 78 observa- tions (lacking February 1992). Monthly winds were statistical in- terpolated using krigiiig minimal variance method and integrated into objective analysis (Bentamy et al. 1996) to obtain the monthly mean. Each grid points used on this analysis had an average of 8.6 satellite observations (SD = 3.6). Grid spatial resolution was 1° x 1° in latitude and longitude (111 km x 92.5 km average for study region). Accuracy reported for ERS scatterometer wind data was analyzed by comparison with in-sitii measurements from the Na- tional Buoy Data Center. Tropical Atmosphere Ocean data and Japan Meteorological Agency buoy data (Graberet al. 1996). The root square mean error was 1.2 m s"' in intensity and 24° in direction (Graber et al. 1996). NSCAT / ERS-2 correlation shown en-or differences about 1.1 m s"' and up to 28° (IFREMER 1998). Direction of the Ekman transport (Ekman 1905, cited by Neu- mann & Pierson 1966) forced by these winds was estimated at selected grid points of study area (Fig. lA) to evaluate favorable larval drift. For the southern hemisphere the direction of pure drift cuiTent at the surface will result 45^ to the left of wind vector, while the vertical integrated transport over Ekman layer will be 90° to left of the wind direction. These angles will start decreasing when water depth diminishes below 40 m where the frictional layer reaches the bottom. For this study, we considered that the transport of the upper surface layer will result 45 ' to the left of the wind direction taking into account the water depth off the coast ranges mostly between 30 to 50 m. This assumption was also coherent with observations on Cancer magister zoeae and mega- lopae drift off Northeastern Pacific (Lough 1976. Reilly 1983) and was applied by McConnaughey et al. (1992) on C. magister re- cruitment variability studies. Winds were assembled by the direction of the resulting surface flow into four groups of 1 80° angle range, where each would force the upper water layer to have, relative to coastline, an onshore or offshore and southward or northward component. For a 0° coast- line oriented N-S, an onshore (offshore) condition will result when wind blows from 45° to 225" (225° to 45°), with maximum effect with SE winds or 135° (NW or 315°: Fig. IB). Southward (north- ward) flow will occur under winds blowing from the angle range 315° to 135° (135° to 315°), with an optimum displacement with NE winds or 45° (SW or 225°). Four coastal regions were asso- ciated to each grid point, where wind directions were broken dow n relative to a defined coastal bearing (Fig. 1 A): ( 1 ) Samborombon- Pinamar (37°07'S, 57°I0'W) 0° angle: (2) Pinamar-Mar del Plata 30° angle: (3) Mar del Plata-Quequen Grande 55° angle and (4) west of Quequen Grande 75° angle. Favorable Uca iiriigiiayensis settlement was considered when winds force upper surface waters to move onshore and in a southward direction. For each region the wind directions that favor this drift condition were plotted through- out the year (insert 1 to 4 on Fig. I ). To a pool of absolute number of events recorded from December to April, the null hypothesis of no difference between regions in the proportion favorable:non fa- vorable events was evaluated with multiple comparisons for pro- portions (Zar 1999). 356 BOGAZZI ET AL. Results Biological Characteristics Maximum density siiowed differences between Samborombon and two other sites (ANOVA: F = 61.98. df= \.'i?>; a posteriori Tukey analysis P < 0.01 1. The mean density showed a latitudinal gradient, being higher at Samborombon, intermediate at Mar Chiq- uita and lower at both Quequen river sites (Kruskall-Wallis: H = 51.075. df = 3, 263, P < 0.01), no differences between Quequen Grande and Quequen Salado were found (Fig. 2). Size frequency distribution differences between sexes were found at Quequen (Fig. 3) where males were larger. Those males were also larger than males at the two other sampled sites (Sam- borombon: X = 1(152 mm. SD = 1.82; Mar Chiquita: x = 10.72 mm, SD = 1.82; Quequen: x = 12.15 mm, SD = 0.89). Females of Samborombon were smaller than those from the other two sites (Samborombon: x = 9.79 mm, SD = 1.38; Mar Chiquita: x = 1 1.09 mm, SD = 1.22; Quequen: x = 10.95 mm, SD = 1.13). Male proportion was higher at all sites (Samborombon: 41. 7f: 38.3m (P < 0.01 ), Mar Chiquita: 36.3f:63.7m (P < 0.05), Quequen: 34.8f:65.2m {P < 0.01)), but there were no differences between sites (F = 2.14, df = 2. 526, P > 0.05). Although some differences were found, ovigerous female percentages showed a similar pat- tern through the summer at all sites (Fig. 4). At all sites, differences in dry weight and carapace width rela- tion between sexes were found, being higher in males and smaller in ovigerous females (Samborombon: ANCOVA: F = 105.2. df = 2. 145, />< 0.01: Mar Chiquita: ANCOVA: F = 31.7, df = 2, 72, P < 0.01; Quequen: ANCOVA: F = 80.3, df = 2, 132, P < 0.01; Tukey ci posteriori, always P < 0.01; Fig. 5). Regarding slope, only differences were found between sites for males, being lower at Samborombon (Regression Analysis for log transformed data: Samborombon; F = 85.3. df = 47. b = 0.09. P < 0.01 ; Mar Chiquita: F = 448.1, df = 30. b = 0.14, />< 0.01; Quequen: F = 160.7, df = 31, b = 0.14, P < 0.01 ). High heteroscedasticity for males at Mar Chiquita and Quequen did not allow comparison. Females showed no difference between sites in slope. ANCOVA analysis showed differences in oviiierous females between sites 100 80 V 60 1- 01 y LLI Q 40 20 T D T B X SAMBOROMBON MAH CHIQUITA ON GHANDE ON SALADO SITES Figure 2. Density expressed as crabs m"" within dense patches (MAXIMUM) and alonj; 500 m of the coast (MEAN). Box plots are constructed with limits of boxes beinj; the 75"' and 25"' percentile; lines represent 10"' and 90"' percentiles. Points inside boxes are me- dians. Boxes underlined are not sl^nitlcanlly different between sites (ANOVA-Tukev lest. /' > 0.05; To MKAN: Kruskall-Wallis test, P > 0.05). Solid box: maximum den.sity and dashed box: mean density. >■ U z w C PJ OS u, > p < w a: 30 [ samborombon" 25 20 15 10 5 0 30 25 20 15 10 5 ■■ MALES I 1 FEMALES Hi MAR CHIQUITA' 0 50 40 30 20 10 J_l IR 0 L« ol:eouen ' n^ 8 9 10 11 12 13 14 15 CARAPACE WIDTH Figure 3. Size frequency distributions (carapace width, mm) of males and females, for each site. Relative frequency is the percentage of crabs of (he total sample. Vertical lines parallel to axis indicate no significant difference between sites (Kolmogorov- Smirnov test, P > 0.05). Results of Kolmogorov-Smirnov test between sexes are indicated by: '' P < 0.01 and "^'": P > 0.05. (non-ovigerous females: F = 2.18. df= 1. 127. P > 0.05; oviger- ous females: F = 11.15. df= 1. 118. P< 0.01) which were higher at Quequen (Tukey a posteriori test: Samborombon-Mar Chiquita: P > 0.05. Samborombon-Quequen: P < 0.01, Mar Chiquita- Quequen: P < 0.01). There was no relationship between carapace width and enei'- getic allocation at any site (Samborombon: t = -1.32. df = 60. r = 0.029. pe = 1.31. P > 0.05; Mar Chiquita: t = 0.55. df = 23. r- = 0.014. pe = 0.88. P > 0.05; Quequen: t = -0.45, df = 60, r- = 0.003, pe = 1.06, P > 0.05). There were no differences between sites in AFDW (Kruskal-Wallis: x' = 3.8, df = 2, 141, P > 0.05 ). There were no differences between sites in the relationship between female carapace width and egg number (ANCOVA: F = 1 .9, df = 2, 143, P > 0.05 Fig. 6). Egg size was not correlated with female size at any site. The early egg size frequency distribution showed smaller sizes in Mar Chiquita (Kolmogorov-Smirnov test: Samborombon-Mar Chiquita: P < 0.01, Samborombon-Quequen: P > 0.05, Mar Chiquita-Quequen: P < O.OI). The relationship between female carapace width and egg numbers did not show significant egg mortality at both sites (between egg stages, Sam- borombon: ANCOVA: F = 0.098, df = 1, 46, P > 0.05: Quequen: ANCOVA: F = 0.58, df = I, 57, P > 0.05), Fiddler Crab Distribution 357 ^ o SAMBOBOMBON 0.6 ■ ■ MAflCHIQUITA • 1 ^ • QUEQUEN 05 * I 1 T Z Q 04 • \ 1- 2 1 cc 1 ^ O 03 •ff ■*■ : 1 ; 0. o ■ * DC 02 CL 2 L • , * M it 2 0 1 ■ a < . : I DECEMBER FEBRUARY JANUARY MONTHS Figure 4. Proportion of ovigerous to non-ovigerous t'emiiles (Ho/H) along rcproductiM' period (December 1998 to February I999|. for each site. Relati\e frequency is the percentage of crabs with respect to total sample. Different numbers indicate differences between sites at f < (1.05 (Non-parametric multiple comparison test, NS: /' > 0.(15; Zar 19991. Segments indicates confidence limits for the proportion of ovigerous females: non-ovigerous females, for each site and month (Zar 19991. The discriminant analysis was unable to discriminate between different groups of Uca iinigiiayensis regardless of their geo- graphical origin (DFA, log transformed: Males: Lambda Wilk = 0.36, F = 1 .65. df = 22, 54, P > 0.05: Females: Lambda Wilk = 0.75, F = 1.48. df = 12. 1 14. P > 0.05). All morphometric characters had some contribution to variance, being DL and PH and CW and MCW the variable with the higher weight to males and females, respectively. Larvae discriminant analysis did not either show significant differences between sites (DFA: Lambda Wilk = 0.83, F = 2.14. df = 4. 41. P > 0.05), However, some differences may occur given that individuals were classified 85.2% at Samborombon while only 42.1"^ those from Quequen. Wind Patterns On the basis of satellite imagery, wind patterns that drive sur- face water southward and onshore were observed in the late spring and summer periods for four regions. However, these favorable conditions did not occur every year, and inter-annual variability is important. From the 46 months analyzed (from December to April) 15, 9, 6 and 4 months were favorable at 1,2, 3, and 4 regions, respectively. There was no significant difference in proportion ( favorable :non favorable events) between regions 2 and 3. and 3 and 4 (Multiple comparison for proportions: region 2:region 3. P > 0.05; region 3:region 4. P > 0.05: insert 1 to 4 on Fig. 1 ). Characteristic mean wind module with favorable directions for regions 1 and 2 was 6.3 m s"' (SD = 0.8 m s"'). DISCUSSION Our results show a clear pattern of decreasing density and abundance of Uca iirKguayensis with latitude over a land distance of less than 300 kilometers. However, there are no evidences of biological differences. Reproduction shows a similar pattern, fe- cundity is not different, eggs and larval sizes as well as the amount of energy allocated to eggs are similar. Only a difference in large adult size frequency distribution is found. Differences in size frequency distribution may be due to re- cruitment or survival, genetic differences (Harlnoll et al. 1993). food availability (Genoni 1985) and/or quality differences or food H X O a; Q 800 o SAMBOROMBON o ■ MAR CHIOUITA cP <»^ ° o QUEQUEN 600 ■ 400 ^■<^ ° ■ O fiC^ ■ 200 JH w^S^ o °*^ ■ o o MALES 0 800 600 e o o 400 ■ 8 -o 200 o.^ <9 o° ■ NON OVIGEROUS FEMALES 0 800 600 400 o ■o 200 ^^ m^ n cfc^ OVIGEROUS FEMALES 10 12 14 CARAPACE WIDTH 16 Figure 5. Relationships between size (carapace width, mm) and dr> weight (mg) of males, non-ovigerous females and ovigerous females, for each site. coiTipetition. Higher density at Samborombon may increase food competition and affect growth rate. Indeed lower body condition (as reflected in dry weight pattern) shown may be the result of higher density. Males and ovigerous females are individuals that require high energetic reserves (Klaassen & Ens 1993). How- ever, better body condition found on females at the southern limit of distribution suggests that the decrease in abundance is not due to physiological constraints. Furthermore, morphometric results did not show geographic differences, suggesting that environmen- tal condition (including diet, habitat etc.) would not be differenc- ing phenotypic expression between sites (see Overton et al. 1997). Sediment characteristics, organic content or humidity are rec- ognized as important factors in determining Uca species distribu- tion (e.g.. Teal 1958). However the comparison of the burrowing crab effect on sediment characteristics, between Mar Chiquita and Samborombon. reveal that there are no differences in the sediment properties (Botto and Iribarne 2000). and suggests that these fac- tors are not the basis for the density differences between sites. Predation by shorebirds has only been observed at Samborom- bon (Iribarne and Martinez 1999). although other sites such as Mar Chiquita are well studied (Botto el al. 1998: Botto et al. 2000). There are four migratory shorebirds {Pliivialis scjtiatarola. P. do- minica. Numeniits phaeopus and Arcitario inteipres) and a tern 358 BOGAZZI ET AL. C/5 O (3 O 6 LU CO A 0 SAMBOHOMBON ■ MAR CHIQUITA ■ 0 0 0 QUEOUEN ■ 0 - 0 0 . 6^0 0 0 ■ ■o °^ ,%%^ / ■ 10 11 12 CARAPACE WIDTH 13 14 15 Figure 6. Relationships between number of eggs (in thousands) and carapace width (mm) of ovifjcrous females. For each site data of egg stage (early, late) and month (December, January, February) were pooled. iCelnchelydon nilotica) that heavily prey on fiddler crabs at Bahi'a Samboroinboii, but not at the other sites (Iriburne and Martinez, 1999). Although this strong predation pressure may explain small sizes at Samborombon. it does not explain the southern limit of the distribution. Size frequency distribution at Mar Chiquita and Quequen show absence of indi\ idual conesponding to early autumn recruitnient (from 5 to 7 mm. see Spivak et al. 1991). All information previ- ously discussed does not explain such difference with Samborom- bon nor the latitudinal gradient density. An alternative explanation can be based on low temperature effect on post-recruitment mor- tality. This hypothesis is based on: ( 1 ) the fact that species of Ucu are known to be tropical (Crane 1973) and (2) its temperature range tolerance is low (see Dezi el al. 1993. Vernberg and Tashian 1959). Although temperature can limit Ucu species latitudinal dis- tribution, limitations of land crab distributions by low temperature does not need to involve mass mortality since cold can limit energy available for growth (Wolcott 1988). Pattern of annual temperature suggests that during winter iDonths the higher latitude sites seem to have a relatively lower minimum (4.7 C at Quequen and 4.S"C at Bahi'a Blanca. 38°43'S; Anonymous 1985. 1992); however, dif- ferences appear to be very small to support such hypothesis (maxi- mum difference 2.7°C, between Samborombon and Quequen). An alternative hypothesis, but not independent, is the lack of supply of larvae to the south. Larvae fate is unknown in this species, though in most ilea species they are exported to sea and will reinvade estuary sometime later (Lambert and Epifanio 1982; Epifanio 1988). Generally, larval movement can be entirely at mercy of currents (Butman 1987). As a consequence, larvae of coastal oiganisms can be transported by nearshore currents hun- dred of kilometers (Bertness 1999). The return of postlarvae (i.e.. megalopae) to the estuary may be the pioblem. Settlement and recruitment process are recognized as important in determining the distribution and abundance of species ("supply side ecology"; i.e.. Lewin 1986) especially for benthic marine invertebrates where new individuals are supplied through settlement of pelagic larvae (Gaines and Roughgarden 1985). In this case, population size may be limited by the arrival rate of larvae. Such "recruitment limited" populations would fluctuate in size as a response to temporal fluc- tuations in larval settlement rate. As a result, recruitment limited populations could be defined empirically as those with significant temporal correlation between settlement rates and local population sizes (Gaines and Lafferty 1995). In metapopulational context, larvae local populations contrib- ute to a common larval pool, from where they are subsequently distributed among local populations (Roughgarden and Iwasa 1986). Larval transport among local populations is a function of the physics of ocean circulation, larvae behavior and duration of planktonic stages (see Gaines and Lafferty 1995). Particularly, wind forced currents have been shown to play an itnportant role in larval transport shorewaid for settleinent at the appropriate titiie (e.g.. acorn barnacles Semibalanus balanoides: Hawkins and Hart- noU 1982; bryozoan Membranipora membranacea: Yoshioka 1982; blue crab Callinecles sapidus: Johnson and Hester 1989). Thus, if recruitment of megalopae to the estuary depends on trans- port forced by nearshore wind patterns, the supply of adults would also be affected by variations in the yearly wind pattern during the critical time when larvae are offshore (Johnson and Hester 1989). Our work evaluated the year round pattern of winds driving a southward and onshore Ekman drift. During late spring and sum- mer the larvae may reside offshore in the surface water and are affected mainly by Ekman transport due to wind stress. This phe- nomenon is ob.served for other species (e.g.. American lobster Hdiiuinis americaniis: Campbell 1989; Cullinectes sapidus: Johnson and Hester 1989; Dungeness crab Cancer magister: Mc- Connaughey et al. 1992). Wind transport influences Uca spp megalopae on larger open bodies of water, whereas tidal events become more important regarding the entry to the estuary (Mense et al. 1995). Our evidence based on wind patterns suggests that supply of larvae is the limiting factor, although the coupling of the coastal ocean and marine atmosphere (exchanging buoyancy and momen- tum) occurs within a time scale of a week (synoptic signal). With I -month resolution, a general trend is still observed. Synoptic sig- nal inclusion in wind monthly mean is confirmed through the comparison between twelve months from 1993 database used, and simulated monthly niean linearly interpolated from synoptic Eu- ropean Center for Medium-Range Weather Forecasts analysis (ECMWF) shows that both data set are correlated between latitude 60°S to 60°N. Correlation coefficient value changes with latitude, but for the study region it is between 0.4 and 0.8 (P < 0.05; IFREMER 1998). Our results show that wind condition in the coastal zone (south- ward and onshore) provides more favorable conditions to settle- ment at the northern sector of the study area than at the southern estuaries. This pattern agrees with a latitudinal decrease in Uca densities fi'om Samborombon to Quequen. Recruitment (as in- ferred from the size frequency distribution) also is continuous at Samborombon (Iribame and Martinez 1999), while it is not evident at southern estuaries. Ancillary information on recruitment be- tween 1993 and 1996 at Mar Chiquita coastal lagoon (Luppi pers. obs.) shows that settlement success agrees with coastward wind pattern inferred from monthly estimates. These data show that recruitment success was high only in Januai'y 1996 (at least 39 megalopae m""), intermediate in March 1993 (21 megalopae m~"). and low in January 1995 (2 megalopae m"") and March 1994 (8 megalopae m~"). Furthermore, during the summer 1997 megalo- pae were found sporadically (Valero et al. 1999). This information coincides with favorable ( 1996) and unfavorable wind pattern con- ditions (for other years). Thus, abundance pattern is consistent with latitudinal gradient in the larvae arrival hypothesis, rather than unfavorable environmental conditions for the biology of this species. Furtheimore. on the basis of results it is likely that Sam- Fiddler Crab Distribution 359 borombon acts as the parental stock while the other are 'satelhte' populations (Hanski 1982). ACKNOWLEDGMENTS This project was supported by the Agenda de Promocion Ci- entifica (PICT 97, N" 07-00000-0 1679B): Universidad Nacional de Mar del Plata (051/94). Fundacion Antorchas. the International zan. Foundation for Science (No. A/2501-2), the Consejo Nacional de Investigaciones CientiTicas y Tecnicas (CONICET) and the Insti- tuto Nacional de Investigacion y Desarrollo Pesquero (INIDEP). We thank T. Luppi for generous collaboration. The results pre- sented here are part of the E.B. thesis to obtain the Licentiate degree in Biolosical Sciences at the Universidad Nacional de Lu- LITERATURE CITED .Aiioii>nious. 1^85. Estadi'sticas climatologica!. 1961-1970. Servicio Me- teorologico Nacional. Buenos Aires. Argentina. 376 pp. Anonymous. 1992. Estadisticas climatologicas 1981-1990. Servicio Me- teorologico Nacional. Buenos Aires. Argentina. 689 pp. Balecli. E. 1964. Caracteres biogeograficos de la Argentina y Uruguay. Hoi. Inst. Biol. Men: 7:107-111. Bernasconi. I. 1964. Distribucion geogratlca de los equinoideos y aster- oideos de la extremidad austral de Sudamerica. Bol. Iits!. Biol. Mar. 7:43-50. Bentamy. A.. Y. Quilfen. F. Gohin, N. Grinia, M. Lenaour & J. Ser\ain. 1996. Determinalion and validation of average wind fields from ERS-I scatterometer measurements. Glolnil Anno.s. Ocean. Sv.vJ. 4:1-29. Bertness, M. D. 1999. The ecology of Atlantic Shore lines. Boston: Sinaver Assoc. 417 pp. Bortolus. A. & O. Iribarne. 1999. Effects of the SW Atlantic burrowing crab Chasmagmithus graniilaiu on a Spanina salt marsh. Mar. Ecol. Prog. Ser. 178:79-88. Boschi, E. E. 1964. Los Crustaceos Decapodos Brachyura del litoral Bo- naerense. Bol. hist. Biol. Mar. 6:1-99. Boschi, E. E., C. E. Fischbach & M. I. lorio. 1992. Catalogo ilustrado de los Crustaceos Estomatopodos y Decapodos marinos de Argentina. Frente Maril. 10:7-94. Boschi. E. E. 2000. Biodiversity of marine decapod brachuirans of the Americas. / Crust. Biol. 20:337-342. Botto. F.. O. Iribarne. M. Martinez. K. Delhey & M. Carrete. 1998. The effect of migratory shorebirds of three southwestern Atlantic Argen- tinean estuaries. Estuaries 21:700-709. Botto, F., G. Palomo, O. Iribarne & M. Martinez. 2000. The effect of southwestern Atlantic burrowing crabs on habitat use and foraging activity of migratory shorebirds. Estuaries 23:208-215. Botto, F. & O. Iribarne. 2000. Contrasting effects of two burrowing crabs [Chasmagnathus granulata and Lica iiruguayensis) on sediment com- position and transport in estuarine environments. E.st. Coast Shelf. Sci. 51:141-151. Butman. C. A. 1987. Larval settlement of soft sediment invertebrates: the spatial scales of pattern explained by active habitat selection and the energetic role of hydrodynamical process. Ocean. Mar. Biol. Ann. Rev. 25:113-165. Campbell. A. 1989. The lobster fishery of southwestern Nova Scotia and the Bay of Fundy. In: J. F. Caddy, editor. Marine invertebrate fisheries: their assessment and management. New York: Wiley-lnterscience. pp. 141-158. Connolly. S. R. & J. Roughgarden. 1998. A latitudinal gradient northeast Pacific intertidal community structure: evidence for an oceanographi- cally based synthesis of marine community theory. Am. Nat. 151:31 1- 326. Conover, W. J. 1980. Practical non-parametric statistics. New York: John Wiley & Sons. 493 pp. Crane. J. 1975. Fiddler crabs of the world. Ocypodidae: Genus Uca. Princeton, NJ: Princeton University Press. 736 pp. Dezi. R. E.. E. M. Rodriguez, C. M. Luquet, C. O. Cervine & G. E Haut. 1993. Preferred thermal regime of the estuarine crabs Uca iiruguayen- sis and Cliasinagnatlms granulata (Decapoda. Brachyura). Comp. Physiol. Ecol. 18:149-155. Eknian. V. W. 1905. On the influence of the earth's rotation on ocean currents. Arkiv. Matem. Astr. Fysik. Stockholm 2(11) (cited in Neu- mann & Pierson 1966). Epifanio. C. E. 1988. Transport ot invertebrate larvae between estuaries and the continental shelf Am. Fish. Soc. Syinp. 3: 104-114. Farrell, T. M.. D. Bracher & J. Roughgarden. 1991. Cross-shelf transport causes recruitment to intertidal populations in central California. Lim- nol. Oceanogr. 36:279-288. Frith. D. W. & S. Brunenmeister. 1980. Ecological and population studies of tiddler crab (Ocypodidae. Genus Uca) on a Mangrove shore at Phuket Island. Western Peninsular Thailand Crustaceana 39:157-184. Gaines. S. D. & K. D. Lafferty. 1995. Modeling the dynamics of marine species: the importance of incorporating larval dispersal. In: L. R. McEdwards. editor. Ecology of marine invertebrate larvae. Boca Ra- ton. FL: CRC Press, pp. 389-412. Gaines, S. & J. Roughgarden. 1985. Larval settlement rate: a leading determinant of structure in a ecological community of the marine in- tertidal zone. Proc. Natl. Acad. Sci. USA 82:3707-3711. Genoni. G. P. 1985. Food limitation in salt marsh fiddler crabs Uca rapctx (Smith) (Decapoda: Ocypodidae). J. E.xp. Mar. Biol. Ecol. 87:97-110. Graber. H.. N. Ebuchi & R. Vakkayil. 1996. Evaluation of ERS-1 scatter- ometer winds with ocean buoy observation. In: Proceedings of OCEANS96. Ft Lauderdale. Florida, pp. 1157-1165. Hair. J. F.. Jr., R. Anderson. E. Tatham & W. C. Black. 1995. Multivariate data analysis. New Brunswick, NJ: Prentice-Hall. 757 pp. Hanski, I. 1982. Dynamic of regional distribution: the core and satellite species hypothesis. Oikos 38:210-221. Hartnoll. R. G.. A. D. Bryant & P. Gould. 1993. Size distribution in spider crab populations spatial and temporal variation. J. Crust. Biol. 13:647- 655. Hawkins. S. J. & R. G. Hartnoll. 1982. Settlement patterns of Semibalanus balanoides (L.) in the Isle of Man (1977-1981). J. E.xp. Mar. Biol. Ecol. 62:119-126. Hengeveld, R. 1992. Dynamic biogeography. Cambridge, UK: Cambridge University Press. 249 pp. IFREMER 1998. Mean surface wind fields from ERS-AMI and ADEOS- Nscat microwave scatterometers. Ocean Circulation and climate con- ference of WOCE, held in Halifax, Canada (May 1998). Iribarne. O. & M. Martinez. 1999. Predation on the Southwestern fiddler crab (Uca uruguayensis) by migratory shorebirds (Pluvialis dominica. P. squatarola. Arenaria interpres and Numenius phaeopus). Estuaries 22:47-54. Johnson. D. & B. Hester. 1989. Larval transport and its association with recruitment of blue crab to Chesapeake Bay. Est. Coast Shelf. Sci. 28:459-472. Klaassen. M. & J. Ens. 1993. Habitat selection and energetic of the fiddler crab ( Uca tangeri). Neth. J. Sea Res. 3 1 :495-502. Lambert, R, & C. E. Epifanio. 1982. A comparison of dispersal strategies in two genera of brachyuran crabs in a secondary estuary. Estuaries 5:182-188. Lewin, R. 1986. Supply-side ecology. Science 234:25-27. Longhurst. A. 1998. Ecological geography of the sea. San Diego. CA: Academic Press. 398 pp. Lough. R. G. 1976. Larval dynamics ot the Dungeness crab. Cancer mag- ister. off the central Oregon Coast. 1970-1971. F/.?/i. Bull. 74:353-375. McConaugha. J. R. 1992. Decapod larvae: dispersal, mortality and ecol- ogy. A working hypothesis. .Anier. Zool. 32:512-523. 360 BOGAZZI ET AL. McConnaughey. R. A., D. A. Armstrong, B. M. Hickey & D. R. Gunder- son. 1992. Juvenile diingeness crab iCaiuer imigislcr) recruitment vari- ability and oceanic transport during the pelagic larval phase. Can. J. Fish. Aqiial. Sci. 49:2028-2044. Menge, B. A. 1991. Relative importance of recruitment and other causes of variation in rocky intertidal community structure. / Ev/j. Mar. Biol. Ecol. 148:89-100. Menni. R. C. & A. E. Gosztonyi. 1982. Benthic and semideniersal fish association in the Argentine sea. Stiidies Neotrnp. Fauna Envirnnm. 17:1-29. Mense. D. J., M. H. Posey. T. West & K. Kincheloe. 199,'i. Settlement of brachyuran postlarvae along the north Carohna Coast. Bull. Mar. Sci. 57:793-806. Neter. J., W. Wasserman & M.H. Kutncr 1990, Applied linear statistical models. Irwin, IL. 1181 pp. Neumann. G. & W. J. Pierson Jr. 1966. Principles of physical oceanogra- phy. New Brunswick, NJ: Prentice-Hall. 545 pp. Overton. J. L.. D. J. Macintosh & R. S. Thorpe. 1997. Multivariate analysis of the mud crab Scylla serrala (Brachyura: Portunidae) from four lo- cations in Southeast Asia. Mar. Biol. 128:55-62. Reilly, P. N. 1983. Dynamics of Dungeness Crab. Cancer magisler. larvae off central and northern California. Fish. Bull. 172:57-84. Rieger, P. J. 1996. Desenvolvimento larval de Uca (Celuca) uruguayensis Nobili 1901 (Crustacea, Decapoda, Ocypodidae) em laboratorio. Nau- plius 4:li-\()}. Roughgarden, J., G. Gaines & H. Possingham. 1988. Recruitment dynam- ics in complex life cycles. Science 241:1460-1466. Roughgarden, J. & Y. Iwasa. 1986. Dynamic of a metapopulation with space-limited subpopulations. Tlieor. Pop. Biol. 29:235-261. Spivak, E. D., M. A. Gavio & C. E. Navarro. 1991. Life history and structure of the world's southermost Uca population: Uca uruguayensis (Crustacea. Brachyura) in Mar Chiquita Lagoon (Argentina). Bull. Mar. Sci. 48:679-688, Stuardo. J. B. 1964. Distnbucion dc moluscos marinos litorales de Lati- noamerica. Bol. Insl. Biol. Mar. 7:78-91. Teal, J. M. 1958. Distribution of fiddler crabs in Georgia salt marshes. &o/<;,i;v 39:185-193. Underwood. A. J. & E. J. Denley, 1984. Paradigms, explanations and generalizations in models for the structure of intertidal communities on rocky shores. In: D, R, Strong, D, Simberloff. L, C, Abele & A. B. Thistle, editors. Ecological communities conceptual: issue and the evi- dence. Princeton, NJ: Princeton University Press, pp. 151-180. Valero, J., T. Luppi & O. Iribarne. 1 999. Size indicator of swimming speed in crab megalopae. J. Shellfish Res. 18:663-666. Vernherg. P. J. & R. E. Tashian. 1959. Studies on the physiological varia- tion tropical and temperate zone fiddler crabs of the genus Uca. I. Thermal death limits and distribution. Ecology 40:589-593. Wolcott, T. G. 1988. Ecology. In: W. W. Burggren & B. R. McMahon, editors. Biology of the land crabs. Cambridge. UK: Cambridge Uni- versity Press, pp. 55-96. Yoshioka, P. M. 1982. Role of planktonic and benthic factors in the popu- lation dynamics of the bryozoan Memhranipora memhranacea. Ecol- ogy 63:457^68. Zar, J. H. 1999. Biostatistical analysis. Englewood Cliffs. NJ: Prentice- Hall. 718 pp. .loiinuil of Shellfish Research. Vol. 20. No. 1. 361-363. 2001. EFFECT OF BROODSTOCK DIET ON REPRODUCTIVE PERFORMANCE OF THE PEPPERMINT SHRIMP, LYSMATA WURDEMANNI JUNDA LIN* AND DONG ZHANG Department of Biological Sciences, Fhnida Institute of Technology, Melbourne. Florida 32901 ABSTRACT Four types of frozen adull Anemia biomass [regular and enriched (with HUFAS and other nutritional supplements) San Francisco Bay and Canadian brands], and frozen hard clam (Meicenuria meirenaria) were evaluated as broodstock diet for the marine ornamental shrimp Lysmata wurdeiuaniu. There is no significant difference among the five diet treatments in fecundity (number of eggs produced), relative fecundity (number of eggs/g female), egg dry weight, egg percent ash, and total length of newly hatched larvae. Egg volume was significantly larger in the San Francisco Bay regular treatment than those of other treatments. Survivorship from zoea 1 to zoea 2 was higher in the San Francisco regular. Canada enriched, and hard clam treatments. KEY WORDS: Broodstock diet, Lxsmaia wiinlemauni. ornainental shrimp, reproductive performance INTRODUCTION Practically all of the marine organisms marketed in the aquarium trade industry are collected from coral reef ecosystems. Extensive and destructive collection has caused great concerns and much efforts have been devoted to develop technologies to culti- vate these species. Peppermint shrimp, Lysmata wiirdeinanni. is one of the most popular species in aquarium trade industry. Complete life-cycle culture of peppermint shrimp has been achieved in the laboratory, using newly hatched /I /7e/);/(/ nauplii as the only food (unpublished data). However, there are still some obstacles in commercial production of the species. For example, there are variations in larval duration (ranging frotn 90 to over 100 days); sometimes embryos are aborted prematurely; and some larvae from captive broodstock have difficulty separating from the old exoskeleton during molting (unpublished data). Im- proving broodstock diet may enhance reproductive performance, thereby increasing the commercial production potential of the species. Frozen adult Arteinia can be used as a supplement or replace- ment of maturation diet for Penaeus semisuleatus (Browdy et al. 1989), P. styUr«stri.s (Bray et al. 1990), and P. vawuimei (Naes- sens et al. 1 997 ). It is possible to boost the Anemia with nutritional supplements, such as highly unsaturated fatty acids (HUFAS). which may stimulate reproductive pert'ortiiance in shrimp. Indi- viduals of Penaeus chinensis (Xu et al. 1992, 1994) and P inJiciis (Cahu et al. 1995) fed with broodstock diets containing higher HUFAs produce better eggs with higher hatching rates. The tissue of clams, squid, or shrimp is rich in n-3 HUFA. arachidonic acid, cholesterol and other sterols, phospholipids, and essential atnino acids, and appears to promote spawning success in penaeid shrimp (see Harrison, 1997 for a review). The present study compares the effects of different types of frozen adult Artemia and hard clam Mercenaria mereenariu as broodstock diets on the reproductive performance of the pepper- mint shrimp, MATERIALS AND METHODS The study was conducted at Florida Institute of Technology, USA from May 1999 to May 2000. Farm-raised adult shrimp used *Corresponding author. E-mail: jlin@fit.edu; Fax: -hi -32 1-674-7238. in the experiment were purchased frotn Oceans Reefs & Aquari- ums. Inc. at Fort Pierce. Florida and maintained in a recirculating seawater system under 14 h light followed by 10 h dark. Water temperature was controlled at 27-29°C and salinity at 31-33 ppt. The shrimp were fed in excess once a day with one of the five test broodstock diets: four types of frozen adult Artemia: regular and enriched with (HUFA, phospholipids, steroids, carotenoids and vitamins) from two sources (Canada and San Francisco Bay): and frozen hard clam (Mercenaria mereeiiaria). Thirty-five shrimp of similar total length (TL. from the end of telson to the tip of ros- trum) (Table I ) seven for each treattnent. were used. The shrimp were fed with the test diet for at least one month prior to collecting the following parameters: TL; wet body weight (WBW); fecundity (number of eggs); egg diameter and dry weight; and larval survivorship from Zoea I (Z,) to Zoea 2 (Z,). Two consecutive spawns from each experimental shrimp were used. Thirty newly hatched larvae (Z,) were selected from the first spawn and TL of each larva was measured under a dissecting tiiicroscope with ocular. Each larva was then placed in a 270-ml transparent plastic cup with 220-ml seawater and fed with newly \\Mc\\eA Artemia nauplii. The number of larvae that survived to Z2 (30-36 hrs. later) was counted and survivorship was calculated. At about 24 h after the second spawning, the entire egg mass was removed gently from the shritiip using stnall forceps and number of eggs were counted. Diameters of 30 eggs were inea- sured under a microscope with an ocular micrometer. Egg volume was calculated using the formulae V = l/6(Trd') for spherical eggs (d is the egg diameter) and V = l/6(TTd, x d,") for prolate sphe- roids (d| is the least diameter and d-, is the greatest diameter). One hundred eggs were separated from the rest of the eggs and all were dried at 70°C for 48 h. The total weight of the 100 eggs was measured to the nearest 0,01 mg using a digital balance and mean dry weight per egg was calculated. The spawned shrimp were dried on blotting paper. TL (to the nearest 0.1 mm), and WBW (to the nearest 0, 1 mg) were measured. Five different-sized shrimp (front 2.81 to 4.50 cm in TL) were used to cotnpare the fecundity and egg volume from consecutive spawns. For each shrimp, fecundity and diameter of 30 eggs were measured from each of the two spawns. One-way ANOVA was used to test the effect of different foods on average (of seven replicate shrimp) fecundity, relative fecundity (fecundity/g WBW). egg volume, dry egg weight, egg density (egg dry weight/egg volume), larval total length and survivorship from 361 362 Lin and Zhang TABLE 1. Effect of frozen Arteniia iiiid clam as broodslock diet on reproductive performance of the peppermint shrimp, Lymiata wiiydemanni (mean ± s.d.. n = 1) San Francisco Bay Canada Clam Enriched Regular Enriched Regular ANOVA Adult total length (cm) 3.S4±(124 4.17 ±0.29 3.95 + 0.51 3.98 + 0.30 3.84 ±(.). 15 ns Fecundity (number of eggs) 1267 + 322 1456 ±429 1203 ±261 1080 ±340 1092 + 224 ns Relative fecundity (#/g female) 1578 ±241 1 309 ± 145 1291 ± 124 1357 ± 139 1304 ±274 ns Egg volume (mm"") 0.0582 ± 0.0049'' 0.0639 ± 0.0064" 0.0565 + 0.003 '2'" 0.0582 ± 0.0029" 0.0589 ± 0.0034'' *:h Egg dry weight (jxg/egg) 30.0 ± 2.3 33.0 ±3.1 30.9 ± 3. 1 30.9 ± 2.4 30.4 ± 2.0 ns Egg percent ash 2.5 ± 1.2 2.3 ± 0.9 2.4 ±0.8 2.4 ± 0.7 2.4 ± 1.0 ns Larval total length (mm) 2.57 ± 0.06 2.55 ± 0.05 2.51 ±0.03 2.57 ± 0.05 2.56 ± 0.96 ns Survivorship from Z, to Z, (%) 75.5 ± 10.2 87.6 ± 15.9 88.1 ±9.2 71.9 ±13.7 86.7 ±4.7 * * Signil'icantl\ diflLTL-nl {!' < 0-05 1. ** Highly significantly diflerenl [P < 0.01 1. ns. Not significantly different [P > 0.05). Z| to Z,, respectively, among the treatments. If the ANOVA result was significant, Fisher LSD method was employed to test for differences among the means. A two-tailed Student's /-test was used to compare the average fecundity and egg volume between the two consecutive spawns. RESULTS There was no significant difference in female total length among the five treatments (Table 1 ). Of the seven reproductive parameters measured, only egg volume and survivorship from Z, to Zt show significant differences among the treatments (Table 1 ). Egg volume from the regular San Francisco brand Anemia is sig- nificantly (P < 0.01) larger than those of the other treatments that do not significantly different from one another. The survivorship from Z| to Z, is significantly higher in the Canada enriched, San Francisco regular, and hard clam treatments than in the San Fran- cisco enriched and Canada regular treatments (Table 1 1. There is no significant difference in fecundity (P = 0.89) between the first (mean ± s.d. = 953 ± 614) and the second (mean ± s.d. = 898 ± 464) spawns. Similarly, the difference in average egg volume be- tween the first (mean ± s.d. = 0.0593 ± 0.0074 mnr^) and second (mean ± s.d. = 0.0533 ± 0.(J074 mm') spawns is not statistically significant (P = 0.23). DISCUSSION Peppermint shrimp, also called red cleaning shrimp, may feed on parasites of fish in its natural environment, like many of its sibling species in the genus (Limbaugh et al. 1961, Debelius 1984, Jonasson 1987). Therefore, its natural diet may be narrow (al- though little is known on the relative importance and composition of parasites in the shrimp's natural diet). We found no significant difference in five of the seven reproductive performance param- eters measured among the five-broodstock diets tested. There is no clear pattern in the two parameters (egg volume and survivorship from Z| to Z,) showing consistent differences among the treat- ments. Larger eggs in the San Francisco regular treatment may be the result of larger (though not statistically significant) total length of the females, as egg volume correlates positively with maternal size in Lysinata wiinlciiuuini (Lin and Zhang, unpublished data). In a sibling species, L, dchelius. larval production is higher for those fed with live Arteniia nauplii and those with enriched (with HU- FAS, vitamins and others) frozen Arteniia nauplii than those fed with regular frozen Arteniia nauplii, which in turn is significantly and substantially (about UW/c) higher than those fed with natural diets (mixture of fresh mussel and polychaete) (F. Simoes, pers. comm., 1998). However, in the same study on another sibling species, L. anihioinensis. enriched frozen Arteniia nauplii and natural diet resulted in the same level of larval production (F. Simoes, pers. comm. 1998). Browdy et al. (1989) demonstrated that using frozen adult Artemia as a dietary supplement can en- hance reproductive performance of Peiiaeiis seniisiilcatiis. The re- productive activity was not consistent, however, and this may be due to possible variability in the nutritional quality of the differ- ent batches of Arteniia used. The enrichment of Artemia is sup- posed to increase nutritional quality and reduce variability among Arteniia batches. Naessens et al. (1997) showed that frozen .4/- teinia may be useful as a supplement or replacement for polycha- etes as maturation diet for Penaeus vanammci. However, brine shrimp biomass enriched with HUFAS did not perform better than the regular Arteniia in their experiments. The authors caution the interpretation of the results and suggest that more research is needed to compare enriched and regular brine shrimp bio- mass and identify potentially active ingredients and their optimal doses. ACKNOWLEGMENTS We thank Feng Cen for her assistance in the laboratory. Florida Sea Grant (Grant #NA76RG-0120) and Maritech, Inc. provided financial support for the project. LITERATURE CITED Browdy, C A. Hadani, T. M. Samocha, & Y. Loya, 1989. An evaluation of frozen Anemia as a dietary supplement for the stimulation of reproduction in penaeid shrimp. Belgium: Euro. Aquaculture .Soc. Bredene. 671-678. 3ray. W. A., A. L. Lawrence & L. J. Lester. 1990. Reproduction of eye- sialk-ablated Penaeus stylirostris fed various levels of total dietary lipid. J. World Aiiuacuhure Society 21(l):41-52. Broodstock Diet on Reprodi'ction 363 Calm. C. 1... G. Cu/,on & P. Quaziiguel. 1995. Effect of highly unsaturated fatty acids, a-tocopherol and ascorbic acid in broodsiock diet on egg composition and development of Pcmiciis uuluiis. Camp. Bitnhcni. Physiol. ll2A:4l7-t24. Debelius. H. 1984. Armoured knights of the sea. Germany: Kernen Verlag. Harrison, K. E. 1997. Broodstock nutrition and maturation diets. In Crus- tacean Nutrition. L. R. D'Abramo. editor. Advances in World Aqua- culture. World Aquaculture Society. Viil. 6:90—401 Jonasson. M. I9H7. Fish cleaning behavior of shrimp. J. Ziiol. LdihI. 21.1: 117-LM, Limbaugh, C. H. Pederson & P. A. Chace, Jr. 1961. Shrimps that clean fishes. Bull. Mar. .Set. Gulf Caithham 11(2): 2.17-2.57. Naessens, E., P. Lavens. L. Gomez. C. L. Browdy. K. McGovern-Hopkins. A. W, Spencer, D. Kawahigashi & P. Sorgeloos. 1997. Maturation performance of Penaeiis vannaim'i co-fed Aricinui biomass prepara- tions. Aquaculturc 153:87-101. Xu, X. L.. W. J. Ji, J. D. Caslell & R. K. O'Dor. 1992. Influence of dietary lipid sources on fecundity, egg hatchability and fatty acid composition of Chinese prawn (Penaeus thincnsis) broodstock. Aquuculture 119: 359-370. Xu, X. L., W. J. Ji, J. D. Castel! & R. K. O'Dor. 1994. Effect of dietary lipids on fecundity, hatchability and egg fatty acid composition of Chinese prawn [Pciuwus vluiwnsis) Marine Fisheries Research 13:1.3-19. Joiinwl KfSlwllftsh Research. Vol. 20. No. I. 365-.^67. 2001. THE TEXAS LIVE BAIT SHRIMP MARKET RYAN GANDY,' TZACHI M. SAMOCHA,' EDWARD R. JONES,' AND DAVID A. MCKEE' ' Texas Agricultural Experiment Station. Shrimp Mariculture Research Facility. 4301 Waldnm Rd. Corpus Christi. Texas 78418: 'Texas A&M University-Corpus Christi. 6300 Ocean Drive. Corpus Christi. Texas 78412 ABSTRACT Decreases in wild catches couplcU wiih increased regulations have attracted interest in the production of farm-raised live bait shrimp. Recent refinements in penaeid culture techniques suggest that the production of live bait is a potentially profitable endeavor. In this study, live bait shrimp dealers in Texas were surveyed by mail. The return rate was 8'/f . Returned surveys were found to be from 33 businesses from various locations over the entire Texas Coast. The information collected provides a small database that can be used to provide a broad and general view of the Texas market demand for farm-raised live bait shrimp. Texas market dynamics were characterized in terms of regional demands and seasonal fluctuations. Most live bait retail suppliers in Texas have been operating for 6-20 years, and during any given year they are able to meet the market demands for two only months. Most retailers expressed acceptance of a farm-raised product if it was of good quality and consistently supplied. KEY WORDS: Texas. Gulf of Mexico. li\e bait shrimp. Faijanlepenaeu.s azteciis. F. dmiiLiniin. Liuipenaeus seiifeiiis. survey Historically, the supply of live bait shrimp for the eastern At- lantic and Gulf of Me.xico coasts has relied on the harvest of wild Farfantepenaeus aztecus (Atlantic brown shrimp), F. chioraruin (Atlantic pink shrimp), and Litopenaeus setiferus (Atlantic white shrimp). The overharvest of this natural resource forced the Na- tional Marine Fisheries Service to take action to maintain a sus- tainable yield of wild-caught shrimp (Warren 1980, McKee 1986, Southern Fisheries Science Center |SFSC] 1992, Nance 1993). The status of the Gulf of Mexico shrimp fishery for F. aztecus, F. duoraniin, and L. setiferus was declared overexploited by the Na- tional Marine Fisheries Service in 1990 (SFSC 1992). The catch fishery is currently seen as overcapitalized, with more effort ex- pended per unit catch than is economically feasible (SFSC 19921. The increasing overexploitation of the wild shrimp fishery has caused the bait shrimp tleet on the Texas coast to redirect its catch emphasis from the historically abundant white shrimp, L. setiferus. to the less abundant brown shrimp, F. aztecus (Texas Parks and Wildlife Department 1995). The Texas Parks and Wildlife Depart- ment's shrimp management proclamation recognizes that the con- tinuing increase in the harvest of small shrimp (>67 count/lb) is ecologically unsustainable and that it will cause shrimp stocks to collapse (Texas Parks and Wildlife Department 199.3). The decline in wild harvests along the Texas Coast was observed, as early as 1986, to have an effect on retail markets (McKee 1986). McKee et al. (1989) observed that the retail demand for live bait shrimp in Texas waters was greater than what the bait shrimp tleet was able to supph. The status of the fishery prompted the establishment of a Gulf of Mexico management plan and prompted the South Atlantic Management Council in 1981 to regulate the fishery along the Gulf of Mexico and eastern coast of the United Stales. Regulations at the state level have also been forthcoming. In 1995, the state of Texas signed into law HB750, which imposed new regulations on bait shrimpers. This bill was designed to provide long-term con- servation of shrimp stocks and to create economic stability in the bait shrimp fishery (Texas Parks and Wildlife 1995). For example, the bill limits the number of future bay bait shrimping licenses, which will eventually lead to a reduction in the bait fishing fieet. Due to the great demand and high dollar value of live bait shrimp, the production of farm-raised live bait shrimp has attracted much attention in the past few years. Recent studies have reported on the profitability of producing cultured live bait shrimp using modern farming techniques (McKee 1986). Limited data were available concerning the economic viability of a live bait shrimp farm (Parker & Holcomb 1973, Quick & Morris 1976, Rossberg & Strawn 1980). The initial research concluded that farming live bait shrimp was unprofitable (McKee et al. 1989). However, recent studies suggest that the production of live bait shrimp is a poten- tially profitable endeavor (Sandifer et al. 1993, Burkott 1994, Samocha et al. 1998). These studies have focused on the manipu- lation of stocking density, the development of new culture tech- niques, and pond management it) improve the profitability of live bait shrimp production. Past studies concerning Texas" demand for live bait shrimp relied on National Marine Fisheries Service landing reports and personal interviews in Texas (McKee 19S6. Burkott 1994). These studies provided only a partial estimation of consumption, produc- tion volume, and a general pricing structure of this market because data were collected mainly from local markets. Assessment of the true potential of a live bait shrimp farming requires an understand- ing of production and infrastructure costs, market demand, and market price data for the major regions of this market. The objectives of the current study were ( 1 ) to collect infor- mation concerning the live bait shrimp business type and market environment in Texas, (2) to analyze the supply and demand re- lationships of live bait shrimp within this region; and (3) to assess the acceptability of farm-raised live bait shrimp by retailers. A mail-out survey was used to collect data on the regional live bait shrimp market demand and supply. In general, the data char- acterize seasonal fluctuations in the supply and demand of bait shrimp. In addition, this survey provides an overall view of the variation in the Texas wholesale pricing structures, quantity de- mand, mortality rates of wild-caught live bait shrimp, average quantity demand of each retail outlet, and acceptance of farm- raised bait shrimp by retailers. These data also provide an under- standing of regional quantity demand, species usage, and the size of live bait shrimp preferred by this market. MATERI.ALS AND METHODS The research was conducted at the Texas Agricultural Experi- ment Station Shrimp Mariculture Research Facility in Corpus 365 366 Gandy et al. Christi. Texas. Mailer information of live hait dealers was obtained from the Texas Parks and Wildlife Department. Survey forms were mailed to the 412 licensed live bait dealers in Texas. The survey contained 31 inultiple-choice questions and two short-answer questions. Questions were presented in an easy-to-read format. The surveys were mailed between July 23. 1996 and August 5, 1996 with a self-addressed stamped return envelope to improve the rate of return. RESULTS AND DISCUSSION Of the 412 licensed Texas bait dealers. 33 returned the survey (8% return rate). The low return rate of this survey suggests some inherent bias due to the use of a small data set. The return data, when grouped by physical location along the Texas coast, shows that the returned surveys were not from one specific area. These surveys were in fact from 33 businesses from various locations over the entire Texas Coast. From this data set. which is limited in volume but broad in scope, it is possible to understand the Texas market from a general perspective. The Texas live bait shrimp retail market is comprised of retail- ers who have been operating in this market for 6-20 y (Fig. 1 ). Texas retailers indicated an increase in live bait shrimp demand over the past 5 y without an increase in supply. The inability of the bait shrimpers to supply the Texas market, observed in the early years by McKee (1986), appears to continue today. Texas retailers indicated that a general lack of supply is expected nearly all year round. Texas retailers expect a semi-consistent supply of live bait shrimp in May. June, and October: however, this supply is not reliable (Fig. 2), although the highest demand for live bait shrimp by fishermen lasts from April to October (Fig. 2). The primary limiting factor in meeting the demand is the bait shrimping laws that limit catches to 200 Ib/vessel/day (TPWD 1995). The sales of live bait shrimp accounted for more than 50% of the sales for 29% of the live bait dealers. Another 29% of the dealers reported live bait shrimp sales that accounted for 25—49% of their total sales (Fig. 3). Live bait shrimp is the key to success for all bait and tackle stores. Lack of live bait also effects the sales of other merchandise in a bait and tackle store. In Texas, the holding capacities are limited to 22-f3 kg (50- 100 lb. ) of live shrimp per retailer. The retailers are restocked daily during the high-demand periods and daily to weekly during low- 20 oi S nn s nil 1-5 6-10 11-20 21-30 41-1 Years In Business Figure I. Texas live bait shrimp retail business age. Jan Feb Mar Apr May June Jul Aug Sept Oct Nov Dec Month -High Demand - « - Supply Fails To meet demand! Figure 2. Consumer demand vs. availability for Texas live bait shrimp market. demand periods. Under the high demand for bait shnmp in this market, supply becomes an important factor. Ninety-four percent of the retailers surveyed responded that three local suppliers sup- ply them; only a small portion of the Texas supply is imported from Florida. The controlling factors involved in the selection of a supplier by Texas retailers are availability and consistency. The ability of a farm-raised product to meet these criteria illustrates its advantage over the wild-caught live bait shrimp. When asked if they would purchase a farm-raised product, 74% of the retailers replied posi- tively, provided the supply is consistent and of good quality (e.g., hardy animals with minimal signs of broken appendages and le- sions). This survey showed that 36% of the live bait shrimp dealers in Texas listed F. aztecus as their species of choice. 24% preferred F. duoramm. 21% had no preference, and 15% preferred L. serifenis. The data further corroborates the eariier study by McKee (1986). which suggested that sport fishermen prefer brown shrimp due to its greater availability and high tolerance to stress. The current study further suggests that the majority of the live bait shnmp are purchased by the gallon (3. 785 I) with a size preference of 60-70 count/lb (132-154/kg) or individual shrimp weight between 6.5 and 7.6 g. The average price for this size of shrimp was $2I.23/g. Inglis and Chin (1966) reported a retailer preference for 80-count/lb shrimp (5.76 g per shrimp). On the other hand. Quick and Monis (1976) suggested a preference for 200-count shrimp (2.27 g per shrimp). Our data corroborates the wholesale and retail pricing trends suggested by eariier studies. 0 ■_ m m ~^ nn [ID ng <2% 2-4% 5-9% 10-24% 25-49% >50% Retail Revenue fronn Live Bait Sales Figure 3. Percent revenue expected by retailers from sales of live bait shrimp. Texas Livr. Bait Shrimp Market 367 The wholesale price for bait shrimp has been documented to fluc- tuate according to supply, whereas the retail price remains constant within a region (McKee 1986). Prior studies have reported an accepted mortality of 10-20% for live bait shrimp after delivery (McKee 1986). This survey found the expected mortality to be higher. Forty percent of the retailers expected the standard IO-247f mortality per shipment, whereas 20Vr expected 25— 4^9'7r mortality (Fig. 4). This great loss of stock translates directly into lower revenues. Although lack of slock has an impact on sales, the weather was cited as the primary controlling factor that negatively impacts retail sales. This study has concluded that, in general, the Texas market for li\e bait shrimp is largely undersupplied during the highest de- mand periods of the year. Whether this is due to a declining re- source or due to tighter catch restrictions is a point of contention. However, the fact remains that this market has a considerable lack of supply. Live bait shrimp sales account for a majority of retail- ers' yearly revenue. This is due to the ability of live bait shrimp to dri\e the sales of other items in bait stores (e.g.. equipment and food) if it is available. Thus, it is easy to see that the market is not realizing its full potential without a consistently available and re- liable product. The acceptance of a farm-raised live bait shrimp, with its flexible product characteristics, would allow it to meet the major demands of the Texas market. 2-4% 5-9% 10-24% 25-49% Live Bait Mortality During Shipment Figure 4. Expected live bait slirimp mortality during shipment. ACKNOWLf:DGMENTS This research was supported hy Grant NA67FD0036 from the National Oceanic and Atmospheric Administration (NOAA). LITERATURE CITED Burkott. B. J. 1994. Management strategies for production of Penaeus setifenis as bait shrimp m outdoor ponds. Corpus Christi. TX: Depart- ment of Biology. Texas A&M University. Inglis. A. & E. Chinn. 1966. The bait shrimp industry of the Gulf of Mexico. Contribution No. 211, Fishery Leaflet No. 582. Galveston. TX: Bureau of Commercial Fisheries. Biological Laboratory. 10 pp. McKee. D. A. 1986. An investigation of the live bait shrimp industry of Texas and the culture and economic potentials for rearing two penaeid species as supplements to that industry. Ph.D. thesis. College Station. TX: Texas A&M University. McKee. D. A.. W. L. Griffm & A. L. Lawrence. 1989. Stocking strategies and an investment analysis for producing Penaeus setifenis as live bait shrimp on the Texas Gulf Coast. J. World Aqimcull. Sac. 20(4):72-80. Nance, J. M. 1993. Analysis of the white shrimp closures in the Gulf of Mexico. NOAA Technical Memorandum. NMFS-SEFSC-321. 12 pp. Parker, J. C. & H. W. Holcomb. 1973. Growth and production of brown and white shrimp from experimental ponds in Brazoria and Orange counties, Texas. Proe. World Maricidt. Soc. 4:215-234. Quick, J. A. & J. A. Morris. 1976. Extensive culture of penaeid siirnnp m a simulated dredge material disposal area. Proe. World Mariciilt. Soc. 7:3(15-326. Rcssberg, K. S. & K. Strawn. 1980. Comparative growth and survival of the brown shrimp cultured with Florida pompano, black drum and striped mullet. Proe. World Marieult. Soc. 1 1:219-225. Samocha, T. M., B. J. Burkott. A. L. Lawrence. Y. S. Juan. E. R. Jones & D. A. McKee. 1998. Management strategies for production of the At- lantic white shrimp Penaeus seliferus as bait shrimp in outdoor ponds. J. World Aquacuh. Soc. 29(2);21 1-220. Sandifer. P. A.. J. S. Hopkins. A. D. Stokes & C. L. Browdy. 1993. Preliminary comparisons of the native Penaeus setifenis and Pacific Penaeus vannamei white shrimp for pond culture in South Carolina. USA./ World Aquacidt. Soc. 24(3):295-303. Southern Fisheries Science Center 1992. Status of the southeastern United States shrimp fishery for 1991. NOAA Technical Memorandum NMFS-SEFSC-306. 75 pp. SPSS. 1997. Statistical Package for the Social Sciences. Graduate Pack. Advanced Class Version 7.0 for Windows 95. Texas Parks and Wildlife Department. 1993. Commercial fishing guide. 1995-1996. Rules and Regulations 1995-1996. Austin. TX: Texas Parks and Wildlife Department. Warren. J. P. 1980. The Texas bay shrimp industry; a description and management model. Ph.D. dissertation. College Station. TX: Texas A&M University. Department of Agricultural Economics. 130 pp. Jowmil ofSlicHJhb Research. Vol. 2U. No. 1. }.W-yi\. 20(11. THE REPRODUCTIVE CYCLE OF GOLDEN KING CRAB UTHODES AEQUISPINUS (ANOMURA: LITHODIDAE) A. J. PAUL* AND J. M. PAUL University of AUisl^a Institute oj Marine Science, Seward Marine Center Laboratory. P.O. Bo.x 730, Seward, Alaska W664 ABSTRACT Female golden king crahs Luhodcs acqiiispmiis were held in captivily to determine the duration of their reproductive cycle. The lime between the first and last egg hatching was 34 days (SD = 16. n = 147) on average. Females molted about 192 days (SD = 72, » = III) after the last egg hatched. Eggs were extruded 2 days (SD = 2, h = 61) after molting. Egg clutches were incubated for an average of 362 days (SD = 78. « = 59). The average amount of time that passed between the production of successive egg clutches was .590 days, or 3..570- degree days of water temperature. KEY WORDS: Litluidcs aequispimis. reproduction, king crab INTRODUCTION The golden king crab. Litlwdes aequispinus Benedict, supports an important fishery in Alaska. This North Pacific species typically lives in deep water, on untrawlable bottoms. Because of their remote deep habitat, and the fact that only males are retained by the fishery, the reproductive cycle of females had not been de- scribed in detail prior to this study. Female golden king crabs carry up to 27,000 large eggs (Jewett et al. 1985) that can hatch in any month of the year (Adams and Paul 1999). Female golden king crabs are found in all reproductive and molting stages throughout the year (Paul and Paul 2()()0a). so the duration between clutches was not apparent frotn gross examination of the egg mass. This laboratory study examined the reproductive cycle of the golden king crab to identify how long it took females to go from the extrusion of one egg clutch to another. This study was accompa- nied by parallel studies on growth of both sexes (Paul and Paul 2000a) and sizes at maturity of tnales (Paul and Patil 2000b). MATERIALS AND METHODS Crabs needed for the study were captured with pots fished at 108 to 152 m depths on the western side of Prince William Sound. Alaska. After capture the specimens were transported to the Seward Marine Center Laboratoi-y by floatplane. No tiiortalities occurred during the transport process. Collections of tnultiparous females were made November 1 1-14.1996; May 1. 1997 and Oc- tober 10-20. 1998. Observation of captives continued until Octo- ber I, 2000. The carapace length (CL) of all crabs was measured to the nearest millimeter for the growth study (Paul and Paul 2000a), and they were tagged with a numbered plastic disk attached to the leg with a plastic cable tie. All females in this study had egg clutches when captured and they had carapace lengths ranging from 104 to 150 mm (X = 132. SD = 7 mm). Each female was held in a separate 800 to 1000 L tank. The seawater in the tanks was ex- changed >100% per hour to maintain the ambient temperature of incoming water from 75 m depth in Resurrection Bay. The water temperature in the tanks was measured daily. The duration of each of the 4 reproductive phases was described as both, number of days and degree-days of water tetriperature. The degree-day intermolt period is considered the sum of the average daily temperature during the intermolt period. Thus for example, if a stage interval *Corresponding author. E-mail: ffajp@uaf.edu was 30 days and the temperature every day was 6°C, the event spanned 180-degree days (30 x 6). The degree-day duration data was calculated because the reproductive phases were nonsynchro- nous with sotne females extruding new clutches in the warm part of the year and others during the cold season. Temperature in the tanks followed the fjord's seasonal cycles ranging from 3.7 to 9.7°C (Fig. 1). Crabs were fed every other day to excess with a repeating cycle of the following foods: whole Pacific herring Clu- pea pallasi Valenciennes 1847. fillet of coho salmon Oncorliyn- cluis kisutch (Walbaum 1792), giant Pacific Octopus Octopus dof- Icini (Wulker 1910), whole squid (species unknown), and whole Alaska northern shrimp Pandahts eons Makarov 1995. STAGE I (n = 147) was the length of time between the hatch- ing of the first and last egg. STAGE 2 (n = 111) started after the last egg hatched and it ended when a female molted. STAGE 3 (n = 61) was the length of time between the day a female tnolted and the day she extruded a new egg clutch. The duration of clutch incubation (STAGE 4) was the length of liine between the extru- sion of a new egg clutch and the hatching of all those eggs (n = 59). Thus, STAGE I and 4 overlap. The duration of the different phases of the full reproductive cycle was not quantified for every female since captives had different collection dates. Tanks were examined daily for the presence of larvae and when the first larvae were seen, females were examined every day until all eggs had hatched. This procedure was used to determine the duration of the larval hatching phase. STAGE 1. Thereafter, fe- males were isolated from males until they molted. Females inust molt prior to mating. These observations estimated the time be- tween egg hatching and molting, STAGE 2. After a female molted a selected hardshell male 2114 mm CL was put into the tank with her. Males of this size are capable of fertilizing females (Paul and Paul 2000b). The time between the molt and the occurrence of the new egg clutch was termed STAGE 3. After ovulation, females were isolated and held until zygotes developed to the 64-cell stage. Groups of at least 100 eggs from each pleopod were then randomly selected and examined under a microscope for cell division to determine percent viability. Fe- males were held from ovulation until hatching to determine the incubation period. STAGE 4. RESULTS STAGE 1. the period between the first and last larvae hatching, averaged 34 days (SD = 16. o = 147. range 8-85) or 202 degree days (SD = 88, range 19-474). On average 192 days (SD = 72, 369 370 Paul and Paul 11 10 - o 3 2 ' liJ 5 - 1996 1997 1998 1999 ' 2000 YEAR Figure 1. I.aboratory seavvater water temperatures during the study. /; = 111. range 5-464) or 1.084 degree days (SD = 428. range ,^6-2.762) passed after the last larvae hatched until the females molted which completed STAGE 2. During STAGE 3 females typically extruded new eggs 2 days (SD = 2. /; = 61. range 1-12) or 15 degree days (SD = 12, range 4-85) after molting. All clutches resulting from these laboratory matings had >80 of their eggs initiating di\ ision. STAGE 4. egg clutch brooding, lasted an average of 362 days (SD = 78. ii = 59. range 40-569) or 2.269 degree days (SD = 570. range 1 14-2.754). The total time passing between egg clutches averaged 590 days or 3.570 degree days. The best fitting regressions relating the CL (mm) of all females that completed STAGES 1^ and their reproductive cycle duration (degree days) showed no apparent relationship between these vari- ables (r- < 0.12). Pandithodes caintschatictis (Tilesius) the female reproductive cycle is synchronous, lasting about one year (Paul & Paul 1990. Paul & Paul 1997). In blue king crab {Pandithodes platypus Brandt) primiparous females may produce egg clutches annually and every two years for larger multiparous females. Like red king crabs, the eggs of primiparous and multiparous blue king crab hatch during the spring plankton bloom (Jensen & Armstrong 1989). Red king crabs hatch their eggs in spring so the larvae can feed on the plankton bloom (Paul et al. 1990). then all ripe females molt and breed soon after hatching is done. With golden king crab, hatching does not need to occur exclusively during the spring plankton bloom because their lecithotropic larvae do not feed and they can tolerate both summer and winter temperatures (Shirley & Zhou 1997. Adams & Paul 1999. Paul & Paul 1999). One striking difference between the reproductive cycle of golden and red king crabs is the amount of time between egg hatching and the female molt. In red king crabs captive females usually molt within two weeks of egg hatching (authors" unpublished observations) vs. about six months for golden females. Cuirently it is not known why these differences exist. Golden king crab females may need more time to produce the large yolk rich eggs that allow their larvae to forgo feeding. However, that idea is speculation at this time. Female Prince William Sound golden king crab first mature around 120 mm CL and typically grow to about 150 mm CL in 5 molts (Paul & Paul 2000a). If they produced an egg clutch every molt, and their reproductive cycles lasted 590 days, females would have to survive at least eight years to produce 5 clutches. The sea water for holding the crabs came from a 75 m depth where temperatures during August to December are 1 to 4 'C warmer than at approximately 150 m where the specimens were captured, versus 1 to 2'C cooler from January to April (Xiong & Royer 1984). Generally warmer conditions decrease intermolt du- ration in crustaceans unless there is thermally induced stress and cooler ones lengthen it. Only in situ studies, or laboratory studies that mimic site specific temperatures, can determine if our degree day estimates for the reproductive cycle duration are appropriate for female golden king crab living in different thermal environ- ments. DISCUSSION In early studies the lime of spawning of golden king crab was described both as seasonal asynchronous and synchronous (Sloan 1985). Some of this reported variability may have been caused by an imperfect understanding of the reproductive cycle. During STAGE 2 (X = 192 days) females carry decaying empty egg capsules on their setae (Sloan 1985) and this condition made it difficult for fishery observers to classify the reproductive status of females. For Prince William Sound golden king crab, molting and hatch- ing events can occur in any month in captive females (Paul & Paul 1999, Paul & Paul 2000a). Thus, their reproductive cycle is mark- edlv different from the aenus PaialitlinJcs. In red kins crab ACKNOWLEDGMENTS This work is a result of research funded by the Alaska Sea Grant College Program grant NA46RG0104 project R/06-36, the University of Alaska, and cooperative agreement NA97FN0129 from the National Oceanic and Atmospheric Administration (NOAA). The views expressed by the authors do not necessarily reflect the views of NOAA or its sub-agencies. L. Clayton, C. Adams and P. Shoemaker helped with the laboratory work. Facili- ties were provided by the University of Alaska Institute of Marine Science's Seward Marine Center Laboratory. This is contribution number 2667 from the Institute of Marine Science. C. Trowbridge and W. Bechtol and the Alaska Department of Fish and Game supplied the experimental animals LITERATURE CITED Adams. C. F. & A. J. Paul. 1999. Phototaxis and geota.xis of light-adapted golden king crab zoeae. Lithoiles aequispiniis (Anoniura: Litliodidae) in the laboratory. J. Cinst. Biol. 19:106-1 10. Jensen. G. & D. Armstrong. 1989. Biennial reproductive cycle of blue king crab, Pcinilithodes platypus, at the Pribilof Islands. Alaska and com- parison to a cogener. P. camischatica. Ciiii. J. Fish. Aquat. Sci 46:932- 940. Jewett. S. C. N. A. Sloan & D. .A. Somerton. 1985. Size at se.xual matunly: and fecundity of the fjord-dwelling golden king crab Lilhodes aequispina Benedict from northern Bntish Columbia. J. Crust. Biol 5:377-385. LminnEs aequispinus Reproduction 371 Paul. A. J. & J. M. Paul. 1997. Breeding success of large male red king crab Pamlitlwdes camlscluiriciis with multiparous mates. / Shellfish Res. 16:379-381. Paul. A. J. & J. M. Paul. 1999. Development of golden king crab Lithoiles uequispiims (Anomura: Lithodidae) larvae reared at different tempera- tures. J. Crust. Biol. 19:42-;5. Paul. A. J. & J. M. Paul. 2000a. Changes in chela heights and carapace lengths in male and female golden king crab, Liihailes aequispinus, alter molting in the laboratorv. .Alaska Fish. Res. Bull. 6:8-15. Paul. A. J. & J. M. Paul. 2000b. Size of maturity in male golden king crab. Lilhades aequispinus. (Anomura: Lithodidae). / Cms!. Biol. 21(2): (in press). Paul. A. J.. J. M. Paul & K. O. Coyle. 1990. Growth of stage 1 king crab larvae oi Paraliihodes camtsclialica (Tilesius) (Decapoda: Lithodidae) in natural communities. J. Crust. Biol. 10:175-183. Paul, J. M. & A. J. Paul. 1990. Breeding success of sublegal size male red king crab Paraliihodes camischalica (Tilesius 1815) (Decapoda. Lithodidae). J. Shellfish Res. 9:29-32. Shirley. T. C. & S. Zhou. 1997. Lecithotrophic development of the golden king crab Lithodes aequispinus (Anomura: Lithodidae). J. Crust. Biol. 17:207-216. Sloan, N. A. 1985. Life history characteristics of fjord-dwelling golden king crab Lithodes aequispina. Mar. Ecol. Proi;. Ser. 22:219-228. Xiong. Q. & T. C. Royer. 1984. Coastal temperature and salinity in the northern Gulf of Alaska. 1970-1983. / Geophys. Res. 89:8061-8068. .loiinuil <-/ Shellfish Research. Vol. 20. No. 1. 373-376. 2001. INTERMOLT DURATIONS OF CAPTIVE JUVENILE AND ADOLESCENT MALE TANNER CRABS CHIONOECETES BAIRDI A. J. PAUL* AND J. M. PAUL Unlversiry of Alaska, Instimie of Marine Science. Seward Marine Center Lulniratory. P.O. Box 730. Seward. Alaska 99664 ABSTRACT This project examined Ilie effect of temperature on internioll durations of juxenile and adolescent male Chionoecetes bairdi Rathbun with carapace widths (CW) of 11 to 105 mm. Juveniles were identified by the absence of spermatophores and adolescents by their presence. The relationship between male maximum CW and interniolt duration exhibited a linear relationship for juveniles described by the equation: Intermolt period (degree-days) = 7.9 (CW mm) + 272; r" = 0.53. As males began to produce spermatophores at about 55 mm CW intermolt durations increased and the analogous equation was Intermolt period (degree-days) = 7.6 (CW mm) + 1,215; r" = 0.14. A 55 inin CW spermatophore bearing male took about 2.5 times longer to molt than a similar sized juvenile. KEY WORDS: Chionoeceres. molting, temperature, tanner crab INTRODUCTION In the tiorthern Gulf of Alaska and the southeastern Bering Sea the Tanner crab. Chionoecetes bairdi. is a ubiquitous benthie in- \ertebiate that is harvested cotniiiercially. The change in carapace size following molting has been described for Gulf of Alaska C. bairdi (Donaldson et al. 1981. Paul & Paul 1996) but the inter- rnolt duration has not. Inf'ortnation on intertnolt duration is critical to understanding recruitment to the population because there is no accepted method to age Tanner crabs (Rosenkranz et al. 1998). Crustacean growth rates are controlled by a nutnber of factors with temperature being one of the most obvious (Fisher 1999). Benthie temperatures in the Gulf of Alaska experience long terni warming and cooling cycles of about 18 to 20 years (Royer 1989). In ad- dition to seasonal changes in temperature El Nii'io events, like that of 1998 (Fig. I), are additional sources of thermal perturbations. The inter-annual differences in the mean monthly bottom tempera- ture where the specimens for this study were captured were on order of 2 to 3''C (Fig. 1 ). Currently there is no information on the effect warm and cold periods have on the duration of molting rates of C. ludrdi. It is logical to assume that thermal conditions will be a tnajor factor regulating the length of the intermolt period. The objective of this study was to determine the intermolt period of juvenile and adolescent male Tanner crabs in degree-days. This study was funded to improve our understanding of the growth process and the relative age of males recruiting to adulthood. In earlier observations of size of maturity and growth in male Tanner crabs (Adairis & Paul. 1983: Paul & Paul. 1996) it appeared that immature tnales took less time to molt than matiuc males of simi- lar sizes but this was not quantifiable with the methods used. We speculated that the molting schedules of mature and immature males would be markedly dissimilar, so we ascertained the matu- rity status of test specitiiens to examine this theory. MATERIALS AND METHODS Male Tanner crabs were collected at the head of Resurrection Bay near Seward Alaska at 35 to 80 m depth using a 2 m otter trawl with 6 mm cod end mesh at a variety of times during the years 1996 to 2000. There were 67 males with carapace widths (CW) 9 to 82 mm that were returned to the nearby laboratory for *Corresponding author. E-mail; ffajpCSuaf.edu the study. The seawater for the Seward Laboratory comes from 75-111 depth in a fjord and its temperature during the study was 3 to lO'C (Fig. I ). The temperature of the incommg water changes w ith season with marked inter-annual variations in monthly values (Fig. 1), Each day the seawater temperature in the tanks was re- corded. Salinity ranged from 31 to 33 ppt. All test animals were held in separate numbered tanks to prevent cannibalism. Males =S45 mm were held in individual 20-L tanks. Males 3=46 mm were held in 100-L tanks and the water exchange rate in ail tanks was 100% per h. Captives were fed to excess every Monday (whole northern shrimp Pandalus eous Makarov 1995), Wednesday (live intertidal mussel Mytdus trossiiliis Gould 1850) and Friday (Coho salmon fillet Oncorhynchus kisiitcli Walbaum 1792). Whenever a male tiiolted for the first time in captivity the date of molting was recorded. After two weeks had passed, and the carapace had hard- ened, its maximutii CW was measured to the nearest 0.1 mm. The new post molt CW of the smallest captive was 1 1 mm and the largest 105 mm CW. A three millimeter coded plastic disk was glued to the carapace of test specimens. The date of each crab's second molt was recorded to calculate the intermolt duration. The new CW was tneasured two weeks later and the vas deferens was retiioved to determine if spermatophores were present in wet mounts using lOOx magnification. The intermolt duration was described in degree-days. Degree- days were calculated by sumtiiing the daily seawater temperatures that occurred during the intermolt period. For example if an event took ten days, and each day the temperature measurement was lO'-'C. then the process would have taken 100 degree days. RESULTS The relationship between male CW and mtermolt duration ex- hibited a linear relationship for juveniles that was described by the equation: Intermolt period (degree-days) = 7.9 (CW mm) -i- 272: r- = 0.53. P<0,OOOI,/i = 53 (Fig, 2A). The relationship between CW (mm) and intertnolt duration for males producing spermato- phores was described by the equation: Intermolt period (degree- days) = 7,6 (CW mm) -t- 1,215: r" = 0.14. P = 0.1832. n = 14 (Fig. 2B). The intermolt duration of mature males was less depen- dent on temperature than that of juveniles. The intermolt durations of males that were producing spermatophores (Fig. 2B. C) were longer than would be expected if the linear tiiodels of intermolt 373 374 Paul and Paul o a: D (- < 50% of Gulf of Alaska males have spermatophores (Paul 1992). A spermatophore bearing male 63 mm CW would require about 7.414 degree-days to grow to 107 mm CW (Table 2). a size close to that of the largest individual that molted in this study. The largest male in the study was 105 mm CW and it molted after 2.001degree-days. The smallest male in the study was 1 1 mm CW and ii molted after 477 degree-days. DISCUSSION It is generally accepted that the amount of time that passes between molts increases with crab carapace size, and this proved to be true for C. hainli. We are not sure why small spermatophore bearing males had longer intermolt periods than similar sized irn- g> 3000 T3 ^ 2500 Q Q 2000 LU O a: UJ 1500 1000 3000 Y = 7.6X + 121S r' = 0.14 n = 14 2500 H 2000 1500 H 1000 500 B \ [ I \ I \ I 1 I 55 60 65 70 75 80 85 90 95 100 105 900 800 h 700 600 h 500 400 300 60 J— 3000 2500 2000 1500 1000 ,.^<^. ^'^' ,*» 3000 2500 2000 1500 - 1000 500 I 1 1 I I I I I I 10 20 30 40 50 60 70 80 90 100 110 CARAPACE WIDTH (mm) Figure 2. Intermult durations in degrees days for juvenile male Chionocetes bairdi that did not have spermatophores in their vas deferens (Panel A, •), spcrnialophdre bearing males (Panel B, ■), and comparisons of both types (Panel C). Chionoecetes molting 375 TABLE 1. The estimated interniolt durations in deRrec da>s for male juvenile C'hiiimieceli's hairdi relative Id carapace width (t'W I. TABLE 2. The estimated interniolt durations in decree days for spermatophore producing; male Chionmceles hairdi relative to carapace width ICW). Initial C\V New C\V Intermold duration i! Degree Initial C\V New CVV Interniolt duration i. Degree ininil' (mm)" (d ■yree davs)'' days (mmf (mm)" (degree days)'' days 3.0 7.5 276 276 62.8 75.7 1692 1692 7.5 12.6 313 598 75.7 90.4 1790 3482 12.6 18.5 358 947 90.4 107.2 1902 53S4 18.5 25.2 32.8 41.6 51.5 408 464 529 603 1 355 1819 2348 2951 107.2 126,2 2030 7414 25.2 32.8 41.6 ■'Growth equation '"From equations from Paul a n Figure 2B id P uil 1996. 51..'^ 62.S 688 3639 ■'Growth equation from Paul and Paul 1996. ''From equations in Figure 2A. mature males. Perhaps spermatophore production is only one ele- ment of a complex maturation process w ith moiphological, bio- chemical and physiological changes that in\olves longer interniolt periods. Currently, no other information on the intermolt durations of juvenile Tanner crabs, or closely related high latitude species, is available to compare to our results. The smallest Tanner crab we have seen in trawl samples is 3 mm CW. Our intermolt duration projections (Tables I, 2) imply that captive Tanner crabs would grow from 3 to 107 mm CW in 11.053 degree-days. In our labo- ratory the average annual seawater temperature is about 6°C, and a male held at 6°C would require approximately 1842 days or 5 years to grow from 3 to 107 mm CW. However, we do not know if laboratory conditions, for example tank size, modified molting schedules. In the laboratory a surplus of food was present while in nature food availability may limit growth rates. It is also possible that the laboratory diet was nutritionally incomplete. This study needs replication with in situ tagging studies to determine if our laboratory molting schedules are applicable to natural conditions. These observations on intermolt durations demonstrate that size and maturity status are important factors to consider when fore- casting molting schedules in Tanner crabs <105 mm CW. Tanner crab males are not harvested until they are 3= 140 mm CW and they can grow to =170 mm CW. Thus, molting rate studies need to be done with larger specimens than we used in this observation. Ad- ditional experimentation with large crabs would be especially im- portant if there are maturation processes other than spermatophore production that influence molting schedules. Juvenile and adoles- cent males typically have relatively small claws while fully mature males have large claws (Stevens et al.l993). All the males in this study were small claw morphotypes using the criteria of Stevens et al. ( 1993). the ratio of chela heighl/CW <0. 17. and most intermolt durations were one year or less. Males continue to molt after reaching maturity (Paul and Paul 1995) presumably because big large-claw males win competitions for mates, and compromised carapaces need replacement. The large claw characteristic devel- ops when males reach 100 to 130 mm CW (Stevens et al. 1993). We studied the consequence of spermatophore presence on molt- ing rates, but not the change to the large claw morphotype, or the attainment of near maximum CW. In another investigation Tanner crabs with CW 5= 1 10 mm had to be held for over two years before they molted (Paul and Paul 1995). None of them were soft-shelled when they were captured, so their intermolt period was longer than two years. In one in sitn study 41% of tagged male C. bairdi > 1 10 mm CW were recaptured after two years and another 1% after three years (Donaldson 1980). These observations (Donaldson 1980. Paul and Paul 1995) suggest that intermolt durations in- crease after males assume the large claw morphotype and approach maximum size. Further growth rate studies with males 105 to 170 mm CW are needed to describe the intermolt durations of these large individuals. ACKNOWLEDGMENTS The study was funded, in part, by cooperative agreement NA97FN0129 from the National Oceanic and Atmospheric Ad- ministration to the Alaska Department of Fish and Game, the Alaska Sea Grant Program grant NA90AADSG066 project R/06- 32 and University of Alaska state funds. The views expressed are the authors and do not necessarily reflect the views of the funding agencies. Facilities were provided by the University of Alaska Institute of Marine Science's Seward Marine Center Laboratory. This is contribution number 2668 from the Institute of Marine Science, University of .Alaska. LITERATURE CITED Adams. A. E. & A. J. Paul. 1983. Male parent si/e. sperm storage and egg production in the crab Chionoecetes hairdi (Decapoda. Majidae). Inl. J. hivert. Repro. 6:181-187. Donaldson, W. E. 1980. Movements of tagged Tanner crabs. Chioiioeceies hairdi liberated off Kodiak Island, Alaska. Alaska Department of Fish and Game. Informational Leaflet Numberl88. 60 pp. Donaldson. W. E.. R. T. Cooney & J. R. Hilsinger. 1981. Growth, age and size at maturity of Tanner crab, Chionoecetes hairdi M. J. Rathbun, in the northern Gulf of Alaska (Decapoda. Brachyura). Cnistnceinia 40; 286-302. Fisher. M. R. 1999. Effect ot temperature and salinity on sl/e at matunty of female blue crabs. Trans. Am. Fish. Soc. 128:499-506. Paul. A. J. 1992. A review of size at maturity in male Tanner {Chionoecetes hairdi) and king [Paralithodes camtschatica) crabs and the methods used to determine maturity. Am. Zool. 32:534-540. Paul, A. J. & J. M. Paul. 1995. Molting of functionally mature male 376 Paul and Paul Chioiwecetes bairdi Ratlibun (Decapoda: Majidae) and changes in carapace and chela measurements. J. Crust. Bio. 15:686-692. Paul, A. J. & J. M. PauL 1996. Male Tanner crab carapace widths at previous intermolt estimated from laboratory growth, Kacheniak Bay, Alaska. Ak. Fish. Res. Bull. 3:132-135. Rosenkranz, G. E.. A. E. Tyler, G, H. Kruse & H. J. Niebauer. 1998. Relationship between wind and year class strength of Tanner crabs in the southeastern Bering Sea. Ak. Fish. Res. Bull. 5:18-24. Royer. T. C. 1989. Upper ocean temperature variability in the northeast Pacific Ocean: is it an indicator of global warming? J. Geophys. Res. 94:18175-18183. Stevens, B. G., W. E. Donaldson, J. A. Hagga & J. E. Munk. 1993. Morphometry and maturity of paired Tanner crabs, Chionoeceles huirdi. from shallow- and deepwater environments. Can. J. Fish. Aquat. Sci. .50: L504- L5 1 6. Jounnil (if Shellfish Research. Vol. 2(1. No. I. 377-3K2. 20(11. EFFECT OF ARTIFICIAL DIETS CONTAINING CAROTENOID-RICH MICROALGAE ON GONAD GROWTH AND COLOR IN THE SEA URCHIN PSAMMECHINUS MILIARIS (GMELIN) GILLIAN McLaughlin' and maeve s. kelly- * ^Department ofZooloiiy. University of Aberdeen. Aberdeen. AB24 2TZ. Scotland. United Kinf^dom: -Scottish Association for Marine Science. PO Box 3. Oban. Argyll PA34 4 AD. Scotland. United Kingdom ABSTRACT Gonadal growth and color were examined in tlie ecliinoid Pscwwwctmms iniliaris when fed diets containing either micro- or niacroalgal supplements. Urchins receiving diets containing the microalgae Ptuteodacryimn Irkormttinn and Tahitian Isoclirxsis sp. showed significantly greater gonad growth at the end of the 12-wk experimental period as compared to urchins ted the artificial diet with no added algae. An improvement in gonad color, compared to the control, was observed for both treatments receiving microalgae; whereas, those fed the macroalgae Uimiiwiui sacchcirinci diet showed no significant improvement in color. The P. tricnnmt:im diet improved gonad color more rapidly than the other algal diets. These results show that cultured microalgae. incor- porated into an artificial diet, have a positive effect on the gonad color of P. iiiiliaris. with promising implications for commercial echinoculture. KEY WORDS: se; a urchin. Psiimnwchinas miliaris. artificial diet, carotenoid pigments, gonad color, echinoculture INTRODUCTION The gonads of seu urchins (Echinodermata. Echinoidea) are a highly sought-after product in many parts of the world, particularly in the Far East and Mediteiranean Europe (Hagen 19961. The global demand for sea urchins has risen sharply over the past two decades, and as a result, most commercial sea urchin fisheries are now considered to be fully or overexploited (Keesing and Hall 1998). Attention is. therefore, turning toward the development of a commercially viable echinoculture industry. Artificial diets have been shown to enhance gonad growth ef- fectively in several commercially important sea urchin species (de Jong-Westman el al. 1995, Lawrence et al. 1997, Barker et al. 1998. Robinson & Colbourne 1998. Fernandez & Pergent 1998). However, in addition to an acceptable quantity of gonad per indi- vidual, the marketplace also requires the gonad to be a bright orange color. Addressing the problem of variability in gimad color is. therefore, an important part of artificial diet design. Carotenoid pigments, the source of red. orange, and yellow coloration in plants and animals are synthesized only by plants and micro-organisms. Animals can. however, alter these molecules by oxidation. The coloration of urchin gonads is a result of selective accutTiulation and chemical modification of carotenoid pigments obtained from their diet (Goodwin 1984). Griffiths & Penot (1976) found the primary carotenoids in the ovaries of StrongyliKentrotus droehachieims were echinenone (79 - 85%), P-carotene, zeaxan- thin, isocryptoxanthin, and small amounts of a fucoxanthin isomer. Echinenone was subsequently shown to be the primary carotenoid in the gonads of I 1 species of urchin from Japanese waters (Tsushima & Matsuno 1990) and of the New Zealand echinoid Evechinus ehlorotieus (Goebel & Barker 1998). Hallenstvet et al.. (1978) described the total purified carotenoid content of whole P. miliaris as fucoxanthinol 68'7f, echinenone 10%, lutein 8%, P-P carotene 10% and p-e carotene 5%^, but did not report the relative amount of gonad tissue present or the color of the gonads of the urchins they analyzed. Griffiths and Perrot (1976) found that gut wall of .S. (Irnehachieiisis contained *Corresponding author. 7-24 times the carotenoid content of the ovaries on a dry weight basis, but that there was little evidence to suggest that fucoxanthin. the main carotenoid of the gut wall, was used in the production ot echinenone, the main pigment of the ovary. They presented evi- dence that echinenone was a conversion product of ingested P-P carotene (via isocryptoxanthin) and that this conversion occurred in the ovary and not in the gut wall. Tsushima et al. (1993) sub- sequently showed p-carotene to be the precursor for echinenone, via p-isocryptoxanthin. in Pseiidoeentrotus depressus. However, Goebel and Barker (1998) have shown that feeding processed car- rot pulp or synthetic p-carotene supplements to Evechinns ehlo- rotieus produced no significant effect on the gonad color. Gonad growth was shown to be significantly enhanced in Psammechinns miliaris when fed on commercially prepared salmon diets, either directly or in polycultured systems (Kelly et al. 1998b; Cook et al. 1998). Although the salmon feed contained the carotenoid aslaxanthin. these urchins produced gonads of a duller orange than the bright and most desirable shades found occasion- ally in local wild populations of P. miliaris. Similarly, Havardsson and Imsland (1999) found no evidence to suggest 5. droebachien- sis could utilize such higher oxidation state pigments as astaxan- thin as metabolic precursors for echinenone. Wild P. miliaris are often found feeding on the macroalgae Lumiiniria saecluirina (Kelly 2000) and a range of encrusting in- vertebrate species (Lawrence 1975). The addition of macroalgae as a source of carotenoids in urchin artificial diets may enhance go- nad color; however, inacroalgae typically exhibit seasonal bio- chemical variations (Black 1950) that could result in variability of the final product. However, microalgae that contain carotenoids can be cultured under controlled conditions and on a large scale using existing technology. The aim of this investigation, therefore, was to determine if incorporating carotenoid-rich microalgae into artificial diets for P. miliaris had a positive effect on gonad color and growth. MATERIALS AND METHODS Sea Urchin Collection and Maintenance Psammechinns miliaris were collected by SCUBA from a depth of 5 m in Loch Creran, Scotland (56°32'20"N; 5°17'00"W). 377 378 McLaughlin and Kelly Twenty-two urchins (horizontal test diameter 20-30 mm), selected at random, were placed into each of 12 10:1 aquaria. Each aquarium had an independent supply of 250 p,m filtered seawater at ambient temperature and salinity. The experiment was con- ducted over twelve weeks from May to August 1999. when the seawater temperature ranged from 9.9-15.3°C. The photoperiod was maintained at a constant 16-h light: 8-h dark cycle throughout the experiment. The urchins were left to acclimatize to the aquarium conditions for a period of sixteen days before the start of the experiment. during which time they were starved to standardize their nutri- tional status (Vadas 1977). The urchins were then fed one of four diets at a constant rate of approximately 3% of their mean body weight per day. Any waste food and feces were removed every second day by carefully siphoning around the urchins. Diet Preparation The diets were allocated to aquaria using a randomized block design, with three replicates of each treatment. The diets were made from commercially available raw materials (Table 1 ) with either additional micro, macro, or no additional algae (as a con- trol). Agar and gelatine were chosen as suitable binding agents (Caltagirone et al. 1992). and the food pellets retained a firm consistency in seawater when subjected to urchin grazing. The microalgae were obtained from the Culture Collection of Algae and Protozoa (CCAP). Dunstaffnage Marine Lab., Oban, UK. The species used were the diatom Pluieodacrylum tricorniitnm Bohlin (CCAP 19/18) and the flagellate Tahitian Ismliiysis sp. (Parke) (T-ISO, CCAP 927/14). The macroalgae Lwnimiria saccharina (L) Lamour was collected locally. The microalgal strains were selected on the basis of their pig- ment profiles and their growth characteristics. The diatom P. tri- conmtum, belonging to the Bacillariophyceae. has as its major carotenoid pigments the xanthophylls fucoxanthin and diadinox- anthin and small amounts «1% of total carotenoid content) of P-^ carotene and cis-p-P carotene (Wright et al. 1991). T-ISO. a golden-brown flagellate (Prymnesiophyceae) similariy contains the carotenoids fucoxanthin. diadinoxanthin, and smaller amounts of P-P carotene (Brown et al. 1993). L. saccharina has fucoxanthin as its primary xanthophyll and violaxanthin as the second major xanthophyll. P-p carotene is the only carotene present, zeaxanthin. neoxanthin. and fucoxanthinol are found at the ().()l-0.02'7r of total carotenoid level (Haugan & Liaaenjensen 1994). The microalgal cultures were grown in a semicontinuous batch cultivation system, maintained at 20°C (± 2°C) and constantly illuminated and aerated. The growth medium was autoclaved sea- T.-VBLE 1. Diet composition. Ingredient Amount Percentage Wet Weight Com oil o.i:g Dried skimmed milk 0.63 g Casein 0.57 g Gelatine 0.37 g Agar 0.53 g Filtered seawater 25mL/20mL Algae 4.5-5g (0 g for control diet) 7. 1 8'* 4.64% 7.95% 6.72% 1.57% 70.84% water enriched with I mL L"' Walnes medium and 0.1 mL L~ vitamin solution. In addition. 0.3 mL L"' of sodium metasilicate (Na,SiO, 0.5H2O) was added to the P. triconnitum cultures. Be- fore harvest, the cell densities were calculated from haemocytom- eter counts; cells were harvested when the densities were approxi- mately 1 X 10^ cells ml I"'. The algae were separated from their culture media using centrifugation (8.000 rpm. 10 min). Approxi- mately 6-9 L of T-ISO and 3-f L of P. triconiiaiim cultures were required to yield a soft algal pellet of 4-6 g wet weight. The L. saccharina was rinsed in filtered seawater and finely chopped be- fore incorporation into the diets. The dry ingredients and the corn oil were weighed into a beaker, the appropriate amount of water added, and the mixture thoroughly stined. The mixture was then heated in the microwave for 15 sec, stiired briefly, and then heated for a further 15 sec. It was then allowed to cool to approximately 35°C before adding the algae. Once set, the prepared diets were cut into l-cm" cubes before distribution to the urchins at a rate of one cube individuaP' day"'. Data Collection and Statistical Analysis Five urchins from each aquarium were sacrificed at the start of the experimental period (sample day 1, 20 May. 1999) and there- after at approximateiN monthly intervals. Whole wet weight (to the nearest 0.001 g) and horizontal test diameter (to the nearest 0.05 mm) were recorded before dissection. Upon dissection, the color of the gonads was assessed immediately by matching it to the closest color in the Pantone® collection of standards (Pantone 1995. Cook 1999). The gonad color was always assessed by the same observer in good natural daylight. Unmarketable shades were dark brown, gray, or black (Pantone colors 4-6, 1545, and 161); acceptable colors were pale yellow and orange (Pantone 155, 156, 713), and the most desirable shades were bright orange (Pantone 123. 136, 149). The wet weight of the gonad, emptied gut, and eviscerated test were also recorded. Whenever possible, the sex of the individual was determined by examination of extruded gametes, because sex can influence gonad color, ripe males tending to paler colors. The gonad index (GI) was calculated as a measure of gonad growth. An alimentary index (Al) was calculated as an additional indicator of sequestered nutrient stores (Klinger et al. 1998. Kelly et al., 1998b). Indices were calculated using the formulae of Kelly et al., 1998a: GI (or Al) = wet weight of gonads (or gut) divided by the wet weight of the eviscerated test and spines, expressed as a per- centage. To allow comparison with GI calculated using total wet weight as the denominator, the GI values quoted here should be divided by a conversion factor of 1.45. The numerical data were log-transformed and tested for nor- mality and homogeneity of variance (Zar 1996) using MINITAB, version 12.1 for Windows. The log-transformed data were ana- lyzed using analysis of variance (ANOVA) when the assumptions were met, and the Kruskal-Wallis nonparametric ANOVA when violations occurred. Tukey"s or the nonparametric Nemenyis (Zar 1996) multiple comparisons tests were employed to assess where the differences occurred in the cases of rejection of the null hy- pothesis. The observed gonad colors were classified as unaccept- able, acceptable, or excellent for marketing. The frequencies of occurrence in each category at each sample date were compared to those expected should the algae have had no effect on color (i.e., those observed in the control group) using the log-likelihood ratio (Zar 1996). MiCROALGAL SUPPLEMENTS IN SEA URCHIN DIETS 379 RESULTS Gomitl (imwth A three-way ANOVA incorporatiiiL' sample dale and trealinent as fixed factors and replicate tanks nested within treatiiients showed that there was a significant increase in GI over the duration of the experiment (F = 64.67. df = 2, 16. P < 0.001 ) (Fig. 1 ). Nested (replicate tanks within treatments) ANOVA was used to examine difference in GI between treatments on any given sample day. At the start of the experiment (sample 1), there were no .significant differences in gonad index between diet groups. At the end of the experiment, both groups of microalgae-fed [P. tricor- intlitm and T-ISO) urchins had significantly higher (F = 4.0. P = 0.013, df = }<) GIs than those fed the control diet. There was no significant difference in GI between urchins fed L. sacchariiui and the control group. The GI produced by P. tricorminiiii became significantly higher than that produced by the control diet after only five weeks; whereas, the effect of T-ISO did not become significant until the end of the trial (after twelve weeks). Alinniilary Indices There were no significant treatment effects on AI at any time during the experiment (sample \: P = 0.734. sample 2: P = 0.097, sample ?>: P = 0.536, sample 4: P = 0.424). There was, however, a significant increase in the AI of all the algae-fed ur- chins after the first five weeks of the trial (L. sacchariua-fed urchins: P < 0.001, P. tricoruutimi-fed urchins: P < 0.001 and T-ISO-fed urchins: P < 0.001 ). The urchins fed the control diet did not show a significant increase in AI until eight weeks into the trial (P < 0.001 ). There was no significant change in test diameter over the duration of the trial iP = 0.198) or among diets on indi\idual sample dates {P = 0.844). Gonad Color and Sex Distribution The gonad color distribution for each treatment at each sample date is summari/.ed in Table 2. At the start of the experiment, there were no differences in color distribution between any of the treat- ment groups and the control group. However, after five weeks (sample 2). the urchins fed P. tricormitiim had a significantly better color distribution than those fed the control diet (Table 3). After twelve weeks, the T-ISO fed urchins also had a significantly better color disti ibiitmn than the control group. The data generated Sample 1 (20/5/99) Sample 2 (23/6/99) Sample 3 (14/7/99) Sample 4 (10/8/99) (3 L, saccharina ■ Isochrysis sp □ P tricomutum □ CONTROL Figure 1. Mean gonad indices at each sample date, with 95% confi- dence intervals. from the control group indicated there was no seasonal change in gonad color over the course of the experiment. The sex ratios on sample days three and four were determined as the urchins spawned on dissection, and a log-likelihood com- parison with the ratio in the control group showed that the male:fe- male ratio for urchins fed the T-ISO diet differed significantly from that of the control on sample days 3 and 4 (Table 4). DISCUSSION Gonadal and Somatic Growth The gonad acts as a nutrient storage organ in Psammechimts iiiiliaris: urchins provided with a nutritious food source demon- strate rapid gonad growth because of a proliferation of nutritive phagocytes (Kelly el al. 1998b). The low GIs of the urchins at the start of this experiment suggest that the wild population at the collection site were of a poor nutritional status. Therefore, it is not surprising that the GI in all treatment groups increased signifi- cantly over the experimental period, during which they were re- ceiving a constant supply of nutritious food. However, by the end of the experimental period, the GIs of the urchins fed the two niicroalgal diets were significantly higher than those fed the con- trol diet, implying that the microalgae imparted additional nutri- tional value for the urchins. The gonad biomass of many commercially important echinoids shows a strong seasonal variation related to their annual reproduc- tive cycle, resulting in a limited season of economically viable harvest (Byrne 1990, Hagen 1998). This experiment was con- ducted during the spawning season of P. miliaris in Scotland (Kelly 2000) during which gamete release normally results in a steady decline in GI. However, instead of decreasing, the GI in- creased steadily over the duration of the experiment, indicating it is possible to have an out-of-season gonad yield from P. miliaris fed artificial diets enhanced with microalgae. Despite an approximate threefold increase during the course of the experiment, the mean values of GI obtained for all groups by the end of the trial ( 14.25-17.64%, /! = 15) remained lower than the maximum values obtained in other studies on this species in Scotland. Kelly et al. (1998b) found that urchins reared in prox- imity to Atlantic salmon (Salmo saUir) in polyculture trials could attain summer GIs of as much as 40%. In laboratory trials. Cook et al. ( 1998) found that P. iiiili(iri\ fed commercially manufactured salmon food attained a mean GI of approximately 35% in August, which continued to grow for the remainder of the trial, reaching a maximum value of almost 60% by December. The lower values obtained in this trial are likely to be a result of the inferior nutri- tional value of the niicroalgal diets, as compared to commercially manufactured salmon food. Although a full comparison of the relative nutritional profile of the algae used was beyond the scope of this study, it is interesting to note that Pliae(>ikicr\lum tricur- luituin and Isocrysis sp. are reported to have a high content of the fatty acid EPA (Otero et al. 1997) and DHA (Fabregas et al. 1995), which may partially account for the enhanced gonad growth of urchins fed these diets. Enhanced gonad growth has been demon- strated for P. miliaris fed diets rich in DHA (Cook et al. 2000). The AI is an additional measure of the nutritional status of echinoids, because it is also a site of nutrient storage (Lawrence et al. 1966, Lawrence & Lane 1982). In this trial, the gut indices of all groups increased significantly during the first five weeks of the trial, further implying that the wild urchins had a low nutritional status at the start of the trial. However, there was no further in- 380 McLaughlin and Kelly TABLK 2. Contingency table summarizing the observed frequencies of gonad color categories. Sample 2 Sample 3 Sample 4 Sample 1 (23rd June aath July llOth August (20th May 1999) 1999) 1999) 1999) u a e u a e u a e u a e Control 60 20 20 80 7 \? 47 40 13 53 40 7 L. sacchariiw 80 \i 7 53 13 33 47 40 13 47 40 13 T-ISO 73 7 20 60 7 33 40 40 20 20 40 40 P. triconiiintni 86 7 7 40 0 60 33 7 60 7 40 53 (u = unacceptable; a = acceptable and e = excellent) ii 15. crease in AI over the remainder of the experiment, reflecting either the limited storage capacity of the gut wall (Lawrence & Lane 1982) or the limiting nutritional value of the experimental diets. No significant somatic (test) growth was observed over the dura- tion of the experiment. However Cook et al. (1998) found that adult urchins fed a macroalgal diet increased in test diameter at an average rate of only 0.02 mm/month. Gonad Color Relatively few studies have been designed to address the im- portant issue of gonad color in sea urchin production specifically. The measurement and communication of color requires careful attention, and although several studies (Bai'ker et al. 1998. Watts et al. 1998) have noted that differences in gonad color occur in echi- noids fed artificial diets, they do not attempt to quantify their comparisons. The use of internationally recognized color standards in the present trial was intended to provide a less ambiguous means of color communication and comparison. The use of synthetic carotenoid pigments to improve flesh col- oration is common in the manufacture of artificial diets for cul- tured salmonid fish (Stoi'ehakken et al. 1987. Foss et ai. 1987). However, the cost of synthetic pigments contributes significantly to the production cost of artificial feeds in salmonid aquaculture (Torris.sen et al. 1990); therefore, it should be avoided if possible in echinoculture. Robinson and Castell (2000) have shown that incorporating a spray-dried form of the niicroalgae DuiuilicUa sciliiui. which is rich in P-carotene, into artificial diets caused a pronounced improve- ment in gonad color in Strongyloccntroliis droebachiensis. Goebel and Barker ( 1998) noted no improvement in the gonad color of £. the form in which the pignient is supplied in the diet also dictates how or if it is later expressed in the gonad. Obviously, it would be of interest to conipare the effect of a range of different pigments on the gonad color of a variety of commercially important echinoid species. A detailed comparison of the pigment profiles of urchin gonads of a "good" and "poor" color for each species niay also help elucidate the pigment requirements of effective, color- enhancing artificial diets. Pigment expression in the gonad niay be linked to factors other than pignient presence alone; for example, lipid content or the presence of gametes. In the spawning season, the ovaries contain more pignient than the testes (Goodwin 1984). Griffiths and Perrott (1976) ha\e shown that carotenoids are passed from the ovaries to the eggs, and Hallenstvet et al. (1978) subsequently showed that they are then transfeiTed to the larvae. It is thought that these pignients play an important protective role in larval development and biological de- fense (Kawakami et al. 1998). Consideration of sex ratios is. there- fore, important in color comparisons between treatment groups. In this study, it was noted that, in general, the fetnale urchins pro- duced gonads of more marketable colors than males, presumably because of the influence of the pigmented eggs as opposed to sperm. At both sample dates three and four, the sex ratios were effectively the same in the control group and the P. tikoiiuaum- fed group, the two groups between which the biggest differences in color were seen. This indicates that the observed improvement in gonad color was not merely attributable to unequal sex ratios in the samples, but was a true effect of the diet. In the T-ISO-fed group; however, there were more males than feniales at sample day three and then more females than niales at sample day four. Because the significant shift in color distribution for T-ISO-fed urchins also occurred between these two dates, this result should perhaps be interpreted with a greater degree of caution. The results of this trial indicate that the microalga P. tiiconni- TABLE 3. Comparison of gonad color hclv\een each treatment group and the control group at the four sample dates of the experiment. Sample 1 (20th May 1999)* Sample 2 (23rd June 1999) Sample 3 (14th July 1999) Sample 4 (10th August 1999) L. scicchiirimi T-ISO P. rricornuluin G = 3.085 G = 2.218 G = 5.166 G = 5.448 G = 3.985 G = 18.756* G G G 0 0.583 20.125* G = 0.903 G = 15,616* G = 29.112* * Indicates a distribution having a significantly different unacceptable: acceptable: excellent ratio from the control group at the given time. G.-nuc.i = X'ooxz = 5.991 (Zar 19961. MlCROALGAL SUPPLEMENTS IN SEA URCHIN DIETS 381 TABLE 4. Comparisons of male:l'enialc ratios hetHeen treatment groups in the trial (after eiglit and tHel\e «eeksl. P. L. T-ISO tricornutum saccharina Control Sample 3 10:5 4:11 S:7 6:4 (14 July 'W) (G = 4.339)* (G = 1.171) (G = 1.(1S5) Sample 4 6:9 9:6 S:7 10:5 (10 Augiisl "99) (G = 4.450)* (G = 0.:91) (G = 1.140) * Indicates Ihe male:feniale ratio is signilKaiitly dilTcrcnl riiiin llia( of the control group at the given date. Gcnuci = X'„„.s,, = 3.841 (Zar 1996). mm is a better choice for inclusion in artificial (diets for sea urchins that the other algae considered. Not only did it result in signifi- cantly better gonad growth; it also significantly improved gonad color distribution. Furthermore, the improvement in color occurred in a much shorter time, having important implications for culture if microalgal diets are to be used as preharvest conditioning diets for wild or polycultured urchins. In addition. P. tiiconiiiliim is e.xceptionally robust in culture and performs well in semiconlinu- ous culture (Otero et al. 1997). It is not clear at present whether the weakly silicified cell walls P. triconutlwn (Bold & Wynne 1985) affects the ability of the urchins to assimilate the pigment from the algae or pigment preservation during diet manufacture. Although reported to contain the same carotenoids as P. tricornutum. the L. saccluuiiui diet did not improve gonad color in comparison with the control. Accui^ate pigment profiles of the algal cultures actually used to create experimental diets would benefit artificial diet de- sign, because the biochemical and pigment composition of algae can change with nutrient and light availability (Fabregas et al. 1995, Fabregas et al. 1998) and season (Stengel & Dring 1998), ACKNOWLEDGMENTS This research was funded by EC FAIR Grant CT98-92I0. We thank the Director for the use of facilities at SAMS and CCAP and for supply of algal strains. LITERATURE CITED Barker. M. F.. J. A. Keogh. J. M. Lawrence & A. L. Lawrence. 1998. Feeding rate, absorption efficiencies, growth, and enhancement of go- nad production in the New Zealand sea urchin Evechinus chloroiicus Valenciennes (Echinoidea: Echinometridae) fed prepared and natural diets. J. Shellfish Res. 17(5): 1583-1590. Black, W. A. P. 1950. The seasonal variation in weight and chemical composition of the common British Laniinariaceae. ./. Mcir. Biol. Assn. U.K. 50:45-72. Bold. H. C. & M. J. Wynne. 1985. Introduction to the algae. New Jersey: Prentice Hall. 395 pp. Brown. M.R.. G. A. Dunstan. S. W. Jeffrey. J. K. Volkman. S. M. Barrett & J. M. Leroi. 1993. The influence of irradiance on the biochemical composition of the Prymnesiophyte Isociysis sp. (Clone T-ISO). / Phycol. 29(5):60I-612. Byrne. M. 1990. Annual reproductive cycles of the commercial sea urchin P. livichis from an exposed intertidal and a sheltered subtidal habitat on the west coast of Ireland. Mar. Biol. 104:275-289. Caltagirone, A., P. Francour & C. Fernandez. 1992. Formulation ol an artificial diet for the rearing of the urchin Paracentrolus lividiis: I. comparison of different binding agents. In: L. Scalera-Liaci & C. Cani- catti. editors. Echinoderm research 1991. Rotterdam: Balkema. pp. II5-11S. Cook. E. J. 1999. F.\iiiiiincLhinns imliaris (Gmelin) (Echtnodermata: Echi- noidea): factors effecting its somatic growth and gonadal growth and development. Ph.D. Thesis, Napier University. Edinburgh. UK. Cook. E. J.. M. V. Bell K. D. Black & M. S. Kelly. 2000. Fatty acid compositions of gonadal material and the diets of the sea urchin Psam- mechimis miliaris: trophic and nutritional implications. J. £.v/). Mur. Biol. Ecol. 255(2):26I-274. Cook. E. J., M. S. Kelly & J. D. McKenzie. 1998. Somatic and gonadal growth of the sea urchin Psammechimts miliaris (Gmelin) fed artificial salmon feed compared with a macroalgal diet. / Shellfish Res. 17(5): 1549-1555. de Jong-Westman. M.. B. E. March & T. H. Carefoot. 1995. The effect of different nutrient formulations in artificial diets on gonad growth in the sea urchin Sirongylocenrrouts droebachieiisis. Can. J. Zool. 73:1495- 1502. Fabregas, J., A. Otero, E. D. Morales, B. O. Arredondo-Vega & M. Patino. 1998. Modification of the nutritive value of Phaeotlactylum trieonni- tum for Artemia sp. in semicontinuous cultures. .Aiiuaeiilliire 169:3^, 167-176. Fabreaas, J.. M. Patino, E. Vecino. F. Chazaro & A. Otero. 1995. Produc- tivity and biochemical composition of cyclostat cultures of the marine microalga Tetrasehnis siiecica. Appl. Mierohiol. Biotech. 43(4):617- 621. Fernandez, C. & G. Pergent. 1998. Effect of different formulated diets and rearing conditions on growth parameters in the sea urchin Paracentro- tus lividus. J. Shellfish Res. 17(5):I57I-158I. Foss. P.. T. Storebakken. E. Austreng & S. Liaaenjensen. 1987. Carot- enoids in diets for Salmonids. V. pigmentation of rainbow trout and sea trout with astaxanthin and aslaxanthin dipalriiitate in comparison with canthaxanthin. Aijnacultiire. 65:293-,305. Goebel. N. & M. F. Barker. 1998. Artificial diets supplemented with ca- rotenoid pigments as feeds for sea urchins. In: R. Moot & M. Telford, editors. Proceedings of the 9th International Echinoderm Conference. San Francisco. Rotterdam: Balkema. pp. 667-672. Goodwin. T. W. 1984. The biochemistry of the carotenoids. vol. II: ani- mals. 2nd ed. New York: Chapman & Hall. Griffiths. M. & P. Perrott. 1976. Seasonal changes in the carotenoids of the sea urchin Strongylocenlrotns droebachiensis. Comp. Biochem. Phys- iol. 55B:435^41. Hagen. N. T. 1996. Echinoculture: from fishery enhancement to closed cycle cultivafion. World Aquaciili. 27(4):6-l9. Hagen. N. T. 1998. Effect of food availability and body size on out-of- season gonad yield in the green sea urchin. Strongyocenrrotus droe- bachiensis. J. Shellfish Rev. 17(5):1533-I539. Hallenstvet, M.. E. Ryberg & S. Liaaen-Jensen. 1978. Animal carotenoids — XIV carotenoids of Psammechiniis miliaris (sea-urchin). Comp. Bio- chem. Physiol. 60B: 173-175. Haugan. J. A. & S. Liaaenjensen. 1994. Algal carotenoids 54. carotenoids of the brown algae (Phaeophyceae). Biochem. Sysremarics Ecol. 22:( I ): 31^1. Havardsson, B. & A. K. Imsland. 1999. The effect of Astaxanthin in feed and environment temperature on carloenoid concentration in the go- nads of the green sea urchins Sirongylocentrotiis droebachiensis Miiller. / World Aquacuh. Soc. 30(2): 208-218. Kawakami, T., M. Tsushima, Y. Katabami, M. Mine, A. Ishida & T. Matsuno. 1998. Effect of p. p-carotene. p-echinenone. astaxanthin. fucoxanthin. vitamin A. and vitamin E on the biological defense of the sea urchin Pseiidocentroiiis depressii.s. J. E.\p. Mar. Biol. Ecol. 226: 165-174. Keesing. J. K. & K. C. Hall. 1998. Review of the status of world sea urchin fisheries points to opportunities for aijuaculture. J. Shellfish Res. 17(5|: 1 597- 1 604. 382 McLaughlin and Kelly Kelly. M. S.. J. D. McKenzie & C. C. Brodie. 1998a. Sea urchin.s in polyculture: the way to enhanced gonad growth? In: R. Mooi & M. Telford, editors.. Proceedings of the 9th International Echinoderm Con- ference, San Francisco. Rotterdam; Balkema. pp. 707-712. Kelly. M. S.. C. C. Brodie & J. D. McKenzie. 1998b. Somatic and gonadal growth of the sea urchin Psammechiniis miliaris (Gmelin) maintained in polyculture with the Atlantic salmon. / Shellfish Res. 17(.S):LS?7- 1562. Kelly, M. S. 2000. The reproductive cycle of the sea urchin P^ainuiecluiiKs miliaris (Gmelin) (Echinodermata: Echinoidea) in a Scottish sea loch. J. Mm: Biol. Ass. UK. 80: 909-919. Klinger. T. S., J. M. Lawrence & A. L. Lawrence. 1998. Digestion, ab- sorption, and assimilation of prepared feeds by echinoids. In: R. Mooi & M. Telford, editors. Proceedings of the 9th International Echinoderm Conference. San Francisco. Rotterdam: Balkema. pp. 713-721. Lawrence, J. M.. A. L. Lawrence & A. C. Giese. 1966. The role of the gut as a nutrient-storage organ in the purple sea urchin iStmngyliHcntiotus droebuchiensis). Phys. Zool. .^9(4):281-290. Lawrence, J. M. 197.'). On the relationships between marine plants and sea urchins. Oceanoi>r. Mar. Biol. Ann. Rev. 1.^21.^-286. Lawrence. J. M. & P. Lane. 1982. The utilization of nutrients by postmeta- morphic echinoderms. In: M. Jangoux & J. M. Lawrence, editors. Echi- noderm nutrition. Rotteerdam: Balkema. pp. 3.^1 -.371. Lawrence J. M.. R. Otaiza, A. L. Lawrence & E. Bustos. 1 997. Enhance- ment of gonad production in the sea urchin Loxechinus albiis in Chile fed extruded feeds. J. World Aqiiaciilt. Soc. 28:91-96. Otero, A., D. Garcia & J. Fabregas. 1997. Factors controlling eicosapen- taenoic acid production in semicontinuous cultures of marine microal- gae. / Aprl. Phycol. 9(5): 465-169. Pantone, 1995. PANTONE color formula guide 1000. Pantone® Inc. Robinson. S. M. C. & L. Colborne. 1998. Roe enhancement trials of the green sea urchin using an artificial food source In: R. Moo & M. Telford, editors. Proceedings of the 9th International Echinoderm Con- ference, San Fransico, 1996. Rotterdam: Balkema. p. 803. Robinson, S. M. C. & J. D. Castell. 2000. Developing suitable color in the roe of cultured green sea urchins {Slrnnftyloccntrotiis ciroehachiensis). Abstract: Proceedings of the Kith International Echinoderin Confer- ence. New Zealand, 2000. Stengel, D. B. & M. J. Dring. 1998. Seasonal variation in the pigment content and photosynthesis of different thallus regions of Ascoplnlhini noilosiini (Fucales: Phaeophyta) in relation to position in the canopy. Phycologia 37(4):259-268. Storebakken, T., P. Foss. K. Schiedt, E. Austreng. S. Liaaen-Jensen & U. Manz. 1987. Carotenoids in diets for Salmondis IV. pigmentation of ■Atlantic salmon with astaxanthin, astaxanthin dipalmitate, and canthax- anthin, Aqiuituliiirf 65:279-292. ToiTissen. O. J.. R. W. Hardy. K. D. Shearer. T. M. Scott & F. E. Stone. 1990. Effects of dietary canthaxanthin level and lipid level on apparent digestibility coefficients for canthaxanthin in rainbow trout iOncorhyn- chus mykiss). Aqiiaculliire 88:351-362. Tsushima, M. & T. Matsuno. 1990. Comparative biochemical studies of carotenoids in sea urchins — I. Comp. Biochcm. Physiol. 96B(4):801- 810. Tsushima. M.. T. Kawakami & T. Matsuno. 1993. Metabolism of carot- enoids in the urchin Pseudocentrotiis depressus. Comp. Biochem. Phys- iol. 106B(4):737-74I. Vadas, R, R. 1977. Preferential feeding: an optimi/ation strategy in sea urchins. Ecol. Monotir. 47:337-371. Watts, S. A., S. A. Boettger, J. B. McClintock & J. M, Lawrence. 1998. Gonad production in the sea urchin Lytechinus variegatiis (Lamarck) fed prepared diets. J. Shellfish Res. 17(5): 1591-1595. Wright. S. W.. S. W. Jeffrey. R. F. C. Mantoura, C. A. Llewellyn. T. Bjornland. D. Repeta & N. Welschmeyer. 1991. Improved HPLC method for the analysis of chlorophylls and carotenoids from marine- phytoplankton. Mar Ecol. Prog. Ser 77 (2-3): 183-196. Zar. J. H. 1996, Biostatistical analysis, 3rd ed. New Jersey: Prentice Hall. TECHNICAL PAPERS Presented at The Sixth International Littorinid Symposinm The Hofstia University Marine Laboratory Priory, St. Ann's Parish, Jamaica July 24-3 K 1999 Editors: Joseph C. Britton Department of Biology Texas Christian University Fort Worth, Texas. USA and Robert F. McMahon Department of Biology The University of Texas at Arlington Arlington, Texas, USA 383 .hninuil i>f Slwllfish Reseanh. Vol, 20. No. I. .^.'^--lyi. 2()U1. LITTORINE FORAGING BEHAVIOR AND POPULATION STRUCTURE ON A WAVE-EXPOSED SHORE: NON-LINEAR RESPONSES ACROSS A PHYSICAL GRADIENT TROY C. ADDY AND LADD E. JOHNSON* Departcmciil dc biologie and GIROQ. Uiiiversite Laval. Quebec. Quebec, GIK 7P4 Canada ABSTRACT Littorines often use topographical features (e.g., holes, crevices) for shelter. We investigated this behavior (% outside of natural crevices and artificial holes during daytime low tides) over a cross-shore wave force gradient. Whereas maximal wave force (MWF) declined linearly with increasing distance from the low tide mark, percent outside increased along this gradient and reached an asymptotic level (70%) 40 m from the low tide mark. A comparison of seasonal averages of percent outside and MWF showed an abrupt transition between high and low values of percent outside with increasing MWF. but comparisons of daily data were more varied, presumably due to the superimposition of other environmental factors (e.g., desiccation). Contrary to earlier descriptive studies and biomechanical predictions, larger littorines were found in locations closer to the low tide mark where MWF were higher, suggesting that size-limitation from hydrodynamic forces does not occur within the size range studied (1-8 mm) when shelters are available. However, the restriction of littorines to areas near shelters can greatly influence community structure by permitting the development of algal assemblages in exposed locations. KEY WORDS: gastropods, grazing, heterogeneity, hydrodynamic forces, intertidal ecology, refuges, and size gradients INTRODUCTION Ecologists have often e.xaniined species over enviroiimeiital gradients in which specific predictions can be made regarding the responses of the species under study. Recent examples include grazing responses of arctic grasses across a snow deposition gra- dient (Wegener & Odasz 1997) and predation on a marine snail across an algal gradient (Alfaro & Carpenter 1999). Although this approach can only provide correlative relationships, it is especially useful for examining factors that are difficult to manipulate ex- perimentally. ■Striking physical gradients occur in intertidal environments, and the distribution of intertidal organisms (e.g., zonation patterns) is often thought to be a function of physical gradients acting either directly on them or indirectly on their predators or competitors (e.g., Garrity 1984, Takada 1996). One well recognized gradient is "wave exposure," which has been typically exatnined at sites along the shoreline, e.g., from a headland into a bay (Lewis 1964, Palumbi 1986). Intuitively, a gradient in wave energy should also exist across the shoreline as breaking waves dissipate their energy, but this gradient has never been quantified in rocky intertidal en- vironments. The biology of intertidal organisms has often been examined with respect to this physical gradient (e.g., Emson & Faller-Fristch 1976, Atkinson & Newbury 1984, Chapman 1994, Britton 1995, Takada 1996). Unfortunately, many studies have only examined the environment at two locations along the gradient, often at the extremes, and then interpolated for intermediate locations. More- over, wave exposure has rarely been quantified and has usually just been subjectively described as "exposed" or "protected." While the most important aspect of water motion is sometimes unclear for a particular situation (Denny et al. 1985, Denny 1988), a number of investigators have developed techniques for measuring maximum wave forces (Jones & Demetropoulus 1968, Palumbi 1984, Bell & Denny 1994). the parameter of water tnotion most related to dis- lodgment. The risk of dislodgnicnt is thought to influence the biology of *CoiTespondmg author. E-mail: ladd.johnson@bio.ulaval.ca littorine snails, the dotninant herbivoi-es of many high- and mid- shore rocky intertidal habitats, and both tiiorphological and behav- ioral traits of littorines have been interpreted as responses to water motion. The.se include larger relative foot area (Atkinson & New- bury 1984, Trussell et al. 1993), shell morphology (Johannesson 1986, Trussell et al. 1493, Trussell 1997, Boulding et al. 1999), and the use of shelters (Emson & Faller-Fritsch 1976, Raffaelli & Hughes 1978, Atkinson & Newbury 1984, Chaptiian 1994). This latter behavioral aspect can be quite itnportant in influencing both the density and size frequence of littorinid populations (Emson & Faller-Fritsch 1976, Raffaelli & Hughes 1978, Atkinson & New- bury 1984, Trussell et al. 1993). While the influence of crevices on both the individual behavior and population structure of littorinids has been explored, there has been a lack of attention to the physical environment that allegedly controls the behavior of these snails. Although past studies may have been qualitatively correct in their estimation of "wave exposure," there remains a need to explore quantitatively both the behavior and the physical environment to develop a predictive understanding of the ecology of littorines and intertidal communities. In this study, tneasures of the physical environtiient are com- bined with ecological data to explore the control of foraging be- havior across an environmental gradient. Specifically, we examine the correlative relationship between maximum wave force and the use of crevices across an intertidal shore. Such behavioral studies of intertidal invertebrates are rare, as active behavior usually oc- curs when the tide is high and researchers are absent. Littorines offer an exceptional opportunity as they are often active during low tide (Jones & Boulding 1999) and are distributed in the mid- and high intertidal zones, thereby permitting prolonged periods for observations. MATERIALS AND METHODS The study was conducted at Poinle Metis (48'-4rN and 68°02'W) located on the St. Lawrence maritime estuary near Mont Joli in Quebec. Canada. Two sites were used: one consisted of a series of bedrock ridges paralleling the shoreline with consolidated boulders and cobbles between the ridges (Fig. 1 ; Site 1 ), the other 385 386 Addy and Johnson Site 1 •10 60 60 70 ] 10 20 30 40 Distance (m) Figure 1. Cross-shore profiles of study sites (note exaggerated vertical scale). All measurements are based on the 0 m low tide mark estab- lished for each site. l>etters represent the locations censused for lit- torine behavior. Arrows indicate the locations of maximal wave force transducers. was a series of less well-defined ridges with intervening tide pools (Fig. 1: Site 2). Elevated rock surfaces at both sites supported few sessile organisms (algae or invertebrates) due to the annual ice scour that occurs along the shores of the estuary (Archanibault & Bourget 1983). Littorine snails, principally Liitorina saxatilis Olivi, but on occasion L. obtiisitia L.. were the most conspicuous animals in this habitat and often aggregated in crevices and holes when not foraging on the open rock surface. The behavior of littorines was studied along a transect perpen- dicular to shore at each site. At Site I. the transect covered ap- proximately 70 m and ran across five distinct ridges. At Site 2. the transect covered approximately 20 m and ran across four ridges. Six specific locations were selected along each transect for obser- vations, each corresponding to a natural feature of the landscape (e.g., a face of a ridge). For the three most seaward locations at Site 1, 30-cm portions of six crevices spaced 0.5-2 m apart were cho- sen at each location (18 total). At Site 2. three crevices were sampled at the most seaward site and at two mid-transect locations (nine total). For each location, tidal height, distance from the spring low tide mark, orientation and inclination of the rock face were measured as well as the dimensions (average of six measures of width and depth) of each crevice (Table 1). Littorines present within the crevices ("inside") as well as the littorines that were present on the exposed rock surface within 25 cm of the crevices ("outside") were counted, and percent outside calculated as I (10x( number outside )/( number inside and outside). The distance of 25 cm was chosen based on earlier observations of browse zones (typically 10-20 cm in width) in the algal mats found at certain locations along the transects (L.E. Johnson, pers. obs.). When other shelters were found within 50 cm of the crevice, only littorines within half the distance between the crevice and the other shelter were counted. Local density was calculated as the number (inside -i- outside) per cm' of crevice volume, calculated assuming a triangular cross-section (i.e. .volume = 0.5 x width x depth x length). In addition to these naturally occurring but variable irregulari- ties in the rock surface, we created standardized irregularities in early May 1998 by drilling sets of holes at all locations along the transects to control for any variation in the form of the shelters across this shoreline. Again, tidal height, distance, orientation and inclination of the rock face were recorded (Table 1 ). Each set of holes consisted of three holes 1 .3 cm in width and I cm in depth (total volume = 4.0 cm'). Each of the holes within the set was TABLE \. Descriptive information on natural crevices and artificial holes used at sites 1 and 2 for examining littorine behavior. Holes: Site 1 Site 2 Location Dist. (m( Height ( m) Face Inclin. Density (no/cm') Dist. (m) Height (m) Face Inclin. Density (no/cm') A 0.1 1.8 South 35 5.4"' 0.3 0.2 hon, 0 3.3^-^ 6 11.7 1.6 North 90 7.6'^^ 1.0 0.7 North 90 5.0'"- C 37.6 2.4 North 50 18.4-' 3.0 1.2 South 65 3.2"^ D 39.8 2.4 South 70 2.5^' 11.5 0.2 West 40 10.5" E 57.2 2.3 North 90 16.4* 16.1 1.8 North 10 10.6-' F 63.6 2.0 South 35 13.0" 17.0 2.2 hori. 0 7.6" Crevice: Site 1 Location Dist. (m) Height (m) Face Inclin. Depth (mm) Width (mm) Vol. (cm') Density (no/cm') A n.i 1.8 South 35 16.1 17.2 41.5 2.1" B 11.7 1.8 North 90 15.2 14.6 33.3 2.1" C 37.6 2.4 North -■^O 15.9 14.1 33.6 2.9" Crevice: Site 2 Location Dist. (m) He ght (m) Face Inclin. Depth (mm) Width (mm) Vol. (cm') Density (no/cm') A 1.1) 0.2 Hon . 0 15.6 10.6 24.S 1.7'' D 11.5 1.7 West 40 17.9 19.3 51.8 2.6^" E 16.0 1.8 North 10 12.5 12.1 22.7 4.0'" Descriptive information on natural crevices and artificial holes used at sites 1 and 2 for examining littorine behavior: distance from the low tide mark, tidal height, orientation and inclination from the horizontal of the rock face, as well as crevice depth, width, volume (sets of holes always had a total volume of 4.0 cm'), and densities (no./cm' shelter) for each location. For densities, similar letters indicate no significant differences between groups for multiple pair-wise analyses of two-factor ANOVA (Table 2) pertbrmed for each site separately. Wave Effects on Littorine Behavior 387 separated by 5 cm in a triangular pattern. Three replicates of tliese sets of holes were established at each location (except one where six were established as a pilot study of the influence of adjacent algal abundance). For each period of observation, the littorines were enumerated as above, except a distance of 15 cm from the holes was used as holes were usually on rock surfaces with few other irregularities and most littorines were found within this dis- tance. Littorines were counted daily during the morning low tide for series of 3-7 days during spring tides on an approximately bi- weekly schedule from May 29 to September 20 in 1998 at both sites and May 30 to August 30 in 1999 at Site 1 only, for a total of 64 days of observations. Limited observations were also made during other periods of the spring-neap tide cycle and during other seasons. Littorines were also visually classified with minimal dis- turbance into 4 different size categories: 1-2 mm. 2-4 mm. 4-6 mm. and 6+ mm (smaller littorines, < 1 mm, were difficult to discriminate against the rock surface). While littorines are usually sized bv shell length, we used shell height as defined by Fletcher (1995) as "the distance of the highest point of the shell above a flat surface upon which the shell is lying with the aperture facing downwards" as this measure appears more relevant to the use of crevices. To estimate the wave energy gradient along these transects, maximum wave force (MWF) was measured over 24 hr simulta- neously at three or four locations that corresponded approximately to a subset of the locations of the shelters (Fig. 1 ). In 1998, only one measurement was taken at each location (Site 1; locations A, B. C. and E; Site 2: locations A. D. and F) while in 1999 the number was increased to four (Site 1 only) to characterize better the MWF at a given location. Maximum wave force was measured using a simple transducer (Bell & Denny 1994) that consisted of a spring attached to a practice golf ball by monofilament line on which a rubber ring was able to slide. Only forces greater than 2 N were measurable by these transducers. The spring and ring were encased in PVC tubing and attached to the rock using a swivel attached to a screw anchored on a local high point. MWF was measured one day after devices were installed or reset. Such mea- surements were made simultaneously with behavioral observations on 18 days in 1998 and 23 days in 1999. Data from both sites in both years were used for seasonal comparisons of MWF and lit- torine behavior, whereas only data froin Site 1 in 1999 were used to make daily comparisons. These measures of MWF were in- tended for relative comparisons among locations and not for esti- mating the forces experienced by the snails. Data were analyzed using simple linear regressions or two-way ANOVAs with multiple pair-wise comparison tests (LS Means based on significant interactions when they occurred) for specific differences within significant sources of variation (SAS v. 6.12; SAS Institute Inc. 1996). When data were non-normal or variance unequal, we also analyzed rank-transformed data to confirm results obtained using non-transformed data. RESULTS Littorines were abundant the entire length of the transects but were 2 to 6 times denser in holes than in natural crevices based on the volume of the shelters available (Table 1 ). At Site 1. densities were higher at locations far from the low tide mark except at 40 m, a location with a steep, south-facing slope and a high number of crevices (Fig. 1 and Table 1). Densities along the shorter Site 2 transect were comparable to locations at similar distances at Site I . but were also higher away from the low tide mark (Table I). Densities in and around crevices were relatively constant (no sig- nificant differences) over the season and between years (Addy 2001). In contrast, after the holes were created in 1998. littorine densities in and around holes increased during the summer season but then stabilized in 1999 (Addy 2001). Littorine size structure varied similarly across both transects with larger (> 4 mm) indi- viduals more abundant at locations near the low tide mark (Fig. 2). The percentage of littorines 2—4 mm remained relatively constant across the sites whereas that of the smallest littorines (1-2 mm) was more variable. A lower percentage of littorines was found outside of holes and crevices at locations closer to the sea at both sites (Fig. 3). Per- centage outside increased away from the low tide mark along both transects but became asymptotic near 40 m along the longer transect (Fig. 3a). This trend occurred for both natural crevices and artificial holes, although the percent outside was much higher around holes, typically double that of natural crevices. Although there were significant differences among different size classes in the percentages found outside of shelters at some specific loca- tions, there were no obvious trends, and no statistical differences were found for pooled data, even when locations closer to the low water mark were analyzed separately from those farther away (Addy 2001). Maximum wave force (MWF) decreased linearly with increas- ing distance from the low tide mark in both years, although maxi- mum wave forces differed between the two years (Fig. 4). When seasonal means of MWF and percentage of littorines outside of holes are compared among locations, a nonlinear relationship is seen (Fig. 5) — the percentage outside of holes becomes exceed- A) 80%- C a> 03 CD Site 1 A(0) B(12) 0(38) D(40) E (57) F (64) Site 2 B) A(0) B(1) C(5) D(11) E(16) F (18) Location (distance [m]) Figure 2. Percentagt i)f each size class at locations of increasing dis- tance from the low tide mark at (A) Site I and (B) Site 2. 388 Addy and Johnson B) 100% 90% 80% 70% ■ 60% 50% 40% 30% 20% 10% 0% A(0) B{12) C(38) D(40) E (57) F (64) Site 2 ' fe . • A(0) B(1) C(5) D(11) E 116) Location (distance [m]) F(1£ Figure 3. Percentage of llttorines found outside of artitlcial holes and natural crevices at (A I Site 1 and (B) Site 2 for locations at increasing distance from the Ioh tide mark. "*" indicates that crevice data were not collected for those locations. Similar letters indicate no significant differences between groups for multiple pair-wise analyses of two- factor ANOVA (Table 2) performed for each site separately. i\ Site 1 (1999) o\. Site 1 8,2 (1998) 20 30 40 50 50 Distance (m) Figure 4. Average maximal wave forces (MWF) as a function of dis- tance from low tide mark. Open circles and dashed line represent average MWF (22 d) for 7 locations (3 at Site 2 and 4 at Site I) in summer of 1998 (/; = 1 for each location 1; tilled circles and solid line represent average MWF (23 d) recorded for four locaticms at Site I in summer of 1999 (h = 4 for each location). Lines represent linear re- gressions which are both signillcantly different from zero (1998: r" = 0.69, F = 11.4,p < 0.(12; 1999: r' = 0.76, F = 44.5,/) < 0.001). 3 O 00% • 90% ■ 80% ■ * Site 1 1998 70% ■ o • A Site 2 1998 60% ■ A • o - Site 1 1 999 50% ■ 40% ■ 30% • 20% • A 10% • o 0% • 0 • A Force (N) Figure 5. The relationship between the seasonal mean of the percent- age of littorines found outside artitlcial holes and the seasonal mean of the maximal wave force (MWF). Kach point represents a different location at Site 1 ( 1998 and 1999) or Site 2 ( 1998). Means for MWF in 1998 (H = 22 d) are from one MWF transducer at each location whereas means in 1999 (h = 23 dl are from four transducers at each location. ingly low (mean < \5'7c) at locations where mean MWF exceeded a threshold of 5.0-5.5 N. Linear regressions of data above and below this obvious threshold were not significantly different from zero. The relationship between MWF and percent outside is less evident when data for each day are compared for each location (Fig. 6; data from four locations at Site 1 in 1999). Generally, locations closer to the low tide mark had lower values of percent outside, but the data were extremely variable at a given location for a given MWF (Fig. 6). Above a value of 8 N, the percentage outside dropped dramatically and remained near zero. The lowest values of percent outside for each location were obtained during a storm in which MWF exceeded 20 N at the most exposed location (Fig. 6). DISCUSSION In temperate environments, the use of crevices is generally thought to be a behavior for reducing the risk of dislodgment by 100%! 90% 80% 70% • 60% 50% 40% ■ 30% 20% 10% 0% ° . o c 8o9« o A ^ A» AA(0) AB(12) • C(38) oE(57) A ^*' ^ A ^ ^, A^ A Jt tt\^^ A ^ / 2 7 12 17 22 Force (N) Figure 6. Comparisons of the percentage of littorines outside artificial holes versus maximal v>ave force iMWF) measured at four locations at Site 1 in 1999. Each point represents the average of three (percent outside) or four (MWF) measurements on a given day (/; = 23 d). Values in parentheses on x-axis are horizontal distance from low tide mark. .Arrows indicate data taken on the dav after a storm. Wave Effects on Littorine Behavior 389 waves (e.g., Emson & Faller-Fritsch 1976, Raffaelli & Hughes 1978). According to this idea, littorines Umit their foraging activi- ties to periods of low tide (Jones & Boulding 1999) and return to shelters to avoid being dislodged by the hydrodynamic forces of the incoming tide. Our results are consistent with this idea — the percentage of littorines foraging (i.e., outside of shelters) was lower in locations of higher maximal wave forces (MWF). An alternative explanation could be that littorines outside of shelters are dislodged at higher rates at locations of higher MWF. which would lead lo a lower percentage of littorines outside of shelters but also to a lower local density unless recolonization occurred quickly. While we occasionally observed abrupt decreases in den- sities at some locations immediately following days of higher MWF, densities were in general quite constant over time (Addy 20011. Thus we interpret this relationship to be due more to be- havioral responses (i.e., shelter-seeking behavior) than to demo- graphic processes (i.e., "emigration" by dislodgment). The relationship between wave force and shelter use is. how- ever, strictly correlative and might be due to other factors. Density, for example, also generally increased with increasing distance from the low tide mark (Table 1). If densities reached a level where shelters were completely filled, then a higher percentage of lit- torines outside of shelters would occur simply due to a lack of available shelters. Although shelters were sometimes very full, this situation only occurred when values of percent outside were at or near zero. For over 90% of the observations at any given location, numbers of snails inside shelters were less than 73% of the maxi- mum observed for that location (T.C. Addy, unpubl. data). In addition, locations with different densities had similar levels of percent outside (locations C and D at Site 1), and locations with similar densities had vastly different levels of percent outside (lo- cations B and F at Site 2). The linear decline in MWF with increasing distance from the low tide mark was not matched by a linear increase in the per- centage of the littorines outside crevices. Instead, a plateau was reached at approximately 40 m, where MWF averaged 5-6 N, which may represent a threshold below which there is a reduced risk of dislodgment. When percent outside and wave forces were compared among individual days (Fig. 6), this general pattern was still apparent, but there was a great deal of variability in the per- centage of littorines outside of shelters at any given location that was not explained by MWF. This variation appears to be related, at least in part to meteorological conditions experienced during low tide (e.g., temperature, desiccation). Other littorines reduce foraging during conditions of high evaporation (e.g. Atkinson & Newbury, 1984, Garrity 1984, Britton 1993), and in tropical en- vironments, desiccation has been suggested as the leading reason for crevice-seeking behavior (Garrity 1984; but see Catesby & McKillup 1998 for the role of predation). We propose that desic- cation plays a secondary role here as other field observations (day vs. night 1 and laboratory experiments (wet vs. dry rocks) support the idea that these littorines avoid conditions of high desiccation (Addy 2001). Reductions in foraging time in locations of higher wave forces could ultimately have impacts on the population structure of lit- torine populations, i.e., smaller sizes and reduced densities. Our findings of larger littorines at locations near the low tide mark (Fig. 2) is contrary to this idea and the results of other studies where, if a size gradient existed, smaller littorines were more common close to the low tide mark (Vermeij 1972, Britton 1993 and references therein, Trussell et al. 1993). This trend has been attributed to the reduced susceptibility of smaller littorines to dislodgment (Trussell et al. 1993) and would be consistent with hydrodynamic theory (Denny et al. 1983). However, we examined relatively small lit- torines, and the trend could be different for larger sizes. Differ- ences in crevice morphology among locations could also explain the trend we observed (e.g.. if there were larger shelters in more exposed locations), but our use of artificial holes specifically con- trolled for this possibility. Thus the underlying cause of the dif- ferences in size distribution remains unknown and may be related to differences in local densities and concomitant competition for food. Regardless, the survival of larger littorines in more exposed locations may depend more on a shelter-seeking behavior that allows littorines to avoid hydrodynamic forces that might dislodge and displace them, thereby obscuring any relationship that might exist between size and the probability of dislodgment if shelters were not present. These littorines may be using crevices much like soldiers use foxholes when under fire. An assessment of the relationship between wave force and littorine density is more difficult because the local abundances we report (i.e., number per unit volume of shelter! do not necessarily reflect the overall density of littorines at these locations or sites. This limitation notwithstanding, we saw increased local densities at locations farther from the low tide mark, which is consistent with the reduced probability of dislodgment there. Other factors (e.g.. food supply, recruitment, densities of shelter) may also be acting. Likewise, marked differences in densities were observed over only several meters (Table 2: Location C vs. D, Site 1 1. suggesting again that other ecological processes are affecting den- sities. Differences were observed between crevices and holes with regards to both local density and the percent outside of the shelters: holes sheltered more individuals for a given volume and those individuals appeared more likely to leave shelters during day-time low tides. Potential explanations for these differences are not ob- vious. The holes were a recent addition to the landscape, and thus were colonized by more actively moving individuals, which may also have been more likely to emerge during low tide. Regardless, the nature of the shelter (e.g., dimensions and location relative to other shelters) appears to have a strong influence on littorine foraging. In this study, we have measured only a population-level re- sponse of littorines along a wave exposure gradient; thus an inter- pretation of individual responses is problematic. Without knowing the identity of individual snails, it is impossible to determine if the fraction of the population found outside of the shelter from day- to-day represents the same group of high risk-taking individuals, alternating subsets of the population, or a combination of the two possibilities. Still, the fact that some littorines stay inside shelter during periods when others leave is puzzling. This observation supports the idea that littorine behavior is likely to be a complex response to both environmental factors and physiological state (Jones & Boulding 1999). and we suggest that the risk of dislodg- ment interacts with more basic drives such as hunger or digestive processing in producing these patterns of beha\'ior. If the hydrodynamic environment is indeed controlling littorine behavior, it may also be indirectly affecting other elements of the intertidal community. At kications with high MWF. where the local densities of littorines and the percentage foraging are lower, green algal mats fomi when shelters are sufficiently sparse (i.e., more than approximately 30 cm apart). Although there may also be direct effects of water motion on these algae (e.g., wave splash reducing desiccation or higher turbulence increasing nutrient trans- 390 Addy and Johnson TABLE 2. Results of two-way ANOVA on the local densities of littorines in and adjacent to siielters (crevices and holes) and percentage of littorines outside of shelters at locations of different distances from the low tide mark at the two sites. Density Site 1 Site 2 Source df MS F P df MS F P Model s 150.7 30.4 . Mar. Biol. Ecol. 31:3I).V 313. Takada, Y. 1996. Vertical migration during the lilc history of the intertidal gastropod Monoclonui lahio on a boulder shore. .Malsl, 2.827 /,. scahra (eight intervals), and 1,155 /,. sulculosa (live intervals). The means and stan- dard errors of |ln( painted) - ln(recaptured)|/|interval in years], which convert to annual survival rate as a proportion of one for the species in the same order, were -4.25 ± 1.12. -7.62 ± 1.29, -5.07 ± (1.75. and -5.02 ± 1.00. by far at Mandu Mandu. in line with thai site having the highest total abundance of snails during our entire study (Fig. 4). Further- more, as for the total snails, the term for strip within site was the smallest, never more than 4%. emphasizing spatial consistency of adjacent strips within each site. For all species of recruits, the variance component for site x quarter term ranked second largest, ranging from 18 to 36%, indicating that the temporal pattern of recruitment differed among sites. As can be seen in Figure 5, the abundance of recruits of all species varied considerably and with different teniporal patterns at the Mandu Mandu site. In summary, gains at an instant in time to the populations of rocky-shore snails were sinall in comparison with the total numbers, spatial variabil- ity was great, and temporal \ariability differed among species. Survival and (irowlli of Lilloriiies at Ningaloo Marine Park The niark and lecapture of painted snails provided quantitative estimates of rates of loss from the populations. We discovered that painted snails disappeared quickly from the mangrove trees, and our data consist of eight sets of painted snails observed between adjacent samplings (Fig. 6). Because the values are plotted on a logarithmic scale, the slopes of the lines connecting the numbers marked and recaptured for the groups of snails retlect the instan- taneous rates of mortality, which we calculated using natural loga- rithms and then conveiled to estimates of annual rate of survival, whose averages are indicated in Figure 6. Clearly, the painted snails, which were mostly large snails rather than small recruits, disappeared at a very high rates so that the annual survival rates were low. with L. cingulata having the greatest survival rate at 1.4% and L filosa the least at 0.05% per year. Besides being low. these estimates are also variable, with the instantaneous rates -i- 1 SE converting to 4.4, 0.2. 1.3. and 1.8% for L cinguluta. L filosa. L. scabra, and L. sulculosa. respectively. Figure 7 shows the survival of painted snails on the rocky shores. Like its congeners on mangroves, L. undulata had a low annual rate of survival of 3%. Similarly, our estimate of survival for the species lowest on the shore. A', millegraiia. was low at 1%. We rarely recaptured any of these snails, except for those that we had painted as small recruits. On the other hand. yv. trochoides, from high on the shore, was a survivor; its annual rate of survival was estimated as 24%. and we kept track of some individuals for 4 y before they disappeared. N. austrulis had an intermediate rale at 14%. All our estimates of survival probably underestimate the true survival because the snails could lose their paint mark. Our census data during the period when we collected recruits provided a sec- ond estimate of survivorship, under the assumption that our col- lections stopped all input to the populations during that time, so that the populations lost snails only due to mortality. The analyses of the regression of ln( abundance) on date in years and the con- versions to annual rate of survival in Table 2 show that the esti- mates were substantially higher than those calculated from the niark and recapture data for both the mangrove and rocky-shore snails but suffered from imprecision in the estimate of the slopes due to the small number of censuses conducted (all the rocky-shore species) and a lack of a distinctly linear decline (L. cingulata. L. filosa. and L undulata). The estimate for M millegrana is probably inappropriate for the reason stated earlier: This species never had any marked animals survive for more than a year. However, within habitats these estimates probably provide an indication of relative rates. Low survival rates have implications for other features of the life history of these snails, and the recaptured snails provided estimates of rate of growth because we could measure the size at which they were painted as well as size at recapture. Table 3 shows a suinmary of the information on growth rates in terms of the fitted paranieters for the von Bertalanffy growth equa- tion. The thin-shelled mangrove littorines all had asymptotic sizes (L'j:) between 23 and 28 mm. considerably larger than the thicker- 1000: 2 100: E Annual Survival Rate (% * N. australis 1 4 o N. millegrana i * N. trochoides 24 * L. undulata 3 1992 1994 Year 2000 Figure 7. Numbers in groups of snails of the four species of rocky- shore littorines known to be painted at se\en times and followed over at least (me time interval at Mandu Mandu. Lines join the data for each group of snails. The total numbers of individuals initially painted and the numbers of intervals providing data were 3.174 Sodilittnrina australis (seven intervals). 570 A', millvgrana (six intervals). 2,452 ,V. trochoides (10 intervals), and 743 Littoraria undulata (seven intervals). The means and standard errors of |lnlpainled) - ln( recaptured))/ linterval in years], which convert to annual survival rate as a propor- tion of one for the species in the same order, were -1.98 ± 0.26. — J.83 ± 1.15. -1.44 ± 0.80, and -3.36 ± 0.96. 400 Black and Johnson TABLE 2. Estimates of annual rate of survival derived from the slope of regression of natural logarithm of abundance of littorines (Y) on year (X) during the interval when recruits were removed regularly from the populations im trees A to E in Mangrove Bay and from the populations at Mandu Mandu. Species Number of Slope ± Standard Error Coefficient of Annual rate of censuses (Probability, b = 0.0) determination survival ( % ) 4 -1.920 ± 1.773 (0.47) 0.75 14.7 4 -1.561 ±0.985 (0.25) 0.75 21.0 4 -1.428 ±0.578 (0.13) 0.73 24.0 3 -1.063 ±0.030 (0.02) 0.99 20.1 3 -0.711 ±0,109(0.10) 0.98 49.1 3 -0.900±0.160 (0.11) 0.97 40.7 3 -0.173 ±0.389 (0.731 0.17 84.1 Liltoraria cin^iilalu Lilloraria filosa Lilloraria scahra Nodilitlorina austratis Nodilitroriiw millegnma Niidilittorina trochoides Linordiiii Hfuliilata shelled rocky-shore snails with values between 10 aniJ 18 nini. With the e.xception of N. millegnma, the rocky shore snails had lower growth rate coefficients (k) than the mangrove species. L. filosa and N. millegnma stood out as having substantially greater rates of growth than any of the other species. Figure 8 provides a visual summary of the information about growth with plots of the average size at age predicted by the von Bertalanffy growth equations. The steepness of the rise of the curves reflects the magnitude of the growth coefficient (k). the elevation of the curves represents the asymptotic size (L^), and the curves for L. filosa and N. millegrana terminate at the time at which the predicted asymptotic length is reached (greater than 10 y for all the other species). There are three groups of growth curves in Figure 8: The three mangrove species grew fastest and reached the largest sizes, the high-shore N. trochoides grew at the slowest rate and reached only moderate size, and the other three rocky shore littorines grew at about the same intermediate rate but reached different asymptotic sizes. As an additional summary of the information on growth. Figure 8 shows the nuniber of years at TABLE 3. Estimates of the parameters of the von Bertalanffy growth equation* for the eight species of littorinid snails studied at Ningaloo Marine Park. Habitat and species Sample size L ^ ± SE k ±SE Mangroves (all tree.s) Liltoraria cingidala Littoraria fdosa Liltoraria scabra Rocky shore (Mandu Mandu) Nodilittorina austrulis Nodilittorina millegrana Nodilittorina trochoides Littoraria nndulata 41 24.1 ±3.85 0.72 ±0,227 340 23.9 ±0.67 1.96 ±0.183 314 28.0 ±0.91 0.89 ±0.098 465 14.6 ±0.11 0.65 ±0.030 88 10.1+0.61 1,.X)± 0.238 632 13.1 ±0.27 0.23 ±0.015 36 18.4 ±0.91 0.51+0.096 * L, = Lx ( 1 - e'"''"), where L^ = asymptotic size, k = growth coef- ficient, and t = age in years. Most of the records of growth for the mangrove littorines cume from the groups marked in October 1991, but data from all intervals were included. The records for the rocky shore came from groups of snails initially niuiked in the winters of 1988 and 1989 and recaptured alter one year, except or Nodilittorina millegrana for which we had recaptures in August after marking a cohort in November the previous year. Snails that survived for two or more intervals contributed growth information for each interval. which the growth equation predicted that each species would reach half the asymptotic size, which is probably a biologically relevant indication of the deniographic processes in these populations. By this measure. L filosa and A', millegrana were the most ephenieral species because they were predicted to reach half size in less than 0.5 y. The other two mangrove species. L. cingitlata and L. scabra. would reach that size in less than 1 y. and L. iinclitlata and A'. aiistralis would take about 1.25 y. N. trochoides was predicted to take nearly 3 y to reach half size. Covariation of Population and Life History Characteristics The simultaneous insestigations of eight species of littorines allow tests for strong patterns of covariation in the estimated de- mographic and life history parameters. The population character- istics available for seven of the species were the coefficients of variation for the abundances (measured on a logarithmic scale) of total snails and of recruits, and the sum of the variance components Years to half asymptotic size 0.75 Littoraria scabra Littoraria clngulata ittoraria filosa Littoraria undulata 0.94 0.35 1.28 1.23 2.87 0.50 4 5 6 Years Figure 8. Predicted size-at-age from von Bertalanffy growth equations for three species of littorines from Mangrove Bay and four species from Mandu Mandu along with the age at which the snails are pre- dicted to reach half their asymptotic size. The plots for Liltoraria filosa and Nodilittorina millegrana end at the time at which the equation predicts that they would reach their asymptotic size: the others take longer than 10 y. The parameters of the von BertalantTy equation, asymptotic size, L^. and growth coefficient (k) are shown in Table i. Demographies of Eight Species 401 for temporal and temporal x site terms. The life history character- istics were the estimates of the growth rate constant and the time to half asymptotic size from fits to the von Bertalanffy growth equation and the two sets of estimated annual sur\ i\al rate. Two patterns stood out. as shown in Figure 9. The first pattern was a negative relationship between the coef- ficient of variation of the abundance of total snails and the mean of the abundance of recruits, with the position of the points for the rocky-shore species above and to the right of the mangrove species (Fig. 9a). On the rocky shore, N. irachoides and N. austndis had higher recruitment and the least variable population of snails, L. w CO c CO *^ •D (0 c o > c o "o o O Z^(J L. undulata • o L. cingulata 100 N. millegrana L. filosa N. australis n o L. scabra a/, trochoides 0 1 2 Mean log(recruits -i- 1) 1 — N. trochoides • Q^^ 20 (C ~ >- C/3 03 re N. australis • > c w3 10 -— c fR • — Annu of pa 0 N. L millegrana •'-■ undulata ... » r^°L. Cingulata filosa o* ° i_ scabra 0 12 3 Years to half asymptotic size Figure 9. Covariation of demographic and life history characteristics of three species of mangrove littorines and four species of rocky shore littorines at Ningaloo Marine Park, (al Overall, the coefficient of varia- tion of adult snails and mean log (recruits + ll had a significant nega- tive correlation (r = -0.731, with the correlation coefficient for rocky- shore species higher (r = -0.971 and positioned farther to the right than for the mangrove species (r = -0.70). (b) Estimates of annual rate of survival of painted snails overall had a significant positive correlation (r = 0.92) with the years to half asymptotic size. The correlation coef- ficients for the three species of Xodililloriiui and the four species of IJtIoniriti were similar (r = 0.95 and r = 0.94, respectively) as viere the correlation coefficients for the three mangrove species and the four rocky-shore species (r = 0,95 and r = 0,91, respectively). undulata had the fewest recruits and the most variable population, and N. millegrana was intermediate. The pattern for the mangrove snails was less obvious, with L scahra having more recruits and less variable populations than L flliisa and L cingulata. These data address the issue of supply-side ecology by pointing to the con- tribution of abundant recruitment in stabilizing the populations. The second pattern was a positive association between annual survival rate calculated from the recapture of painted snails and the predictions of years to half asymptotic size (Fig. 9b). There was a significant positive correlation involving all seven species, but it was unclear whether the important grouping of the points was associated with the genus of the snails or with the habitat in which they lived because L. undulata from the rocky .shore was directly in line with its congeners from the mangroves. DISCUSSION This decade-long study of eight species of littorines at Ninga- loo Reef revealed three important features of the population ecol- ogy of these species. The most conspicuous feature revealed was that, at the spatial and temporal scales used in this investigation and for the sets of species in the two habitats, the species' abun- dances varied largely independently, as indicated by the different patterns in the variance components (Fig. 2-5). Thus, even though all the species have planktonic larvae and our sites were spread over only 40 km of coastline, patterns of abundance of all snails and of recruits differed over space and time. There were, however, some conspicuous exceptions to this general feature In some com- ponents of the patterns. The abundances of L filosa and L scabra on the four groups of trees did vary in concert over time, but the patterns differed between these species. The differences in abun- dance of the rocky-shore littorines were dominated by differences among sites because of the consistently high numbers of all snails and of recruits at Mandu Mandu, where the four species also varied In concert. Finally, for total abundances, the variance component at the spatial scale of adjacent sampling units, represented by tree (group) and strip (site), was consistently the lowest for all species except for L scabra. where It was the greatest. However, even these patterns are inconsistent across species, so, overall, patterns of spatial and temporal variability are unique for each species. The second important feature involving abundances of these liltorines was the relation between supply of recruits and the vari- ability of the total abundances, especially for the rocky-shore spe- cies (Fig. 9a): The greater the number of recruits, the less variable was the number of snails. There is another aspect about the im- portance of the supply of recruits related to the source of the recruits. As can be seen In Figures 2 and 4, all the species except L. scabra. N. australis. and N. trochoides were absent from a group of trees or a site on at least one occasion. Because all the species recovered from being absent at a group or site, recruits must come from elsewhere, emphasizing the role of the planktonic larvae and the open nature of the populations. Furthermore, these absences of species from some sites at some times point out how misleading surveys at single sites and times could be about the distribution and abundance of these species. As a case in point, Reid (1986) spe- cifically indicated that Littoraria spp. were not found at Mangrove Bay, presumably on the basis of surveys when the snails were scarce. The third aspect about population ecology concerns the diver- sity of turnover rates exhibited by the eight species of littorines studied at Ningaloo Reef (Fig. 9b). The dynamics of populations of 402 Black and Johnson L. filosa and N. millei^nina operated much faster than other spe- cies, with fast rates of growth combined with low rates of survival suggesting an annual life cycle, as has been interpreted for other littorines (e.g.. Underwood & McFadyen 1983 for L acutispiru and Williams 1992 for L. mahae). N. nochoides was at the other extreme, with high survival and a slow rate of growth and probably being long-lived like L. neritoides (Hughes & Roberts 1981). On average, the mangrove species were more ephemeral than the rocky-shore species, but both groups vary considerably, and it is impossible to determine from our small sample and the ambiguous position of Z.. undiilata whether the patterns in Figure 9b are driven by habitat or generic relationship. To our knowledge, this is the first comparative study of demography in species of Littoraria and Nitdilittorina. The variability and the apparent patterns highlight the potential of these snails for testing general ideas about life- history evolution. ACKNOWLEDGMENTS This work was funded by the Australian Research Council and the Marine Sciences and Technology Grants Scheme. Jeremy Fitz- patrick, Michael Forde, Darcy Hebbert, Lou Johnson, Moss Johnson. Susan Johnson. Alan Kendrick, James Murray, Robyn Watts, and Kelley Whitaker helped with the censuses. The West- ern Australian Department of Conservation and Land Management provided accommodation in the Ningaloo Marine Park. LITERATURE CITED Behrens Yumada, S. 1989. Are direct developers more locally adiipled than planklonic developers? Mar. Biol. 103:403—11 1. Davies, M. S. & G. A. Williams. 1998. Epilogue. Hxclnihiologia 378:243- 246. Heller. J. 1990. Longevity in molluscs. Makicohii^ui 31:259-295. Hughes, R. N. & D. J. Roberts. 1981. Comparative demography of Lil- rnriihi ri(ili.s. L. nigroHneciUi. and L. neriundes on three contrasting shores in North Wales. J. Anim. Ecol. 50:251-268. Johnson, M. S. & R. Black. 1999. NodilUtorina nodosa (Gray, 1839) is a plastic morphotype of Nodilittorina aiisjrcilis (Gray. 1826). J. Moll. Stud. 65:111-119. McQuaid. C. D. 1996a. Biology of the gastropod Littorinidae \. Evolution- ary aspects. Oceauofir. Mar. Biol. Ann. Rev. 34: 233-262. McQuaid, C. D. 1996b. Biology of the gastropod Littorinidae II. Role in the ecology of intertidal and shallow marine systems. Oceiinogr. Mar. Biol. Ann. Rev. 34:263-302. Powell, E. N. & H. Cummins. 1985. Are moUuscan maximum lifespans determined by long-term cycles in henthic communities? Oecologia 67:177-182. Reid, D. G. 1986. The littorinid molluscs of mangrove forests in the Indo- Pacific Region: the genus Lillorariti. London: British Museum (Natural History). Reid, D. G. 1989. The comparative morphology, phytogeny and evolution of the gasttopod family Littorinidae. Phil. Trans. Roy. Soc. London B 324:1-110. Rosewater, J. 1970. The family Littorinidae in the Indo-Pacific. Part I. The Subfamily Littorininae. Indo-Pacific Mollusca 2:417-506. Siegel, S. 1960. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill. Tilman. D. 1989. Ecological experimentation: strengths and conceptual problems. In: G. E. Likens, editor. Long term studies in ecology: ap- proaches and alternatives. New York: Springer- Verlag. pp. 136-157. Underwood, A. J. & E. J. Denley. 1984. Paradigms, explanations and generalizations in models for the structure of intertidal communities on rocky shores. In: D. R. Strong, D. SimberlotT, A. B. Ahele & B. Thistle, editors. Ecological communities: conceptual issues and the evidence. Princeton: Princeton University Press, pp. 151-180. Underwood, A. J. & K. E. McFadyen. 1983. Ecology of the intertidal snail Littorina aculispira Smith. J. E.xp. Mar. Biol. Ecol. 66:169-197. Williams, G. A. 1992. The effect of predation on the life histories of Littorina ohtusata and Littorina marine. J. Mar. Biol. Assoc. UK 72: 403-116. Wilson, B. R. 1993. Australian marine shells. Kallaroo, Western Australia: Odyssey Publishing. JoKnial iij Slii'lljlsh Research. Vol. 20, No. 1, 403^09, 2UU1. PREDATION BY THE PILE PERCH, RHACOCHILUS VACCA, ON AGGREGATIONS OF THE GASTROPOD LITTORINA SITKANA ELIZABETH G. BOULDING.* DEBORAH PAKES, AND STEPHANIE KAMEL Dcpiirniicnt aj Zoology. Uiiin'r.sity of Giiclpli. Guclph. Ontario NIG 2WI. Canada ABSTRACT The pile perch, Rhacnchilus vacea (Embiotocidae). is abundant on the Pacific Coast of North Ameiica and is known to crush hard-shelled prey with its heavy pharyngeal teeth. We investigated whether pile perch predation has the potential to limit or regulate populations of Littorina .1 1 mm shell width). Indeed, some fish showed a significant preference for large snails {shell width >6.3 mm) over small snails (3.35 mm < shell width < 4.0 mm). Our 1998 observations with SCUBA during daytime high tides showed that the density of pile perch foraging in the intertidal averaged 0.1 19 (SE = 0.0205; n = 20) individuals per square meter (estimated fork lengths 5^0 cm). However, the intertidal distribution of pile perch that were actually consuming prey was highly aggregated. In the field, we investigated whether predation by this fish on snails deployed onto boulders was density dependent. The fish swam parallel to the shore and located and consumed 4()'/f of the patches of L. siikiiiui that we deployed (;? = 70) within 50 min. This foraging behavior resulted in density-dependent predation on the deployed snails in only one of the three tidal cycles (or 2 of the 12 days) of our experiment. We offer several proximate reasons for the low frequency of density-dependent predation found in this study and conclude that the pile perch may not be an important regulating factor for L. siikiiiui populations at this site at the present time. However, the very high predation rates we observed suggest that this predatory fish is an important limiting factor at this site. KEY WORDS: aggregative response, limiting factor, population regulation INTRODUCTION The factors that determine the distribution and ahiiiidaiice of marine invertebrates are still not well understood. Predation is known to be important in limiting the abundance of littorinid gas- tropods on rocky shores. Paine (1994) reviewed examples where predators had been excluded from rocky shores and concluded that such exclusions usually resulted in dramatic increases in the abun- dance of the competitive dominant. Many of the classic exclusion experiments he reviewed have involved slow-moving predators, such as gastropods or starfish. However, slow-moving predators are unable to move rapidly to high-density aggregations of their prey and are generally less tolerant to emersion than their prey (Newell 1970). As a result, they usually have a very local effect at the lower end of the intertidal distribution of their prey. In contrast, highly mobile predators, such as crabs and fish, can forage throughout the intertidal during a single high tide period and have been observed to feed more heavily on high-density aggregations of their prey (Boulding & Hay 1984, Behrens Yamada & Boulding 1996). Sinclair (1989) argues that only density-dependent factors can provide the negative feedback that can regulate a population at equilibrium. Consequently, highly mobile predators may be dis- proportionately important as regulating factors on rocky intertidal shores. Two highly mobile predators, crabs and fish, can prey very heavily on littorinid gastropods on wave-sheltered shores of the northeastern Pacific, and it is not clear which predator type is more important. McCormack (1982) deployed Littorina sitkana into fenced areas of the intertidal and observed predation rates of up to 54'7f per daytime high tide period, which she attributed entirely to predation by pile perch (Rlmcocliiliis vacca). Behrens Yamada and Boulding (1996) tethered L sitkana in the intertidal zone and found predation rates of up to 77% per high tide period. They *Cortespoiiduig author. H-mail: boulding@uoguelph.ca recovered diagnostic shell fragments that had been peeled by crabs from 40% of the dead snails of shell length 15-17.5 mm. Boulding et al. (1999) found predation rates on tethered L sitkana to be up to 25% per high tide period and used shell fragments to attribute at least 52% of the predation of their largest si/e class to predation by crabs. The daily rates of predation reported above are too high for any L. sitkana population to sustain given their known reproduc- tive rates (Behrens Yamada 1989). We hypothesize that these high predation rates represent a strong density-dependent response to local high-density patches of prey and present experiments here to test our hypothesis. Foraging excursions into the intertidal during high tide have been well documented for large subtidal crabs such as Cancer productus (Boulding & Hay 1984. Robles et al. 1989, Behrens Yamada & Boulding 1996, Boulding et al. 1999). There is experi- mental evidence that predation by this crab species can be density dependent (Boulding & Hay 1984). However, much less is known about the foraging behavior of the pile perch other than its extreme morphological specializations for crushing hard-shelled prey (Brett 1979). One of the major components consistently found in stomach analyses of pile perch are gastropod molluscs (Ellison et al. 1979. Haldorson & Moser 1979, Hueckel & Stayton 1982, Laur & Ebeling 1983. Stouder 1987). The only other surf perch that is common in the intertidal zone on the west coast of Vancouver Island and is known to eat gastropods is the striped perch {Enilnoloca lateralis) (Lamb & Edgell 1986, E.G. Boulding Pers. Comm.). Striped perch stomachs sometimes contain gastropod mollusks, but soft bodied inverte- brates such as amphipods, bryozoans, and isopods are much more prevalent (Haldorson & Moser 1979). The different feeding pref- erences of the two species are correlated w ith their mouth size and structure. The pile perch has extremely large pharyngeal plates, heavy blunt (pavement-like) pharyngeal teeth, and well-developed associated musculature for specialization in crushing and grinding (DeMartini 1969). The striped perch, however, has only moder- 403 404 BOULDING ET AL. ately si/ed pharyngeal plates; smaller, more pointed pharyngeal teeth; and moderately developed musculature and is specialized for feeding on large whole prey (DeMartini 1969). For these reasons, we decided to focus on determining whether pile perch predation on patches of L. silkami deployed onto a rocky intertidal shore was density dependent. MATERIALS AND METHODS Study Area Our experimental site was located on a wave-sheltered boulder shore in front of Bamfield Marine Station (48°50'. 125"08') in Barkley Sound, Vancouver Island. British Columbia, Canada. We chose the site because it was one of the sites {"Pebble Beach A," hereafter Pebble Beach) used by McCormack (1982) in her study of the role of pile perch in maintaining shore-level size gradients in L. sitkaini populations. Laboratory Feeding Experiments Our laboratory feeding experiments were conducted in round fiberglass tanks at Bamfield Marine Station. The experimental tanks were either 1.22 or 1.8.^ m in diameter and were filled with free-tlowing sea water to a depth of 0.914 m. The tanks were covered with plywood but were left open along a 20-mni crack to allow for low levels of light. The flow rate of seawater into the tanks was at least 1 l/s, and the water had a temperature of 1 0 ± 1°C and a salinity of 31S?f. The pile perch were collected by fishing with a hook baited with a whole mussel {Mytilus trossulusl Upon landing, the hook was carefully removed with pliers, and the fish was placed in a bucket of sea water and rapidly transported to 2.44-m diameter round holding tanks. The fish were then held for at least a week and maintained on a diet of M. rrDssulits. The total length of the fish collected ranged from 275-3.^() mm. L \itkaiui were collected from Seppings Island near Bamfield Marine Station. The large size-class of snails was alst) collected from nearby Second Beach. They were held in perforated contain- ers in free-tlowing sea water until used in the experiments. The size of the minimum diameter (shell width) of the snails was determined by passing them through a series of brass soil test sieves stacked so that the mesh size decreased from the top sieve to the bottom sieve. Three different feeding experiments were conducted using one fish in each experimental tank. In the first experiment (Consump- tion Amount) we offered each of four pile perch 100 extra-large (>6.3 mm shell width) L sirkaiiu placed at the bottom of the tank. After 24 h we recorded the number of live snails above, at, and below the water line and then removed the live snails and the shell fragments from each tank. This experiment was continued for nine replicate days. We chose not to use repeated measures analysis of variance (ANOVA) because otir primary objective was to compare the consumption rates of the different fish rather than trends over time. Instead, we used a one-way ANOVA with FISH as the factor and then used Bonferroni pairwise comparisons to determine whether some individual fish ate more than others. We used ver- sion 5 of the statistical package SYSTAT for this and all other statistical analyses (SYST.AT 1992). In the second experiment (Size-Selection) we offered each of four pile perch 100 extra-large and 100 small (3.35 mm < shell width < 4.0 mm) L. sitkaiw. After 24 h we recorded the number of live snails of each size-class above, at, and below the water line. We then removed the live snails and the shell fragments from each tank. This experiment was conducted for 6 days. To analyze the data we used three-way ANOVA with DAY. FISH, and SIZE as the factors and included only the interaction we thought most important. FISH x SIZE, in our model. One fish did not eat and was excluded from further analysis. In the third experiment (Consumption Rate) we opened the crack in the plywood cover to 10 cm and offered each of four pile perch 20 large L sitkaiw (4.7.5 mm > shell width < 6.3 mm). We then used a stopwatch to record the time it took the fish to crush each snail. We noted that only certain fish would feed while being observed. Field Transects In 1998 we deployed two .50-m transect lines at Pebble Beach (below the Fucks zone) parallel to the depth contours at 0.8 m and 1.1 m above 0.0 datum (Canadian Hydrographic Services). The transects were surveyed for pile perch and striped perch densities. Counts were done using SCUBA during high tides of 3.0 m or greater such that the upper transect (1.1 m) was in at least 1 .9 m of water. Two observers swam along the transect line, one on each side, for about 15 min recording fish densities and size classes within 2 m of their side of the line. Fish size-classes were divided into 5-cm intervals from a minimum length of 5 cm to a maximum length of 40 cm. A total of five surveys were done in July 1998. We did four similar transect surveys between June 28 and July 1, 1999, except that these transect lines were at 0,0 m and 1.3 m above 0.0 datum (Canadian Hydrographic Services). In 1999, we tested how good the divers were at estimating the total length of the fish by placing models of different sizes of fish underwater at distances from 2-6 m from each diver and asking her to estimate its total length. We estimated the abundance and mean size of mollusc prey at Pebble Beach by counting and measuring the molluscs in five lO-cm X 10-cm quadrats sampled at random along a tape measure placed on the beach, parallel to the water's edge, at low tide. The sampling was repeated at 1 .5 m, 2.2 m, 2.8 m, and 3,2 m above 0,0 m datum. Because only one L. sitkana was found in the quadrats, an additional search was done to find L. sitkaiui by walking along the entire beach at low tide and turning over rocks for a total of 2 h. The abundances were then compared with estimates of L. sit- kana density done by McCormack at the same site in 1981. Field Experiments In 1 999 we used SCUBA to deploy three high-density and three low-density replicate patches of snails each day at high tide for a total of 12 days. A container with either a high-density patch (/; = 50 snails) or a low-density patch of L. sitkana Oi = 5 snails) was chosen at random from the mesh bag carried by the divers. The divers opened the container, and the snails were placed on boulders along the 1.3-m depth transect line. We chose boulders that were flat on top, about 0.3-0.6 m in diameter and 0.3-0.5 m high, and cleared them of macroalgae. This clearing was done to reduce variance among replicates caused by the snails hiding more effec- Pile Perch Predation on Littorina sitkana 405 lively on some rocks than on others. After 40-30 iiiin the divers returned to the boulder, collected all the live snails and shell frag- ments from the boulder and surrounding area, and placed them in labeled containers. All shell fragments we collected from om experimental boul- ders could be attributed to predation by fish and not by crabs. In our laboratory experiments we noted that pile perch preying on snails immediately spat out small broken pieces of shell body whorl. However, they swallowed the columella and later excreted it in a mucus-coated pellet onto the tank floor. All shell fragments found in our field experiment were small broken pieces, and no columellas were seen. We also did not see pelleted fragments of columellas, probably because fish retreated to the subtidal before defecating. This is distinct from shell fragments formed as a result of predation by crabs, which typically have either an intact col- umella with a spiral gouge up the body whorl or broken pieces of columella (Zipser & Vermeij 1978, Lawton & Hughes 1985, Be- hrens Yamada & Boulding 1996). Upon returning to the laboratory we counted the number of replicate patches in which we found shell fragments. We also counted the percentage of live snails that we had deployed that were recovered. The data were analyzed with contingency table analysis and with ANOVA. Contingency Table .Analysis The different dates of the expeiiment were categorized mto three separate spring (or large) tide series: A, B, and C. Tide A refers to dives from June 28 to July 1, Tide B is from July 12 to July 14, and Tide C is from July 29 to August 3. 1999. We assumed predation by pile perch when we recovered shell frag- ments on or near the rocks where snails were deployed. First, tidal series A-C were coinpared with each other in a pairwise fashion to test for differences in the mean percentage of the deployed patches that experienced predation using Fisher's Exact Test. Tidal series B and C showed no differences in mean percentage predation per day: they were therefore lumped together. Differences in mean percentage predation between high- and low-density patches were then compared within the two new groups using Fisher's Exact Test. ANOVA We coinpared the mean percentage of snails we recovered in the high patches vs. the low patches across the 12 days using a factorial ANOVA model with date and treatment as the factors. We then used Fisher's Least Significant Difference test to deter- mine which pairs of means were significantly different. RESULTS iMharatory I'xpvrimcnts The four pile perch in the consumption experiment ate an av- erage of 33.3 ± 6.3 snails per day (mean ± SE). The largest fish (total length = 330 mm) ate significantly more snails {P < 0.006) than one of the smallest fish (total length = 280 mm). We ob- served that when the L. sitkana were placed in the bottom of the tanks they eventually crawled up to or above the water line, which made them inaccessible to the fish. We therefore reanalyzed the data considering only the consumption rate of the snails found at or below the water line but got very similar patterns. The fish ate 70 >. re ■a 60 ^ ^ 50 c (1) 40 re (0 30 ^ 0) 20 F 3 10 * ■ small snails D large snails a 25 cm 23.5 cm Figure ratory cantiv 21.5 cm Fish size 1. Size-selection for larger snails {Littorina sitkana) in the labo- consuniplion experiment, -indicates a fish which ate signifi- more large snails (Bonferroni pairwise comparisons, P < 0.05). 32-619f of the snails available with the largest fish again eating significantly more than the smallest fish iP = 0.033). Only the largest of the three remaining pile perch (total length = 330 inm) in the size-selection experiinent showed a significant preference for the extra-large size -class of L. sitkana over the smallest size-class (Fig. 1). There was a significant interaction between FISH and SIZE {P < 0.00 1 ) hut no significant effect of DATE {P = 0.319). The laboratory consumption experiment verified that the pile perch were very efficient predators on the snails, but only one fish would feed while being watched. One 250-mm fish required only 42.9 ± 8.02 sec (mean ± SE, ;; = 19) to consume each large L. sitkana (shell length 10 mm). The same fish became even more efficient at eating the snails by the next day and required only 19.2 ± 5.61 sec (mean ± SE: n = 20) to consume each snail. Field Transects The average density of pile perch in the intertidal at high tide was moderately low and was concentrated in the middle intertidal zone. In 1998, the average density of pile perch in the 0.8- and 1.1 -m transects combined was an order of magnitude higher than the average density of striped perch (Table 1 ). In 1999, the den- sities of pile perch and striped perch in the 1 .3-m transect were similar to those observed in the 0.8- and 1.1 -m transects in the previous year (Table 1). However, in 1999 we also surveyed a 0.0-m transect where we observed no striped or pile perch (Table TABLE I. Density of pile perch and striped perch in 1998 and 1999 in transects at Pebble Beach near Bamfield Marine Station. Mean Year Species density SE h" 1998" Pile perch ().ll9/m- 0.02(15 20 Striped perch ().()2().Vni- 0.0055 20 igggKc Pile perch 0. 1 24/ni- 0.0221 16 Striped perch ().()27,Vnr 0.00818 16 ■" Densities are taken from an average of 0.8 m and 1 . 1 m above 0.0 datum transects. ''Densities are taken Inim 1.^ m above 0.0 datum transect. *" No fish were seen at 0.0 m; n = 6. '' n is number of transects surveyed. 406 BOULDING ET AL. I ). Most of the fish that we observed in both years were small. In 1998 the average length of the pile perch was 12.0 mm, and the average length of the striped perch was 11.2 mm (Table 2). In 1999. the average length was slightly larger for both species, with the pile perch averaging 15.2 mm and the striped perch averaging 13.5 mm (Table 2). In 1998 most fish were small, but some pile perch were more than 30 cm in length (Fig. 2). We found that experienced divers were good at estimating the total length of model fish. All three divers were able to place the models into the correct 5-cm size-class 95% of the time. Our quadrat sampling of Pebble Beach estimated that the cur- rent densities of L. sitkana and other potential prey species for the pile perch were very low. The most common prey species found were the gastropods L. scutulata and Tegula funehralis. and even these were not abundant (Table 3). L. sitkana was especially rare, with only one found in the entire quadrat survey and only three more found after walking the entire beach. The density of mussels (A/, trossulus) was very low, but the abundance of barnacles (Bahi- iius i^laudula) was high. Field Experiments Contingency Table Analysis A total of 407(1 of all the patches deployed in this experiment (;i = 70) were found by fish within 50 inin. We found that the number of patches where we found shell fragments (i.e., predation) was significantly higher in the second and third tidal series than the first (A vs. B: P < 0.004. df = \.n = 41: A vs. C: P < 0.036, df = I, /I = 52). The percentage of patches in each tidal series where we found shell fragments is shown in Figure 3. The second and third tidal series were not significantly different from each other in the number of patches that experienced predation and were thus lumped together for analysis of treatment effects (B vs. C, P = 0.37, df = \.n =47). In the first tidal cycle, significantly more of the high-density treatments experienced predation (Table 4). In the second and third tidal series, there was no difference in the percentage of the patches that experienced predation between the high- and the low-density treatments (Table 4). The mean percent- age predation in the low- and high-density treatments is shown in Figure 3 for each tidal series. In all patches where shell fragments were found, an average of 81 ± 0.06'/f (/; = 26) of the snails deployed were eaten. ANOVA A significant DATE x TREATMENT interaction was found for the percentage of recovered snails (Table 5); therefore, we present TABLE 2. Mean size of pile perch and striped perch in 1998 and 1999 in transects at Pebble Beach near Bamfield Marine Station. CO > T3 in _c •D > V tl) o 15 «^ .o 120 100 80 60 40 20 0 - ■ ■ pile perch □ striped perch ll^L < = 5 cm 11-15 cm 21-25 cm 31-35 cm 6-10 cm 16-20 cm 26-30 cm 36-40 cm Fish size Figure 2. Size-distribution of the pile perch, Rhacnchilus vacca, and the striped perch, Embiutuea lateralis, seen in intertidal transects in 1998. only the interaction means (Fig. 4). We recovered a significantly lower proportion of snails from the high-density patches than from the low-density patches on 2 of the 12 days of the experiment (June 28 and July 13). We observed an inverse density-dependent trend on 3 days, but these were not significant (Fig. 4). The divers made some additional observations on foraging pile perch. In 1998, divers at Pebble Beach watched two 35-cm pile perch consume a patch of deployed snails and noted that they required only 9 min to consume a total of 17 snails. As previously mentioned, in 1999. divers left the patches for at least half an hour after they had deployed the snails. However, by the third spring tidal series they noticed that some fish began to circle the divers before all the snails were deployed. Although divers were not present during the entire time that the snail patches were deployed, any predation that was observed was always by pile perch. DISCUSSION This study shows that pile perch were the major consumers of the large L sitkana we deployed in the Pebble Beach intertidal during diurnal high tides. Striped perch were much less common in the field transects at Pebble Beach, and we did not observe striped perch eating the snails we deployed. The pile perch has previously been implicated as an important predator of littorinid gastropods on wave-sheltered shores of the Northeastern Pacific (McCormack 1982. Boulding et al. 1999). However, there has been some ques- tion about its importance because at some sites predation rates by TABLE 3. .\bundance and mean size (shell length) of mollusc prey present at Pebble Beach near Bamfield Marine Station' Tidal height" Prey species'^^ Mean density (m"-) ± SE Mean size (mm) + SE Mean size H Year Species (cm) SE n 1 0 0 + ^ 36 5.21 +6.44 50 1998" Pile perch 12.01 1.33 237 Tef^tiUi funehralis 1.6± 1.16 1 2.60 ± 1 .20 8 Striped perch 11.17 1.14 41 2.2 m Liltorina scutulata 15.6 + 4.09 5.77 ± 5.90 78 1999" Pile perch 1.^.2 1.63 99 2.8 m Littorina scutulata 8.4 ± 2.42 5.54 ± 1.96 42 Striped perch 13.5 3.06 T> 3.2 m Littorina scutulata 4.4 ± 1.80 6.09 + 0.50 ->-) ° Lengths are taken trom an average ol 0.8 ni and Mm above 0.0 datum transects. " Lengths are taken from 1.3 m above 0.0 datum transect. " A total of five quadrats were done per tidal height. "Tidal height above 0.0 datum '' Also one Littorina sitkana shell length 9.0 mm. Pile Perch Predation on Litturina sitkana 407 c V E (0 V n Q. 00 80 1 1 60 ■ low density D high density ABC Tidal series Figure 3. Percenta<;f of siiiiil patches deployed where shell fragments were found at Pebble Beach in hi(;h-densit> and low -density treat- ments for tidal series A (June 2S through July I, 1999). B (July 12 through July 14. 19991, and C (July 28 through August 3, 1999). *Sig- nitlcant differences between densities (Fisher's Least Significant Dif- ference lest, P < (1.05). pile perch have been estimated to be close to zero (Behrens Ya- mada & Boulding 1996). In our experiment the pile perch found 40% of the deployed patches of snails. Within the patches found hy the fish, the consuniption rate averaged 81%. This suggests that pile perch have the potential to severely limit the size of L. sitkana populations. We doubt that predation rates on the natural snail population on Pebble Beach are nearly so high because the popu- lation is currently at a very low density. However, natural popu- lations of L. sitkana do occur at higher densities at other locations {McCormack 1982), and pile perch predation rates on these popu- lations may be high. Adult pile perch exhibited a very high foraging efficiency when feeding on large L sitkana. The pile pei'ch consumed snails at a very high rate (19.1 sec per snail). The largest fish showed a size preference for large snails, which makes it unlikely that certain size-classes of L. sitkana have a size refuge from pile perch. In addition, large pile perch were moderately abundant in the inter- tidal zone at high tide at Pebble Beach and reached fish densities of 0.1 19/m-. This is an order of magnitude higher than the only previously reported densities for pile perch of 0.0039-0.0 l()9/nr in kelp beds off southern California (Ebeling et al. 1980). These data suggest that the pile perch could be an important limiting factor for the population of L. sitkana at this site. TABLE 4. Number of high- and of low-density patches of Uttorina sitkana deployed at Pebble Beach in 1999 that were recovered with and without shell fragments resulting from predation by pile perch. Tide series Patches Patches with without P fragments fragments df \alue' A,' High-density (50 snails) 0 Low-density (5 snails) 4 B*' & C^ High-density (50 snailM 10 Low-density (5 snails) 14 12 1 7 0.037 14 1 9 0.248 •"Trials performed from June 28 through July 1. 1999. ''Trials performed from July 12 through July 14. 1999. ' Trials performed from July 28 through August 3. 1999. '' Probability values are from Fisher's E.\acl Test for particular tide series. TABLE 5. Analysis of \ ariance of percentage of snails recovered." Factors are DATE (Date of Experimenl) and DENSITY (Density of Snails). Source SS df MS P value DATE 8.469 11 0.770 3.827 0.001 DENSITY 0.760 1 0.760 3.777 0.058 DATE X DENSITY 4.833 11 0.439 2. 184 0.032 Error 9.25? 46 0.201 0.201 " Percentage of live snails (Linorina sitkana) recovered after exposure to predation during one high tide period. Data were transformed with an angular transform before analysis. When a pile perch found a patch of deployed snails in our field experiment, it ate the majority of the snails. Density-dependent predation can result from a predator finding high-density patches of prey more often or from the predator spending more time in high-density patches once they are found (Krebs & Davies 1991). Consumption rates were high in all patches that were found sug- gesting that the fish were not spending more time in high-density patches than in low-density patches. Therefore, any density- dependent predation we observed probably resulted from the fish finding high-density patches more easily than low-density patches. No work has been done specifically on the searching behavior of the pile perch, but they are known to be visual predators that feed during the day (Ellison et al. 1979). Many other benthivorous fishes that feed diurnally are know to use vision almost exclusively to find prey (Keenleyside 1979). Our analysis by tidal series found significant density-dependent predation in the first tidal series (A) but not the second (B) and third (C). We doubt the lack of significance in tidal series B and C is due to low statistical power because sample sizes were smallest in tidal series A. Further)iiore. we also observed a low frequency of density-dependent predation when we analyzed our experiment with ANOVA, detecting it on only 2 of the 12 days. There may be several reasons why our study did not frequently detect density-dependant predation of L. sitkana by pile perch. In tidal series C, the divers noticed several large pile perch follow- ing them that consumed the snails as .soon as they were deployed. I low density IH high density June 28 June 30 July 12 July 14 July 29 Aug 2 June 29 July 1 July 13 July 28 July 30 Aug 3 Date Figure 4. Mean percentage of snails (Uttorina sitkana) recovered from the low- and high-density treatments on six different sampling dates. *Dates on which a significantly higher percentage of snails was recov- ered from the low-density treatment (Fisher's Least Significant Dif- ference test, /' < 0.(15). 408 BOULDING ET AL. This suggests that the fish hud become habituated to the divers and used their presence as a cue to find the patches of snails. This behavior may have made it difficult to detect density-dependent predation. Another explanation for the low frequency of density- dependent predation is that the high consumption rate of snails by pile perch in both the high- and the low-density treatments made it difficult to detect subtle density-dependent predation. Our ability to detect density dependence was further reduced by our inability to distinguish snails that had been consumed from those that had simply migrated away from the top of the boulder. A single miss- ing snail would affect the percentage recovered in the low-density treatment much more strongly than in the high-density treatment. For example, two snails migrating from the low density-treatment where a total of only five were deployed would result in a 60% recovery of snails, whereas two snails migrating from the high- density treatment would result in a 96% recovery. In addition, the low-density treatments may have been in fact high density when compared with the present ambient density of possible prey species at this site. Almost no L. sitkana were found in our quadrats at Pebble Beach. Two other gastropod species were present at Pebble Beach, L. scutulata and Tegula funebndis, but neither are preferred prey of pile perch. Adult Tegula are rejected by pile perch (McCormack 1981 ). whereas the L. scutulata present on this beach averaged only 3.21 mm in shell length and are likely too small to be profitable. We suspect that the intense predation by pile perch at Pebble Beach keeps the densities of their preferred prey very low. Indeed, the low densities of potential prey items found there in 1999 cause us to question why the fish are venturing in the intenidal at all. The pile perch may be feeding on the barnacles there as was observed by our divers in 1998. Barnacles were reported to make up 56-75% of the stomach contents of large pile perch foraging on an artificial reef off the Washington coast (Hueckel & Stayton 1982). Given the high predation rates we observed, we find it highly unlikely that any of the deployed snails would have survived if left out for another hour. The ambient density of prey may have been exceptionally low at Pebble Beach in 1999 that an unusually high proportion of the deployed snails were found and eaten. McCor- mack (1981) found much higher densities of L sitkana at Pebble Beach in 1980 ( 10 snails/nr at 1.5 m tidal height and 970 snails/m- at 2.2 m in 1980) than we did (Table 3). We hypothesize that, rather than being regulated at an equilibrium density, there may be large fluctuations over time in the density of this L. sitkana popu- lation. This study did not find predation by the pile perch on deployed patches of littorinid prey to be consistently density dependent. Consequently, we lack strong evidence that this predator has the potential to be an important regulating factor for the L. sitkana population at this site. Nevertheless, its high success rate at locat- ing patches, its high foraging efficiency, and the high mortality rates we observed for the deployed prey suggests that the pile perch is an important limiting factor for the snail population at this site. ACKNOWLEDGMENTS We thank M. Gray, T. Hay, and T. Lundrigan for their sugges- tions for improving the manuscript, T. Hay and E.N. Hay for collecting fish. L. Kusumo for technical assistance, S. Servant and S. Zaklan for SCUBA observations, the Director and staff of Bam- field Marine Station for field support, and the Huu-ay-aht First Nation for access to our study sites. Financial assistance was pro- vided by N.S.E.R.C. (Canada) research and equipment grants to E.G. Boulding. LITERATIRE CITED Behrens Yaniada. S. 19S'). Are direct developers more locally adapted than planktonic developers? Mtir. Biol. 103:403—11 1. Behrens Yaniada, S. & E. G. Boulding. 1996. The role of highly mobile crab predators in the intertidal zonation of their gastropod prey. J. Exp. Mar. Biol. Ecol. 204:59-83. Behrens Yamada, S. & E. G. Boulding. 1998. Claw morphology, prey size selection, and foraging efficiency in generalist and specialist crabs. J. Exp. Mar. Biol. Ecol. 220:191-211. Boulding, E. G. & T. K. Hay. 1984. Crah response to prey density can result in density-dependent mortality of clams. Can. J. Fish, .\quut. Sci. 41:521-525. Boulding, E. G., M. Hoist & V. Pilon. 1999. Changes in selection on gastropod shell size and thickness with wave-exposure on nonheastem Pacitlc shores. / E.\p. Mar Biol. Ecol. 232:217-239. Brett, J. R. 1979. Some morphological and behavioural adaptations of pile perch {Rhacochilus vacca) feeding on mussels (Mytilus cchilis). Can. J. Zool. 57:658-664. DeManinI, E. E. 1969. A correlative study of the ecology and comparative feeding mechanism morphology of the Embiotocidae (surfperches) and evidence of the family's adaptive radiation into available ecological niches. Wassmau J. Biol. 27:177-247. Ebeling, A. W., R. J. Larson, W. S. Alevizon & R. N. Bray. 1980. Annual variability of reef-fish assemblages in kelp forests off Sama Barbara. California. Fish. Bull. 78:361-377. Ellison. J. P.. C. Terry & J. S. Stephens, Jr. 1977. Food resource utiliza- tion among five species of embioticids at King Harbor. Califor- nia, with preliminary estimates of caloric intake. Mar. Biol. 52:161- 169. Haldorson. L. & M. Moser. 1979. Geographic paUerns of prey utUi- zation in two species of surfperch (Embioticidae). Copeia 1979:567- 572. Hueckel. G. J. & R. L. Stayton. 1982. Fish foraging on an artificial reef in Puget Sound. Washington. Mar. Fish. Rev. 44:38-44. Keenleyside, M. H. A. 1979. Diversity and adaptation in fish behaviour. In: D. S. Famer, editor. Zoophysiology and Ecology. Volume 11. New York: Springer-Verlag. Krebs, J. R. & N. B. Davies. 1991. Behavioral ecology: An evolutionary approach. 3rd ed. Blackwell Scientific Publications. Lamb. A. & P. Edgell. 1986. Coastal Fishes of the Pacific Northwest. Madeira Park, B.C.. Canada: Harbour Publishmg. Laur. D. R. & A. W. Ebeling. 1983. Predator-prey relationships m surf- perches. Env. Biol. Fish 8:217-229. Lawton. P. & R. N. Hughes. 1985. Foraging behavior of the crab Cancer pagioiis feeding on the gastropods Nucclla lapillus and Littorina lit- lorca: comparisons with optimum foraging threory. Mar. Ecol. Progr. Ser 27:14.3-L54. McCormack, S. M. D. 1981. The maintenance of shore-level size gradients in an intertidal snail (L. sitkana). Vancouver: M. Sc. Thesis Universtiy of British Columbia. Pile Perch Predation on Liuorina sitkana 409 McCormuL-k. S. M. D. 1982. The maintenance of shore-level size gradients in an intertidal snail (Littorimi sirkmui). Ot'colo^i^iii 34:177-183. Newell, R. C. 1970. The biology of intertidal animals. Kent, U.K.: Marine Ecological Survey Ltd. 555 pp. Paine. R. T. 1994. Marine rocky shores and community ecology: an ex- perimentalist's perspective. In: O. Kinne, editor. Excellence in ecology. Vol. 4. Luhe, Germany: Ecology Institute. 152 pp. Robles. C. D. A. Sweetham & D. Dittman. 1989. Diel variation of inter- tidal variation hy Cancer prodticnis in British Columbia. J. Nat. Hist. 23:1041-1049. Sinclair. A. R. E. 1989. Population regulation in animals. In: J. M. Cherrett. editor. Ecological concepts. Oxford: Blackwell. pp. 197-241. Stouder, D. J. 1987. Effects of a severe-weather disturbance on foraging patterns within a California surfpcrcli guild. J. Exp. Mar. Biol. Ecol. 114:73-84. SYSTAT for Windows. Statistics. Version 5 Edition. Evanston, IL: SYSTAT, INC., 1992. 750 pp. Zipser, E. & G. J. Vermeij. 1978. Crushing behavior of tropical and tem- perate crabs. J. E.\p. Mar. Bin). Ecol. 31:155-172. Jinaiial of Slwllfish Rc.seanh. Vol. 20. No. 1.411-414, 200]. RADULAR MYOGLOBIN AS A MOLECULAR MARKER IN LITTORINID SYSTEM ATICS (CAENOGASTROPODA) CARLOS BRITO,' PAULINHA LOURENCO,' ROBERTO MEDEIROS,' JOSE F. REBELO.' HANS DE WOLF,- KURT JORDAENS." AND THIERRY BACKELJAU" * ' Depaitiucut of Biology. University of the Azores. Riia da Mae de Dens 58, Apartado 1422. P-9502 Poiita Delgada. Azores. Portugal: 'Department of Biology. University of Antwerp (RUCA). Groenenhorgerlaan 17 L B-2020 Antwerp. Belgium; 'Royal Belgian Institute of Natural Sciences. Vaiilicrstraat 29. B-IOOO Brussels, Belgium ABSTRACT Radiilar myoglobin (Mb) was investigated in 2SX specimens of 10 littorinid species using vertical polyacrylamide gel electrophoresis (PAGE) and isoelectric focusing (lEF). Within the genus Lillorina the two most basal species, L slriuia and L. keenae. have Mb patterns that correspond to those of the genera Lituirana and Nodilittorina. while the sibling species L. sculututa and L plemi have identical Mb profiles that consistently differ from those of L. littorea. L. saxatilis, L. compressa and L circiina. In contrast to previou.s claims, Mb does not consistently separate the sibling rough periwinkles Lillorina saxalilis and L. arccma. These data suggest (1) that the NodilillorinalLiltoraria Mb profile in L. siriala is not unique within the genus Lillorina and therefore does not refute the assignment of L siriala to this genus, and (2) that L saiiiilata and L plena occupy a separate position compared to the other species of the subgenus Lillorina. This latter result supports the suggestion that L. sciilulala and L. plena may constitute a separate subgeneric taxon. Finally, the lEF Mb profiles of NodHiltorina hauaiiensis and Cenchriiis miiricalus were nearly identical to the Nodilillorinal Lilloraria Mb pattern. Yet. PAGE of Mb in Cenchriiis imiricanis suggests a tentative Mendelian polymorphism. It is concluded that littorinid Mb mav not be a useful marker to distinguish closely related species, but rather provides information on 'higher level' systematics. KEY WORDS: Caenogastropoda. isoelectric focusing. Litlonnidae. myoglobin, protein electrophoresis, systematics INTRODUCTION With only two published studies prior to 1998, radular myo- globin (Mb) has not been widely used in littorinid systematics and population genetics (Wiuni-Andersen 1970. Jones 1972). This is not unexpected since the genetic background of Mb variation in periwinkles remains obscure and controversial (Olabarria et al. 1998). Nevertheless, it has been shown that simple protein elec- trophoretic surveys of Mb can provide useful data for littorinid systematics and population genetic analyses (De Wolf et al. 1998. Medeiros et al. 1998. Olabarria et al. 1998). In this context. Me- deiros et al. ( 1998) observed that within the genus Lillorina there was considerable intra- and interspecific Mb differentiation. This allowed, for example, to separate the sibling rough periwinkles Lillorina (Nerilreina) sa.xatilis and L. (N.) arcana. In contrast, the Mb patterns of two Lilloraria species and three Nodililtorina spe- cies were almost, if not completely, identical. Interestingly, the Mb profile of L. striata was similar to that of Lilloraria spp. and Nodilillorina spp., but differed conspicuously from that of Lil- toriiia spp. This latter result could be interpreted in two. ways: either L. striata is not a Litloriiui or the Mb profile of L. siriala represents a plesiomorphic condition within Lillorina. The present contribution is a follow-up of the work by Me- deiros et al. (1998). In particular, we will: (1 ) test the reliability of Mb as species marker to differentiate between L. saxalilis and L. arcana. (2) compare the Mb profile of L striata with those of three other basal Lillorina species, which are supposed to represent its closest relatives (L. keenae. L. scutulata and L. plena) (Reid 1990, Reid 1996. Reid et al. 1996). and (3) evaluate the electrophoretic Mb monomorphism in Lilloraria and Nodilillorina (and related genera) by resolving Mb patterns in two additional species {Nodil- illorina hawaiiensis and Cenchriiis miiricalus). *Corresponding author. E-mail; brito(a'alf.uac.pt Throughout this article we will follow the taxonomy and no- menclature proposed by Reid (1989. 1996). We will use the ab- breviation 'L' for the name Lillorina. whereas the name Lilloraria will be written in full. MATERIALS AND METHODS Electrophoretic profiles of radular Mb (and other structural proteins in the radular muscle) were surveyed in 288 periwinkles representing 10 species (Table 1 ). After collection specimens were transported alive or in liquid nitrogen to the laboratory, where they were stored at -80°C. Sample preparation was as described by Medeiros et al. (1998) and adapted in order to reduce possible artificial Mb variation caused by oxidative denaturation (e.g. Di lorio 1981. Righetti 1983). Therefore individual radular tissue ho- mogenates were prepared by thawing frozen snails, crushing their shells and dissecting the radular muscle in cold distilled water. The radular muscle was then blotted on filler paper and homogenized in a 0.1% KCN (w/v) in 20% (v/v) glycerol solution, in a ratio of 20|xl solution per mg tissue. KCN converts Mb to cyanometmyo- globin, which is more stable and prevents denaturation to hemichromes (Atassi 1964. Di lorio 1981). Crude homogenates were subsequently centrifuged for 30 min at 27200 x g (15000 r.p.m.) at 4°C. The resulting supernates were stored at -80°C until used for electrophoresis. Vertical polyacrylamide gel electrophoresis (PAGE) was per- formed in 80 X 80 X 0.75 mm gels ('Mini Protean II' apparatus of Biorad) with a gel strength of 7% and using a discontinuous buffer system with Tris/HCI pH 9.0 as gel buffer and Tris/Glycine pH 9,0 as tray buffer (Backeljau 1989). Otherwise, PAGE procedures and conditions were as described by Medeiros et al. (1998). The pro- tocols of these authors were also followed to perform horizontal isoelectric focusing (lEF) in pH gradients 3-9 and 4-6.5. PAGE and lEF gels were stained for general proteins (including Mb) with Coomassie Brilliant Blue and specifically for Mb with a benzidine recipe (Medeiros et al. 1998). 411 4i; Brito et al. TABLE 1. List of littorinid species and collection sites screened for Mb variation Species Locality N Melarhaphe ueiiloides Pico. Azores 10 {Linnaeus, 1758) Cenchrilis muricatus Isla Margarita. Venezuela 15 (Linnaeus, 1758) Nodililtorina hanaiicn.sis Hawaii \5 Rosewater & Kadolsky. 1981 Littorina (Linilirtorinti) slriuia Sao Miguel. Azores 13 King & Broderip. 1S32 Pico, Azores 15 Terceira, Azores 15 Madeira 4 Liaorina (Ptanilillorina) San Simeon, CA. USA 15 keenae Rosewater, 1978 Leo Carrillo Beach, CA, USA 15 Morro Bay. CA. USA 15 Liltnnna iLiuoriint) plciut Leo Carrillo Beach. CA. 15 Gould. 1849 USA Morro Bay. CA. USA 15 Liltoiina {Littorina) .sciitulutci Morro Bay. CA. USA 4 Gould. 1849 Littorina (Neritrema) Port Bheal an Duin. Ireland 15 compressa Jeffreys. 1 865 Trebeurden. France 15 Littiiriua {Neritrema) arcana Ravenscar. UK 15 Hannaford Ellis. 1978 Rohin Hoods Bay. UK 15 Littorina {Neritrema) sa.xatilis Venice. Italy (type loc.) 15 (Olivi, 1792) Robin Hoods Bay. UK 15 Ravenscar. UK 15 Siio Miguel. A/ores 15 RESULTS All specimens of L. saxatilis. L. cinema and L. compressa had the same monomorphie Mb profile, both with PAGE (Fig. I ) and lEF (Fig. 2). These profiles corresponded with the L. saxatilis profile reported by Medeiros et al. (1998), while the alternative Mb profile, said to be typical of L. arcana (Medeiros et al. 1998: figs 2. 4), was not observed here. Both PAGE (not shown) and IFF revealed that L. scutulata and L. plena have identical Mb profiles, which resemble that of L. saxatilis, except for the fact that with lEF the whole L saxatilis Mb profile was slightly shifted toward a higher pH (Fig. 3). L. keenae. on the contrary, had a very different Mb profile which was shared with L. stiiata. Yet. with IFF L. striata revealed an addi- Fij;ure 1. Benzidine staining of PAGE profiles of radular Mb In L. arcana (a), L. compressa (cl and L. saxatilis (s). tional Mb fraction al a pi of about 4.6. w hich w as not observed in the other littorinids studied here (Fig 3). On the other hand. L striata and L. keenae showed a more or less strong Mb band at a pi around 6.3. This band was lacking in the subgenera Littorina and Neritrema. but was present in N. Iiawaiiensis. C. muricatus and M. neritoides. For the remainder the IFF profiles of N. ha- waiiensis and C. muricatus were similar to those of L. keenae and L. striata, although they did not show the Mb fraction at pi 4.6 of L. striata (Fig. 4). Finally, similar to the results of Medeiros et al. ( 1998). the Mb profile of A/, neritoides was indistinguishable from that of L. striata with PAGE (not shown), but appeared to be distinct with IFF (Fig. 3). Surprisingly, in contrast to the apparent Mb monomorphism of C. inuricatus revealed by lEF, PAGE of the same individuals yielded a tentative Mendelian polymorphism reminiscent of a mo- nomeric protein coded by a single locus with two alleles (Fig. 5). DISCUSSION The present analyses show that the alleged species specific Mb differentiation between L. saxatilis and L. arcana (Medeiros et al. 1998) is not foolproof, since the L. saxatilis profile also occurs in L. arcana. Hence, as Medeiros et al. (1998) screened only one population of L. arcana (from Great Castle Head near Dale Fort. UK) and did not apply the KCN protocol to reduce artificial Mb variation, it seems worthwhile to screen a new batch of animals from this population in order to confirm the existence of two electrophoretic Mb types in L. arcana. The fact that L. scutulata and L. pleiui have identical Mb pat- terns is not surprising, given the close relationship between both species (Mastro et al. 1982, Murray 1982, Reid 1996, Rugh 1997). It is however, interesting that the Mb profile of these two species differs from that of L. saxatilis. L. arcana. L. compressa. and L. littorea (this latter by inference from Medeiros et al. 1998). while A B cccsssss ssssaaaa sssaaccc ^SB? ^S^S ^W "•?■ ^ *55R 0^ "^^ ' "Sg* -«-^ <^ppHP*iP ^P'^ ^9*'^iMiii Figure 2. lEF profiles (pH 4-6.5) of radular Mb in L. arcana (a), L. compressa (c) and L. saxatilis (s), stained witii Coomassie Brilliant Blue (A) and benzidine (B). Myoglobin in Littorinid Systematics 413 upcas krn •4 .6 Figure 5. Benzidine staininj; of I'AfiE profiles of radular Mb in C. murkatus (m) and N. hawaiieitsis (111. Note the suggestive Mendelian- like variation in C. muricalus. 6 .3- Figure 3. Benzidine staining of IFF profiles (pH 4-6.5) of radular Ml) in L. arcana (a), L. compressa (c), L. keeiiae (k), M. iieritoides (nl, L plena (p), L. striata (r), /„ saxatilis (s). and L scutulata (u). The arrow indicates the special band at pH 4.6 in /„ striata; the triangle indicates the band at pH 6.3. the Mb profiles of L keenae and L. striala differ e\en more con- spicuously from the Mb profile of these four species. Instead the Mb profiles of L. keeiuie and L. striata are similar to those of Nodilittorina spp.. Littoraria spp. and Ct'iiciiritis sp. reported here and by Medeiros et al. (1998). These data are consistent with current taxonomic practice to place L striata and L. keenae in the sepai"ate monotypic subgenera Liralittorina and Plan ilittorina (Reid 1996). such that the Mb pattern of L. keenae could have been derived f)om that of L. striata by the loss of the supposedly auta- poinorphic Mb band at pi 4.6 in L. striata. The Mb profile of L. scutulata and L plena would then represent a synapomorphy dis- tinguishing both species from the other Littorina and Neritrema species, while Mb profiles of L. littorea. L. saxatilis. L. compressa and L. arcana would unite the subgenera Littorina and Neritrema. Finally, the Mb patterns reported by Medeiros et al. (1998) for L arcana (if correct). L. fahalis and L ohtn.sata may involve still further derived states. Given that in the consensus phylogeny of the genus Littorina (Reid 1996: Fig. 119) L. scutulata and L pleiui form an independent clade (making the subgenus Littorina para- phyletic). our Mb data support the suggestion that these latter two species may constitute a separate subgeneric group. Although this scenario fits the currently accepted phylogeny of the genus Littorina. the Mb patterns are also compatible with the alternative idea that L striata may be not a Littorina. Yet, in that case L. keenae. whose assignment to Littorina has never been challenged, would become the most basal branch of the genus, suggesting that a Nodilittorina/Littoraria-Wke Mb profile (like in L. striata and L keenae) is not a priori inconsistent with the genus Littorina. However, since electrophoretic mobilities are not reli- able to infer homology and/or identity, the present Mb data are essentially phenetic. Hence apparent similarities may not neces- sarily indicate common descent, particulaily not in a molecule like Mb which may be subject to functional constraints that may cause homoplasy. Nevertheless, the Mb data do suggest that a sequence analysis at the amino acid and nucleotide level could help to un- derstand littorinid phylogeny and functional ecology. Such studies have been done for abalones (Suzuki et al. 1997) and anaspids (Rinaldi & Ophir 1998) Finally, the IFF Mb data on N. hawaiiensis and C. iniuiccuus. add to the Mb monomorphism in the genera Littoraria and Noilil- ittorina (Medeiros et al. 1998) and extend this observation to the genus Cenchritis. However, they contradict the alleged species specific Mb variation in Littoraria and Nodilittorina as reported by Jones (1972). who used PAGE. Probably this discrepancy is due to technical issues and/or hidden Mb heterogeneity. Indeed, when we applied PAGE in C. inuricatus. we also detected variation that was not uncovered by IFF. Hence, combined with the observations of Medeiros et al. (1998) on the hidden PAGE differentiation in M. neritoides that was resolved by lEF. it is obvious that PAGE and IFF are complementaiy in the analysis of littorinid Mb variation. However, whether the C nniricatus patterns really involve a Men- delian polymorphism, needs further scrutiny in view of the con- fusing evidence on the genetic backgiound of littorinid Mb (Ola- ban-iaet al. 1998). In conclusion, the present data confirm Medeiros et al.'s ( 1998) claim that littorinid Mb are useful systematic markers that can provide the same kind of information as do the haemoglobins in freshwater snails (Bailey et al. 1986) and fishes (e.g. Basaglia & Callegarini 1987. Macaranas et al. 1996. Rizzotti & Gioppato 1999), or the haemocyanins in terrestrial gastropods (e.g. Symond- sen & Walton 1 994) and crustaceans (e.g. Mangum 1 996, Mangum & McKenney 1996). Nevertheless, littorinid Mb are less suited to differentiate closely related species, but on the other hand seem quite informative for higher level systematics. Unfortunately, the present knowledge on littorinid Mb is still far too scanty to fully exploit them for ecophysiological. population genetic, taxonomic and phylogenetic investigations. ACKNOWLEDGMENTS Figure 4. Coomassie Brilliant Blue staining of lEF profiles (pH 4-6.5) of ,V. hawaiieiisis (h). L. keenae (k), C. inuricatus Ini), and /„ striata (r). We are indebted to J. Grahame, P. Mill. C. Van Osselaer, C. Wilding, and B. Winnepenninckx for their help in obtaining speci- mens. The constructive comments of two anonymous referees gieally improved the manuscript. This research was supported by the MAST 3 program of the European Commission under contract number MAS3-CT95-0042 (AMBIOS) and by F.J.B.R. grant 2.0023.94. H.D.W. is postdoctoral fellow at the F.W.O. 414 Brito et al. LITERATURE CITED Atassi, M. Z. 1964. Properties of components of myoglobin of ihe sperm whale. Nature 202:496-498. Backeljau. T. 19S9. Electrophoresis of albumen gland proteins as a tool to elucidate ta.xonomic problems in the genus Arioii (Gastropoda, Pulmo- nata). J. Med. Appl. Matacoi l:29-il. Bailey. J. B.. E. H. Michelson & W. L. Paraense. 1986. Differentiation of the sibling species Bioniphalaria occidenlaUs and BioinphaUiriti tena- gopliila by the electrophoretic patterns of their hemoglobin. Mem. Inst. Oswaldo Cruz 81:319-322. Basaglia, F. &. C. Callegarini. 1987. Electrophoretic and isoelectrophoretic characteristics of hemoglobins of Italian ictalurids. Camp. Biochem. Physiol. B 86:269-271. De Wolf. H., T. Backeljau & R. Verhagen. 1998. Lack of significant esterase and myoglobin differentiation in the periwinkle. Lilloriiui striata (Gastropoda. Priisobranchia). Hydrohioloiiia 378:27-32. Di lorio, E. E. 1981. Preparation of derivatives of ferrous and ferric he- moglobin. Meth. Enzymol. 76:57-72. Jones, M. L. 1972. Comparisons of electrophoretic patterns of littorine snails of Panama: an attempt to define geminate species. XVIIe Con- gres International de Zoologie, Monte Carlo, 25-30 Septembre 1972. Theme 3:1-10. Macaranas, J. M., L. Q. Augustni & A. E. Eknath. 1996. Multiple hemo- globins in three tilapiine species of the genus Orochromis and in eight strains of O. niloticiis. Ai/iiucult. Res. 27:597-601. Mangum, C. P. 1996. Subunit composition of polymeric hemocyanins in the decapod crustaceans: differences between sibling species. Physiol. Zool. 69:568-585. Mangum, C. P. & A. L. McKenney. 1996. Subunit composition of the crustacean hemocyanins: divergence in incipient speciation. Biol. Bull. 191:33-41. Mastro, E., V. Chow & D. Hedgecock. 1982. Littorina scutulata and Lit- torina plena: sibling species status of two prosobranch gastropod spe- cies confirmed by electrophoresis. The Veliger 24:239-246. Medeiros, R., L. Serpa, C. Brito, H. De Wolf, K. Jordaens, B. Winnepen- ninckx & T. Backeljau. 1998. Radular myoglobin and protein variation within and among some littorinid species (Mollusca: Gastropoda), Hy- di-obioloi>ia 378:43-51. Murray, T. 1982. Morphological characterization of the Littorina scutulata species complex. The Veliger 24:233-238. Olabarria. C, J.-M. Timmemians & T. Backeljau. 1998. Electrophoretic heterogeneity within and between flat periwinkles (Mollusca: Gas- tropoda) along an intertidal transect at Ria Ferrol, northwest Spain. Hydrohiologia 378:11-19. Reid. D. G. 1989. The comparative morphology, phylogeny and evolution of the gastropod family Littorinidae. Phil. Trans. R. Soc. Land. B 324:1-110. Reid. D. G. 1990. A cladistic phylogeny of the genus Littorina (Gas- tropoda): implications for evolution of reproductive strategies and for classification. Hydrohiologia 193:1-19. Reid, D. G. 1996. Systematics and evolution of Littorina. London: The Ray Society. Reid. D. G.. E. Rumbak & R. H. Thomas. 1996. DNA, morphology and fossils: phylogeny and evolutionary rates of the gastropod genus Lit- torina. Phil. Trans. R. Soc. Land. B 351:877-895. Righetti, P. G. 1983. Isoelectric focusing: theory, methodology and appli- cations. Amsterdam: Elsevier. Rinaldi, A. C. & R. Ophir. 1998. Phylogeny of anaspid taxa as inferred from amino acid sequences of monomeric myoglobins. Lsrael J. Zool. 44:3-8. Rizzotti. M. & F. Gioppato. 1999. Fish haemoglobins: the order Clupei- fornies. Rev. Fi.sh Biol. Fish. 9:71-87. Rugh, N. S. 1997. Differences in shell morphology between the sibling species Littorina scutulata and Littorina plena (Gastropoda: Prosobran- chia). The Veliger 40:350-357. Suzuki, T.. M. Shirai & T. Furukohri. 1997. Molecular phylogeny of aba- lones inferred from cDNA-derived amino acid sequences of indoleam- ine dioxygenase-like myoglobins. Bull. Mar. Sci. Fish. Kochi Univ. 17:7-13. Symondsen. W. O. C. & M. P. Walton. 1994. Electrophoretic separation of pulnionate haemocyanins: a simple taxonomic tool. ./. Moll. Stud. 60: 351-354. Wium-Andersen. G. 1970. Haemoglobin and protein variation in three species of Littoriita. Ophelia ^■.Ibl-lli. Joiinuil oj Shellfish Research, Vol. 20, No. I. 4LS_422, 2001. HERITABII.ITY OF SHELL TRAITS IN WILD UTTORINA SAXATILIS POPULATIONS: RESULTS ACROSS A HYBRID ZONE MONICA CARBALLO,' CARLOS GARCIA," AND EMILIO ROLAN-ALVAREZ' * 'Di'partLiinciiUi de Biocjiimiica. Gcnetka e Immmologia. Facidtad de Ciencias. Universidad de Vigo. 36200 Vigo. Spain; -Departamento de Biologia Fundamental. Facidtad dc Biologia. Universidad de Santiago de Compostela. 15706 Santiago de Compostela. Spain ABSTRACT There is a hybrid zone in Linoiimi stixanlis from exposed Galician rocky shores (northwest Spain) where two ecotypes, ridged and banded and smooth and unhanded are found to be associated with distinct habitats and shore levels. On the mid-shore, however, the two ecotypes and a variable percentage of hybrids can be found in sympatry. Furthermore, many shell traits present a clinal distribution along the vertical environmental gradient within ecotype, although very little knowledge about its genetic basis is available. This species is viviparous, and every female has a brood pouch containing many embryos. This makes it possible to use the shelled embryos from the brood pouch of every female to estimate the heritability of morphological traits (by full-sib correlation and offsprine-mother regression). We took seven samples of 20 females across two vertical transects and found significant heritabilities for most shell traits and a multivariable factor score. The average narrow -sense heritability was 0.102. and the average broad-sense hentability was 0.489. although the present estimates can be somewhat biased. The high correlation observed between the genotypic variance (estimated by full-sibs correlation) and the additive v;u-iance (estimated by offspring-mother regression) as well as the similar results obtained after transforming the variables correcting for scale effects support the existence of a common additive genetic component in both kinds of heritability. Genetic variability significantly differed between shore levels. Interestingly, hybrids did not show an increase in genetic variability with respect to both pure ecotypes. which perhaps suggests that natural selection is eroding their variability. These results complement a previous claim about the adaptive shell polymorphism in Galician L. scixatilis populations KEY WORDS: natural selection, hybridization, clinal variation, genetic variability, shell pohmorphism. embryo shell INTRODUCTION Lituiiiiw sa.xatilis (Olivi) is a polymorphic species from North Atlantic intertidal shores (Fretter & Graham 1980). The species inhabits a wide range of different ecological niches, from estuaries to rocky shores or even salt marshes (Raid 1996). Coupled with these habitat differences there is a known morphological and be- havioral population differentiation, which has been traditionally interpreted an as example of adaptive polymorphism (Janson 1983, Reid 199.^. reviewed in Reid 1996). However, to be adaptive, a polymorphism must be first heritable (Endler 1986), which is usu- ally a priori assumed in natural populations without further e.xperi- mental evidence. A few studies have tried to estimate the propor- tion of genetic vs. phenotypic variability by rearing families in the laboratory (Johannesson & Johannesson 1996, Ward et al. 1986, Ward et al. 1991, Warwick 1983, Warwick et al. 1990), but this approach is troublesome, very time consuming, and, for .some ecotypes (sensu Turesson 1922), very difficult (see Warwick et al. 1990). Yet, Newkirk and Doyle (1975) have suggested applying classical quantitative genetics inethods to estimate genetic vari- ances for quantitative traits in wild populations of this and other species with similar life-history characteristics. Because L saxa- tilis females usually carry a brood pouch with many embryos (Fretter & Graham 1980, Reid 1996), pregnant wild females can be used as different families of full sibs, from which full-sib corre- lation and offspring-mother regression allow us to estimate the genetic variances for different shell traits (Falconer & Mackay 1996). Newkirk and Doyle (1975) found significant genetic dif- ferentiation between L sa.xatilis populations allocated along an estuarine gradient. They found the lower genetic variability in those traits and population more affected by natural selection. They suggested that their polymorphism was maintained by natu- ral selection. *Corresponding author. E-mail: roUm@uvigo.es On exposed Galician rocky shores, two sympatric ecotypes can be found associated with two different habitats (Johannesson et al, 1993): The ridged and banded form (RB) is usually found on upper shores among barnacles, whereas the smooth and unhanded torm (SU) is typically found on lower shores among mussels. On the mid-shore, both barnacles and mussels overlap forming a patchy habitat in which both pure forms and some (5^0%) phenotypi- cally intermediate forms (HY) can be found even copulating to- gether. Every ecotype (RB, HY, and SU) partially aggregates in the mid-shore, showing particular preferences for different micro- habitats (Otero-Schmitt et al. 1997). In fact, they do not mate randomly on the mid-shore, showing a noticeable but still incom- plete assortative mating among ecotypes (Johannesson et al. 1995). The two ecotypes differ for many morphological, physiological, and behavioral traits (Erlandsson et al. 1998. Johannesson et al, 1993, Johannesson et al. 1997, Rolan-Alvarez et al. 1996, Rolan- Alvarez et al. 1997). Particularly, they show striking differences in shell si/e, morphology, and color polymorphism. Besides, this sharp ecotype distribution (an ecotone sensu Endler 1986), a more typical clinal distribution (sensu Endler 1986) across the shore gradient, can be found for many morphological traits within an ecotype (Cruz 2000). A genetic basis behind the presence of shell bands and ridges has been shown because females of each morph produce progeny of the saine characteristics under laboratory con- ditions (Johannesson et al. 1993). However, there is no such proof for other quantitative traits or for quantitative estimates of the relative importance of genetic variability behind other shell differ- ences. This information is fundamental if we want to understand fully the evolutionary forces that have modeled the observed eco- tone and clines in shell traits along the environmental gradient. We used the Newkirk and Doyle (1975) approach to estimate the heritability for different shell traits in wild populations of the Galician hybrid zone. We analyzed replicated transects along the vertical gradient and compared the results by full-sib correlation and mother-offspring regression. Because we are studying the 415 416 Carballo et al. trends in genetic variability on a zone where hybridization occurs (Johannesson et al. 1993. Rolan-Alvarez et al. 1996), we would expect that hybrids from mid-shore present some increase of their genetic variance due to hybridization for those traits being geneti- cally differentiated between upper and lower shore populations. This latter null hypothesis assumes that the hybrid populations are stable and fertile (Johannesson et al. 199.')). However, the results seem to contradict this expectation, suggesting that natural selec- tion may be eroding the genetic variability in the hybrids or, al- ternatively, that the assumptions of such a model are false. MATERIALS AND METHODS Sampling Adult L. sci.mtilis were sampled in June 1998 from two differ- ent localities. Centinela and Seni'n (separated by 3 km). They were picked along two vertical transects in intertidal rocky shore areas and five different samples (quadrats) within the transect (Fig. 1). Upper shore and upper mid-shore samples and lower shore and lower mid-shore samples covered the whole range of pure RB and SU populations, respectively. Upper mid- and lower mid-shore samples were closer to mid-shore samples (1-3 m) than to their respective upper or lower shore samples (4-10 m). In each sample, 100 snails (larger than 3 mm) within a 30-50 cm diameter circle were obtained and stored at -70°C. Later, they were randomly chosen from the sample, dissected, and sexed. The first 20 females with embryos within their brood pouch were used for the morpho- logical study. The mid-shore sample included similar frequencies of RB. SU. and HY individuals. However, due to the iow fre- quency of natural hybrids, it was not possible to obtain 20 pregnant HY females per sample (see Fig. 1). To avoid an unbalanced experimental design, hybrid data were excluded from some analy- ses (see below). Finally, 120 females (families) and 360 embryos of RB. 120 females and 360 embryos of SU. and 24 females and 72 embryos of HY were employed. Morphological Variables The shell of the female and the three largest embryos within their brood pouch were placed in an apertural top view and digi- tized to make several measurements (Fig. 2): shell height (SMI), aperture height (SM3), aperture width (SM4), and width of first spire whorl (SM7) following Cruz. Carballo and Rolan-Alvarez (submitted); distance from the center of the first whorl to the apex (Y) and ratio of the first whorl (R) following Newkirk and Doyle ( 1975): and top shell diameter (SD) and top aperture width (AW) following Rolan-Alvarez et al. ( 1996). Adult RB individuals were measured with an error of 0.0096 mm (eight magnifications re- quired), adult SU individuals were measured with an error of 0.0042 (12.5 magnifications required), and both RB and SU em- bryos were measured with an error of 0.0016 mm (80 magnifica- tions required) by a Leica MZI2 stereoscopic microscope. We statistically analyzed the parameter T ( Y/R) and S (SM4/SM3). the latter with small modifications from Newkirk and Doyle (1975). We also used the distance from the end of first whorl to the apex (SM15) and the parallel to the maximum aperture length (SM16). Genetic analyses were accomplished on individual variables but also on a multivariable factor score obtained by principal compo- nent analysis. Classical statistical analyses and multivariable meth- ods were done by the SPSS/PC package (version 9.0. 1 ). Genetic Analysis The indi\ idual variables from embryo data were analyzed with a two-way nested analysis of variance (ANOVA) including the factor shore level (fixed; upper, upper-mid, lower-mid, and lower), the factor localin- (random; Senfn and Centinela). and the factor faiiulx (random; nested within the shore level x locality interac- SHORE LEVEL: Centinela Upper shore Upper-mid shore Mid shore Low-mid shore Lower shore 20 RB 20 RB Senin 20 RB 20 RB 20 RB, 20 SU and 1 1 HY 20 RB, 20 SU and 13 HY □ □ C3 Barnacle Belt Patcliy Mid Shore Mussel Belt •yyi-mW-:-:-:':': y>''yMW-:>><:< 'Z(ym>>y >>><>> 15-25 m 3 Km Figure 1. Sampling design. Every sample (quadrat) represents 20 adult females and the three largest embryos within their brood pouch, except in mid-shore samples where we tried to obtain 20 ridged and banded (RBI, smooth and unhanded (SU), and phcnot>picall\ intermediate (HY) specimens. Upper shore samples were obtained on the barnacle belt, and lower shore samples were obtained on the mussel belt. Mid-shore samples where obtained in the patchy barnacle/mussel zone. Estimating Genetic Variances 417 Figurt 2. The morphological measurements studied on a smooth and unhanded (SU) individual. SMI. shell height; SM3. aperture height; SM4, aperture width; SM7. vvidth of the first spire whorl; SMI5. distance from the end of first whorl to the apex: SM16, the parallel to the maximum aperture length; R and Y following Ncwkirk and Doyle (19751; SD, top shell diameter; AW, top aperture width. tion) on upper and lower shore samples. A second nested ANOVA was performed on mid-shore samples with the factor )norph (fixed; RB and SU). the factor locality (random; Seni'n and Centinela), and the factor family (random; nested within the iiinrph x loculirv interaction). We assumed that embryos from the same brood pouch were true full sibs following Newkirk and Doyle ( 1975). These authors analyzed 1 2 independent groups of brood pouch embryos for an esterase polymorphism (with four alleles) and found that the em- bryos were full sibs in all cases except one (but see Newkirk & Doyle 1979). This assumption, if false, implies an underestimation of the heritability (it could be up to twice the observed amount). To minimize the risk of heterogeneity in progeny we used the three largest embryos (similarly sized) in every female. Classical genetic analyses of full sibs (ANOVA model I) and offspring-mother regression were done following Falconer and Mackay (1996). The covariance between full-sib is one half of the additive variance plus one quarter of the dominance variance plus the variance due to the common environment, whereas the covariance between off- spring and mother is exclusively one half of the additive variance. The former relationship can be used to obtain an estimate of the broad-sense heritability. whereas the latter allows us to obtain an estmiate of the narrow-sense heritability. However, complex in- teractions and maternal effects may somewhat bias these estimates. For example, the phenotypic value of the mother for the character in question can directly influence the same character in the off- spring. This effect can produce biases simultaneously in both full- sib correlation and offspring-mother regression. Another possible source of bias is the genotype-environment interaction, due to some environments causing intrinsically larger genotypic \ariabil- ity than others. Snails from different shore levels usually have different aver- age sizes (Johannesson et al. 1995). We have accomplished dif- ferent statistical corrections to minimize this scale effect. First, we have preferentially used coefficients of variation rather than the heritability to compare results from different methods and samples (Sokal & Rohlf 1995, Houle 1992, Falconer & Mackay 1996). Furthermore, we have incorporated a transformation that corrects the scale effects between samples, using the linear regression of the mean against the standard deviation in every sample as sug- gested by Falconer and Mackay ( 1996). The transformation is X = log (X + a/b). where X is the transformed variable, x is the original variable, and a and b are the intercept and the slope of the linear regression of the mean on its standard deviation. The comparison between mother and offspring measurements may be biased by other causes because the analyses assume that the two sets of data are obtained in individuals of the same age and for the same trait. The former is not true in our case, causing at best a scale effect (allometry) between mother and offspring. To avoid this, we repeated the analyses using the transformation suggested by Lynch and Walsh (1998) to correct for allometry with respect to size. The transformation uses the algorithm Y = a*X'', where a and b are the intercept and the slope, respectively, of the regression of the trait w ith the shell height. The residual of such a regression for every data point can be used to study the variation of the trait corrected for allometric effects (Packard & Boardman 1987). RESULTS The two-way nested ANOVA for individual variables pre- sented in Table 1 show that the factor shore level was significant for most traits, whereas the factors locality and the interaction were significant only in a few cases. The factor /«m//v was sig- nificant in all cases, which may perhaps be due to genetic effects (see also below). In most cases we tried to correct for heterosce- dasticity by using the square root or logarithmic transformations, succeeding for SMI, SM4, SD, and S. Therefore, for other traits, type 1 error must be taken with caution. Similarly, we found dif- ferences in the factors inorpli and interaction for most traits in mid-shore samples. Again we used square root and logarithmic transformation to correct for heteroscedasticity. being unsuccessful only with S. Similar results were obtained with the transformation 418 Carballo et al. TABLE 1. ANOVA on embryo measurements for upper/lower and mid-shore data sets. SMI SM3 SM4 SM7 SM15 SMI 6 SD AW T S Upper/lower Shore Shore level 6.78*** 30.30*** 23.31*** 4.39** 24.80*** 0.05 11.40*** 59.85*** 7.99*** 4.28** Locality 1.18 0.34 18.59*** 0.89 0.78 1.83 0.43 0.04 76.05*** 80.68*** Interaction 1.42 1.59 1.31 2.08 6.46*** 1.72 0.43 2.81* 9.61*** 12.75*** Family (interaction) 3.94*** 4 95*** 3.04*** 4 24*-^^ 2.15*** 3.19*** 5.89*** 3.49*** 1.59*** 1.52** Mid-shore Morph 13.54*** 38.13*** 30.96*** 0.00 13.16** 5.96* 20.79*** 70.64*** 13.37 5.72* Locality 3.50 0.81 6.08* 4.35* 18.28*** 0.78 2.57 2.56 1.38 20.52*** Interaction 3.05 2.05 1.05 8.42** 4.71* 5.69* 5.03* 1.95 0.14 2.19 Family (interaction) 5.16 * 4.88*** 5.02*** 4.37*** 2 7*** 3.94*** 5.96*** 1.80*** 1.04 0.95 Abbreviations: SMI = shell height; SM3 = aperture height; SM4 = aperture width; SM7 = width of the first spire whorl: SMI 5 = drstance from the end of the first whorl to the apex; SMI 6 = the parallel to the maximum aperture length; SD = top shell diameter, AW = top aperture width; T = parameter (Y/R); S = SM4/SM3. In the upper/lower data set the fixed factor is shore level (upper, upper-mid. lower-inid. and lower shore samples), whereas in the mid-shore data set the fixed factor is morph (ridged and banded vs. smooth and unhanded). In both cases the ANOVA incorporate the random factor localin (Senfn and Centinela), the corresponding interaction, and the nested factor /«mi7v (within the interaction). The analyses used 20 families per sample and three embryos per family. Hybrid families were excluded to avoid unbalanced designs. of Falconer and Mackay ( 1996) (coiTecting tor scale effects among samples) as well as traits divided by shell height (not shown). This suggests extensive morphological differences across shore levels and morphs, as well as among families, for most shell traits. The genetic coefficient of variation and the heritability are presented for the measured variables across the en\ ironmental gra- dient in Table 2 for both full-sib and offspring-mother regression methods. The averages for the two localities are presented, with the significant cases designated by lowercase letters (with signifi- cant cases determined after sequential Bonferroni corrections un- derlined) in Seni'n (a) and Centinela (b). respectively. Furthermore, estimated heritabilities were much larger in full-sib analyses (broad-sense heritability) than in offspring-mother analyses (nar- row-sense heritability). This may be caused by some contribution of maternal plus dominance effects to the full-sib resemblance but not to the offspring-mother resemblance. A more useful statistic to compare genetic variances between samples and methods may be the coefficient of variation, which partially corrects for scale ef- fects. Interestingly, the coefficient of genetic variance (in full sibs) and the coefficient of additive variance (in mother-offspring re- gression) were highly correlated for most shell traits across shore levels (Table 2), suggesting a common component (probably the additive) in both kinds of estimates. Besides, the estimates seem to fall into two categories, upper or upper mid-shore samples (typical RB samples from upper shore), with very low coefficients of ge- netic variation and samples from mid- to lower shore with larger coefficients of genetic variation (Table 2). Scale effects may not cause these differences because those populations with the small- est variances had the greatest means for all measured variables. On the mid-.shore another trend emerged: The coefficient of genetic variation of hybrids is never clearly larger than the corresponding estimates for both RB and SU ecotypes (Table 2). The values for hybrids are usually smaller or intermediate between pure ecotypes, which is exactly the opposite of that we would expect under a null hypothesis of hybridization for neutral genetic variability. The transformation of Falconer and Mackay ( 1996) for full sibs and the transformation of Lynch and Walsh for offspring-mother regres- sion, which corrected for scale effects among samples and for embryo-adult allometry. respectively, rendered similar results in all variables: significant heritabilities for some traits and the same two trends mentioned above. We could successfully reduce all the variables by principal component analysis of females. The first principal component (PCD represented 79.9% of the overall variability. The relation- ship between the two main components (PCI and PC2) is repre- sented in Figure 3 for all females included in the analyses. The PCI acts like a discriminant function between RB and SU morphs, having a clear biological meaning (a summary variable for all the morph differences). Figure 4 shows the coefficient of genetic vari- ance (estimated from full-sih correlation) and the coefficient of additive variance (estimated from offspring-mother regression) for PCI across shore levels. The estimates across sampling points were highly correlated in = 14; /■ = 0.832, P = 0.000). which suggests that there is an important component of the covariance between relatives (probably the additive) in common (Fig. 4). Fur- thermore, this relationship was maintained when we used the Fal- coner and Mackay (1996) transformation to correct for scale ef- fects among samples Ui = 14: ;• = 0.863. P = 0.000). This supports the differences between shore levels in genetic/additive variances, although at present we cannot exclude some minor bias due to genotype-environment interaction or maternal effects. Again, the genetic variance estimates by full-sib correlation were larger than the additive variances estimated by offspring-mother regression, suggesting that maternal or dominance effects signifi- cantly contribute to the resemblance between relatives in the former. Interestingly, we observe the same two trends discussed above (Fig. 4) that upper and upper-mid populations presented significantly smaller genetic variances and that hybrids did not show an increase of genetic variances as would be expected under neutral hybridization. DISCUSSION Many studies have repoiled morphological, behavioral, and physiological polymorphism in littorinids (reviewed in Reid 1996). In some cases, laboratory experimentation, computer simu- lations, and natural studies have shown that natural selection may account for some of this polymorphism (Strushaker 1968, Janson Estimating Genetic Variances 419 TABLE 2. The coefficient of genetic variation and the broad-sense heritability (in parentheses) b> full-sib correlation and the coefficient of additive variation and the narrow-sense heritability (in parentheses l b) offspring-mother regression for different traits across the environmental gradient (shore level). Analysis Level VIorph SMI SNL3 SM4 SM7 SM15 Offspring-mother regression Upper RB 9.06 (-0.004) 6.20 (0.000) 5.74 (-0.011) 7.90(0.001) 26.61 (-0.003) Upper- mid RB 6.15 a (0.019) 5.69 a (0.015) 5.91 (0.014) 7.53 a (0.031) 23.90(0.016) Mid RB 11.78(0.034) 11.00(0.020) 10.67 (0.023) 13.91 (0.067) 32.82 (0.004) Mid HY 11.06(0.028) 10.34(0.011) 9.84 (-0.006) 12.23(0.064) 27.67 b (0.055) Mid SU 1 1.22 a (0.066) 10.03 a (0.065) 9.66 (0.058) 14.97 a (0.107) 38.64 (0.024) Lower- mid su 13.19ab(0.122) 12.02 a b (0.138) 11.11 ab(0.118) 15.89 a b (0.155) 35.42 (0.040) Lower SU 9.88 a (0.063) 9.43 a (0.077) 10.51 (0.023) 13.36(0.075) 39.21 (0.029) Full-sib correlation Upper RB 12.46 a b (0.308) 8.02 a b (0.382) 7.47 ab (0.398) 1(1,02 a b (0.422) 37.57 a (0.244) Upper- mid RB 8.01 ab (0.374) 7.43 ab (0.370) 7.53 ab (0.411) 10.38 b (0.272) 34.99 b (0.191) Mid RB 13.19 ab (0.666) 12.40 a b (0.650) 1 1.94 ab (0.667) 16.62 ab (0.526) 41.92 a^ (0.407) Mid HY 12.54 a b (0.625) 12.11 a b (0.555) 1 1.08 ab (0.639) 15.39 a b (0.438) 41.33 b (0.224) Mid SU 13.59 ab (0.503) 12.15 ab (0.480) 11.81 ab(0.481) 17.99 ab (0.513) 52.76 a b (0.300) Lower- mid su 14.90 ab (0.642) 13.51 ab(0.647) 12.51 ah (0.644) 18.29 ab (0.582) 49.74 b (0.241) Lower SU I2.30a(0.401) 11.50 a (0.436) 13.27 a ((1190) 16.14ab(0.509) 49.83 a b (0.416) Correlation between methods 0.960'* 0.992** 0.928** 0.990** 0,888** TABLE 2. Continued. Analysis SM16 ^ R SD AW T S Offspnng-molher regression 6.42(0.006) 9.38 b (0.017 7.91 a b (0.040) 5.07(0.006) 6.71 (-0.(108) 6.27(0.160) 3.86(0.154) 6.46(0.015) 8.61 (0.012) 6.46 a (0,031) 5.32 a (0.027) 7.80(0.018) 7.16(0.105) 3.17(0.223) 11.14 b (0.054) 12.44(0.013) 18.46 b (0.072) 9.95 (0.039) 9.28 (0.020) 7.80(0.156) 3.46 (-0.152) 8.43(0.019) 7.64 (0.042) 20.66(0.072) 8.88 (0.020) 9.15 (-(.1.011) 7.58(0.219) 3.44 (0.046) 11.18(0.036) 7.44 10.024) 22.29 b (0.088) 9.69 (0.062) 8.70(0.039) 5.85(0.041) 3.38 (-0.240) 12.42 b (0.100) 16.52 ab(0.085) 13.63 ab(0.140) 1 1.12 a b (0.1 18) 1 1.13 b (0.061 ) 8.94 (-0.173) 4.89 (0.093) 8.04 b (0.060) 12.82(0.0681 11.04 a b (0.126) 8.74 a (0.078) 10.94 a (0.072) 9.23(0.195) 8.35 a b (0.040) Full-sib correlation 9.05 b (0.251) 13.40 a b (0.22 9) 10.60 a b (0.327) 6.48 ab (0.405 1 10.75(0.071) 10,25 (0.(J89) 5.83 a (0.1 75) 9.16 a b (0.240) 13.68(0.088) 10.19(0.101) 6.50 a b (0.482 ) 11.09 a (0.234) 11.20(0.124) 4.99(0.119) 13.66 a b (0.476) 17.44 ab (0.446) 17.35 ab (0.57; ) 1 1.03 aj. (0.684) 12.92 a (0.357) 12.36 a (0.097) 6.461-0.054) 11.54 b (0.296) 13.31 ab(0.441) 17.45 a b (0.549) 9.91 ab (0.661 ) 12.17 b (0.327) 11.59 b (0.186) 6.24 (-0.022) 13.61 ab (0.487) 14.76 a b (0.391) 17.62 ab (0.518) 1 1.73 ab (0.568) 13.58 b (0.167) 11.30 (-0.089) 6.07 (-0.061) 14.48 ab (0.565) 21.31 ab(0,390) 15.81 ab (0.595) 12.16 ab (0.717) 13.46 ab (0.521 ) 13.16 a (0.201) 6.91 a (0.241) 10.89 a (0.295) 18.44(0.219) 13.39 ab (0.488) 10.46 a (0.447) 13,07 a (0.460) 13.48 b (0.187) 10.27(0.148) Correlation between methods 0.984** 0.792** 0.818** 0.989** 0.932** ().73(l'* 0.672** In every sample we measured 20 adult females and the three largest embryos per female. The estimates are the mean across localities. The letter "a" represents significant heritability estimates, by analysis of variance (ANOVA) for Centinela samples (P s. 0.05), and the letter "b" represents significant genetic variability estimates (by ANOVA) for Senin samples [P s 0.05). The F tests differed between full sibs (df, = 19; df, = 40) and offspring-mother (df, = 1; df, = 18) data sets (see also MATERIALS .AND METHODS). Significant cases after sequential Bonferroni multilest corrections for the 14 different estimates within trait (Rice 1989) are underlined. **Significant Pearson correlation (/> < 0.01 ) between the coefficient of genetic variance (in full sibs) and the coefficient of additi\e \anancc I in mother-offspring regression). Abbreviations: SMI = shell height; SM3 = apenure height; SM4 = aperture width; SM7 = width of the first spire whorl; SMI 5 = distance from the end of the first whorl to the apex; SM16 = the parallel to the ma.ximum aperture length; Y = distance from the center of the whorl to the apex; R = ratio of the first whorl; SD = top shell diameters; AW = top apenure width; T = parameter (Y/R); S = SM4/SM3, 1983. Boulding 1990, Rolan- Alvarez et al. 1997). but only in a few cases is there data showing a genetic basis for the studied traits (Strushaker 1968. Murray & Clark 1966, Reitnchen 1979. Bould- ing & Hay 1993, Johannesson & Johannesson 1996), The low frequency of studies showing genetic variability for the phenotypic polymorphism in the wild is frustrating because it is a serious drawback of any adaptive hypothesis. Besides, there is a theoret- ical good chance to be successful (nearly all examined quantitative traits have shown genetic variation; Falconer & Mackay 1996), Here, we have attempted quantitative genetic estimates (heritabili- 420 Carballo et al. O RB 4 + + HY • Sli o C/1 :i- ^ ■ 0 o n o o B ^o r/o . ■ . o CO o O o U 1 - -1 • o? Oo%)0 o o O 0 0 o 0 -2- ■ -I ^ o -3. 0 -2-10 1 2 3 PCI Factor Scores Figure 3. Relationship between principal component (PC) I and PC2 factor scores in adult females. ties) from wild populations of Galician L saxalilis populations. The results clearly show that there is some genetic variability under the shell traits measured (Tables I and 2; Fig. 4). We found a mean narrow-sense heritability of 0.102 (mean of significant offspring-mother heritabilities) and a mean broad-sense heritabil- ity of 0.489 (mean of significant full-sib heritabilities) for the shell traits studied in our populations of L. suMitills. However, our quan- titative estiniates niay be soniewhat biased by nongenetic effects. The resemblance between mother and offspring may underesti- mate the heritability due to scale effects of individual variables because adult and embryo measurements differ in one order of magnitude. Besides, if a small percentage of the embryos studied were half sibs rather than full sibs. these heritability estimates would be underestimated as well. On the other hand, the resem- blance between full sibs may overestimate the heritability because it also includes the common environmental component. Maternal 1.0 0.8 • Full-Sibs Offspring-Mother 0.6. 0.4, < )■ , 1 [^ X J-^^^- - { 0.0 - IMI.. '■■( '"W^-^J^^ 1/3 O .2 > -4— > c O SU, SUl.^ SUm HY„ RB„ RB, .„, RB, Figure 4. Trends of genetic variances (and their standard deviations) across shore levels and samples for principal component (PC 11 from full-sib correlation and offspring-mother regression analyses. SU, smooth and unhanded; HY, phenotypically intermediate: RB, ridged and banded; L, lower shore; L-m, lower mid-shore; M, mid-shore: L'-m, upper mid-shore; U, upper shore. effects could bias both estimating methods simultaneously, but this would usually occur in traits that are directly related to an em- bryo's fitness rather than to morphological traits. Maternal effects are assumed to be important mainly in vertebrates (Falconer & Mackay 1996). Due to the possible bias mentioned above, perhaps a more realistic narrow-sense heritability estimate wiuild be the average between both estimates (0.296). Similar heritabilities (slightly larger) were found for morphological traits in other lit- torinids (Boulding & Hay 1993) and Drosophila (Houle 1992). The coefficient of genetic variation obtained for T and S were of similar magnitude than that obtained by Newkirk and Doyle (197?) in L. siixatilis populations from Halifa.x. These results complement our previous adaptive interpretation with respect to the Galician shell trait polymorphism by disruptive natural selec- tion favoring each morph at different shore levels (Johannesson et al. 1993. Rolan-Alvarez et al. 1997). The additive and genetic variability of different shell traits varied across the environmental gradient (the hybrid zone; see Table 2 and Fig. 4). In fact, upper and upper mid-shore populations showed lower additive or genetic variability than the rest of popu- lations (Table 2; Fig. 4). These differences could be partially caused by phenotypic plasticity, a term used to include all nonge- netic changes in the phenotype (genotype-environmental interac- tions, for example) that improve individual fitness. There are. in fact, important differences between upper and lower shore levels for many physical and biological parameters (Johannesson et al. 1993. Rolan-Alvarez et al. 1997). Phenotypic plasticity has been also shown in other littorinids (Kemp & Bertness 1984, Boulding & Hay 1993) and may have a nontrivial role in many L. saxatilis polymorphisms (Boulding 1990). However, the differences in ge- netic variances across shore levels were observed after correcting for scale effects (using both the coefficients of variation and the logarithmic transformed variables). This transformation may re- move or at least diminish most interaction terms (including geno- type-environment interactions) contributing to the resemblance between relatives (Falconer & Mackay 1996). Furthermore, we have obtained a significant correlation between the coefficient of genetic (estimated from full-sibs correlation) and the coefficient of additive variation (estimated from offspring-mother regression). These facts may suggest that the differences in coefficient of varia- tion across shore levels may not be significantly biased by pheno- typic plasticity, maternal effects, or other factors because the only expected component in common between these two kinds of esti- mates is the additive variance component. In summary, upper shore populations showed smaller genetic variability than the rest of populations. We can hypothesize that natural selection (which can erode the genetic variability of a population; see Newkirk & Doyle 1975. Falconer & Mackay 1996) may act stronger (at least in summer) on upper shores than on mid/lower shores (perhaps due to sun and desiccation stresses as well as crab predation; see Ro- lan-Alvarez et al. 1997). Nevertheless, to properly quantify genetic variances and phenotypic plasticity, we should do a breeding labo- ratory experiment to partition the familiar variance in different causal components (e.g., additive and dominance components, ma- ternal effects, or genotype-environmental interactions). Another interesting trend is observed on the mid-shore, where hybrids had coefficients of genetic variation or coefficients of additive variation intermediate or even lower than the same esti- mates in both pure morphs (Table 2; Fig. 4). These differences between morphs (RB vs. SU) were maintained when comparing populations from different habitats (upper vs. lower shore) and Estimating Genetic Variances 421 populations from the same habitat (RB vs. SU from mid-shore), which suggests that the differences may not be caused by geno- type-environment interactions. The observed trend was the oppo- site of that we would expect under hybridization of genes fixed at different shore levels or ecotypes (RB \'s. SU ecotypes from mid- shore), assuming that hybrid populations are stable and at least partially fertile (Johannesson et al. 1993. Johannesson et al. 1995. Rolan-Alvarez et al. 1997. Rolan-Alvarez et al. 1999). One expla- nation for the low frequency of hybrids would be that hybrids are rather true F, hybrids unlikely to produce viable progeny. In such circumstances, the population of hybrids would be replaced every generation and would have lower genetic variability. However, embryos from female F, hybrids are actually F, hybrids; thus, we have estimated our genotypic variances at least from F^ hybrid populations. In summary, the results suggest that several factors may be eroding the genetic variability of the hybrid populations (probably a complex mixture of hybrid genotypes). The main fac- tor may be natural selection. v\ ith hybrids adapted to microhabitats slightly different from those of both pure ecotypes. This process of adaptation will necessarily cause some loss of the original genetic variability in the hybrid population (Falconer & Maekay 1996). A similar effect has been described in other hybrid zones: although endogenous hybrid unfitness may be the most common process on hybrid zones (tension zones sensu Barton & Hewitt 1989). hybrid amelioration (sensu Ritchie & Hewitt 1995) could explain these results if some hybrid genotypes are favored with respect to others on the mid-shore (eroding their overall genetic variability). We are presently involved in laboratory experimentation attempting to ob- tain F| hybrids and quantitative genetic estimates for the same shell traits presented above, which may help to resolve many of the questions regarding the dynamics of this hybrid zone. ACKNOWLEDGMENTS We thank A. Caballero and two anonymous referees for helpful corrections and suggestions on the manuscript. We thank the XUNTA de Galicia (XUGA 30I05B98) and Universidad de Vigo (64102C856 and 64502C925) for research funds. M.C. thanks the University of Vigo for a research grant. LITERATURE CITED Barton. N. H. & G. M. Hewitt. 1989. Adaptation, speclntion and hybrid zones. Ncinire 341:497-503. Boulding. E. G. 1990. Are the opposing selection presures on exposed and protected shores sufficient to maintain genetic differentation between gastropod populations with high interniigration rates? Hxdriihiolo^ia 193:4I-?2, Boulding. E. G. & T. K. Hay. 1993. Quantitative genetics of shell form of an intertidal snail: constraints on short-term response to selection. Evo- tulion 47:576-.'i92. Cruz. R. 2000. Natural selection across a vertical intertidal gradient: the Galician hybrid zone of L. saxatilis. PhD. thesis. University of Santiago de Compostela. Spain, (in Spanish). Cruz. R. M. Carballo & E. Rolan-Alvarez.Behavioral and morphological basis of the reproducti\e isolation in an intertidal snail. Evoliaion (In press). Endler. J. 1986. Natural selection in the wild. Princeton: Princenlon Uni- versity Press. Erlands>,on. J.. E. Rolan-Aharez & K. Johannesson. 1998. Migratory dif- ferences between ecotypes of the snail Lilloriiui .'M.\{irills on Galician rocky shores. Evol. Ecol. 12:913-924. Falconer. D. S. & T. F. C. Maekay. 1996. Introduction to Quanlitative Genetics. Essex, England: Longman. Fretter. V. & A. Graham. 1980. The prosobranch molluscs of Britain and Denmark. Part 5- Marine Littorinacea. J. Moll. Stud. 7:242-2(i3. Houle. D. 1992. Comparing evolvability and variability of quantitjuive traits. Generics 130:195-204. Janson. K. 1983. Selection and migration in two distinct phenolypes of Littorina sa.xalilis in Sweden. Oecnlogia 59:58-61. Johannesson. B. & K. Johannesson. 1996. Population differences in be- haviour and morphology in the snail Littorina .sii.\iitili.'>: phenotypic plasticity or genetic differentiation? J. Zool. 240:475^93. Johannesson. K.. B. Johannesson & E. Rolan-Alvarez. 1993. Morphologi- cal dilTerentiation and genetic cohesiveness over a microenvironmental gradient m the marine snail Littorina sa.xalilis. Evolution 47:1770- 1787. Johannesson. K.. E. Rolan-Alvarez & A. Ekendahl. 1995. Incipient repro- ductive isolation between two sympatric morphs of the intertidal snail littorina sa.xatilis. Evolution 49:1180-1190. Johannesson. K.. E. Rolan-Alvarez & J. Eriandsson. 1997. Growth rate dilfciences between upper and lower shore ecotypes of the marine snail Littorina sa.xatilis (Olivi) (Gastropoda). Biol. J. Linn. Soc. 61:267-279 Kemp. P. & M. D. Bertness. 1984. Snail shape and growth rates: evidence for plastic shell allonictrs in Litioriiia liitorea. Proc. Natl. .Acad. Sci. USA 81:811. Lynch. M. & B. Walsh. 1998. Genetics and analysis of quantitative traits. Sunderland: Sinauer Associates, Inc. MuiTay, J. B. & B. Clarke. 1966. The inheritance of polymorphic shell characters in Partula. Genetics 54:1261-1277. Newkirk, G. F. & R. W. Doyle. 1975. Genetic analysis of shell-shape variation in Littorina sa.xatilis on an environmental dine. Mar. Biol. 30:227-237. Newkirk. G. F. & R. W. Doyle. 1979. Clinal variation at an esterase locus in Littorina sa.xatilis and L. oblusata. Can. J. Genet. Cytol. 21:505- 513. Otero-Schmitt. J.. R. Cruz, C. Garcia & E. Rolan-Alvarez. 1997. Feeding strategy and habitat choice in Littorina saxatilis (Gastropoda: Proso- branchia) and their role in the origin and maintenance of a sympatric polymorphism. Ophelia 46:205-216. Packard. G. C. & T. J. Boardman. 1987. The misuse of ratios to scale physiological data that vary allometrically with body size.. In: M. E. Feder. A. F. Bennett. W. W. Burgren & R. B. Huey. editors. New directions in ecological physiology. Cambridge. UK: Cambridge Uni- versity Press, pp. 216-239. Reid. D. G. 1993. Barnacle-dwelling ecotypes of three British Littorina species and the status of Littorina nci>lccia Bean. J. Moll. Stud. 59:51- 62. Reid. D. G. 1996. Sytematics and evolution of Littorina. Dorcherster: Dorset Press. Reimchen. T. E. 1979. Substratum heterogeneity, crypsis. and colour poly- morphism in an intertidal snail [Littorina mariae). Can. J. Zool. 57: 1070-1085. Rice. W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223- 225. Ritchie. M. G. & G. M. Hewitt. 1995. Outcomes of negative heterosis. In: D. M. Lambert & H.G. Spencer, editors. Speciation and the recognition concept. Baltimore: John Hopkins LIniversity Press, pp. 157-174. Rolan-Alvarez. E.. J. Eriandsson. K. Johannesson & R. Cruz. 1999. Mecha- nisms of incomplete prezygotic reproductive isolation in an intertidal snail: testing behavioural models in wild populations. J. Evolutionary Biol. 12:879-890. Rolan-Alvarez. E.. K. Johannesson & A. Ekendahl. 1995. Frequency- and density-dependent sexual selection in natural populations of Galician Littorina sa.xati!is Olivi. Hydrohiolotiia 309:167-172. Rolan-Alvarez, E.. K. Johannesson & J. Eriandsson. 1997. The mainte- 422 Carballo et al. nance of a dine in the marine snail Liltorina saxcirilis - the role of home site advantage and hybrid fitness. Evolution 51:1838-1847. Rolan-Alvarez. E.. E. Rolan & and K. Johannesson. 1996. Differentiation in radular and embryonic characters, and further comments on gene flow, between two sympatric morphs of Liuorina saxatilis (Olivil. Ophelia A5:\-\5. Sokal, R. R. & F. J. Rohlf 199?. Biometry. New York: Freeman & Co. Struhsaker, J. W. 1968. Selection mechanisms associated with intraspecific shell variation in Linorina picta (Prosobranchia:Mesogastropoda). Evolution 22:459-480. Turesson, G. 1922. The genotypical response of the plant species to the habitat. Hereditas 3:211-350. Ward, R. D., T. Warwick & A J. Knight. 1986. Genetic analysis of ten polymorphic enzyme loci in Littorina suxutilis (Prosobranchia:Mol- lusca). Heredity 57:233-241. Ward, R. D., T. Warwick & A. J. Knight. 1991. Further genetic analysis of polymorphic enzyme loci in Littorina saxatilis (Prosobranchia: Mol- lusca). Heredity 66:151-158. Warwick, T. 1983. A method of maintaining and breeding members of the Littorina saxatilis (Olivi) species complex. J. Moll. Stud. 48:368-370. Warwick, T., A. J. Knight & R. D. Ward. 1990. Hybridisation in the Littorina sa.xatilis species complex (Prosobranchia: Mollusca). Hydro- biologia 193:109-116. .loiirihil oj Slwllfish Research. Vol. 20, No. 1. 423-426. 2001. DIET IN MANGROVE SNAILS: PRELIMINARY DATA ON GUT CONTENTS AND STABLE ISOTOPE ANALYSIS J. T. CHRISTENSEN,' * P.-G. SAURIAU," P. RICHARD," AND P. D. JENSEN' ^Department of Marine Ecology. Institute of Biological Sciences. University of Aarhus. Finlaudsgade 14. DK-H2UU Aarhus N. Denmark: 'Centre de Recherche en Ecologie Marine et Aqiiaculliire de L'Houmeau (CREMA) (CNRS-IFREMER UMR 10) BP 5, F-17137 L'Homueau. France ABSTRACT Microscopic analy.si.s of gut contents perlbrmed on three Lirtoiariu species from mangrove forests in Thailand revealed differences In diet ainong species. Analysis of carbon and nitrogen stable isotopes was used as an alternative way of tracing food sources. Rhizophora leaves, scrapings from both leaf and prop-root surfaces, and local particulate organic matter (POM) were well separated on the basis of their S''C and 8'^N values. In contrast, the three Litioraria species exhibit considerable overlap and scatter in both carbon and nitrogen isotope ratio values, suggesting that the snails are opportunistic feeders sharing similar food resources. The wide range of 8"C values of Lirroniria (-17.2^t to -Ib.ifa} is consistent with carbon assimilation from multiple sources (epiphytes from leaves and prop roots, suspended POM. and Rliizophora detritus). Liuoraria intermedia and L. pallescens. the smallest species, had similar 8'^C values, whereas L. seahra was significantly more '^C depleted. A diet of microalgae and cork cells from prop roots could explain this pattern, w ith L seahra. being larger, consuining relatively more cork cells. However, only a few of the L. seahra and L. iinermedia individuals had 5'^N values consistent with such a diet, and the remaining L. seahra and L. intermedia and all L. palleseens individuals were too depleted, indicating that these individuals must derive a significant amount of their food froin a strongly '"^N-depleted source. Such a source is present on Rliizophora leaf surfaces (6"N = 0.30 ± 0.05; n = 2). Some very low values of Liuoraria 8''^N. down to -l%c. indicate that some individuals have assimilated a yet unknown, highly ''^N-depleted food source or that other unknown fractionation processes are involved. A7;)' WORDS: Liuoraria. diet, mangroves, stable isotopes. 8"C. 8"N INTRODUCTION Mangrove snails of the littorinid genus Littoraria are found throught)ut the tropics with an especially large nutnber of species in the Indo-Pacific region (Reid 1986. Reid 1989). Indo-Pacific species all have pianktonic larvae and spend their adult lives on stems, prop roots, and leaves of the trees. Like most littorinids they feed on biofiltns. and the inangroves presumably act merely as substrates. Littamria scabra (L., 1758). L. intermedia (Philippi, 1846). and L. pallescens (Philippi, 1846) have been studied in Thailand, where their stomachs have been found to contain fungal hyphae and spores, among other items (Christensen 1998). L. seahra and L. intermedia live on stems and prop roots of mangroves with L. scabra found lower and usually on the seaward edge of the forests. L. pallescens is found mostly on the leaves of mangroves where L intermedia is also occasionally seen. L. pallescens feeds on the leaf surface without damaging the leaf epidermis, although crescent- shaped necrotic marks may be seen on leaves where the snails rest during the day. However, Ohgaki (1990) reported radular marks on leaves of Rhizophora stylosa. The snails move up and down with the tides and come in direct contact with the water only when releasing offspring. The zonation of the three species may be seen as degrees of terrestrialization of snails with marine origins, and it is therefore interesting to compare their diets. Direct microscopic analysis of gut contents may reveal the range of food items ingested by animals, but the method has cer- tain drawbacks. First of all. the method will depend strongly on the skills of the observer with respect to identifying fragmented ob- jects. Indeed, most of the material in the gut contents of animals may not be visually identifiable. Hard structural tissues may be more easily identified than soft tissues, resulting in a bias in the *Corresponding author. relative importance of itetns, and this bias may be further enhanced by differences in degradability. Finally, gut contents analysis does not reveal the extent to which food items are actually assimilated. Stable isotope analysis of animal tissues can provide alternative information on sources of food. The stable carbon isotope profile of animal tissues resembles that of the food taking into account a fractionation of about l%c per trophic level. Thus, for food sources differing in carbon isotope profile, the isotopic composition of the tissue may ideally indicate the relative contribution of each source to the diet. With respect to nitrogen, animals are usually enriched in '^N (about i%c} relative to the diet (Michener & Schell 1994). The purpose of this study was to provide some preliminary information on diet characteristics of these species and to serve as a possible starting point for more detailed analyses. MATERIALS AND METHODS The samples analyzed were collected on different occasions; however, all samples were taken during the south-west monsoon (July-September) at the island of Phuket. Thailand. Gut contents analysis was performed on a sample (Sample 1 ) of snails collected on R. apiculata mangroves at Chalong Bay and dropped into 70% ethanol (five individuals of each species were analyzed). Another sample (Sample 2) v\'as collected on Avieennia marina (L seahra and L intermedia) at Tang Khen Bay and Rhizophora (L. pal- lescens) at Chalong Bay and fixed in 70% ethanol after cracking the shell ( 10 individuals of each species). The contents of the stomachs were removed under a dissecting microscope, smeared onto mi- croscope slides, and imbedded in glycerol gelatin. From each snail three smears were prepared, and from each smear three fields were scored by ocular grid (81 intersections per field). Objects at each intersection were classified (i.e.. 729 scores per individual). Snails were collected for carbon and nitrogen stable isotope analysis from R. apiculata mangroves at Chalong Bay and kept cool for 3 days in plastic containers during transport to Denmark 423 424 Christensen et al. before being frozen. For analysis tlie soft tissue was separated from tlie shell and operculum and lyophilized prior to being ground with a mortar and pestle. Prior to analysis the tissue was acidified ( 107f. HCl) to ensure the removal of any carbonate debris, rinsed with distilled water, and then freeze-dried again. Analysis of carbon and nitrogen isotopes was also performed on stomach contents re- moved from snails that had been fixed in 709f ethanol (Sample 1), but only nitrogen results are presented. Potential food sources were collected from R. upiciilata man- groves at Chalong Bay. Rhizophora leaves were picked from trees. dried at 60°C. and around for stable isotope analysis. Leaf surfaces were also carefully scraped with a surgical blade without damag- ing the epidermis. Scrapings from about 50 leaves were pooled in each sample and dried al 60°C before analysis. Wetted surfaces of prop roots and stems were lightly scraped with surgical blades. These scrapings were taken from random spots distributed within the range occupied by the snails, and the obtained material was shaken in a bottle with distilled water and filtered through 250-|jim and 63-(j.m filters and finally onto pre-combusted Whatman GF/ C-filters (Whatman Intl. Ltd., Maidstone, UK). The three fractions were dried at 60T along with nonfractionated scrapings. Particu- late organic matter (POM) was filtered from the waters of the bay onto precombu.sted Whatman GF/C filters and dried at 60°C. Samples were prepared and analyzed as reported by Handley et al. (1991. 1993). The mass speclrometric analyses were done on a Europa 20-20 IRMS with an ANCA-SL sample converter (PDZ Europa, Cheshire, UK). A routine precision of appro.ximately 0.1%c for both C and N for invertebrate samples have been ob- tained. Stable isotope ratios are reponed in standard 8 notation as SI = ["^s,in,pic^Mandard) " '] X 1.000, lu uults of pcr mil, where I is the element in question, and R is the ratio of the heavy to the light isotope. Standards were a CO^-C standard previously cali- brated against the universal Pee Dee Belemnite standard and at- mospheric nitrogen. RESULTS The three Liltomria species exhibited clear differences in com- position of the stomach contents with respect to identifiable com- ponents (Fig. 1 ). However, the major part of the stomach contents could not be identified microscopically. Cork cells from man- groves contributed significantly to the stomach contents of L. sea- bra and was also the dominant identifiable item in the stomachs of L. intermedia. In L. pallescens. fungal hyphae and spores were the most prominent identifiable objects, but these items were also present in the other two species. Diatoms, other algae, and cyano- bacteria were present in all species, but only in L. scabra and L. inlermedia could algae in any significant amounts be identified. Kruskal-Wallis tests revealed significant differences among spe- cies for all food items except diatoms (Sample 2): hyphae: K = 15.8, P < 0.001; spores: K = 2\3. P < 0.001 ; other algae: K = 10.9, P = 0.004; cork cells: K = 19.1, P < 0.001 (d.f. = 2 in all cases). Stomach contents from snails that had the shell cracked had a higher diversity of identifiable items than snails that had been dropped directly into 70*^^ ethanol. but also in the latter (Sample 1 ) were there clear differences between species, with cork cells being most prominent in L scabra and fungal hyphae most prominent in L. pallescens. Isotopic signatures of the soft tissue of the three Littoraria species exhibited considerable individual variation (Fig. 2). In L. DL. scabra 20- BL. intermedia 16 • OL. pallescens 10 • ^ 5 ■ 0- r^ , — r-^ r^ ^t\ .iri Hyphae Spores Diatoms Other algae Cork cells Figure \. The occurrence of identifiable items in stomachs of three Littoraria species expressed as percentages (mean + standard error) of the total number of ohjecis observed under an ocular grid in micro- scopic smears. Items include fungal hyphae and spores, diatoms, other algae (including cyanobacteria). and cork cells from mangroves. Ani- mals fixed in 7(l'7f ethanol after cracking of the shell (Sample 2). Lit- toraria pallescens collected on Rhizophora apiculata: L. scahra and /„ intermedia collected on .Avicennia marina. scabra the mean 8"C value was -24.22';f(, and the range was -26.34 to -22.677cc (/; = 16). L intermedia had a mean 8"C of -22.51'7rr and a range of -24.82 to -20.219;^r (n = 16), and L pallescens had a mean 8'-'C of-22.45%c and a range of -24.86 to -\l.TI7c< (n = 15). A Kruskal-Wallis test revealed significant differences among species (K = 9.21; P = 0.01; d.f. = 2), with L. scabra being significantly more "C depleted than the other two species. Also, nitrogen isotopic signatures of the three snail species were highly variable. L. scabra had a mean S'-'^N of 1.89%o and a range of -4.64 to b.\\%( (n = 16). L. intermedia had a mean S'^'N value 1.43"?f and a range of -6.97 to 6.80%f {n = 16), and L. pallescens had a mean 8'^^N of -1.80'%t and a range of -6.12 to 2.00'>?r (;i = 15). Differences among species were significant (K = 9.42; P < 0.01 ; d.f = 2). and L pallescens was significantly more '''N depleted than L. scabra and L intermedia: the latter showed the largest individual variation. Isotope analyses performed on stomach contents of snails fixed in 707f ethanol resulted in a mean 8'''N of 1.33%f. (0.22 to 3.52%250-|j.m fraction of the prop-root scrapings were the most de- pleted. Leaf scrapings were highly depleted in ''^N compared with the other sources. A few of the L. scabra and L intermedia indi- \iduals had 8''^N values consistent with a diet of mixed mangrove and POM. whereas the remaining L .scahra and L intermedia and all the L. pallescens were too '"^N depleted. DISCUSSION Although only small amounts of the stomach contents could be identified, the differences among species with respect to the iden- tifiable fraction of the diet appear to reflect true differences in diet. Whether the differences are solely a result of differences in com- Diet in Mangrove Snails 425 R - 63-250 //m A A non-fractionated . > 250 ^m ^ £ 6 - 4 - ^ 0.2 - 63/;m / 1 A / •• A D 2 0 ■ -2 ■ -4 - O o — 1 — 1 — 1 — 1 — 1 — I — 1 — 1- i + -H 1 1 h — 1 — 1 — 1 — • Rhizophora leaves O Leaf surface scrapings A Prop root scrapings a POM ■ i. scabra ♦ /.. intermedia At. pallescens •30 -28 -26 ■24 -22 'C (%o) Figure 2, tissue of (means d food for . Stable isotopic carbon and nitrogen signatures of the soft tliree species of l.ittorariu collected on Rhizophnra apiciilala : standard error of nieansl and of some potential sources of these snails presented as individual sample measurements. position of the substrates upon which the snails feed (i.e.. a result of zonation) or of the snails" ability to actively select among avail- able items cannot be discerned on the basis of available data. Neither can the extent to which different items contribute to as- similated matter in the three species. Cork cells, fungal hyphae. and spores are major identifiable structural components of the diet, but whether they contribute significantly to the snails" energy bud- get is unknown. Fungal material in the diet was also reported by Kohlmeyer and Bebout (1986) in L an\;iilifcia and by Newell and Barlocher (1993) in L. irrorala. The difference between snails that had their shells cracked prior to fixing and those fixed whole with respect to diversity of items is probably a result of continued breakdown of easily degradable food items in those snails dropped whole into ethanol. This un- derlines the importance of rapid and effective fixing of material for stomach analysis. The diet analyses demonstrate that there are differences between species as far as structural components in the gut contents are concerned, but these differences should not be overinterpreted because only a fraction of the gut contents can be identified. The considerable overlap and scatter in isotope ratio values suggest that the snails are opportunistic feeders and that they to some extent share food resources. The wide range of Liltoraria 8'-'C values (-26.3 to 17.3%c) suggest carbon assimilation from multiple sources (epiphytes from leaves and prop roots, deposited POM. and Rhi:.(iplwia detritus). L. intemwcliii and L. iicillescens. the smallest species, had identical mean 8''C values, whereas L. scahra was significantly more ' ''C depleted. A diet of microalgae and cork cells from prop roots could explain this pattern with L scabra. which is larger and consumes relatively more cork cells. However, all three species were on average more '"^N depleted than these food sources. L. pallescens had a significantly lower mean 8'^N value than the other two species, and it is clear that it does not derive its food directly from the mangrove leaves upon which it lives. It must derive a significant amount of its food from a strongly ''^N-depleled source. Such a source was present in scrapings from leaf surfaces (8''^N = 0.3 ± 0.05%f; ii = 2). but it is as yet unknown what it represents and how it is related to the diet of the snail. Rodelli et al. (I9S4) reported stable carbon isotope ratios in plants and animals from Malaysian mangrove forests. They found a 8'''C of -27.29ff in R. apiciilata. which is acceptably close to our value of-29.6'/rr (n = 2). In L. melanostoina they found a 8''C of -24.6%f (mean of 3) comparable to that of L. scabra in our study, and a value of -21. 5 (mean of 10) in L. iindidata (a rock-dwelling species). They did not report nitrogen isotope ratios. L. irrorala in a North Carolina salt marsh had S'-'C values of -16.6 to -15.17rf and 8'-'N values of 2.2 to 3.7%r (diet O.I to 3.8'^r) (Curtin et al. 1995), and in a Kenya mangrove forest the herbivorous snail Terebraliu palustris had a h^^C value of -24.23%(. similar to that of its presumed mangrove leaf diet (8'''C = -24.28%f). and a S'-'^N signal consistent with the normal pattern of enrichment relative to the diet (Marguillier et al. 1997). If we assume that the on average very depleted '^N signatures of the studied Liltoraria tissues are not in conflict with the generally observed 39?r enrichment per trophic level in animals (Owens 1987. Michener & Schell 1994). we must conclude that all of the three Liltoraria species have assimilated a yet unknown and very ' ''N-depleted food source, but other unknown fractionation pro- cesses may be involved. The fact that the snails were kept alive for 3 days before being frozen could be invoked as a source of '""N depletion. However, it is difficult to identify a mechanism leading to this result. First of all. such a mechanism should affect indi\ iduals differently because not all individuals were highly depleted. Furthermore, starvation is known to lead to ''^N enrichment of the tissues (e.g.. Hobson et al. 1993), and, finally, stomach contents had equally '^N-depleted signatures, again with L pallescens being on average the most depleted. The ' N signatures of stomach contents were obtained from snails that were killed without delay by dropping them into 70'7'fethanol, a procedure that is expected not to affect nitrogen isotope ratios. Mangrove snails may stay inactive for days during dry periods, so being deprived of food for 3 days is not entirely unnatural for them. A purely hypothetical scenario explaining the '""N depletion is that nitrogen excreted by the snails (uric acid) is recycled (e.g.. through fungi), involving fractionation (depletion), and that the fungi subsequently are ingested by the snails. Supporting such a hypothesis is the fact that excretion products (ammonia, urea, and uric acid) are depleted compared with the dietary source and the tissues of the animals excreting them (Cannes et al. 1997) and that fungi are most prominent in the diet of the most '^N-depleted species (L. pallescens). Fungi are among the microorganisms known to degrade uric acid (Kieslich 1976). Microorganisms like the cyanobaclerium Anabaeiia grown on nitrate or ammonia show large fractionations and '"^N depletions (Macko et al. 1987). The results underline the usefulness of multiple isotope analy- ses. A more limited conclusion would have been reached had the analysis been based on gut contents and stable carbon isotopes alone. The wide range of isotope ratios within species further stresses the importance of large samples in food chain studies. Considerable bias or loss of information may result if one attempts to deduce trophic relationships based on pooled samples of three or four individuals. During the last three decades, most isotopic stud- ies have been based on small samples under the assumption that there is insignificant variation between individuals within species. Our data demonstrate that this is not always the case and that variation can be large, possibly due to small-scale heterogeneity in the occurrence and accessibility of food items. Detailed studies are needed to explain the highly '""N-depleted tissues and the range of variation in N isotope ratios in these snails. 426 Christensen et al. ACKNOWLEDGMENTS We wish to thank the director of the Phuket Marine Biological Center, Thailand. Dr. Prawin Limpsaichol. and Mr. Sonichai Bus- sarawit for the opportunity to work at the laboratory. Dr. C.-K. Kang prepared all the samples (decalcification and weighing) and did some initial mass spectronietric analyses while Dr. C. M. Scrimgeour of Dundee University. Scotland, performed the final mass spectrometric analyses on our samples. We also thank two anonymous reviewers for valuable criticism. LITERATURE CITED Christensen. J. T. 1998. Diet in Linoraria. Hydmbiolo^iu 378:235-236. Cumn. C. A.. S. Y. Newell & H. W. Paerl. 1995. The role of standing dead Spartina aherniflora and benthic microalgae in salt marsh food webs: considerations based on multiple stable isotope analysis. Mar. Ecol. Prog. Ser. 121:99-116. Cannes. L. Z.. D. M. O'Brien & C. M. Martinez del Rio. 1997. Stable isotopes in animal ecology: assumptions, caveats, and a call for more laboratory experiments. Ecology 78:1271-1276. Handley. L. L., C. M. Scrimgeour. S. F. Thornon & J. I. Sprent. 1991. Determination of the natural abundances of the stable isotope of ''^N and "C by mass spectrometry: a simplified manual method for the preparation of N, and CO,. Fimct. Ecol. 5:119-124. Handley. L. L.. M. J. Daft. J. Wilson. C. M. Scrimgeour, K. Ingleby & M. A. Sattar. 1993. Effects of the ecto- and VA-mycorrhizal fungi Hydiuigiitin carnewn and Glomus cicirimi on the '^N and '^^C values of Eiicalyptu.s globulus and Ricinus communis. Plant. Cell Environ. 16: 375-382. Hobson. K. A.. R. T. Alisauskas & R. G. Clark. 1993. Stable-nitrogen isotope enrichment in avian tissues due to fasting and nutritional stress: implications for isotopic analysis of diet. Condor 95:388-394. Kieslich, K. 1976. Microbial transformations of non-steroid cyclic com- pounds. Stuttgart: Georg Thieme Publishers. Kohlmeyer. J. & B. Bebout. 1986. On the occurrence of marine fungi in the diet of Linoraria angulifera and observations on the heha\ lour of the periwinkle. P. S. Z. N. I. Mar. Ecol. 7:333-343. Macko. S. A.. M. L. Fogel (Estep). P. E. Hare & T. C. Hoering. 1987. Isotopic fractionation of nitrogen and carbon in the synthesis of amino acids by microorganisms. Cliem. Geol. [Isotope Geoscience Section) 65:79-92. Marguillier. S.. G. van der Velde. F. Dehairs. M. A. Hemminga & S. Rajagopal. 1997. Trophic relationships in an interlinked mangrove- seagrass ecosystem as traced by 6"C and 5'^N. Mar. Ecol. Prog. Ser. 151:115-121. Michener. R. H. & D. M. Schell. 1994. Stable isotope ratios as tracers in marine aquatic food webs. In: K. Lajtha & R. H. Michener, editors. Stable isotopes in ecology and environmental science. Oxford: Black- well Scientific Publications, pp. 138-157. Newell, S. Y. & F. Barlocher. 1993. Removal of fungal and total organic matter from decaying cordgrass leaves by shredder snails. J. E.\p. Mar. Biol. Ecol. 171:39^9. Ohgaki. S. 1990. Food items of the mangrove Littoraria (in Japanese). Chiribotan 21:51-53. Owens. N. J. P. 1987. Natural variations in '^N in the marine environment. Adv. Mar. Biol. 24:389-451. Reid. D. G. 1986. The littorinid molluscs of mangrove forests in the Indo- Pacific region. London: British Museum (Natural History). Reid. D. G. 1989. The comparative morphology, phylogeny and evolution of the gastropod family Littorinidae. Philos. Trans. Roy. Soc. Lond. B 324:1-110. Rodelli. M. R., J. N. Gearing. P. J. Gearing, N. Marshall & A. Sasekumar. 1984. Stable isotope ratio as a tracer of mangrove carbon in Malaysian ecosystems. Oecologia 61:326-333. JoKi-mil of Shellfish Research. Vol. 20. No. I. 427-430. 2001. SHELL SIZE VARIATION IN LITTORINA LITTOREA IN THE WESTERN SCHELDT ESTUARY HANS DE WOLF,' * RONNY BLUST,' AND THIERRY BACKELJAU" ^Universin' of Antwerp (RUCA). Groenenhorgerlaan 171 B-2020. Antwerp. Beli;liiiii: 'Royal Belgian Institute of Natural Sciences (KBIN). Vautierstraat 29. B-1000. Brussels. Belgium ABSTRACT Liuorimi litwrea was collected along a salinity gradient in the Scheldt estuary, located in the South of the Netherlands. Its morphological population structure was investigated to see whether salinity was correlated with shell size and shell weight. Shell size did not increase along the salinity gradient, as wa.s expected, but rather showed a clear size transition between two salinity ranges (i.e., lO-lO^ir and 21~30%t). Animals attain their largest size within a salinity range ot'2l-307tc. Relative shell weight did not vary consistently with salinity. KEY WORDS: environmental stress, estuary. Litioniui liUnrca. salinity, Scheldt river, shell size INTRODUCTION Althotii;li littorinids show high levels of intraspecifie shell polyiiiorphisiiis (see Reid 1996 and references therein), Liuorina liuorea (Linnaeus, 175S). the largest species in the genus, shows relatively little morphological variation (Janson 1987). However, morphological differences were noted between populations of L. litrorea on the West Somerset coast (Crothers 1992). Along this coastline, a weak cotTelation was found between shell shape, as expressed by the shell length/aperture length ratio, and wave ex- posure (Crothers 1992), The differences were explained by differ- ential growth and/or survival rates in response to the effects of wave exposure (Crothers 1992). In contrast, Janson (1987) found almost identical shapes between exposed rocky and boulder shore specimens. Apparently, the only consistent shell variation is found between marine and sheltered brackish forms, with the latter being smaller and thinner-walled (see Reid 1996). This variation is sup- posed to be ecophenotypic (Reid 1996) because L littorea is a planktonic developing animal that is presumed to ha\e a high dispersal and gene flow potential (Janson 1987. Reid 1996). tnini- mizing the likelihood of selection as a possible impetus for the observed shell variation (Chapman 1995). Nevertheless, predation experitnents with the oystercatcher Haematopus iistrulegus have shown that the aperture size of L. littorea may be susceptible to selection, although field observations have never confirmed these experimental results (Robertson 1992). L. tilloreii is widely distributed, occurring in the eastern (White Sea to southern Portugal) and western (Labrador to Virginia) At- lantic (Reid 1996). Unlike other littorinids. it does not solely occur on hard substrates but is also able to crawl over sand and soft mud (references in Reid 1996). This ability, along with its planktonic development and its tolerance to low salinity conditions (9.5%c), enables it to penetrate far into estuaries (Reid 1996). In the Scheldt estuary, situated in the south of The Netherlands. L littorea is found from Vlissingen (i.e.. rivers' mouth) to Bath. 50 km inward from the mouth, where it occurs along a gradually decreasing salinity gradient, ranging from marine to brackish (Fig. 1 ). The Scheldt estuary thus forms an ideal setting to test whether salinity is indeed correlated with the shell morphology of L lit- torea. If, under brackish conditions, L litrorea has a smaller, thin- *Corresponding author. E-mail: dewolffe'ruca.ua.ac.be ner-walled shell, then we expect a shell-size, weight gradient in the estuarv following the salinity gradient. MATERIALS AND METHODS On 8 August 1 998. L littorea was collected at seven sites along the western Scheldt estuary, covering its entire range in the west- ern Sheldt (Fig. 1). These sites included, in order of increasing salinity: Bath. Waarde. Hansweert. Hoedekenskerke. Ellewouts- dijk. Borssele. and Vlissingen (Fig. 1). One population was col- lected at each site. Each sample consisted of 40 animals. Each of the 280 specimens was morphometrically characterized. Five shell traits were measured to the nearest 0.05 mm using a caliper: shell height (HS). shell width, aperture height, aperture width, and shell- top height (De Wolf et al. 1997). In addition, total wet weight (i.e.. shell + soft body parts) and body wet weight (soft body parts) were determined to the nearest mg. and all individuals were sexed on the basis of the presence or absence of a penis. A seven-by-two contingency table was constructed to test whether the sex distribution differed from site to site, employing the Metropolis algorithm to obtain unbiased estimates of the exact P value (Miller 1997). Morphometric patterns were investigated by means of a two-way multivariate analysis of variance (MANOVA). contrasting the fixed factor "sex" with the random factor "sampling site." Morphological patterns were further inves- tigated by means of a standard canonical discriminant analysis (CDA). Finally, an analysis of covariance (ANCOVA) of shell weight was pert'ormed using shell height as a covariate. Of special interest in this analysis is the interaction with the covariate because it tests whether the slopes of the shell weight on shell length are homogenous for the seven sites. Except for the contingency table analysis, all statistical analyses were performed using the software package Statistica v. 5.0 (Statsoft 1995). RESULTS At each site, except for Hoedekenskerke. males outnumbered females (Fig. 1 ). The number of males differed significantly at the different sites (P = 0.0003). The results of the two-way MANOVA are summarized in Table 1. A significant part of the total variation can be explained by the random factor "sampling site" (Table I ), whereas the fixed factor "sex" did not contribute significantly, nor did the interaction (Table 1 ) 427 428 De Wolf et al. 30 1 , 25 ^\ 20 \ 15 ^\J 10 Vlissingen Borssele Hansweert 4-E _L_ Figure 1. Sampling area, sites, and shells typical of each of the sampling sites (sites 1-7). Pie diagrams represent number of males (black zone) and females (white zone) collected at each site (/i = 40). The line graph represents mean salinity values and standard deviations along the Scheldt estuary on the basis of seasonal measurements from 1990-1997. These results are illustrated in Figure 2. where the mean shell height and standard deviations for males and females are plotted for each sampling site. Mean shell height values for both sexes overlap but simultaneously reveal a structuring at the sampling site level. Mean shell height of specimens collected at the less marine- like sites (i.e.. salinity range 10-20%c: Fig. 1 ) are on average smaller compared with the shell height of speciinens collected more downstream at a salinity range of 21-309i( (Fig. 1). How- ever, due to individual variation this observation is merely a trend because post-hoc Sheffe tests failed to significantly discriminate both groups. Given that, with respect to the measured shell characteristics. TABLE 1. Results of the lv»o-v\ay M.ANOVA. contrasting the random factor "sampling site" and the llxed factor "sex." Effect VVilks" \ dfl df2 P value .Site 0.18.^218 42 1.227 <0.00()l Sex 0.957741 7 261 0.1230 Site X Sex 0.X44134 42 1.227 0.3448 males do not differ significantly from females, a single CDA was peiformed without considering the factor "sex." The mean values of the first two canonical variables (CV) are used to plot all sam- pling sites, as shown in Figure 3. The first CV describes 64.83% of the total variation and is mainly an expression of shell height (Table 2; HS = -1.1 1614). Shell height decreases with decreasing CVl values, discriminating the different sampling sites, so that specimens collected at the least marine-like sampling sites are in general smaller than specimens collected at more marine-like sites (Fig. 3). The second CV describes an additional 19.83% of the total variation and is mainly an expression of the shell weight (Table 2: SW = 1.91514). Shell weight increases along the posi- tive CV2 axis. Four groups can be distinguished along both CV axes: Bath and Waarde, consisting of specimens with small and light shells; Hansweert. consisting of specimens with small and relatively heavy shells; Ellewoutsdijk and Vlissingen. consisting of speciinens with intermediate-sized shells and intermediate shell weights, and Borssele and Hoedekenskerke. consisting of speci- mens with larger and heavier shells. Hence, relative shell weight does not follow the salinity gradient. This is also illustrated in the ANCOVA. where the regression slopes of the shell weight on shell height are not homogeneous at the different sampling sites (Table Shell Size Variation in L. Littorea 429 O 5000 (U 5 E E b a 1000 shell height male ■ shell height female total weight shell weight TABLE 2. Standardufd iidflKiinls for llie first Iho caniuinicul variable.s (CV) in a canonical discriniani analysis cuntrastin;' all ei)>lit sampling sites. Dependent Variable Shell height Shell width Apenure height Aperture width Shelltop height Shell weight Eigenvalue Explained variation. ' CVl -I. 11614 0.28414 -0.7 1.504 0.68485 0.57448 -0.58913 1.839.35 W.83 CV2 -1.92037 0.81833 0.03916 - 1 .47668 0.61353 1,91514 0.56258 19.83 Sampling site Figure 2. Mean shell height (H.S). total weight (TVVl (i.e.. soft body weight + shell weight), shell weight (.S\V), and standard deviations lor males and females of Lilloriiia littorea collected at the seven sites. 3) and in Figure 2. where mean shell weight, mean total weight, and mean shell height are plotted at the different sampling sites. DISCUSSION As was predicted. L. lillorea had smaller and lighter shells in more brackish conditions. However, shell size did not decrease clinally away from the sea. following the salinity gradient. Instead, specimens were either small (approximately I9 mm). No intermediate-sized animals were found at sites having intermediate salinity levels (i.e., ±l5%c). If salinity is indeed an environmental factor that affects shell si/.e in L lillorea. it seems that its effect is either present or absent with a threshold salinity value of approximately 15-20%c. Above this salinity threshold, animals are able to attain a large shell; below this value, animals never have comparable shell sizes. In either case it must be clear that, at brackish sites, L. liilnrca is likelv to > o c ra (D 21 - 30 X. 20%. 1 Hansweert o Borssele Hoedekenskerke i o o 21 - 30 %o Ellewoutsdijk (^ 1 Vlissingen 1 10X. Waarde Bath o o encounter less favorable living conditions, which might result in a decrease of gixiwlh and/or survival rate (Crothers 1992), shifting the less marine-like populations toward smaller-sized individuals. It must also be noted, however, that the largest shells were not recorded at Vlissingen (i.e.. highest salinity) but at Borssele (i.e., second highest salinity). At Borssele, specimens were collected in the direct vicinity of a nuclear power plant. Possible water tem- perature differences, due to the outflowing cooling water, might affect the species shell growth. Indeed, larger shells can be pro- duced at higher temperatures due to the lower energy cost of calcification because calcium carbt)nate dissolves less well at higher tempeiatures (Graus 1974, Clarke 198.3). However, on a macrogeographical scale. L. littorea attains its largest size in the northern parts of its distribution range (i.e., cooler waters) (Reid 1996). Clearly, further experimental work is needed to clarify and explain the possible effects of water temperature and salinity on shell deposition in L. lillorea. In the case of shell weight, the presumed relationship with salinity was even less clear. Moreover, it seemed unlikely that relative shell weight differences can be explained by salinity dif- feiences. The top right positioning of the Hansweert population in the CDA graph (Fig. 3) indicates that specimens with the highest relative shell weight occur under the third lowest salinity condi- tions. TABLE 3. Results for the interaction of the dependent variable shell weight (SW) and the covariate shell height (HS) in the ANCOVA (i.e., test for parallelism) and corresponding regression equations at each of the seven sampling sites. SS df MS F Value P Value Effect 3,811,414 Error 8,454,196 6 635,235.6 20.06198 264 31663.7 <0.000l Site Regression equation r^ Mean CVl (64 83%) Figure 3, Mean values for the first two cannonical variables. Bath HS Waarde HS Hansweert HS Hoedeken.skerke HS Ellewoutsdijk HS Borssele HS Vlissingen HS 1.0.59.5 + 0.5614 SW 1,133.3 -H 0.4672 SW 913.2 -I- 0.5549 SW 1 .258.3 + 0.3239 SW 1,082.9 + 0.4553 SW 1,398 + 0.2897 SW 1.169.3 + 0.3844 SW 0.8928 0.8832 0.8485 0.8452 0.9103 0.9078 0.8737 Abbreviations; SS = Sum of squares; MS = Mean square. 430 De Wolf et al. Sexual dimorphism is common in the genus Lirron'iui. with females being larger than males (Reid 1996). which is presumed to be related to growth and/or longevity differences (Reid 1996). However, sex-related shell height differences have not been found in L. littorea (Reid 1996), even though sexual selection for female shell size has been documented, with males preferring to mate with larger, and thus more fecund, females (Erlandson & Johannesson 1994). The fact that we did not find sex-related shell size differ- ences is thus in agreement with what was previously found. How- ever, our sex assignment, made on the basis of the presence or absence of a penis, leads to a sex ratio of almost 2: 1 (male/female), which differs markedly from a previously published sex ratio of 1:1 (Daguzan 1977). The fact that we found twice as many males might be related to the presence of imposex — the development of male sex characteristics on females (e.g.. a penis and/or vas def- erens) (Bauer et al. 1997) — and/or intersex — the disturbance of the phenotypic sex determination between the gonad and genital tract — which is known to occur in L. littorea (Bauer et al. 1997). Imposex has never been recorded in L. littorea. In contrast, inter- sex gradually transforms the female pallial tract such that the fe- male pallial organs are supplanted by a male prostate gland, and a seminal groove and a small penis occur (Bauer et al. 1997). The fact that penis shedding and/or regression also occur in L. littorea (Deutsh & Fioroni 1992) makes it even more difficult to distin- guish between an intersex female and a male with a shed or re- gressed penis. Therefore, it could be that some specimens that were classified as males were in fact intersex females. However, female intersex expression occurs only in juvenile stages or during sexual immaturity (Bauer et al. 1997). As a consequence, intersex females are expected to have a smaller shell. Hence, if intersex females were included in the male population, they would not have increased the mean male shell height, masking possible shell size differences with the presumed larger females, but rather would have decreased the mean shell height of the male population. In any event, the occurrence of penis shedding, penis regression, and intersex make basing sex assignment in L. littorea on the presence of a penis unreliable. Similarly, the presence or absence of a pros- tate gland is also an unreliable sex-determining character (Bauer et al. 1997). Finally, salinity is not the only environmental factor that may be correlated with the morphology of L. littorea. In an estuary, which is structured by a complex of gradients, a wide variety of natural and human-induced stresses are present that may affect the shell moijihology of estuarine gastropods. In this respect, in the period 1981 to 1983. the Scheldt estuary was ranked among the inost heavily polluted estuaries around the world for both the dis- solved as well as the particulate metal phase (Bayens 1998). Dis- solved metal concentrations measured at Hansweert and Vlissin- gen differ significantly, with Vlissingen being less polluted (Rijksinsituut voor Kust en Zee. RIKZ. pers. comm.). Concentra- tions of volatile organic compounds are high and decrease along the estuary as well (De Wulf et al. 1998). Clearly, pollution is an important potential stressor in the Scheldt estuary, and its effect on the morphological population structure of the estuarine organisms must be investigated. ACKNOWLEDGMENTS The authors would like to thank Gilles Watel (RIKZ) for kindly pro\ iding us with the salinity and dissolved heavy metal measure- ments taken in the western Scheldt. Peter Mill and an anonymous referee improved this manuscript greatly. This research was sup- ported by a RAFO project under the contract number RAFO/1 DEWOH KP98. HDW is a Postdoctoral Fellow of the Fund for Scientific Research - Flanders (Belcium) (F.W.O.). LITERATURE CITED Bayens. W. 1998. Evolution of trace metal concentrations in the Scheldt estuary (1978-1995). A comparison with estuarine and ocean levels. Hydrobiologia 366:157-167. Bauer. B.. P. Fioroni. U. Schulte-Oehlmann. J. Oehlmann & W. Kalbfus. 1997. The use of Littorina littorea for tributyltin (TBTl effect moni- toring - results from the German TBT survey 1994/1995 and laboratory experiments. Em-iron. Poll. 96:299-309. Chapman. M. G. 1995. Spatial patterns of shell shape of three species of co-existing littorinid snails in New South Wales. Australia. J. Moll. Stud. 61:141-162. Clarke. A. 1983. Life in cold water: The physiological ecology of polar marine ecosystems. Oceanogr. Mar. Biol. Ann. Rev. 21:341^53. Crothers, J. H. 1992. Shell size and shape variation in Littorina littorea (L.) from west Somerset. In: J. Grahame. P. J. Mill & D. G. Reid. editors. Proceedings of the third international symposium on Littorinid biology. London: The Malacological Society of London, pp. 91-97. Daguzan. J. 1977. Analyse biometrique du dimorphisme sexuel chez quelques Littorinidae (MoUusques. Gasteropodes. Prosobranches). Haliotis 6: 1 7-10. Deutseh, U. & P. Fioroni. 1992. The shedding of the penis in Littorina littorea: some new aspects. In: J. Grahame. P. J. Mill & D. G. Reid. editors. Proceedings of the third international symposium on Littorinid Biology. London: The Malacological Society of London, pp. 309-31 1. De Wolf H.. T. Backeljau. R. Medeiros & R. Verhagen. 1997. Microgeo- graphical shell variation in Littorina striata, a planktonic developing periwinkle. Mar Biol. 129:331-342. De Wulf J.. H. Van Langenhove. M. Everaen & H. Vanthournout. 1998. Volatile organic compounds in the Scheldt estuary along the trajectory Antwerp-Vlissingen: concentration protiles. modeling and estimation of emissions into the atmosphere. Wat. Re.':. 32:2941-2950. Eriandson, J. & K. Johannesson. 1994. Sexual selection on female size in a marine snail. Littorina littorea. J. Exp. Mar. Biol. Ecol. 181:145-157. Graus. R. R. 1974. Latitudinal trends in the shell characteristics of marine ga.stropods. Lethaia 7:303-314. Janson. K. 1987. Allozyme and shell variation in two marine snads (£/;- torina. Prosobranchia) with different dispersal abilities. Biol. J. Linn. Soc. 30:245-256. Miller. M. P. 1997. R\C. A program for the analysis of contingency tables. Flagstaff AZ. Reid, D. G. 1996. Systematics and evolution of Littorina. London: The Ray Society. Robertson, A. 1992. The oystercalcher. Haemalopus osiralegiis. as a se- lective agent on littoral gastropods. In: J. Grahame. P. J. Mill & D. G. Reid. editors. Proceedings of the third international symposium on Littorinid biology. London: The Malacological Society of London, pp. 15.V161. .lotirmil of Shellfish Research. Vol. 20, No. 1. 431-439, 2001. CARBONATE PROCESSING BY INTERTIDAL GASTROPODA ON JAMAICAN LIMESTONE SHORES ALLUWEE DOBSON-MOORE AND JOSEPH C. BRITTON* Department of Bi(>li>!i\-. Texas Cliristian University. Fart W'ortli. Texas 76129 ABSTRACT At least eight species of Littorinidae occur sympatrically on a limestone platform on the northern coast of Jamaica with little evidence of competitive displacement. Most of these, plus a cobble shore gastropod {Planaxis nucleus), were studied with respect to the amount of carbonate each removes from the shore while feeding. The feces of P. nucleus contain 92.8% carbonate. On the platform, Nodilinorimi riisei fecal pellets contained 88.7% carbonate, N. ziczac 74.9%. N. ongustior 67.0%, N. dilatata 88.3%, Tectarius anumii 91.3%. Cenchriris muricatus collected from the rocky substratum 74.8%, and C. muricalus collected from the maritime shrub Rluichiciillis umericana 18.6%. Although the rock-dwelling Littorinidae ingest different quantities of carbonate, there is no clear relationship between the amount of carbonate ingested and the position each species occupies on shore. Fecal pellet counts were made for all species. The mean numbers of pellets recovered from the rectums were: P. nucleus (48.7), A", riisei (8.56), N. ziczuc (26.1). N. angusliori i4.\). N. dilatata (22.2). T. anlcmii (22.8), and C. muricatus (35.3). The mean size and weight of pellets for each species were: P. nucleus (1.73 mm", 0.040 mg). N. riisei (1.21 mm", 0.028 mg), N. ziczac (1.43 mm". 0.066 mg), N. angustior {0.959 mm-, 0.024 mg). A", dilatata (1.76 mm", 0.052 mg), T. antonii (3.23 mm", 0.118 mg), and C. muricatus (4.45 mm", 0.112 mg). The bioerosive impact of each species was assessed by calculating the amount of carbonate removed from the shore per individual and per species based on density and an estimated 48-h defecation cycle. Impacts were expressed in terms of both a single defecation cycle and annually. Collectively, the Littorinidae are estimated to remove at least 2.850 kg of carbonate from the approximately 500 x 24-m limestone platform annually. KEY WORDS: Lirtoriiiidae. fecal pellets, carbonate, rocky shore, Jamaica INTRODUCTION At least eight species of Littorinidae (Lang et al. 199^!, Minton & Gochfeld 2001) occupy a limestone platform on the northern coast of Jamaica, from near sea level to as much as 4 m above it. Most of them (except the tidepool-dwelling Nodilittohna inespel- hiin) occupy very similar niches in broadly overlapping ranges from the splash zone to as much as .30 ni inland from the sea, with little evidence of competitive displacement. Several previous stud- ies have attempted to elucidate the roles of biotic and abiotic factors in controlling the distribution of sympatric littorinid species on hard-shore habitats, mostly with inconclusive results. Britton (1992) showed that the resistance to evaporative water loss and tolerance to thermal stress had little correlation to the position on the shore occupied by each species. He suggested, however, that dessication tolerance may be a more important limiting factor than thermal stress in dictating species position on shore. Lang (1995) found that shell sculpture and spire height in three littorinid species were unrelated to shore position and shell nodulosity. The presence of nodules on shells is widely assumed to confer a temperature regulatory advantage (Vermeij 1973); however, they could not be shown to do so. In a comprehensive survey of several ecological parameters associated with Jamaican rocky-shore littorinids, in- cluding position and direction of repose, substratum and body temperatures, nature of attachment to the substratum, microhabitat (within crevices or not), and availability of moisture in association with the substratum, Lang et al. (1998) found few significant cor- relations between these parameters and the positions species oc- cupy on the shore. Furthermore, there were few species-specific differences for most of the parameters examined. Substratum moisture, however, was modestly predictive of which species might be expected on a portion of the shore, with N. riisei always occupying wetted surfaces at the seaward face of the platform. *Corresponding author. Cenehritis muricatus always on the highest and driest areas at the back of the platform, and the other species at various, broadly overlapping positions in between. Location on a shore is an important factor in determining the degree of stress a species will encounter, ranging from thermal and dessication exposure to what foods are available to the species (McMahon 1990, Norton et al. 1990. McMahon & Britton 1991). On the Jamaican platform shore, N. riisei (Morch. 1876), A', ziczae (Gmelin, 1791 ), and A', angustior (Miirch, 1876) occupy the upper eulittoral zone, although the latter may also range into the lower supralittoral fringe. N. dilatata (d'Orbigny, 1846) and Tectarius antonii (Philippi, 1846) have a broad distribution across the mid- to lower supralittoral fringe, whereas C. nnirieatus (Linnaeus, 1758) preferentially occupies the upper supralittoral to the margin of maritime vegetation (Lewis 1960. Bandel 1974a, Lang et al. 1998). Planaxis nucleus (Bruguiere, 1789) (Prosobranchia; Planaxidae). the only nonlittorinid examined in this study, does not occur on the platform but is found instead on large immobile limestone boulders awash in suif on a nearby cobble shore, iden- tified locally as Jingle Beach. The Littorinidae are predominantly herbivorous gastropods that rasp the substratum for food by means of a radula (Reid 1989, Norton et al. 1990). On the Jamaican rocky shore, there are dif- ferences in the quantity and quality of food sources from the splash zone to the highest parts of the limestone platform. It seems plau- sible, therefore, that the herbivores occupying different vertical zones have different food preferences. Low-zoned species such as N. riisei. in the presence of abundant moisture and epilithic algae, might rely less upon potentially harder-to-reach endolithic cyano- bacteria. In contrast, the high-zoned 7". antonii. living exclusively on rock, has little opponunily to feed upon epilithic or epiphytic algae. The higher-zoned C. nnirieatus may graze upon the stems and leaves of maritime vegetation, but when on rock, it, like T. antonii. must focus instead on a diet of endolithic algae and/or lichens. Peckol and Guamagia (1989) showed that the gut of T. 431 432 DOBSON-MOORE AND BRITTON aiilimii contains considerable amounts of endolithic cyanohacteria. which occur abundantly throughout the vertical range of this spe- cies. We will show that C. muricatus has an alternate food source among the maritime angiosperm vegetation. In this study, we hypothesize that diet may reflect one aspect of resource partitioning among intertidal gastropods. Rock-dwelling species that must rely heavily upon endolithic algae for food (e.g.. C. inuikatus and T. antonii) should also have a large quantity of non-nutritious calcium carbonate in the feces. Conversely, species living among epilithic algae (e.g.. N. riisei and N. ziczac) may focus on this more easily obtained food when it is available. If so, the fecal carbonate fraction should be much less than that associ- ated with the high-zoned species. Members of the Littorinidae and Plana.xidae are ideal models for an analysis of fecal content because they form the feces into discrete, compact, and relatively glutinous pellets that are retained in a discrete rectum until eliminated from the body. Fecal pellets of the Jamaican rocky shore Littorinidae have a similar morphol- ogy, mostly ranging from ovoid to drop-shape. They contrast sharply with the nan'ow. spindle-shaped pellets of the Planaxidae (Bandel 1974b). The pellets of those species living on rocky sur- faces consist mainly of fine detritus, including mineral flakes and rock fragments that they scrape from the surface while rasping off algae (Bandel iy74b). We hypothesize, therefore, that the fecal pellets of different species of herbivorous intertidal Gastropoda contain carbonate in different proportions according to their food choices and/or method of grazing. If such differences can be detected, this might suggest one method by which these species partition resources on this shore. We also assess the bioerosive impact of these species. The number and weight of the fecal pellets per individual as well as measurements of the size of fecal pellets for 50 individuals of each species were used to estimate the amount of carbonate re- moved from shore by these gastropods. MATERIALS AND METHODS Six species of Littorinidae (Gastropoda: Prosobranchia: Lit- torinidae) were collected from a micro-karsted limestone shore platform near the Hofstra Marine Lab at Priory. St. Ann's Parish, on the northern coast of Jamaica. Lang et al. ( 1998) described this shore and the distribution of the littorinids on it. N. dilahita. T. antonii. and C. muricatus are mid- to high-zoned nodulose species, whereas N. ziczac. N. angtistior. and N. riisei produce relatively smooth, zebra-striped shells and occupy mid- to lower levels of the platform. A seventh species examined for this study, P. nucleus (Gastropoda: Prosobranchia: Planaxidae), was collected from large immobile boulders at Jingle Beach, a cobble shore located about 500 m west of the limestone platform. All species were collected from a limestone substratum, either in repose or crawling upon it. C. nuiricatiis. however, also oc- curred on a small maritime shrub. Rhtuliiciillis ainericuna. Indi- viduals from this substratum were also collected and treated sepa- rately during the laboratory phase of this study. Collected indi- viduals of all species were preserved in 70% ethyl alcohol within seconds of collection to minimize the loss of fecal pellets by natu- ral defecation. Fecal pellets were removed from 740 individuals (i.e.. 80-130 individuals per species) (Table I). Prior to dissection, shell mea- surements (shell height, shell width, aperture height, and aperture width) of each individual were determined by means of digital calipers. Each shell was then broken, the soft body was removed, and the rectum and its contained fecal pellets were excised from each body. Fecal pellets removed from the rectum were placed on a pre-weighed drying boat, their wet weight was determined, and the number extracted from each individual was counted. The height and width of 250 fecal pellets (i.e.. five from each of the first 50 individuals dissected) were determined for each species by means of a microscopic monocular micrometer. Be- cause there was no reliable, accurate way to measure depth, pellet size was expressed as area, not volume. Means and standard de- viations of fecal pellet area were calculated for all species, and a Kruskal-Wallis single factor analysis of variance (ANOVA) by ranks was performed to determine species-specific differences, if any. Linear regressions were performed to assess the within- species relationship between fecal pellet number and body size, as indicated by shell height. If these analyses indicated a significant relationship, the slopes of the regression lines could be compared by an analysis of covariance (Zar 1984) to determine if there were significant differences between species. Fecal pellets were placed in a 55°C drying oven until a constant dry weight was attained. The means and standard deviations for fecal pellet dry weight were calculated on a per-pellet basis for all species, and these values were compared between species by means of a Kruskal-Wallis single factor ANOVA by ranks. The pellets from each individual were then assessed for meeting a critical dry weight (i.e.. a minimum quantity of fecal material TABLE L Summary statistics for number of individuals dissected, total number of fecal pellet samples, number of samples containing only one individual, number of samples containing more than one individual li.e,, a pooled sample), and the mean, standard deviation, and range for the number of individuals contained in the pooled samples. Number of Total Number of Number of pooled Number of individuals in a pooled sample ind ividuals number sampli es Hii m omv dissected of samples one indiv idual samples Mean Standard deviation Range P. nucleus n3 37 0 37 3.05 0.88 2-5 N. riisei 80 3 0 3 26.7 12.2 16-40 N. ziczac 122 38 1 37 3.27 2.77 2-13 N. angustior 80 6 0 6 13.3 2.16 11-17 N. dilatata 81 20 0 20 4.15 1.18 2-6 T. antonii 130 56 4 52 2.42 0.67 2-5 C. muricatus on pl^ :ints .54 45 42 3 2.67 0.58 2-3 C. muricatus on rocks 62 .Vs 12 23 2.17 0.49 2^ Carbonate Processing by Jamaica Gastropods 433 necessary to assure a sufficient, measurable sample size after acid treatment). The critical dry weight was determined to be about 4.5 mg. If the fecal pellets from an individual weighed less, they were combined with others from individuals of the same species (i.e.. pooled) until the critical weight was reached or exceeded. It was necessary to pool fecal pellets for all species (Table 1 ). but many individuals of C, imiricatus and T. antonii had sufficiently large and/or numerous fecal pellets that attained or exceeded critical weight. Each dried fecal sample attaining critical weight (i.e.. either (he rectal contents of one individual or a pooled sample of pellets from several individuals within a species) was washed with de-ioni/,ed water onto a 25-mm nitrocellulose filter attached to a Millipore microfiltration apparatus. The water was drained from the appa- ratus with a vacuum pump, leaving the fecal pellets on the filter. Then. I M nitric acid was added to the apparatus and allowed to react with the calcium carbonate in the fecal sample until all ef- fervescence ceased, indicating conversion of calcium carbonate to carbon dioxide accordinc to the followinc reaction: ■Ca(N03),,^^, + C0,,t;, + H,0,^„ CaCO, ,s, + 2 HNO3 ,^^, After effervescence ceased, another dose of nitric acid was added to ensure that the reaction was complete. The filter paper contain- ing the fecal sample was thoroughly washed with de-ionized water, removed to a pre-weighed drying boat, and dried in an oven to a constant weight. The difference between the dry weights prior to and after the acid treatment is considered to be the amount of calcium carbonate removed by the acid treatment plus an assumed negligible amount of other acid-volatile or acid-soluble organic materials. This difference will be referred to subsequently as the fecal carbonate fraction or, simply, ingested carbonate. The fecal carbonate fraction was compared between each spe- cies and between the two subsets of C. nuiiicatus (i.e., those col- lected from limestone and those collected from plants). Means and standard deviations, expressed as proportions, were calculated for all species and were compared by a Kruskal-Wallis single factor ANOVA by ranks. Two species. N. riisei and N. aiii^Kstior. re- quired extensive pooling to produce a sample that reached the critical mass needed for analysis (Table I ). Thus, they were rep- resented by a small number of pooled samples in = 3 and /; = 6, respectively). Accordingly, all species except these two were again compared by a Kruskal-Wallis single-factor ANOVA by ranks to assess the impact caused by extensive pooling and to eliminate any bias caused by it. C. nuiiicatus and 7". aiilonii included both unpooled and pooled samples. Such data were compared within each species (and within each species according to habitat for C. miiricatiis) by means of a t test to assess the impact of pooling. The parametric r test was selected for comparison of within-species proportions because both normality and equality of variance tests were within required limits. All statistical tests were performed using the Sigma Stat statistical package. On the basis of fecal carbonate estimates determined in this study, we also estimated the annual bioerosive impact of these species on the shore ( I ) per individual on the basis of an estimated 48-h defecation cycle (the duration from ingestion to defecation); (2) per individual per year, also on the basis of an estimated 48-h defecation cycle; and (3) total carbonate removed from this shore per species per year on the basis of density estimates by Metz (1997). Because the densities of each species varied with respect to both species and location on the shore, bioerosive impacts were calculated on the basis of mean densities within the occupied range o\ each species, not for the entire platform. The 48-h duration of the littorinid defecation cycle was never tested empirically. Labo- ratory behavior of these animals suggested, however, that the def- ecation cycle, although mfiuenced by numerous extrinsic factors including duration of repose, is probably no longer than a mean 48 h and is likely much shorter. Thus, the carbonate removal calcu- lated t\)r each species using the 48-h estimate is probably a con- servative assessment of the bioerosive impact of the littorinids. No attempt was made to estimate carbonate removal by P. nucleus and C. muiiciitiis from plants because there were no estimates available concerning their density on shore. RESULTS Fecal pellets were removed from between 80 and 130 individu- als of each species (Table 1 ). Between 20 and 56 samples, con- sisting of fecal pellets from one to several individuals, were pro- cessed for all species except A', riisei and N. angiistior. Because the latter produced such small pellets, it was necessary to combine those from several individuals to obtain a sample sufficient for analysis. Thus, 80 A', riisei yielded only three pooled samples for analysis, and 80 A', angiistior yielded only six. Mean fecal pellet size, expressed as area (Table 2), and mean fecal pellet weight (Table 3) were calculated for each species. N. angiistior produced the smallest fecal pellet in terms of both mean size and mean weight per pellet. C. inuricatus produced the largest pellets with respect to mean size, but T. antonii had the largest pellets with respect to mean weight. The pellets produced by both of these species were significantly larger than those produced by the other species with respect to size (Table 2) and weight (Table 3). C. nuiiicatus pellets also differed significantly in size from those of T. antonii. but not significantly, with respect to weight. Intraspecific variation in pellet size was relatively great for all species, as reflected by high standard deviations (Table 2). Dunn's pairwise comparisons following a Kruskal-Wallis single-factor ANOVA by ranks also indicated fecal pellet size to be signifi- cantly different between most species (Table 2). Variation in pellet weight within each species was slight to moderate, as indicated by standard deviation values (Table 3), and there were no significant TABLE 2. Comparison of fecal pellet size, expressed as area (mm"), among all species by a Kruskal-Wallis single-factor ANO\ '.A by ranks (a = 0.05. H = I117.L df = 6. P = <0.00L * = signitlcani difference). CM, Cenchritis inuricatus; NA, Nodililtoriiia anguslior: ND, ,V. dilatala; NR, A', riisei; NZ. A*, ziczac; PN, Planaxis nucleus; TA, Teclarius antonii. Mean fecal pellet area Standard (mm-) deviation CM NA ND NR NZ PN C. nntricatus 4.45 1.61 N. angiistior 0.959 (1.460 N. ciilatata 1.76 0.672 N. riisei 1.21 0.554 N. ziczac 1.4.^ 0.837 P. nucleus 1.73 0.586 T. antonii 3.23 0.837 434 DOBSON-MOORE AND BRITTON TABLE 3. Comparison of fecal pellet weight (mg) among all species by a Kruskal-Wallis single factor ANOVA by ranks (a = 0.05, H = 161.7, df = 6, P = <0.001, * = significant difference). Weight per Standard pellet (mg) deviation CM NA ND NR NZ PN C. iiiuricatus 0.112 U.044 N. angiistior 0.024 0.003 N. dilalaia 0.052 0.011 N. rissei 0.028 0.005 N. ziczac 0.066 0.022 P. nucleus 0.040 0.009 T. tmlanii 0.118 0.019 Ahhrevialions as in Table 2. differences in pellet weight between N. angustior. N. dikitcita. N. riisei. N. ziczac. and P. nucleus. Among the species of Littorinidae. the mean number of fecal pellets per individual generally increased as mean shell size in- creased (Table 4; Fig. 1 ). The smallest littorinid {N. riisei) pro- duced the fewest number of pellets, and the largest species (C. iiuiricalus) produced the most pellets among the littorinids. P. nucleus produced significantly more fecal pellets than any of the littorinids. but each pellet was significantly smaller than that pro- duced by any littorinid of comparable size. Linear regressions were performed to compare the number of pellets as a function of size (shell height) for each species (Table 5). Despite the general trend of increasing numbers of pellets with increasing body size for all species except N. dilatata, the predictive value for all regressions was poor, as reflected by very low adjusted r values, ranging from a low -0.002 for C. nniricatus to a high of 0.37 for N. ziczac. Accordingly, the slopes of the regression lines were not compared between species. Samples of T. antonii, C. muricatus collected from plants, and C. nmricalus collected from rock included both single-individual and pooled (several individuals) pellet samples. We performed / tests performed between the individual and pooled subsets of each group, which detected no statistically significant differences be- tween any of them. The proportion of carbonate (strictly speaking, acid- volatile or acid-soluble material, represented mostly by calcium carbonate) in the fecal pellets varied greatly (Table 6). C. nniricatus fecal pellets collected from plants contained the smallest carbonate fraction (18.6%), whereas T. antonii and P. nucleus contained the largest (93.1 and 92.8%, respectively). There were also three other distinct groupings: (1) M riisei and N. dilatata (88.7 and 88.3%. respec- tively), (2) N. ziczac and C. muricatus from rocks (74.9 and 74.8%, respectively), and (3) N. ani^nstior (67.0%). Dunn"s pairwise com- parisons following a Kruskal-Wallis single-factor ANOVA by ranks also indicated that the fecal carbonate content was not sig- nificantly different between those species with the highest amount (T. antonii. P. nucleus. N. rii.'iei. and N. dilatata) and not signifi- cantly different between those with low to middle amounts (A'. ziczac. C. nniricatus from rocks, N. riisei. N. dilatata. and N. angustior) (Table 6). The carbonate content of fecal pellets from C. muricatus found on plants (18.6%) differed significantly from that of conspecific individuals taken from rock as well as from that of all other species except N. angustior (67.0%). When N. riisei and N. angustior were removed from the data set and the Kruskal-Wallis single factor ANOVA by ranks was performed again without them (Table 6). two new significant dif- ferences appeared between N. dilatata and C. muricatus from rocks and between N. dilatata and N. ziczac. Two large, upper-shore species. C. nniricatus (from rock) and T. antonii. processed the greatest mean amounts of carbonate per defecation cycle (3.07 and 2.44 mg. respectively) and per year (0.546 and 0.435 g, respectively) (Table 7: Fig. 2). The two small- est species. N. riisei and M angustior. processed the smallest mean amount of carbonate per defecation cycle (0.21 and 0.23 mg, re- spectively) and per year (0.037 and 0.040 g. respectively). Ex- pressed on an individual basis. N. dilatata also seems to have only a modest impact, processing only 1.12 mg per defecation cycle or 0.199 g per year. These relationships are much different when population densi- ties are considered. Population densities of the Littorinidae on this shore ranged from 80 indi\'iduals/in- for N. ziczac to 736 individu- als/m- forM dilatata (Table 7). C. muricatus. T. antonii. N. riisei. and N. angustior had densities of 128. 208, 272. and 480 individu- als per m". respectively, within occupied ranges. On the basis of these densities, N. dilatata lemoves the largest amount of carbon- ate from the platform annually, 146.5 g/nr (Table 7; Fig. 3). T. antonii and C. muricatus from rocks had the second and third highest values, removing 90.5 g/nr/y and 69.9 g/m'/y. respec- tively. All other species had much lower values for annual car- bonate removed: N. angustior with 19.2 g/m"/y. N. ziczac with 16.9 g/m'/y, C. muricatus on plants with 15.6 g/m'/y. and N. riisei with 10.1 g/m-/y. DISCUSSION Although the Littorinidae and Planaxidae are primarily herbi- vores, the focus of their feeding varies widely, often according to habitat. Species that occupy the algal gardens of the lower eulit- toral may ingest the algae upon which they live. For example. Littorina littorea is attracted to and feeds upon Uha lactuca. and L ohtu.sata is attracted to this species plus Fucus serratus and .Ascophvlluni spp. (Watson & Norton 1985. Watson & Notion 1987). Other littorinids browse on the surfaces of seaweeds or seagrasses. consuming epiphytic algae, fungi, or microorganisms such as diatoms, protozoa, cyanobacteria. and bacteria (Norton et al. 1990). Several littorinids. especially members of the genus Littoraria. browse on the surfaces of mangroves (Reid 1986. Christensen 1998). Others, such as Bembicium auratinn, feed on TABLE 4. Comparison of number of fecal pellets per individual among all species by a Kriiskal-\\ allis single-factor ANOVA by ranks (a = 0.05. H = 341.8, df = 6, P = <0.(IOI, * = signit'icant difference). Mean number Standard of pellets deviation CM NA ND NR NZ PN C. nuiricalus 35.3 17.1 N. angustior 14.1 7.88 N. dilatata T) T 10.9 N. riisei 8.56 5.37 N. ziczac 26.1 16.8 f nucleus 48.7 16.1 r. iUUi'nii 22.8 7.81 Abbreviations as in Table Carbonate Processing by Jamaica Gastropods 435 1^0 > f 60 pellets per Ipn 6— -H ^ , CM ^ ber of fecal o o ND * '-NZ ■ 1^ 1,0 S 0 it: NR 10 11 12 13 14 15 16 17 18 19 20 21 22 Mean shell height (mm) Fi)>urf I. Mean numbtr of ftcal pellets per individual as a I'unction of shell heij-hl (mm). Krror bars are given onlv in the p()siti\e directions for clarity. Open circle. Planaxidae: filled circles, l.ittorinidae. CM, Cenchrilis imiricaliisi NA, \odilittoriiia angiistior; ND, A', dilalala: NR, N. riisei; NZ, A', ziczac: PN, Planaxis nucleus; TA, Teclarius aiUonii. the microflora in the mud beneath the mangroves (Branch & Branch 1981a). Perhaps the majority of littorinids that occupy rocky shores browse directly on rock surfaces and ingest detritus, small live animals, egg capsules, sand grains and fragments of rock, lichens, and epilithic and endolithic algae (Norton et al. 1990). Species occupying the eulittoral on rocky shores probably focus on epilithic and/or epiphytic algae with those inhabiting similar habitats sometimes seeking different foods. For example. the stomach contents of L. plaiiaxis. a eulittoral snail of granitic shores in California, include microphytes such as green algae, blue-green algae, and abundant quantities of diatoms, the latter probably the primary food source (Foster 1964). On the same shore. L sTiitiilaki often occurs on and grazes macrophytic algae such as Cliulophoiv. Pclvetia. and Rhodofilossuin (Dahl 1964). Higher upon the shore where drier conditions are more com- mon, epilithic and epiphytic algae usually disappear, often to be replaced by a surface film of encrusting black lichens, which are grazed upon by several species of littorinids including N. unijhs- ciata in Australia (Branch & Branch 1981b), On the highest zones of tropical limestone shores, however, endolithic algae, especially cyanobacteria. are present (Peckol & Guarnagia 1989) and provide high-zoned littorinids a food source in an otherwise nearly barren habitat (North 1954, Hodgkin et al. 1959, Underwood 1984a. Un- derwood 1984b, Underwood 1984c). A few rocky-shore littorinids. such as C. miiricatiis. range to the fringe of maritime vegetation and derive additional nutrition by grazing upon their stems and leaves. Planaxidae occupv bedrock or cobble shores and, like the Lit- torinidae, employ a radula to scrape the hard surfaces upon which they live. P. nucleus lives in a high-energy, mid-intertidal habitat among boulders or large rocks usually surrounded by loose, mobile stones. Planaxids generally have a large, powerful foot that grips the substrate tenaciously, preventing dislodgemenl by the waves or rolling stones (Houbrick 1987). Most previous studies of intertidal gastropod fecal content have emphasized the nutritious organic fraction, not the inorganic com- ponent. Accordingly, most of these studies employed either ashing or isotope analysis to estimate the organic fraction (Black et al. 1988, Norton et al. 1990, Newell & Barlocher 1993. Christensen 1998). For example. Black et al. (1988) examined the fecal content of six species of western Australian intertidal herbivorous gastro- pods and one chiton by an ashing technique. That portion of the feces not removed by ashing represeiited the non-organic feces fraction, composed primarily of calcium carbonate, the rock type upon which these molluscs lived. Expressed as percentages, the feces of the pulmonate limpet Siphonaria kurracheenis contained 69% carbonate, the feces of the acmaeid limpet CalliseUa onychitis contained 77% carbonate, the feces of L. nnifasciata contained 73% carbonate, the feces of N. ausinilis contained 80% carbonate, the feces of the nerite Nerita atramentosa contained 83% carbon- ate, the feces of the acmaeid limpet Patelloida alticostam con- tained 84% carbonate, and the feces of the chiton Clavarizona liirrosa contained 89% carbonate. Some of these species also pos- sessed distinctly different radulae. For example, the cusp size of the S. kurracheenis radula was much smaller and that for N. atra- nienlosa and much larger than other species with a similar body size. Black et al. (1988) concluded that such differences might influence competitive interactions of species occupying the same shore level, although they recognized that either specific behaviors or the nature of the rock surface might further mediate such inter- actions. The vertical ranges of several of the Litlorinidae of the Jamai- can limestone platform overlap to a greater degree than some of the Australian species studied by Black et al. (1988). a situation that should only enhance the possibility of competitive interac- tions. There is little evidence, however, of such interactions among the Jamaican littorinids (Britton 1992. Lang et al. 1998. and this study). Furthermore, the nature of the radulae of the Jamaican species, although displaying some species-specific differences, are much more similar (Bandel & Kadolsky 1982) than those of the species studied by Black et al. ( 1988). The proportions of carbon- ate contained in the feces of both Australian and Jamaican inter- tidal molluscs, however, were very similar, an indication that both groups rely to a considerable degree on food attached to or within the substratum. It also possibly implies that such a food source is not especially limiting for the molluscs feeding upon it. P. nucleus was found to have a very high fecal carbonate frac- tion (92.8%). consistent with expectations for individuals living in a high-energy, limestone boulder environment devoid of epilithic macrophytes or epiphytes. The mobile stones that dominate this beach are poorly suited for either algal attachment or growth. Likely food sources for P. nucleus include either diatoms washed up on the rocks by waves or endolithic algae, the latter necessi- tating ingestion of considerable rock carbonate while grazing for TABLE 5. Linear regressions predicting number of fecal pellets contained within rectum as a function of shell height (mm) for all species. Linear regression equation Adjusted r^ PUiiuiMis luaieiis Noilililtorina riisei N. zic-ac N. aiigustior N. clilcitata Tectarius antonii Cemiirilis muricatus NP = 8.33 + (3.00 X SH) 0.039 NP = -4.93 + (1.14 X SH) 0.058 NP = -28.3 + (3.60 X SH) 0.336 NP = -10.8 -f (2.03 X SH) 0.028 NP = 43.4 - (1.63 X SH) 0.007 NP = -14.2 -I- (2.61 X SH) 0.232 NP = 14.7 + (1.13 X SH) -0.002 Abljreviiilions: NP = number of pellets; SH = shell height. 436 DOBSON-MOORE AND BrITTON TABLE 6. Comparison of the fecal carbonate fraction expressed as a proportion among all species by a Kruskal-Wallis single-factor ANOVA by ranks (a = 0.(15; H = 2(11.7; df = 7; P = <0.001; * = significant difference! and among all species except N. rissei and .V. angustior by a Kruskal-Wallis single factor ANOVA by ranks (a = 0.05; H = 194.4; df = 5; P = <0.001; t = significant difference). Mean carbonate fraction Standard deviation CiVI-P CM-R NA ND NR NZ PN C iiniriciitits on plams 0.186 0.133 ^ C. miiiicatiis on rocks 0.748 0.057 *t — N. anguslior 0.670 0.142 N. dilarata 0.883 0.043 *t t N. riisei 0.887 0.058 * N. ziczac 0.749 0.049 *t P. nucleus 0.928 0.049 *t *t T. unlonii 0.931 0.040 *t *t Abbreviations as in Table 2. except CM-P. C. muricatus from plants; CM-R. C. muricatus from rocks. this food source. Our data support the latter as a primary food source for P. nitckus. The only other planaxid for which diet has been studied is the Indo-Pacit'ic P. siilcaiiis. which grazes upon epilithic algae but lives in a different habitat (i.e., bedrock shores where such algae are frequently abundant) (Rohde 1981). The low-carbonate fraction in the fecal pellets of C. niiiiicatus collected from plants (18.6%: Table 6) is likely the result of their most immediate diet prior to collection. If these individuals were grazing upon the plants or. more likely, on the epiphytic flora upon them, then the fecal carbonate fraction should be lower than that of conspecifics feeding exclusively upon the rocky substratum. The longer the duration of grazing upon plants, the greater the expected reduction of carbonate in the fecal pellets. In fact, C. muricatus from plants had the second highest standard deviation of all spe- cies tested (Table 6), indicating considerable variation among in- dividuals, as should be expected. It has also been obser\ed that C. iiuiricanis moves readily to and from R. amerwanii branches and the rocky platform (JCB. unpublished data). Thus, the carbonate present in the fecal pellets of specimens collected from plants is inteipreted as carbonate ingested when the individual grazed upon limestone. Conversely, some C. muricatus individuals collected from rock could have a disproportionately low fecal carbonate content due to having I'ecently grazed on the plants. N. ziczac and A^. riisei occupy approximately the same low shore habitat, characterized by frequent wave splash, occasional tidal i)iundation, and some epilithic and much endolithic algae on which to graze. The former is largest foreshore littorinid and pro- duced the largest fecal pellets of those studied from this habitat (Tables 2 and 3). The latter, on the other hand, is the smallest species studied, but its fecal pellets were slightly larger and heavier than those of A', angustior. facilitated perhaps by a larger storage space within more globose shell. The feces of both species con- tained considerable carbonate, N. ziczac with a mean of 74.9% and N. riisei with a mean of 88.7%. The differences, although slight, might indicate a tendency toward resource partitioning, with N. ziczac perhaps taking more epilithic algae than N. riisei. On the other hand, these differences are not significant, perhaps due to the small number of pooled samples of N. riisei. The three pooled samples of A', riisei (Table 1) exhibited approximately the same variation as that for 38 samples for A', ziczac (SD of mean pro- portions = 0.058 and 0.(149, respectively; Table 6). The greatest variation among samples occurred with A', angus- tior. which also produced the smallest fecal pellets in terms of size (Table 2) and weight (Table 3), likely the result of a slender shell providing a small internal volume. This species was represented by only six pooled samples. Its feces contained a mean carbonate TABLK 7. Means and standard deviations of quantities of carbonate processed per individual based on an estimated 48-h defecation cycle and derived estimates of mean quantities of carbonate processed annually per individual and the total amount removed annually per species per ni" on the Jamaican shore based on density estimates cited in Metz(1997). 48-h defacation cvcle Derived estimates Mean amount of carbonate M ean amount Mean Total amount processed/indl vidual Standard reniov ed/indiv idual/y density removed/v (mg) deviation clc. Species arranged from Ioh shore (left I to high shore (right). CM. Cenchritis muricatiis: N.\, Nodilittorina aiigus- tior; ND, A', dilatata; NR, S. riisei: NZ, A', ziczac; PN. Planaxis nucleus: TA, Tectarius anlonii; CM-P. C muricatus from plants: C'M-R, C. iiiuriialus from rocks. fraction of 67.09f , with a staniiard deviation, expressed as a pro- portion, of 0.142 (Table 6). The range of A'. aiit>iistic}r on the platform overlaps the frequently wetted upper distribution of both N. ziczac and N. riisei but also extends several meters landward into the lower range of N. dilatata where the rocks can remain unwetted by seawater for many days (Lang et al. 1998). One should expect the diet of A', angiistior. of necessity, to include a considerable quantity of either endolithic algae or lichens, the former, at least, accompanied by a considerable quantity of car- bonate. These samples, however, show that N. angiistior fecal pellets possess the smallest proportion of carbonate of all species tested except for C. muricatus from plants. Because N. anguslior is not known to graze upon maritime vegetation along this plat- form (JCB. personal observation), such behavior is not a potential cause for the lower carbonate content of the feces. If N. angustior focused upon the epilithic lichens, which are abundant within its range, fecal carbonate might be somewhat diminished. Such a diet might also suggest possible specialization in the radular structure, but Bandel and Kadolsky ( 1982) found the radulae of M ziczac. N. angustior. N. dilatata. and N. riisei to be very similar. It seems unlikely, therefore, that N. angustior has such a specialized diet, but whether it does or does not is simply conjecture and requires further testing. When all species were compared, the Knjskal-Wallis single factor ANOVA by ranks showed no significant differences among the four species with the highest fecal carbonate fraction (93.1- 88.3%: T. antonii. P. nucleus. N. riisei. and A', dilatata: Table 6). There were no significant differences found among N. riisei. N. dilatata. N. ziczac. N. angustior. and C. muricatus on rocks (88.7- 74. 8*^). but there were significant differences between C. murica- tus on plants (18.6%) and all other species except A', angustior (67.0%). There were also significant differences within C. muri- catus on plants (18.6%), those on rocks (74.8%), and P. nucleus (92.8%); and between C. muricatus from plants, those from rocks, and T. antonii (93.1%). When N. riisei and N. angustior data were remo\ed and the ANOVA was perfomied again without them, all of the significant differences previously detected remained unchanged, and two new significant differences appeared: one between A', dilatata and C. muricatus from rocks and the other between N. dilatata and N. ziczac (Table 6). Apparently, the large variation within the N. angustior samples obscured these differences when all eight groups were compared. In contrast, the large variation within the C. muricatus samples from plants did not alter the ANOVA re- sults, probably because the mean proportion of carbonate was so much lower than the others that the high variation did not obscure any other significant differences. Our original hypothesis, that fecal carbonate should \ary with position on shore and increase toward the landward portion of the platform, was poorly supported by the data. Another pattern emerged, albeit unexpected and subtle, which might also be in- dicative of resource partitioning (Fig. 4). Depending upon shore position three pairs of littorinid associates are present, one of each pair having a relatively high fecal carbonate content and the other having somewhat to significantly less. On the high shore. T. an- tonii had the greatest amount of fecal carbonate, whereas rock- dwelling C. muricatus had the sixth highest amount and signifi- cantly less than that of T. antonii (Table 6). It is apparent from this study that C. muricatus. unlike T. antonii, feeds both upon endo- lithic algae and the epiphytic biota clinging to certain maritime plants, hence mediating the carbonate fraction of its feces. On the midshoi-e, N. dilatata had the fourth greatest fecal carbonate frac- tion, and N. angustior had the seventh, a nonsignificant but dis- tinctive difference. On the low shore. N. riisei had the third great- est fecal carbonate fraction, and N. ziczac had the fifth greatest, again a nonsignificant but distinctive difference. Perhaps the spe- cies pairs at the lower shore levels focus on somewhat different food sources, as does the high-shore pair. Alternately, and perhaps more parsimonious with these data, either food may be nonlimiting throughout the platform habitat, such that neither competition nor 160 140 120 oj 100 g 80 E 40 « 20 NR NZ CM-R NA ND Species Figure 3. Estimated annual carbonate removed per ni" from the Ja- maican limestone platform shore by littorinid species, based on density estimates from Melz ( 1997). No density values were available for either Planuxis nucleus or Cenchritis muricatus from plants. Species ar- ranged from low shore (left) lo high shore (right). CM, Cenchritis muricatus: NA, Nodilittiirina angustior: NT), .V. dilatata: NR, A', riisei: NZ. A', ziczac: PN. Planaxis nucleus: T.\. Tectarius antonii: CM-P, C. muricatus from plants: CM-R, C. muricatus from rocks. 438 DOBSON-MOORE AND BRITTON NR NZ NA ND TA CM-R CM-P Species Cobble Beach rTVTi Low Shore ■■■ Mid Shore ivx^ High Shore Figure 4. Mean carbonate fraction in fecal pellets with species grouped according to habitat from low shore (left) to high shore (right I. CM, Cenchritis nuincatiis; NA, Nodililtorina aiigustior: ND, N. dilalala: NR, N. riisei: NZ, N. ziczac: PN, Planaxis nucleus. TA. Tec- tarius anlonii; CM-P, C. muricatus from plants; CM-R, C muhcatus from rocks. resource partitioning is an issue or each species is broadly oppor- tunistic with respect to food selection, grazing on epilithic primary production when available and on endolithic primary production when epiphytes are reduced in abundance. The platform littorinids have a significant bioerosive impact on this tropical limestone shore. The combined grazing of C. iniiri- catKs and T. antonii strips considerable carbonate (160.4 g/m"/y) from the upper shore. This translates to the removal of about 1 .472 kg annually along the approximately 500-m-long platform and across an upper shore band about 24 m wide. Although N. dilutuia is a moderate-sized species with each individual having only a modest bioerosive impact, it also has the highest population den- sity (a mean of 736 individuals/rn"). Thus, it significantly impacts the central platform, removing limestone at the rate of 146.5 g/m"/y (Table 7; Fig. 3). and. because its range covers a band about 17 m wide, it removes a little more than 1.245 kg annually from this shore. The lower shore, although having more littorinid spe- cies, experiences the least amount of littorinid bioerosion. both per species and in total (Table 7). Several factors contribute to this diminished impact. The lower shore is horizontally limited, being 7 m wide at best and hardly more than 5 m wide if the range of N. angustior is excluded. Population densities of A', ziczac are very low. minimizing its impact. Although both N. ans;itstit>r and N. riisei can be very abundant locally, dense populations are patchy and often separated by areas of sparse density. Their small size also reduces the bioerosive impact of each individual. Finally, the presence of epilithic algae on some parts of the lower shore may provide an almost carbonate-free alternate food source. Thus, the total annual carbonate removal rate, combining the grazing impact of N. riisei, N. ziczac. and N. anguslior. is 46.2 g/m"/y. or almost 140 kg annually. It should be noted, however, that other, much larger, grazing gastropods, especially Neriui tesselata. N. pelorontii. and N. versicolar. and one common chiton, Acaiilho- pleiira graindata. occupy the lower shore (Minton & Gochfeld 2001 ) and have the potential to remove considerably greater quan- tities of carbonate than the low shore littorinids. For the entire platform, all littorinid species account for the removal of a little more than 2.850 kg of limestone annually. In conclusion, all species studied ingested considerable quan- tities of carbonate, probably indicating that all rely to some degree upon endolithic algae for food. C. muricatus was the only species clearly identified as also having an alternate food source, probably epiphytes on the branches, stems, and leaves of the maritime shrub R. anicriciuni. Although the rock-dwelling Littorinidae ingested different quantities of carbonate, ranging from 67-93'/f . there was no clear relationship between the amount of carbonate ingested and the position each species occupied on shore. However, species pairs that cohabitate on particular shore levels possessed notice- ably different mean fecal carbonate fractions, leaving open the possibility that some form of resource partitioning might be in- volved. There was no indication, however, what it might be. The quantity of carbonate in P. nucleus fecal pellets was similar to that found in the several littorinid species studied, but the size and morphology of P. nucleus fecal pellets (spindle shaped) were sig- nificantly different from those found among the Jamaican Littorin- idae (ovoid shaped). The Littorinidae occupying the rocky inter- tidal platform remove significant quantities of carbonate from the shore each year and should be considered important bioerosive species. ACKNOWLEDGMENTS We sincerely appreciate the help provided in Jamaica by the Resident Directors of the Hofstra University Marine Laboratory. Deborah Gochfeld and Dwayne Minton (1998). and Patricia Han- kenson and Mike Lewis (1999). Jens Tang Christensen kindly contributed additional A', angustior and N. riisei specimens from the limestone platform. Robert McMahon and David Hicks pro- vided numerous useful comments and suggestions, many of which we have incorporated in the final draft. We are especially indebted to Eugene Kaplan and the late Edgar Ross. who. together, had the vision and foresight to transform a small, minor, resort property into a teaching and research laboratory readily accessible to a broad spectrum of the academic community. The Texas Christian University Department of Biology provided some financial assis- tance for this project. Portions of this study first appeared as a TCU Department of Biology Master's Thesis by the senior author, which also includes additional data not presented here. Bandel. K. 1974a. Studies of Littorinidae from the .Allantic. Velige 92-114, Bandel. K. lL'74b. Faecal Pellets of Aniphineura and Prosobranchia (Mol- lusca) from the Caribbean Coast cif Columbia. South America. Senck- eiiber.i'iiina Mariliiiiu. 6(1): 1-31. Bandel. K. & D. Kadolsky. 1982. Western AUanlic specie.s of N. (Gas- LITERATURE CITED 17: Iropoda: Prosobranchia): comparative morphology and its functional, ecological, phylogenetic and taxonomic implications. Veliger 25:1—12. Black. R.. A. Lymbery & A. Hill. 1988. Form and function: size of radular teeth and inorganic content of faeces in a guild of grazing molluscs at Rottnest Island. Western Australia. J. Exp. Mar. Biol. Ecol. 121:23-35. Branch. G. M. & M. L. Branch. 1981a. Competition in Bembicium auratum Carbonate Processing by Jamaica Gastropods 439 (Gastropoda) and its effect on microalgal standing stock in mangrove muds. Occologiii 46:106-1 14. Branch. G. M. & M. L. Branch. 1981b. Experimental analysis of intraspe- cific competition in an intertidal gastropod. Lirtmina imifiisciciiu. Aiisr. J. Mar. Freshwaler Res. 32:573-589. Britton. J. C. 1992. Evaporative water loss, behavior during emersion, and upper thermal tolerance limits in seven species of eulittoral-fringe Lit- tonnidae (Mollusca:Gastropoda) from Jamaica. In.- J, Grahame. P. J. Mill & D. G. Reid. editors. Proc. Third Int. Symp. Littorinid Biol. London: The Malacological Society of London, pp. 69-83. Chnslensen, J. T. 1998. Diet in. Lmoraria. Hydrobiohfiiii 378:235-236. Dahl. A, L. 1964, Macroscopic algal foods o( Liinniui pliiii(i.\i.s Philippi and Litiuriim scutuhiui Gould. Veliger 7:139-143. Foster. M. S. 1964. Microscopic algal food of Littorina plaihi\is Philippi and Lillorina .':cii!iilata Gould (Gastropoda:Prosobranchiata). Veligcr 7:149-152. Hodgkin. E. P.. L. Marsh & G. G. Smith. 1959. The littoral environment of Rottnest Island. J. Rin. Soc. W. .Aii.<:t. 42:85-88. Houbrick. R. S. 1987. Anatomy, reproductive biology, and phylogeny of the Planaxidae (Cerithiacea: Prosobranchia). Smithsonian Contrib. Zool. Number 445. Lang. R. C. 1995. Ecological aspects of the distribution of Jamaican lit- torines and an empirical study of the thermoregulatory value of shell sculpture (Master's Thesis). Fort Worth: Texas Christian University. Lang. R. C. J. C. Britton & T. Metz. 1998. What to do when there is nothing to do: the ecology of Jamaican intertidal Littorinidae (Gas- tropoda:Prosobranchia) in repose. Hydrohiolngiu 378:161-185. Lewis. J. B. 1960. The fauna of rocky shores of Barbados. West Indies. Can. J. Zoo/ 38:391-435. McMahon. R. F. 1 990. Thermal tolerance, evaporative water loss, air-water oxygen consumption and zonation of intertidal prosobranchs: a new synthesis. Hydrobiologia 193:241-260. McMahon. R. F. & J. C. Britton, 1991. The relationship between vertical distribution, rate of evaporative water loss, behavior during emergence. and morphometries in six species of rocky shore gastropods from Prin- cess Royal Harbour. Western Australia. In: F. E. Wells. D. I. Walker. A. H. Kirkman & R. Lethbridge. editors. Proc. Third Int. Mar. Biol. Workshop: The marine flora and fauna of Albany. Western Australia. Volume II. Perth: Western Australia Museum, pp. 675-692. Metz. T. L. 1997. An empirical assessment of shell size, nodulosity. and surface area in relation to the position of Jamaican Littorines on shore (Master's thesis). Fort Worth: Texas Christian University. Minton. D. & D. J. Gochfeld. 2001. Is life on a tropical shore really so hard? The role of abiotic factors in structuring a supralittoral molluscan assemblage. / Shellfrdi ft'.s. 20:477-t83. Newell. S. Y. & F. Barlocher. 1993. Removal of fungal and total organic matter from decaying cordgrass leavels by shredder snails. J. Exp. Mar. Biol. Ecol. 171:39-49. North. W. J. 1954. Size distribution, erosive activities and gross metabolic efficiency of the marine intertidal snails Lirinrina plaiui-xis and L scutulata. Biol. Bidl. 106:185-197. Norton. T. A.. S. J. Hawkins. N. L. Manley. G. A. Williams & D. C. Watson. 1990. Scraping a living: a review of littorinid grazing. Hydro- biologia 193:117-138. Peckol. P. & S. Guarnagia. 1989. Zonation and behavioral patterns of the intertidal gastropods N. (Tectiniiius) antoni (Philippi. 1846) and Nerita versicolor Gmelin. 1791. in the Bahamas. Veliger 32:8-15. Reid, D. G. 1986. The littorinid mollusca of mangrove forests in the Indo- Pacific region, the genus Littoiana. London: British Museum (Natural History). 216 pp. Reid. D. G. 1989. The comparative morphology, phylogeny. and evolution of the gastropod family Littorinidae. Philos. Trans. R. Soc. London B. Biol. Sci. 324:1-110. Rohde. K. 1981. Population dynamics of two snail species. Plana.xis sul- catiis and Cerithium mimiliferuin. and their trematode species at Heron Island. Great Barrier Reef. Oecologia 49:.344-352. Underwood. A. J. 1984a. The vertical distribution and abundance of inter- tidal microalgae on a rocky shore in New South Wales. J. Exp. Mar. Biol. Ecol. 78:199-220. Underwood. A. J. 1984b. Microalgal food and the growth of the intertidal gastropods Nerita atrunientasa Reeve and Beinbiciwn nanum (Lama- rck) at four heights on shore. J. Exp. Mar. Biol. Ecol. 79:277-291. Underwood. A. J. 1984c. Vertical and seasonal patterns in competition for microalgae between intertidal gastropods. Oecologia 64:211-233. Vermeij. G. J. 1973. Morphological patterns in high-intertidal gastropods: adaptive strategies and their limitations. Mar. Biol. 20:319-346. Watson. D. C. & T. A. Norton. 1985. Dietary preferences of the common periwinkle Littorina littorea. J. E.xp. Mar. Biol. Ecol. 88:193-211. Watson. D. C. & T. A. Norton. 1987. The habitat and feeding preferences of Littorina obtusata (L.) and Littorina muriae Sacchi et Rastelli. J. Exp. Mar Biol. Ecol. 1 12:61-72. Zar. J. H. 1984. Biostatistical analysis. 2nd ed. Englewood Cliffs: Prentice- Hall. 718 pp. ./(niival nl Shc-mish Rcscanh. Vol. 20. No. 1. 44l-i46. 2001. ELLOBIIDAE— LOST BETWEEN LAND AND SEA ANTONIO M. DE FRIAS MARTINS Departmeiil of Biology. University of the A{-orc'.s. Apaitiulo 1422 9501 -HOI. Poiita Dclgacla Scio Miguel. Azores. Portugal ABSTRACT The Ellobiidae are a diverse group ot arcliaeopuliiionate .snails living mostly near the sea. Their shell length ranges from barely 2 mm to over 100 mm. Six structural types of reproductive system set the boundaries for the subfamilial division: Pythiinae. Ellobiinae, Carychiinae. Cassidulinae. Pedipedinae. and Melampodinae. This division is in general supported by four structural types of nervous system and by the internal morphology of the penial comple.x. The halophile ellobiids are coinmonly associated with the supralidal fringe of mangroves and salt marshes; they are also important components of the supratidal biota of the mobile rocky shore. The various species occupy different portions of the shore. In rocky habitats, in addition to horizontal /onation, the ellobiids also partition their vertical distribution. The purpose of this study is to describe the /.onation patterns of the halophile ellobiids of the cobble shores of the Ai,-ores and. in less detail, of the associated malacofauna. KEY WORDS: Ellobiidae, systematics. morphology, anatomy, ecology, cobble shores, A(,-ores INTRODUCTION The Ellobiidae have long been recognized as a cohesive ta.xo- noniic unit (Lamarck 1809) and were illustrated in the most im- portant iconographic works of the nineteenth century ( Reeve 1 84 1 , Reeve 1877. Kiisler 1844. Kobelt 1897-1901), Classification was based exclusively or primarily on shell morphology. Pfeiffer (18.'i6, 1876). elaborating on shell descriptions of all species known at the time and adding many of his own, provided an important monographic treatment for the group. Zilch's (1959) subfamilial arrangement was also based on shell characteristics. Radular (Odhner 1925) and anatomical features have since been included in taxonomic research and have contributed greatly to the clarification of the relationships of the various taxa (Morton 1955. Martins 1996a, Martins 1996b. Martins 1998), Ellobiids are important members of the supratidal communities of marshes and mangroves, and their contribution to the ecological balance of those environments has been soinewhat recognized in studies dealing with their general biota (Berry i96.\ Brown 1971. Sasekumar 1974. Murty & Balaparameswara Rao 1977, Li & Gao 1985). More specific studies have been done on Melampus biden- tatus and M. coffeus (Hausman 19.^2. Holle & Dineen 1957. Mor- rison 1958. Mortison 1959, Golley 1960. Grandy 1972. Russell- Hunter et al. 1972. Orton 1976). The various species of halophile ellobiids occupy different por- tions of the shore, their distribution being loosely related to their Figure 1. Shell, reproductive system, and nervous system of selected species of Ellobiidae. exeniplifyinj; the structural patterns characterizing the various suhramiiies, \. Pythiinae Pythia scarahaeiis (L., 1758); B. Ellobiinae Ellobium aiirismidac (E., 1758): C, Carychiinae Carychium minimum (0,F. Miiller. 1774); D, Cassidulinae Cassidiila aurisfelis (Bruguiere, 1789); E, Pedipedinae Pedipes pedipes (Bruguiere. 1791); F, Melampodinae Melampus coffeus (E., 1758), a. cerebro-pedal connective; h. pleural ganglion; c, pleuro-parietal connective; d, parietal ganglion; e, pedal ganglion; f. parieto-visceral connective; g. visceral ganglion; h, cerehro-pleural connective; i, cerebral ganglion; j. ovotestis; k, her- maphroditic duct; I. posterior mucous gland; m. bursa; n, bursa duct; o, posterior vas deferens: p, vagina; q, prostate gland; r, anterior mucous gland; s, albumen gland. (Adapted from Martins iy96a.) 441 442 Martins Figure 2. Ellobiids of the cobble shores of the Azores. A, Aiiriciiliiiella bidentata: B. Myosotella myosolis; C, Ovalella rulcani: D, Pedipes pedipes; E, Pseiidomelampiis exiguus. Scale bars = 1 mm. taxonomic affiliation. The Pythiinae {Pythia and. to a lesser extent. Myosotella) venture farther inland and may live in an almost ter- restrial environment, whereas the Pedipedinae prefer the upper intertidal. Within the Melampodinae. some species of Mehimpus are found usually in mangroves and marshes, whereas others and Tralia prefer cobble shores. The ellobiids abound also on cobble shores, and attempts have been made to quantify their horizontal and \ertical distributions on such environments (Martins 1980, Martins & Cunha 1992). Mor- ton et al. ( 1998) summarized the ecology of ellobiids on the cobble shores of the Ac^-ores. This paper will briefly introduce the Ellobiidae as a morpho- logically and anatomically highly diverse family and. reinterpret- ing data from Martins ( 1980), present a model for their distribution along the cobble shores of the Ai;ores. MATERIALS AND METHODS Morphology and Aiialomy The ellobiids range in shell size from barely 2 mm (Leuconop- sis) to 100 mm (Ellohiiim): most species, however, do not exceed 25 mm. The smallest taxa inhabit moderately exposed rocky shores, whereas the largest species prefer the still backwaters of mangrove swamps. Currently, six subfamilies of the Ellobiidae are recognized: Pythiinae, Ellobiinae, Carychiinae. Cassidulinae, Pedipedinae, and Melampodinae. Although shell morphology and internal anatomy have been considered in delimiting the subfamilies, these are based primarily on characteristics of the repri)ductive and nervous sys- tems (Morton 1955, Martins 1996a, Martins 1996b, Martins 1998). The dentate appearance of the shell aperture establishes a gen- eralized ellobiid profile. However, the elaboration of the apertural teeth is far from uniform and, although relatively constant at the genus level, shows variability within and similarities across the subfamilies, which is not consistent with the corresponding inter- nal anatomy. For example, a tridentate inner lip with the middle tooth strongest is seen in Myosotella (Pythiinae). Ellohium (Ello- biinae). and Tralia (Melampodinae); a naiTOW. elongated aperture, typical oi Melainpiis, is also found in Microtialia (Cassidulinae) and Pseudomelampus (Pedipedinae); a double columellar tooth, typical of Pedipes. is also found in some Aiiriciilastra (Ellobiinae) as well as in Cieedonia (Cassidulinae). The anatomy of the reproductive and nervous systems shows a series of patterns that lead to more consistent assemblages (Fig. I ) (see also Martins 1996a. Martins 1998). The subfamilies can then be characterized as follows: ( I) Pythiinae: monaulic, entirely glan- dular pallial gonoducts and long visceral nerve ring (pleural. parietal and visceral ganglia and respective connectives) with both parieto-visceral connectives of about the same length; (2) Ellobii- nae: diaulic. entirely glandular pallial gonoducts and long visceral nerve ring with the right parieto-visceral connective much shorter than its left counterpart; (3) Carychiinae: monaulic. entirely glan- dular pallial gonoducts, the prostate gland anterior to the mucous gland, whereas in all other subfamilies both glands are located side-by-side; nervous system as in the Ellobiinae; (4) Cassidulinae: monaulic. entirely glandular pallial gonoducts. the distal portion of the bursa duct transformed into a nonglandular, well-developed vaginal atrium where the oviduct empties; nervous system with visceral and esophageal (cerebral and pedal ganglia, cerebro-pedal TABLE 1. Species of mollusks mentioned in the text and abbreviations used in Table 2. Species Halophile pulmonates (ellobiids I AuricuUnella bidentata (Montagu. 1808) Myosotella myosolis (Draparnaud, 1801) Ovalella viilcani (Morelet, 1860) Pedipes pedipes (Bruguiere, 1789) Pseudomelamptis e.xigniis (Lowe, 1832) Terrestrial pulmonates Eiieontihis fiilvKs (O. F. Miiller. 1774) Lauria anconostoma (Lowe, 1831) Leiostyla fuscidula (Morelet, 1860) Oestophora barhula (Rossmassler. 1838) Oxyehilus drapamaudi (Beck. 1837) Punctum azoricuw (de Winter. 1988) Systellomatophoran Onchidella celtica (Cuvier. 1S17) Prosobranchs Alvania mediolilloralis Gofas, 1989 Assiminea eliae Paladilhe, 1875 Cingiila trifaseiata (Adams. 1798) Fossarus ambiguus (Linnaeus, 1758) Littoiina striata King & Broderip. 1832 Manzonia unifasciata (Dautzenberg, 1889) Melaraphe neritoides (Linnaeus, 1758) Paliidinella littorina (delle Chiage, 1828) Patella gomesii Drouet, 1858 Peringiella orwmnuscae (Gofas, 1990) Thais haemastoma (Linnaeus, 1767) Bivalves Cardita calyculata Linnaeus, 1758 Lasaea adansoni (Gmelin. 1791) Myselta bidentata (Montagu. 1803) ■Abbreviation Ab Mn Ov Pp Pe Ef La Lf Ob Od Pa Oc Am Ae Ct Fa Ls Mu Mn PI Pg Po Ts Cc Lr Mb Ellobiidae 443 TABLE 2. Number of specimens of ellobiids and presence (x) of associated malacofauna per m' along transects on five cobble shores of the Azores. Transect A 3 - 8 - - - - - - - 3 4 22 23 - 105 105 60 40 1 3 1 75 290 10 140 600 250 50 ~> 1 - 16 50 - 170 600 55 75 13 _ - 3 2 _ 15 63 23 7 2 X X X X X X X X X X X X X X _ _ - X X X X X X - X X X X _ _ X X Species 16 15 14 13 12 U 10 y 8 7 6 5 4 3 2 1 0 -I -2 -3 -4 Mn 6 S 2 12 4 6 y 2 - Ov - ___----- Pe - __-_--_- Pp - _------- Ab - __------ Pa X _____--- - --_- Ef X -------- - -_-- Lf X X------- - _-_- La X _-_-__-- _ -_-_ Ae - -xxxxxxx x xxxx x Mn - -------X - --XX X Ts - -------X - -XX- Fa - -------------- PO _ -------------- Pg - -------------- Mn 17 10 70 Ov 3 Pe _ _ - Pp _ - _ Ab La X Oh X Od X Ac X X X Mn Ls PI Ct - - - Fa Transect C Mm 270 550 440 110 170 23 5 - - - Ov 15 6 12 5 10 5 1 15 2 Pe - 120 1,140 120 700 720 210 15 20 2 Pp - - - - 25 85 75 3 20 6 Mn -------- X X Ls - - - - X X X X X Ct -------- Transect B 70 80 90 130 20 - - - - 1 10 13 140 1.200 4.850 2.680 590 65 1 2 2 1 30 430 47(1 - - - - 1 2 10 270 950 1,100 320 35 1 - — _ 5 75 50 60 70 12 1 X X X X X X X X X X X X - - X X X X - X X X X X - X X X X X - - X X X X - - - X X X - - - - X - -------- - -XXX Fa - - - - - - - - - X - X X Transect D Mm 15 10 15 25 15 57 30 45 75 41 42 45 7 11 2 - - - - Ov -------- 1- 3 26 80 140 95 3 - - - Pe - - - - - 15 33 5 5 21 25 100 65 50 20 25 - - - Pp ------ 1- - - 7 31 106 105 85 80 15- Mn ----XXXX X XXXX x x x xxx Ls X - -xxx X X X xxx PI ------- _ - - - X X X X X X -- Fa -_-__-- - - - _ - - - X X X XX Coiuiinicd 444 Martins TABLE 2. continued Species 16 15 14 \i 11 10 -3 -4 Mn Ov Pe Pp Ah Ob Mn Ls PI Ct Mu Fa Am Lr Mb Cc Ps 20 40 15 3U 70 Transect E 85 40 72 15 - - - - -1 5 23 400 255 515 500 15 47 80 700 1.730 1 .860 1.440 200 1 - - 60 485 1.250 1.760 1 .080 1 .000 _ _ _ _ ■) 5 ~) _ 60 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X - - X - X X X X X - - X X X - X X X X X - - - X X - - - _ _ _ - X X X X X High tide level at 0; remaining numbers are meters away from 0. with negative numbers being seaward. See text for transect identification and Table I for species abbreviations (data from Martins 1980). connectives and pedal commissure) rings moderately wide, ap- proximately of equal size; (5) Pedipedinae: monaulic to incipient semidiaulic, pallial gonoducts with glandular coverage only on the proximal (posterior) half; visceral nerve ring extremely concen- trated, right connectives of the esophageal nerve ring shorter than their left counterparts; and (6) Melampodinae: advanced semidi- aulic pallial gonoducts. the residual glandular portion concentrated proximally (posteriorly); visceral nerve ring extremely concen- trated, connectives of the esophageal nerve ring of approximately equal length. Additional diagnostic features have been sought in the internal morphology of the penis, which have in many cases corroborated the assemblages just presented (Martins 1996a. Martins 1998). Ellobiids on Cobble Shores The ellobiids are common inhabitants of the supratidal of cer- tain rocky habitats. Cobble beaches are a common feature of the Aforean rocky shore (Morton et al. 1998) and are the habitat of five species of halophile ellobiids (Fig. 2). The distribution of these species along the cobble shores has previously been recorded (Martins 1980). Following is an elaboration on those data to present a model for their distribution. The distribution of the as- sociated malacofauna will also be presented (Table I). Five of the original 11 transects of l-m" quadrats (Martins 1980) were selected on the basis of their relatively similar length to provide a more or less homogeneous set of data (Table 2). The locations of the transects and the sampling dates are as follows: Transect A: West of Ponta do Alcaide, Silveira, Terceira, Azores. February. 1976. Transect B: Calhau da Pata, Caminho de Baixo, Sao Mateus, Ter- ceira, Azores. March, 1976. Transect C: Populo, Sao Miguel. Azores. January. 1976. Transect D: Ponta de Sao Pedio, Vila Franca do Campo, Sao Miguel, Azores. February, 1976. Transect E: Atalhada. Lugoa. Sao Miguel. Azores. March, 1976. Cobble shores are. in principle, exposed to at least moderate wave action. All but transect C have moderate to high exposure to wave action, with transect C being more protected. Kite diagrams were constructed to illustrate the distributions of the species. For ellobiids. data (as percentages) from the five transects were combined, and the percentages present at each height above or below high tide level were used to construct the kite diagrams. For other species, presence/absence was simply totaled for all five quadrats at a particular height, for a maximum of five. RESULTS The distribution of the ellobiids and associated malacofauna along the cobble shores of the Azores is summarized in Figure 3. DISCUSSION The five species of ellobiids exhibit a definite pattern of dis- tribution along the cobble shore. Myosotella myosotis lives highest up. almost assuming a terrestrial habitat, indicated by the presence of tenestrial pulmonates at these levels. It is followed by Ovatella Yulcani and Pscudiniickiiupiis cjii^uiis. which closely overlap. Ad- ditional information (Martins 1980) showed that, although over- lapping along the transect, both species are somewhat separated veilically in the cobble accumulation. The former prefers the upper levels, whereas the latter is more abundant toward the bottom, frequently found under half-buried porous rocks. Pedipes pedipes. although overlapping with the latter two species, aggregates closer to a level just above the high tide level, a distribution more nar- rowly exhibited by Aiiricidiiiellii hidentata. Here, too, there ap- pears to exist vertical zonation, with P. pedipes prefeiring the upper levels, whereas A. hidentata lives underneath the rocks (Martins 1980). The single specimen of P. pedipes recorded from D9 (quadrat 9 of transect D) could be interpreted as accidental presence because the next closest recording is another single speci- men in A6. However, transect D is in a highly exposed area, and Ellobiidae 445 If. 15 14 13 i: 11 1(1 ' 1 t—i- rri/r/'cs /'('(///'I's Emoiiiiiit^tliliJii^ Lfio^tiilti ^ii>Jiiu!ii Aunculnii'llit hniciitatti A^^tiiiiiitii fhut' Wrlllftlf'ili lltl.'/il,'./:- ^^^^^ Lillonitit ^tniiln T'-iimvi/W/iJ >i(/'i v/jMi/ri. ii riiliiJiiifilii lill.v Ali^iiua iiu'ititviili'iihi^ /.((-i7nJ tiilfin-piii Mi(-.'/;,i hulcnlalii Cifi iiiUl ailmiihiln Pillt'lltl ■;u/ij('s/l Tiling hiir ^luhulrllaidlhi: I Figure 3. Distribulion of elloblids and associated iiialucol'aiina alonj; the cobble shores in the Azores. Combined percentage (ellobiids) and cumulative presence (associated malacofauna) on five transects of different lengths. Dotted line (()( marks the high tide limit, located at the landward end of quadrat 0. Associated malacofauna not recorded in quadrat -4. (Adapted from Martins 1980.1 strong hydKidynamisni could explain this apparciillv unusual pies- ence, an interpretation supported hy the equally more landward distribution of Pseudomelampiis exiguus in transect D. The ellobiids of a boulder shore of Hong Kong, although a different species assemblage, exhibited similar distribution pat- terns (Martins & Cunha 1992). Their horizontal distribution was more clearly understood when their vertical distribution in each quadrat along the transect was examined: In general, the horizon- tally widely distributed species maintained some preference for an upper vertical level as they approached the sea. The data for the A(5-ores, presented here, lack transect profiles and vertical quanti- fication, not allowing for inferences of this kind. However, as already referred to by Martins (1980). in many instances similar behavior occurs in the A^'orean species. Worth mentioning, within the associated malacofauna, is the wide horizontal range of Assiminea eliae. restricted to Terceira Island, contrasting with the more marine preference of another assimineid. PahidineUa littohna. which is present on both islands. The two littorinids present in the studied areas exhibit their commonly observed relative distribution pattern, with Melaraphe iwritoides extending farther inland than Litloiina striata (Morton et al. 1998). ACKNOWLEDGMENTS I thank Joseph C. Britton and Robert McMahon for welcoming a study on ellobiids at a littorinid symposium. The comments of an anonymous reviewer were most appreciated. Support for the pre- sentation of this paper was provided by Funda^ao Calouste Gul- benkian and Funda^'ao para a Ciencia e a Tecnologia (Portugal). Berry. A. J. 1963. Faunal zonation in mangrove swamps. Hull. Nat. Mas. Singapore 32:90-98. Brown. D. S. 1971. Ecology of Gastropoda in a South Atrican mangrove swamp. Proc. Malac. Soc. London 39:263-279. Golley, F. B. 1960. Ecologic notes on Puerto Rican Mollusca. The Nai{tili{s 73:152-155. Grandy. J. W.. Iv, 1972. Winter distribution of Melumpu.\ hulcuiuiiis Say on a Cape Cod salt marsh. The Nautilus 85:106-109. Hausnian. S. A. 1932. A contribution to the ecology of the salt marsh snail Melampus bidentatus Say. Am. Naturalist 64:541-345. Holle. P. A. & C. F. Dineen. 1957. Life history of the salt-marsh snail Melampus bidentatus Say. The Nautilus 70:90-95. LITERATURE CITED Kobelt. W. 1897-1901. Die Familie Auriculacea. U16):l-76, pis. 10. II, 13 [18971:77-228. pis. 12. 14-.30 1 1898]; 229-268. pis. 31. 32 [1900]: 269-316. pi. 33 [1901]. In: F. H. W. Martini & J. H. Chemnitz. Sys- tematisches Conchylien-Cabinet. Nurnberg: Verlag von Bauer & Raspe (Emit Kiisterl. Kiister. C. H. 1844. Die Ohrschnecken (Auriculacea). 1( 16): 1-77, 9 pis. In: F. H. W. Martini & J. H. Chemnitz. Systemati.sches Conchylien- Cabinet. NiJrnberg: Verlag von Bauer & Raspe (Julius Merz). Lamarck. J. B. P. A. 1809. Philosophie zoologique; ou exposition des considerations relatives a I'hisloire n;iluielle des animaux. Duminil Lesueur. Paris. 2 vols. Li, F.-X. & S.-H. Gao. 1985. The ground-dvvellmg molluscan faunas of 446 Martins mangrove swamps in Hong Kong and Xiamen. In: B. Morion & D. Dudgeon, editors. The Malacofauna of Hong Kong and southern China II. Hong Kong: Hong Kong University Press, pp. 449^55. Martins, A. M. F. 19S0. Notes on the habitat of five halophile EUobiidae in the Azores. Publica(j6es do Museu "Carlos Machado." Ponta Del- gada: Centenario da Funda^ao do Museu. 1976. 32 pp. Martins. A. M. F. 1996a. Relationships within the EUobiidae. In: J. D. Taylor, editor. Origin and evolutionary radiation of (he Mollusca. Lon- don: Oxford University Press, pp. 285-294. Martins. A. M. F. 1996b. Anatomy and systematics of the western Atlantic EUobiidae (Gastropoda: Pulniona(a). Mciliicologia 37(21:163-332. Martins, A. M. F. 1998. The anatomy of Cassidula Ferussac, 1821 and a case for the revival of the Cassidulinae Odhner, 1925. In: B. Morton, editor. Proceedings of the Third International Conference on the Ma- rine Biology of the .South China Sea, Hong Kong. 1996. Hong Kong: Hong Kong University Press, pp. 25—42. Martins, A. M. F. & R. M. P. T. T. Cunha. 1992, On the ecology of Hong Kong EUobiidae (Gastropoda, Pulmonata). In: B. Morton, editor. Pro- ceedings of the Fourth International Marine Biological Workshop: The marine flora and fauna of Hong Kong and Southern China, Hong Kong, 1989. Hong Kong: Hong Kong University Press, pp. 417-^29. Morrison, J. P. E. 1958. Ellobiid and other ecology in Florida. Tin' NauUhis 71:118-124. Morrison, J. P. E. 1959. The primitive history of some salt-marsh snails. Bull. Am. Malac. Union, Report for 1958, 25:25-26. Morion, B., J. C. Britton & A. M. de Frias Martins. 1998. Coastal ecology of the Acj-ores. Ponta Delgada: Sociedade Afonso Chaves. 249 pp. Morton, J. E. 1955. The evolution of the EUobiidae with a discussion on the origin of the Pulmonata, Proc. Zool. Soc. London 125:127-168. Murty, A. S. & M, Balaparameswara Rao, 1977. Studies on the ecology of Mollusca in a South Indian mangrove swamp. J. Moll. Studies 43:223- 229. Odhner. N. H. 1925. MunniiUi juancii.sis. n. sp. nebst Bemerkungen uber der Systematik der Ellobiiden. Arkiv f. Zool. Stockholm 17A:1-15. Orton. A. K. 1976. Distribution of the salt marsh snail {Melampus biden- tiitiis Say) in relation to the vegetative cover of a Rhode Island salt marsh. Master of Science thesis. University of Rhode Island. 57 pp. Pfeiffer, L. 1856. Monographia Auriculaceoruni viventium. Kassel, Ger- many: T. Fischer. 209 pp. Pfeiffer, L. 1876. Monographia Pncumonoponiorum viventium. accedente fossilium enumeratione. Supplenientum tertium, Monographiae Au- riculaceorum parte secunda auctum. Kassel, Germany: T. Fischer, pp. 241-479. Reeve, L.A. 1841-1842. Conchologia systematica. London: Longman, Brown, Green and Longmans. 1:1-195, pp. 1-129 [1841]; 2:1-337, pis. 130-300 [18421. Reeve, L. A. 1877. Conchologia Iconica. Monograph of (he genus ,4;(- ricnlii. Vol. 20. London: J. Reeve, pp. 1-6. Russell-Hunter, W. D., M. L. Apley & R. D. Hunter. 1972. Eariy life- history of Mflamptis and the significance of semilunar synchrony. Biol Bull. Wood.s Hole 143:623-656. Sasekumar, A. 1974. Distribution of macrofauna on a nialayan mangrove shore. / Animal Ecol. 43:51-69. Zilch, A. 1959 Gastropoda. In: O. H. Schindewolf, editor. Handbuch der Paliiozoologie, 6, Berlin: Borntrager. pp. 1-200. Joiiriuil ofShcllfhh Research. Vol. 20. No. 1. 447-452. 2001. WHEN TO MOVE AND WHERE TO GO: MOVEMENT BEHAVIOR OF THE TROPICAL LITTORINID CENCHRITIS MURICATUS (LINNAEUS, 1758) DEBORAH J. GOCHFELD' - AND DVVAYNE T. MINTON" * ^Hiifstni Uuiversiiy Murine Lcihoniroiy. Prioiy, St. Ann. Jcinuiica. West Indies: 'Deparuncnt of Phannac<)i>n()sy. University of Mississippi. Mississippi 3H677; Department of Zoology. University of Hiiwaii. Honolulu. Hawaii 96822 ABSTRACT Movemenl behavior in intertidal molluscs is related to the intensity of physical and biological stress. Tropical littorinids are exposed to e.\treme environmental conditions, and movement at night or on a tidal cycle may alleviate desiccation and heat stress, as would returning to a sheltered location following foraging. We examined the movement behavior of individually marked Cenchritis muricuuis on a Jamaican shore for 30 days. On b days, selected from different parts of the lunar cycle, snail locations were monitored at 3-b intervals. Movemenl did not occur on a diurnal, tidal, or lunar cycle. Although C. muricutus preferentially rested in crevices, there was no evidence of homing behavior. Snails resting on exposed rock surfaces were four times more likely to move than snails resting in more sheltered niicrohabitats. In general, movement in C. nuiricalus appears to occur in direct response to wetting: >89'7r of movements occurred within 12 h after rainfall or heavy dew. We believe this behavior is primarily a response to desiccation stress, but it may also facilitate foraging. KEY \\ORDS: movement beha\ior, movement periodicity, microhabitat. tropical littorinid INTRODUCTION Rocky shore intertidal gastropods display a variety of move- ment behaviors, ranging from homing to a specific location on the rock to periodic migrations up and down the shoreline. Homing to a particular scar on the rock following foraging excursions is well known in limpets (Underwiiod 1979. Branch 19S1. Chelazzi et al. 1988, Delia Santini et al. 1995) and chitons (Chelazzi et al. 1988, Little 1989). Vertical migrations or periodicity in foraging activity occurring on a diurnal or tidal cycle have been demonstrated in neritids (Underwood 1979, Little 1989), limpets (Hartnoll & Wright 1977, Branch 1981, Little 1989, Gray & Naylor 1996), and littorinids (Little 1989, Ohgaki 1989, Norton et al. 1990). In ad- dition, seasonal migrations, often associated with reproductive events, have been observed in some limpets (Underwood 1979, Chelazzi et al. 1988. Delia Santini et al. 199,5) and littorinids (Chelazzi et al. 1988, Takada 1992). Most studies of movement behavior in littoral gastropods as- sume that desiccation stress is a primary influence on the evolution of the timing of movement ( Little 1 989 ). For gastropods occupying tropical rocky shores, which are exposed to direct sunlight and extremely high temperatures, individuals should move when the risk of desiccation and heat stress is lowest, such as during rain or high tides or during the night when temperatures are cooler. Tropi- cal high-shore gastropods may avoid potentially lethal daytime conditions by use of a limited cyclic activity period and by selec- tive use of microhabitat (Garrity & Levings 1984) as well as by withdrawing into their shells and cementing the shell to the rock with a mucus holdfast (McMahon & Britton 1991). Supralittoral gastropods also display a variety of morphological and physiologi- cal adaptations, which enable them to persist under extreme con- ditions of heat and desiccation (McMahon 1991, Britton 1992). The Caribbean littorinid Cenchritis muricatiis (Linnaeus, 1 758) occupies the highest position in the supralittoral fringe of rocky shorelines (Fraenkel 1968, Lang et al. 1998, Minton & Gochfeld, 2001 ). The zone occupied by C. muricatus is subjected to extreme environmental conditions of heat and desiccation and is rarely *Corresponding author. inundated by seawater. Few studies have examined the ecology of C. muricatus (Lang et al. 1998), and little is known about its behavior patterns, although Lang et al. ( 1998) noted that this spe- cies spends most of its time in repose, and Kaplan (1988) sug- gested that C. muricatus tnust migrate downshore to forage at night when it is cooler. The objectives of this study were to investigate the movement behavior of C. muricatus over a range of temporal scales. Unlike many earlier studies, the present study spans an entire lunar cycle to discriminate between movement patterns based on diunial, tidal, or lunar cues. MATERIALS AND METHODS This study was earned out on an uplifted Holocene limestone reef on the north coast of Jamaica near the Hofstra University Marine Laboratory at Priory. St. Ann. This north-facing platform rises up 10 3 m above sea level and ranges frotn 10-25 m wide in the study area. The substrate is predominantly unvegetated and heavily microkarstified. Snails were selected from ten sites within an area 100 m long. On 23 July 1998. five adult individuals of C. muricatus (length: 17.1 ± 0.6 mm; width: 13.2 ± 0.7 mm) were selected from each of ten sites, marked, and returned to their original locations. Snails were marked using colored nail polish or enamel paint, followed by an application of clear enamel to prevent the color from wear- ing off. Snails from the same site were painted different colors so that we could identify individuals. After 3 days we found as many of the marked snails as we could and placed a numbered coin on the rock beside them to denote their present location. For each snail, we measured the nearest distance and compass heading to the ocean and the micro- habitat in which it was located. Distance measurements were de- termined using a fiberglass measuring tape. Distances recorded in this study represent the shortest distance measured between the snail's previous and present locations. Because it is unlikely that snails travel in straight lines, these values represent the minimum distances traveled and are therefore underestimates of the rate of moxement. We used four microhabitat classifications: surface (on the exposed rock surface), pit (depression in the rock of sufficient size to house only a single adult individual), crevice (depression in 447 448 GOCHFELD AND MiNTON the rock of sufficient size to house more than one adult individual), and vegetation (either on or under vegetation). In addition, a dab of paint corresponding to the color of the snail was placed on the rock at the snail's initial location to determine whether the snail showed evidence of homing to this site. During subsequent observations, the following data were col- lected: distance moved from the previous location {denoted by the coin), compass direction moved, and microhabitat of the new lo- cation. The coin was then mo\ed to the snail's new location. Observations were made daily, generally between 8 am and 12 PM, from 26 July to 24 August 1998. corresponding roughly to a lunar cycle. In addition, during 6 days randomly selected from within this period (27 July, 29 July, 4 August, 8 August. 19 Au- gust, and 22 August), observations were made at approximately 3-h intervals for 24 h. To test the hypothesis that movement in C. muricatiis was restricted to nocturnal periods, a chi-square test was performed on a 2 x 2 contingency table that compared the number of nocturnal vs. diurnal observation periods during which move- ment vs. no movement was recorded. Tidal periodicity in move- ment behavior was analyzed by correlating the relative tide height, measured with an automated data logger based at the Hofstra Uni- versity Marine Laboratory, with movement patterns for the dates on which 24-h observations were conducted. The tide data for 27-28 July were omitted from this analysis due to a malfunction of the data logger. Lunar periodicity in snail movement behavior was examined by converting the lunar phase into a relative value be- tween -i-l (full) and -I (new) and coixelating this value with the percentage of snails moving on each day over the lunar cycle. For this analysis, snails that were observed to move multiple times during the 24-h observation periods were included only once. Rainfall was monitored using a plastic rain gauge. Microhabitat Availability At each site, microhabitat availability was determined by hap- hazardly placing 0.0625-m" quadrats on unvegetated substrate in the zone in which C. muricatus occurred (;i = 3-12 quadrats per site, depending on the width of this zone). The type of microhabitat (crevice, pit, or surface) present at each of 16 points within the quadrat was identified. All replicates (« = 63) were pooled to determine mean proportions of each microhabitat available to C. muricatus. The overall preference of C. muricatus for each micro- habitat was determined by comparing the number of snails occu- pying each microhabitat with the availability of that microhabitat using a chi-square test. Additionally, for each snail, its occurrence in the three microhabitats was compared with the availability of each microhabitat by a chi-square test to investigate its apparent microhabitat preference. Although the number of individual snails associated with vegetation are reported, percent cover of vegeta- tion was not quantified, and therefore the preference of snails for vegetation was not analyzed. Wetting Experiment To determine whether the quantity or salinity of wetting influ- enced the movement of C. muricatus. snails were exposed to dif- ferent quantities of fresh water or to seawater. On a hot sunny day. 16 l-m~ quadrats were haphazardly laid on a relatively flat area of shoreline in the genera! vicinity of. but not actually in, our study sites. Quadrats were randomly assigned to four treatments: no water (control). 250 ml fresh water, 500 ml fresh water, or 250 ml sea water. These volumes of water were equivalent to 0.25 mm and 0.5 mm oi rainfall or wave splash. All C. muricatus within each quadrat were counted initially. Water was applied to the treatment quadrats using a plastic bottle with a piece of plastic wrap attached to the top with a rubber band. Several small holes were punched in the plastic wrap, which allowed the water to be sprinkled evenly over the substrate. Once the quadrats were wetted, we determined the percentage of snails that had moved after 10 min. Control quadrats were not wetted but were inonitored in the same way. A one-way analysis of variance (ANOVA) was used to examine the effect of wetting on snail movement. Significant results were then analyzed using Tukey's pairwise comparisons with an overall a = 0.05. RESULTS Altogether. 64 periods were sampled, and 37—46 marked snails (42.6 ± 2.47) were located during each observation period. We made 2,726 total observations on 48 different snails (56.7 ± 13.6 observations per snail). Rates and Direction o] Movement The majority of our observations (91%) were of snails that did not move. When C. muricatus did move, they were able to move relatively rapidly. The maximum rate of movement recorded was 38 cm/h. with a mean overall rate of movement of 5.0 ± 5.5 cm/h (/I = 175). Individuals of C. nmricatus did not display any con- sistent preferred direction of movement (Bonferroni-corrected Rayleigh's test for random angular dispersion, a = 0.001; maxi- mum R = 5.9; n = 48 snails: minimum P = 0.005). When the movements of all snails were pooled and the direction of the ocean was standardized, there was no evidence of directional movement (Rayleigh's test, R = 29.5; n = 366 observations; P > 0.05) (Fig. I). We found no evidence of homing in C. muricatus. either to the initial marked "homes" or to any subsequent resting sites. On the 270 90 180 Figure L Frequency distribution of directions of all movements made by Cenchritis muricatus durlnj; all observation periods {n = 366 obser- vations). Radial lines represent Id and 20 snails, respectively. Direc- tions of movement at each site have been standardized so that the direction of the ocean is zero. MovFiMENT Behavior in a Tropical Littorinid 449 other hand, all snails remained within 1-2 m of the sites from which they were originally released. In fact, I y later (24 July 1999). 14 (28%) of our marked snails were found v\ ithui .^ m of the areas from which they were first collected. Mu yemeiil I'eriodicily lndi\ iduals of C iiniiicaliis were equally likely to move during the day as at night during our 24-h observation periods (x~ <0.(K)1 ; df = \: P = 0.99) (Fig. 2). There were some periods of high snail movement on three dates, and in each case the period of high snail movement occurred during the hours following a rainfall or. in the case of the night of 22-23 August, a period of unusually heavy dew. We observed no correlation between the percentage of snails moving and the level of the tide (Pearson correlation, /• = -0.249; t = -1.5426; n = 38: P > 0.1) (Fig. 2). There was also no correlation between the percentage of snails moving and the phase of the moon (/• = -0.229: t = -1.245: n = 30; P > 0.2) (Fig. 3). Again, there were several dates on which a high percentage of snail movement was recorded, and each of these corresponded with a preceding rainfall (Fig. 3). There was a relationship between movement and rainfall. Over- all, 89.19f of snail movements occurred within 12 h following a raintall. The percentage of snails moving during observation pe- riods within 12 h following rainfall (50.5 ± 23.0%; ii = 15 peri- ods) was significantly greater than during observation periods that were not preceded by a rainfall w ithin 1 2 h ( 1 .7 ± 4. 1 1 %; « = 45 periods: unpaired two-sample t test. / = 8.12; df = 14; P < O.OtJOl). To detennine whether snail movement decreased with increasing time since rainfall, all observations within 12 h follow- ing rainfall for the 24-h observation periods were examined. The percentage of snails moving was greatest shortly following a rain- July 27-28 1000 100 1800 2200 200 600 1000 Time of day Figure 2. Movement of Cenchritis muricatus during tlie six 24-li ob- servation periods. The percentage of snails moving (solid circles) and relative tide height (open circles) are shown. Tide data from 27-2S .luly were omitted due to equipment malfunction. Solid vertical lines indicate times of rainfall: dashed vertical line indicates an unusually heavy dew. c > o E to c c O O) a. 100. 90. 80. 70. 60. 50. 40- 30. 20 10. 0 23 o 27 July 31 >|»|«|*;«^| ■#I%»I* -Ir^ o 12 16 20 August 3 O Figure 3. Movement of Cenchrilis muricatus over the lunar cycle from 26 July to 24 August 1998. For dates on which there were multiple obser- vation periods, the percentage of snails moving during each observation period is indicated. Solid vertical lines indicate dates on which rainfall occurred; dashed vertical line indicates an unusually heavy dew. fall and showed a significant linear decrease over time (regression, y = 76.3 - 4.54x; r" = 0.35; P = 0.044). Although the percent- age of snails that was active decreased following rainfall, there was no change in the rate of movement as the time from rainfall in- creased (regression, y = 26.6- 1.18x:)~ = 0.16:P = 0.19). Snail movement appears to cease more rapidly during daytime than at night: however, our low sample sizes did not permit us to test this statistically. Effect of Microhabitat on Location and Movement When all observations from all snails were pooled, C. rmirica- lus exhibited a strong preference for crevices and pits over exposed surface (x" = 5324.21: df = 2: P < 0.001 ). Observations during which snails were found on or under vegetation (/i = 426) were excluded from this analysis. Of the remaining 2.296 observations, 54.2% were in crevices, 28.0% were in pits, and 17.8% were on the surface, in spite of the fact that only 14.9% of the area consisted of crevices, 8.5% was pitted, and 77.1% was exposed surface. When chi-square tests were performed on microhabitat preferences for each snail independently, 45 of 46 snails showed significant pref- erences for crevices and pits over exposed surface (all at P < 0.001). In 31 cases, crevices was the major contributor to these significant chi-square values, and in 14 cases, pits was the major contributor. In the absence of rainfall, snails were equally likely to move from all microhabitats (2.5. 1.1. 0.96. and 1.4% from exposed surface, pits, crevices and vegetation, respectively; x" = 2.74; df = 3; P > 0.1 ). Following a rainfall, snails were more likely to move from a resting position on the surface (24.3%) than from pits (9.2%). crevices (9.8%), or vegetation (9.9%; x' = 60.7; df = 3; P < 0.001). Wetting Experiment There was a significant effect of wetting on the percentage of snails moving (one-way ANOVA: F = 9.88: df = 3: P = 0.001 ). However, a Tukey's multiple comparisons test indicated that the 450 GOCHFELD AND MiNTON type of wetting treatment (250 ml vs. 500 ml fresh water vs. 250 ml sea water) had no effect on the percentage of snails moving. The percentage of snails moving in all treatment quadrats com- bined was 63.9 ± 23.0%, as compared with 0 in the control quad- rats. DISCUSSION This study demonstrated that individuals of C. mwicatus do not exhibit diurnal, tidal, or lunar periodicity in their movement pat- terns. Movement in C. mwicatus appears to occur in response to wetting of the substrate by rain or dew. Individuals of C. iniiricatiis occurred preferentially in pits and crevices, and this selection for sheltered microhabitats and a propensity to move only under moist conditions suggest that these behaviors may have evolved as means to reduce desiccation potential in this tropical high-shore species. Lang et al. ( 1998) observed that C imincatus spends most of its time in repose, and this study supports that observation. When C. miiricatiis did move, it did not show evidence of directional move- ments, and it did not move long distances. During our study, we were able to locate all but two of our marked snails, and all were found within 2 m of their original locations. One year later, we were able to find 14 (289r) of our marked snails, all of which were within 3 m of their initial locations. This pattern of short, nondi- rected movements — a random walk, in effect — appears to maintain the distribution of C. nniriciitus along the shore, both over the short and long term, as suggested by Chelazzi et al. ( 1988). Underwood ( 1977) also found that three species of intertidal gastropods moved in random directions, at least over a I- to 3-day period, and Pe- traitis (1982) observed that Littorina littorea moved randomly ex- cept following dislodgment, when movement was aimed at return- ing the individual to its initial zone on the shore. We did not observe any evidence of diurnal, tidal, or lunar migration in C. muricatus during this 4-wk study. Nocturnal for- aging has been implicated in littorinids (Voltolina & Sacchi 1990), and Kaplan (1988) suggested that C. nniiicatus might migrate downshore to forage on algae in the intertidal zone at night, when temperatures are cooler and the potential for desiccation is re- duced. However, this was not the case in our study. One possible explanation for the absence of nocturnal migrations in C. nnirica- tus may be that its food source, which most likely consists of endolithic and/or epilithic algae (Norton et al. 1990, McQuaid 1996). is adequate in the supralittorai fringe. The absence of tidal rhythmicity to movement in C. uiiiriccitns is not unexpected. The supralittorai fringe zone occupied by C. muricatus is highest on the shore and is never inundated by tides except during severe storms, none of which occurred during our study. Therefore, C. muricatus does not receive an exogenous cue from the tide itself, which might initiate moxement, and there does not appear to be an endogenous cue for tidal periodicity in move- ment behavior. Zann (1973) also noted the absence of tidal peri- odicity to movement in supralittorai gastropods that were not sub- jected to regular tidal action. Although clearly not ubiquitous, patterns of diunial or tidal periodicity of movement are known from littorinids. Ohgaki (1989) noted that Nodilittorina exigua in Japan exhibits upward migrations with the high tide, and foraging activity in L. littorea is also associated with the tide (Newell et al. 1971). L. scutulata exhibits diurnal periodicity in foraging activity (Voltolina & Sac- chi 1990). On the other hand. Little (1989J reviews foraging ac- tivity in a variety of littoral molluscs and lists only three species that exhibit foraging activity with no regular pattern relating to tides or time of day; all three of these species are littorinids. Unlike many earlier studies on movement in intertidal mol- luscs, our observations on C. muricatus spanned an entire lunar cycle. Nevertheless, we did not see any evidence of movement periodicity associated with the lunar cycle during our .study. In some species of littorinids, reproductive behavior has been linked to lunar cues (Hughes & Roberts 1980. Ben-y 1986). In at least one species of high-shore littorinid. there is evidence of a seasonal downshore migration associated with spawning (Ohgaki 1988). C. muricatus may display a similar seasonal reproductive migration, which would not necessarily have been observed during our study. In fact, in June-July 1998 and 1999, large aggregations of C. muricatus were observed low on the shore at the water's edge, well below this species" typical supralittorai fringe habitat (unpublished data). Upshore migrations were subsequently ob.served in mid- July. It is suspected that this migration is associated with repro- duction; however, this has not yet been confirmed. There was no evidence of homing to specific sites on the rock in C. muricatus. However, the fact that snails were found at or near their initial locations a year later, following a massive downshore migration, suggests that there may be a mechanism by which in- di\iduals are able to return to the same general location on the shore. Other littoral molluscs that migrate vertically with the tide are able to return to the same approximate locations following downshore foraging movements, and this is probably regulated by a complex of stimuli (LInderwood 1979, Chelaz/i et al. 1988). Movement in C. muricatus. as in several other species of su- pralittorai gastropods, appears to occur exclusixely when the rock is wet. The results of the wetting experiment indicate that the volume and salinity of the water was unimportant in eliciting movement behavior. Because C. muricatus occupies the highest zone on the rocky shore in Jamaica, this area is rarely reached by wave splash and is never inundated by tides. On the other hand, rainfall and dew can wet this zone on a reasonably regular, if unpredictable, basis. If rainfall occurs in midday, the rock can di'y quite rapidly, but at night the rock may remain damp for several hours. Although we did not measure relative humidity. C. muri- catus may also move when relative humidity of the air is very high (R. McMahon. pers. comm.). Individuals of C. muricatus appear to take advantage of the wet rock for their foraging excursions. Foi'aging when the rock is wet may be an adaptive mechanism to reduce desiccation potential, which is normally severe in this zone. In addition, rainfall may facilitate foraging by softening the epilithic biofilm. the probable food source of C. muricatus. Chapman (1994a) reported that Nodilittorina pyramidalis. a high-shore species from Australia, also feeds opportunistically when the rock is wet, either by rain or tide, and becomes inactive when the rock dries. The limpet Sipho- luiria pectiiiala forages when the substrate is very damp or wet. and when the relative humidity drops below 759'r during foraging, the limpet stops moving and attaches to the rock even if it has not returned to its home scar (Ocana & Emson 1999). This is presum- ably an adaptive mechanism for limiting water loss until the limpet can return to its home scar (Ocana & Emson 1999). Individuals of C. muricatus preferentially occupied pits and crevices as opposed to exposed surfaces on the rock. Gastropods occupying the supialittoral fringe zone on tropical shores are ex- posed to direct sunlight and extreme temperatures during their prolonged periods of emersion. The preferential use of pits and Movement Behavior in a Tropical Littorinid 451 crevices by C. miiricanis may be an adaptation to mediate extreme environmental conditions. Individuals of C. muricatus occupying crevices tend to have lower ambient temperatures than those on exposed rock surfaces (Lang et al. 1998). Crevices and pits may also provide some degree of shade in a habitat otherwise fully exposed to direct sunlight, and snails occurring in these micro- habitats may be able to trap more moisture. Snails found on or under vegetation may also experience reduced desiccation poten- tial as a resuh of shading and reduced contact with the hot rock, although they may also be farther from their food source. Preferential use of certain microhabitats, particularly crevices, has been reported in several other littorinids (Garrity 1984, Un- derwood & Chapman 1992). although explanations for such pref- erences vary. Our data are consistent with other studies of tropical littorinids, which found that pits and crevices provided protection from heat stress and desiccation (Garrity 1984, Peckol et al. 1989). To the contrary, studies of temperate littorinids suggest that the use of pits and crevices do not protect against heating or desiccation (Chapman 1994b I but may provide a more abundant food source (Voltolina & Sacchi 1990) or protection from wave action (Raf- faelli & Hughes 1978). Except during severe storms, C. iniiriiatiis at our study site in Jamaica is not at risk from wave action, but desiccation potential on this tropical shoreline is high and may pose a serious risk to individuals of this species (Lang et al. 1998). Littorinids occurring in crevices may also be protected from pre- dation pressure (Catesby & McKillup 1998); however, predation on C. muricatus is probably a relatively minor selection pressure at our study site. Individuals of C. muricatus on exposed rock surfaces were more likely to move when the substrate was wet, whereas indi- viduals in more sheltered microhabitats (pits, crevices, or vegeta- tion) were more likely to remain stationary. This difference in behavior as a function of microhabitat may reflect the more ex- treme desiccation potential on exposed rock surfaces, and moving from an exposed surface to a sheltered habitat may be advanta- geous. Alternatively, snails on exposed surfaces may be more en- ergetically stressed due to increased heat and desiccation potential and may be more inclined to resume foraging when the rock is wet than are snails in more sheltered microhabitats. When the rock was dry, very few snails moved, regardless of their resting microhabi- tat. Foraging activity in C. murictilus appears to be an opportunistic strategy in which snails are able to take advantage of favorable conditions during which desiccation potential is reduced. For mol- luscs found in the supralittoral fringe on tropical shorelines, des- iccation potential may be an important selective pressure, although other abiotic factors may be equally or more important because most highshore gastropods are able to tolerate greater levels of desiccation than they ever experience in their natural habitats (R. McMahon, pers. comm.). Substratum temperatures of the exposed rock surface in the supralittoral fringe at our study site in July reached a high of 45''C (Minton & Gochfeld in press), ^A'hereas the heat coma temperature of C. muricatus is 42.1°C (Britton 1992). Therefore, C. muricatus is exposed to potentially lethal tempera- tures for extended periods of time and has evolved behavioral mechanisms that minimize its potential for desiccation and heat stress, including preferential use of sheltered microhabitats and movement only when the rock surface is damp or wet. ACKNOWLEDGMENTS We thank Dr. Eugene Kaplan, Sandy Walters, Sandy Ross, and the staff of the Hofstra University Marine Laboratory/Columbus Beach Cottages for providing lodging and facilities during the course of this study. Dirk served as a dutiful field assistant during our nocturnal observation periods. This project was supported by a grant to D.M. from Concholoaisls of America. Inc. LITERATURE CITED Berry. A. J. 198fi. Semi-lunar and lunar spawning periodicity in some tropical littorinid gastropods. / Moll. Stud. 52:144-149. Branch. G. M. 1981. The biology of limpets: physical factors, energy How and ecological interactions. Ocemwgr. Mar. Bid. Ann. Rev. 19:23.S- 380. Bnlton. J. C. 1992. Evaporative water loss, behaviour during emersion, and upper thermal tolerance limits in seven species of euliltoral-fringe Lit- torinidae (Mollusca: Gastropoda) from Jamaica. In: J. Grahame. P. J. Mill & D. G. Reid, editors. Proceedmgs of the Third International Symposium on Littorinid Biology. London: The Malacological Sl((.s7. J. Eciil. 19:83-9?. Chapman, M. G. I994h, Small-scale patterns of distribution and size- structure of the intertidal littorinid Litlorinu unifasciiita (Gastropoda: Litlorinidae) in New South Wales. Aust. J. Mar. Frcshw. Res. 45:635- 652. Chelazzi. G.. S. Focardi & J.-L. Deneubourg. 1988. Analysis of movement patterns and orientation mechanisms in intertidal chitons and gastro- pods. In: G. Chelazzi & M. Vannini. editors. Behavioral adaptations to intertidal life. New York: Plenum Press, pp. 173-184. Delia Santmi, P., G. Santini & G. Chelazzi. 1995. Factors affecting van- ability of foraging excursions in a population of the limpet Patella viilgata (Mollusca, Gastropoda). Mar Biol. 122:265-270. Fraenkel. G. 1 968. The heat resistance of intertidal snails at Bimini. Ba- hamas; Ocean Springs, Mississippi; and Woods Hole. Massachusetts. Physiol. Zool. 41:1-13. Garrity, S. D. 1984. Some adaptations of gastropods to physical stress on a tropical rocky shore. Ecol. 65:559-574. Garrity. S. D. & S. C. Levings. 1984. Aggregation in a tropical neritid. Veliger21:\-t. Gray. D. R. & E. Naylor. 1996. Foraging and homing behaviour of the limpet. Patella viilgata: a geographical comparison. J. Moll. Stud. 62: 121-124, Hartnoll. R. G. & J. R. Wright. 1977. Foragmg movements and hommg m the limpet Patella vulgata L. Anim. Behav. 25:806-810. Hughes. R. N. & E. J. Roberts. 1980. Growth and reproduction rates of Littorina neritoides (L.) in North Wales. / Mar. Biol. Ass. U.K. 60: 591-599. Kaplan. E. H. 1988. A tleld guide to Southeastern and Caribbean sea- shores: Cape Hatteras to the Gulf Coast. Flonda and the Caribbean. Boston: Houghton Mifflin. 425 pp. Lang. R. C. J. C. Britton & T. Metz. 1998. What to do when there is nothing to do: the ecology of Jamaican intertidal Littorinidae (Gas- tropoda: Prosobranchia) in repose. Hydrobiol. 378:161-185. Little. C. 1989. Factors governing patterns of foraging activity in littoral marine herbivorous molluscs. ./. Moll. Stud. 55:273-284. McMahon. R. F. 1991. Upper critical temperature limits, tissue tempera- tures dunng emersion and evaporative cooling at elevated temperatures 452 GOCHFELD AND MiNTON in Western Australian intenidal gastropods: evidence for a physiologi- cal barrier at tlie high tide mark. In: F, E. Wells, D. I. Walker, A. H. Kirknian & R. Lethbridge. editors. Proceedings of the Third Interna- tional Marine Biological Workshop: The Marine Flora and Fauna of Albany, Western Australia. Volume II. Perth: Western Australian Mu- seum, pp. 661-673. McMahon, R. F. & J. C. Britton. 1991. The relationship between vertical distribution, rate of evaporative water loss, behaviour during emer- gence and morphometries in six species of rocky shore gastropods from Princess Royal Harbour, Western Australia. In: F. E. Wells, D. I. Walker, A. H. Kirkman & R. Lethbridge, editors. Proceedings of the Third International Marine Biological Workshop: The Marine Flora and Fauna of Albany, Western Australia, Volume II. Perth: Western Australian Museum, pp. 675-692. McQuaid, C. D. 1996. Biology of the gastropod family Littorinidae. II. Role in the ecology of intertidal and shallow marine ecosystems. Ocewiogr. Mar. Biol. Ann. Rev. 34:263-302. Minton, D. T. & D. J. Gochfeld. 2001. Is life on the tropical shore really so hard?: the role of abiotic factors in structuring a supralittoral niol- luscan assemblage. / Slielljl-.h /?e.v. 20:477-183. Newell, R. C. V. I. Pye & M. Ahnsanullah. 1971. Factors affecting the feeding rate of the winkle Liriorina liitorea. Mar. Biol. 9:138-144. Norton. T. A., S. J. Hawkins, N. L. Manley, G. A. Williams & D, C. Watson. 1990. Scraping a living: a review of littorinid grazing. Hydro- bioi 193:117-138. Ocana, T. M. J. & R. H. Emson. 1999. Preliminary observations on the timing and geometry of foraging activity in the intertidal puhnonate limpet Si/ilionuria pectinahi. J. Mar. Biol. Ass. U.K. 79:459—165. Ohgaki, S. 1988. Vertical migration and spawning in Nodiliitorina e.xigua (Gastropoda: Littorinidae). J. Elhol. 6:33-38. Ohgaki, S. 1989. Vertical movement of the littoral fringe periwinkle NoJi- liriorina e.iii;iia in relation to wave height. Mar. Biol. 100:443-448. Peckol. P., S. Guarnagia & M. Fisher. 1989. Zonation and behavioral patterns of the intertidal gastropods Nodiliitorina (Tectininnsl antoni (Philippi. 1846) and Nerila versicolor Gmelin. 1791, in the Bahamas. Veliger },l:)i-\5. Petraitis, P. S. 1982. Occurrence of random and directional movements in the periwinkle, Liilorina liitorea (L.). J. E.xp. Mar. Biol. Ecol. 59:207- 217. Raffaelli. D. G. & R. N. Hughes. 1978. The effects of crevice size and availability on populations of Liilorina rndis and Liilorina neriloides. J. Annn. Ecol. 47:71-83. Takada, Y. 1992. The migration and growth of Liilorina hreviciila on a boulder shore in Amakusa, Japan. In: J. Grahame, P. J. Mill & D. G. Reid. editors. Proceedings of the Third International Symposium on Littorinid Biology. London: The Malacological Society of London, pp. 277-279. Underwood, A. J. 1977. Movements in intertidal gastropods. / E.xp. Mar. Biol. Ecol. 26:191-201. Underwood. A. J. 1979. The ecology of intertidal gastropods. Adv. Mar. Biol. 16:111-210. Underwood, A. J. & M. G. Chapman. 1992. Experiments on topographic influences on density and dispersion of Liilorina unifasciala in New South Wales. In: J. Grahame, P. J. Mill & D. G. Reid, editors. Pro- ceedings of the Third International Symposium on Littorinid Biology. London: The Malacological Society of London, pp. 181-195. Voltolina, D. & C. F. Sacchi, 1990. Field observations on the feeding habits of Liilorina scululala Gould and L. silkana Philippi (Gastropoda. Prosobranchia) of southern Vancouver Island (British Columbia, Canada). Hxdrohiol. 193:147-154. Zann, L. P. 1973. Relationships between intertidal zonation and circatidal rhythmicity in littoral gastropods. Mar. Biol. 18:243-250. .lounuil ,:fShcllthh Research. Vol. 20. No. 1. 453-157. 2001. A MOLECULAR ASSAY IDENTIFIES MORPHOLOGICAL CHARACTERS USEFUL FOR DISTINGUISHING THE SIBLING SPECIES LITTORINA SCUTULATA AND L. PLENA PAUL A. HOHENLOHE' * AND ELIZABETH G. BOULDING" ^ Friday Harbor Laboratories. University oj Wasiiini>ton. 620 University Rd. Friday Harbor. Washington 98250; 'Department of Zoology. University of Giielph, Gnelpli, Ontario NIG 2W I. Canada ABSTRACT Sibling specie.s Litwrina sculidciui and L. plena are difficult to distinguish in the field. Here we present a new molecular liiul and use it lo evaluate the discrete and quantitative morphological characters that have been proposed as diagnostic. We collected 38.5 snails of both species from 1 1 sites in Washington state and used restriction enzyme digestion of a PCR-amplified, 480 bp fragment of the mitochondrial cytochrome b gene to distinguish the species. This new molecular assay produces species-specific restriction fragment patterns that correspond with identification of males by penis morphology. To evaluate the usefulness of morphological characters, we scored three discrete shell characters (presence of basal band, presence of basal ridge, and size of checker pattern) as well as tentacle coloration. The four discrete characters differ significantly between the two species, though none is completely diagnostic. Tentacle coloration is the most reliable character and may be combined with the shell characters for successful identifi- cation. The two species also differ significantly in overall size and in three out of five size-independent shell shape measurements, with L scunilata having larger, taller-spired shells with narrower apertures. However, shell shape does not separate the species well because of intraspecific variation, and it is unlikely to be useful for species identification. Further analysis suggests that at least some of this intraspecific variation is genetic rather than environmental. The distributions of the two species overlap broadly in Washington, though only L plena was found in exposed outer coast habilals, contrary to previous work. KEY WORDS: LiUonna scuhilahi. L. plena, species identification, sibling species, molecular systematics INTRODUCTION The taxoiiomic history of the Littorina scundata species com- plex, a group of sympatric intertidal prosobranch gastropods in the Northeastern Pacific, has been complicated by morphological siniilarity across species and variation within species. Currently two sibling species are recognized, L. scitntUita [.sen.sii siricta) Gould 1849 and L plena Gould 1849, which are distinguished on the basis of reproductive characters, including penis, pallial ovi- duct, and egg capsule morphology (Murray 1979, Mastro et al. 1982, Reid 19961. These characters, however, are difficult to use for the non-destructive field identification that is necessary for many ecological studies. Reproductive characters cannot be used for juveniles, and we have found the dissection necessary for ex- amining pallial oviduct morphology to be difficult, especially in small specimens. Other diagnostic morphological characters have been proposed. Three discrete shell characters have been described: a pale basal band (Murray 1982, Reid 1996, Chow 1987, Rugh 1997) and a narrow basal ridge (Rugh 1997), both found more often in L plena, and the pattern of checkers on many shells, which tend to be smaller in L plena than in L. scutulata (Reid 1996, Rugh 1997). Rugh (1997) was able to use these three shell characters alone to correctly identify 17 male specimens of both species from southern California. Reid (1996) described differences in tentacle color- ation: L scunilata individuals tend to have "transverse black bands and flecks," while L. plena tend to have a "broad, unbroken black stripe with transverse flecks, or all black." Murray (1982) de- scribed a set of discriminant functions of four quantitative shell measurements that correctly classified 961 of specimens. Further principal component analysis by Murray (1982) showed L sciitii- *Corresponding author. E-mail: hohenlo(a'u. washington.edu lata shells to be generally taller with narrower spire angles and shorter aperture openings relative to shell height. Chow (1987) combined three quantitative shell measurements with number of whorls, presence of a basal band, and presence of tesselation in another discriminant function analysis. This analysis correctly classified 92% of specimens, but only when using snails from one habitat; combining specimens from different habitats introduced too much intraspecific variation to allow correct classification. Chow (1987) also found L. scutulata shells to be larger, narrower, and less likely to have a basal band, which agreed with past work. Other characters used with varying success to distinguish these species include spiral sculpture on the shell and radular characters (Reid 1996, Mastro et al. 1982). Mastro et al. (1982) found eight polymorphic allozyme loci at which the two species differ in their allele frequencies. However, none of these loci was diagnostic. No other molecular studies to date have identified a reliable molecular character at a polymor- phic locus that distinguishes the two species. The previous studies of moiphological differences used speci- mens that were positively identified using reproductive characters, thus excluding pre-reproductive animals. Here we present a mo- lecular technique for identifying individuals of all ages using mi- tochondrial DNA. We use this tool to evaluate the reliability of characters that can be observed on intact animals: the three discrete shell characters described above, tentacle coloration, and quanti- tative shell shape differences. MATERIALS AND METHODS We collected 385 snails of both species from 1 1 areas around Puget Sound and the outer coast of Washington state in January 1998 (see Fig. 1) and kept them alive until DNA extraction. Ani- mals of all sizes, including juveniles, from within a randomly chosen, small (-1 m"^) area of rocky shore habitat were collected. Individuals anesthetized in 1% MsCU seawater solution were 453 454 HOHENLOHE AND BOULDING British Columbia 151 Figure 1. Map of western Washington state showing collection sites in Puget Sound and on the outer coast. Pie diagrams show relative abun- dance of Z.. plena (dark! and /.. scutulata (light) with the total sample size for each site. scored for three discrete shell characters (Fig. 2a-d): basal band present or absent, basal ridge present or absent, and checkers large, small, or absent. Large checkers lie in five to 12 spiral rows, depending on shell size, while small checkers number more than 10 spiral rows. Tentacle coloration was scored in one of seven categories: dark transverse bands, spots and bands, spots, dark central stripe, central stripe with bands, no color, or all dark (Fig. 2e-f: see also Reid 1996). In all further analyses, the first three categories were grouped as "transverse bands." and the last four were grouped as "central stripe." Individuals' sex was also re- corded and males scored as L. ple)ia or L. scutulata penis type. L. scutulata penes are relatively short with a terminal bifurcation, while L. plena penes are longer, often coiled, with a bifurcation near the base (Reid 1996). Animals were then sacrificed for molecular analysis. Using the extraction protocol and PCR primers described in Kyle and Boul- ding ( 1998), a 480 bp fragment of the mitochondrial cytochrome-b gene was amplified for each individual. These were then digested for two hours using the restriction enzyme Alu 1 and the digests run on a 2% agarose gel. To predict restriction sites for the two spe- cies, we examined .^6 L. plena and 18 L. scutulata haplotype sequences from Kyle and Boulding (2000) (Genbank accession nos. AF077238-AF077291 ). Identification of these sequences by Kyle and Boulding (2000), however, depended on a single L. plena sequence from Reid et al. (1996) (Genbank accession no. U46815), so our analysis also functioned to confirm the identification of those sequences. We expected fragments of 161, 233. and 86 bp for L. scutulata and 109. \5. and 356 bp for L. plena. Eight quantitative shell measurements were taken from each ^ 1 (e) (f) Figure 2. Discrete characters, (a) A pale basal band and (b) basal ridge are found more often in L. plena, (c) A pattern of large checkers typifies L. scululala. while (dl small checkers are found more often in L. plena, (e) Two tentacle coloration patterns found in L. scutulata: transverse bands (left) and bands and spots (right), (f) Two tentacle patterns found in L. plena: a broad central stripe with bands (left) and without bands (right). individual using a dissecting microscope connected to computer imaging softwaie (Fig. 3). Since six of these measurements are lengths, they were combined into three non-dimensional ratios to remove the effect of overall shell size as follows: Relative aperture height = Apeilure height / Shell height; Whorl ratio = Whorl n-2 / Whorl n-1; and Aperture shape = Short axis / Long axis. Sta- tistical tests and discriminant function analysis used these three ratios as well as spire angle and aperture angle (Fig. 3). In an attempt to improve differentiation of the two species based on shell morphology, further discriminant function analyses were done using different sets of variables. In the first, size was included specifically by adding shell height to the three ratios and two angles, providing six variables for discriminant analysis. In the second revision, the original six linear measurements were not combined into ratios as above, but rather normalized by the Distinguishing Sibling Species 455 Spire Shell height Aperture height Aperture: Long axis Short axis Aperture angle Figure 3. Quantitative shell measurements. Shells were viewed through a dissecting scope and video camera connected to a computer and measurements were taken using imaging software. The two angles were measured in degrees and the six lengths in mm. method of Clarke et al. ( 1999). The effects of size were removed from each linear measurement by dividing the geometric mean of each specimen's measurements: geometric mean = 10^[(log,||(X| ) + logmtx,) + . . . + log|,|(X(,))/6] where x, through x,, are the original linear measurements. This provided two angles and six normalized linear measurements for discriminant analysis. The third attempt to improve the discrimination used only the specimens from six sites at which both species were found (see Fig. 1 ) and analyzed the three ratios and two angles as above. Finally, another analysis was performed on data limited to three sites of similar habitat following Chow (1987). These three sites are the three southern Puget Sound sites shown in Figure I . All are protected shores at which both species were found. RESULTS Restriction enzyme digestion with Alii I produced two discrete fragment length patterns as expected: one pattern with three closely spaced bands (L. scutulata) and another with two widely spaced bands (L. plena; the 15 bp fragments typically migrated off the gel ). These corresponded precisely with identification of males bv penis morphology (n = 48 Z,. scutulata and 71 L. plena). This supports the Identification of sequences in Kyle and Boulding (2000) and suggests that these restriction sites are consistent across haplotypes within each species. The following analyses are based on these 119 males as well as females identified using only re- striction digest and males identified by penis morphology alone. Sample sizes vary because of damage to some shells during DNA extraction. The frequency of each of the four discrete characters differs significantly between the species (Table I ) consistent with Rugh (1997) and Reid (1996). The differences remain significant fol- lowing Bonferroni correction of the original p-values from four separate contingency table analyses (Rice 1989). However, no single character completely separates the two species. The shell characters were often not visible because of shell erosion from wave action or fungal or other epiphytic growth, creating a poten- tial bias toward identification as L. scutulata from the first two characters. Many undamaged shells also lack any checker pattern. Tentacle coloration, because it was always scorable in live ani- mals, was the most reliable discrete character. Following Bonferroni correction of the original p-values from six separate two-tailed /-tests (Rice 1989). the species differ sig- nificantly in four shell measures (Table 2). L. scutulata shells are significantly larger than L. plena shells. Of the five size- independent shell measures, the two species differed significantly in spire angle, whorl ratio, and aperture shape. These results are consistent with those of Murray (1982) and Chow (1987), con- firming that L scutulata shells are larger, narrower, and taller- spired with narrower apertures. Relative aperture height, though not significantly different, also follows the trend found by Murray (1982). Three of these shell measures, shell height, whorl ratio, and aperture shape, remain significant when considering only the six sites at which both species were found (Table 2). The differ- ence is lost for spire angle. However, the difference in relative aperture height reverses and becomes statistically significant when only these sites are considered. Combining specimens from all the collection sites, these five TABLE I. Species differences for four discrete characters. Species were identified by restriction enzyme digest and penis morphology. Data given are number of specimens (percentage) in each category. The p-values are Bonferroni corrections (Rice 1989) of separate x'-tests. Character State L. scutulata /,. plena p-value Basal band present 18(12.4) 124(51.7) <0.00I absent 127(87.6) 116(48.3) Basal ridge present 23(15.9) 149(62.1) <0.00I absent 122(84.1) 91 (37.9) Checker pattern large 122(84.1) 25(10.4) <0.001 small 5(124) 93 (38.8) absent 18(12.4) 122(50.8) Tentacle color transverse bands 130(89.7) 10(4.2) <0.001 central stripe 15(10.3) 230(95.8) 456 HOHENLOHE AND BOULDING TABLE 2. Quantitative shell measurements. Numbers given are mean (standard deviation), and the p-values are Bonferroni corrections (Rice 1989) of separate two-tailed f-tests. For each measure, the first row includes specimens from all sites, and the second row includes only specimens from the 6 sites at which both species were found. Measurement L. scutulata L. plena all sites: n= 142 n = 210 6 sites: n = 104 n = 146 p-value Shell height (mm) 8.1 (1.6) 6.3(1.3) <0.001 6 sites 7.4(1.2) 6.3(1.2) <0.001 Spire angle (deg.) 54.2 (4.0) 55.5(5.8) 0.045 6 sites 54.4(4.1) 54.5(5.5) >0.5 Aperture angle (deg.) 24.0(2.4) 23.8(2.1) >0.5 6 sites 23.9(2.4) 23.9 (2.2) >0.5 Relative ap. Height 0.516(0.038) 0.519(0.039) >0.5 6 sites 0.528(0.034) 0.513(0.039) 0.01 Whorl ratio 0.595(0.163) 0.555(0.045) <0.001 6 sites 0.607(0.188) 0.568 (0.002) 0.045 Aperture shape 0.703 (0.038) 0.726(0.037) <0.001 6 sites 0.697 (0.040) 0.721 (0.037) <0.001 she!) measures were used in a discrimi)iant t'unetio)! analysis with- out much success: the function correctly classified only 69'7r ot L sciitukitii and 66% of L. plena specimens. This seems to be the result of overlapping intraspecific variation for all of the charac- ters. Attempts to impixne the discriminant function analysis were marginally successful. Because L. scutukita shells were signifi- cantly larger, including shell height improved posterior classifica- tion to 169c for L. scutulata and 8 1 '7c for L. plena. Normalizing the linear measuiements by the geometric mean (Clarke et al. 1999) inipioved classification of L. scutulata to 827r. but reduced suc- cessful classification of i.. plena to 62%. Limiting the analysis to the six sites where both species were found made no improvement over the original dataset: 69% for L. scutulata and 68% for L plena. Finally, restricting the analysis to the three protected Puget Sound sites only slightly improved classification; 66% for L. scutulata and 74% for L plena. These results are summarized in Table 3. The geographic ranges of these two species overlap signifi- cantly in Washington, although L scutulata is found only in mod- erately exposed to sheltered areas of Puget Sound while L. plena is found from sheltered sites in Puget Sound to the exposed head- TABLE 3. Varying success of discriminant function analyses of quantitative shell measurements. Correct posterior classification percentages are given along with total sample size of specimens for each species. Variable combination L. scutulata L. plena 2 angles. 3 ratios 69 (n = 142) 66 (n = 210) 2 angles, 3 ratios, shell height 76 (n = 142) 81 (n = 210) 2 angles, 6 linear measurements normalized by geometric mean 82 (n = 142) 62 (n = 210) 2 angles, 3 ratios from sites with both species 69 (n = 104) 68 (n = 146) 2 angles, 3 ratios from protected habitats only 66 (n = 38) 74 (n = 99) lands of the outer coast (Fig. 1 ). This conflicts with Reid ( 1996), who found only L. scutulata in exposed habitats. DISCUSSION L. plena individuals tend to have a pale basal band, basal ridge, and small checker pattern on their shells and a broad central stripe on their tentacles, while L. scutulata individuals lack the basal band and ridge, have a larger checker pattern, and have transverse bands and spots on their tentacles (see Fig. 2). In addition, L. scutulata shells are larger and taller-spired with narrower aper- tures. However, we do not believe the discrete shell characters alone to be sufficient, as did Rugh (1997). nor do we believe shell shape differences to be diagnostic, as did Murray (1982). Though the species differ significantly in these characters, intraspecific variation and shell damage may confound identification. No combination of quantitative shell measurements in a dis- criminant function analysis piovided reliable identification. The most successful analysis used shell size explicitly, which could potentially bias the identification of different ages of snails, so its utility in ecological studies would be limited. Neither method of removing the effects of size matched the discrimination ability of Mun-ay (1982) or Chow (1987). suggesting that interspecific differences are truly confounded by intraspecific variation. One attempt to eliminate some intraspecific variation by examining only sites at which both species were found did not improve the results, so this dataset does not show any evidence for character displacement in shell shape. Limiting the analysis to a single habi- tat type was not successful either. Since this analysis attempted to reduce environmental phenotypic variation, the result suggests that at least some of the intraspecific variation observed is geneti- cally based. One likely explanation for our different results is that these previous studies used only reproductive, hence larger, animals. Species differences may become more apparent as the snails grow (Reid 1996; Hohenlohe pers. obs.), and ecological applications typically require identification of animals of all ages. Murray (1982), Chow ( 1987), and Rugh (1997) primarily used specimens from California, and geographic differences may also play a role. Chow ( 1987) also combined discrete and quantitative characters in a single analysis, which was not done here. We found L plena occupying a wider range of habitats in Washington, from sheltered Puget Sound sites to the exposed outer coast. In contrast. L. scutulata was found only on sheltered to moderately exposed shores. This result conflicts with some previ- ous work (Reid 1996) but is consistent with other data on species distributions (Hohenlohe 2000). This discrepancy is investigated further in Hohenlohe (2000). These species can be distinguished non-destructively by com- bining the characters discussed here. Male penis morphology can be easily examined by holding the snail upside down, underwater, under a dissecting microscope. For females and non-reproductive males, tentacle coloration is the most reliable character and can be combined with shell characters on undamaged specimens. For posi- tive identification of all ages and both sexes, restriction enzyme di- gestion of cytochrome b with Alu I is straightforward and reliable and provides a diagnostic character independent of moiphology. ACKNOWLEDGMENTS This work was supported by a NSF Graduate Fellowship to P.A.H. and an NSERC Research Grant to E.G.B. Thanks to A. Tie for assistance with molecular work and to D.G. Reid and A.J. Kohn for helpful comments. Distinguishing Sibling Species 457 LITERATURE CITED Chow. V. 1987. Morphological classification of sibling species of Liltorinii (Gastropods; Prosohranchia); discretionary use of discriminant analy- sis. Veliger 29:359-366. Clarke, R. K., J. Grahame & P. J. Mill. 1999. Variation and construct in the shells of two sibling species of intertidal rough periwinkles (Gas- tropoda: Linorma spp.). Land. J. Zool. 247:145-154. Hohenlohe. P. A. 2000. Larval dispersal, gene flow, and speciation m the marine gastropod genus Littorinu. Ph.D. dissertation. University of Washington. Kyle, C. J. & E. G. Boulding. 1998. Molecular genetic evidence for parallel evolution in a marine gastropod. Litloriiiii siibroumdata. Proc. R. Soc. Loiiil^ H 265:1-6. Kyle, C. J. & E. G. Boulding. 2(J00. Comparative population genetic struc- ture of marine gastropods (Litmniui spp.) with and without pelagic larval dispersal. Marine Biol. 137:835-845. Mastro. E., V. Chow & D. Hedgecock. 1982. Litionmi sLiilulata and L plena: sibling species status of two prosobranch gastropod species con- firmed by electrophoresis. Velii>er 24:239-246. Murray. T. 1979. Evidence for an additional Liiwiina species and a sum- mary of the reproductive biology of Liriinina from California. Veliger 21:469-474. Murray. T. 1982. Morphological characterization of the Lituiiina scutulata species complex. Veliger 24:233-238. Reid, D. G. 1996. Systematics and evolution of Littmiiia. London: The Ray Society. Reid, D. G., E. Rumbak & R. H. Thomas. 1996. DNA, morphology and fossils: phylogeny and evolutionary rates of the gastropod genus Lit- torina. Phil. Tran. R. Soc. Loml. B 351:877-895. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223- Rugh, N. S. 1997. Differences in shell morphology between the sibling species Littorina .scutulata and L plena (Gastropoda: Prosohranchia). Veliger 40:350-357. Jimniul of Shfllfish Research. Vol. 20. Nu. I. 4.^4-467. 2001. ACUTE THERMAL TOLERANCE IN INTERTIDAL GASTROPODS RELATIVE TO LATITUDE, SUPERFAMILY, ZONATION AND HABITAT WITH SPECIAL EMPHASIS ON THE LITTORINOIDEA robp:rt f. McMahon Deparuucnt i>f Biology. The University af Texas al Arliiiiiton. Box 1949H. Arlint^ron. Texas 76019 ABSTRACT Acute thermal tolerance wa.s determined as mean heat coma temperature (HCT) tor samples ot 30 to 36 individuals each (range = 17-71) of 60 .species of intertidal snails from rocky shores (40 species), mangroves (17 species) and .salt marshes (3 species) at 10 collecting sites across a latitudinal range of 12.,'!°- 54.6° (54.61°N - 35.00°S). HCT was quadratically correlated to latitude, inerea.sing with decreasing latitude above 25°, but remainmg constant in tropical latitudes below 25°, indicating that thermal tolerance of tropical species is not as latitudinally influenced as m subtropical and temperate species, or, alternately, that tropical species have attained physiologically near ma,\imal levels of thermal tolerance. Mean HCT was greater in the Littorinoidea than in the other six superfamilies tested, perhaps accounting its dominance of high-shore habitats where increased thermal tolerance reduces water loss due to evaporative cooling. Similarly, the high-shore superfamilies. Cerithoidea and Neritoidea. also had high mean HCT values. Increasing thermal selection with shore height was reflected in increasing HCT among low-, mid- and high-shore species. HCT did not differ among littorinoidean and nonlittonnoidean snails on low- or high-shores, but was elevated in mid-shore littonnoideans, suggesting retention of the elevated thermal tolerance of this latter group's high-shore ancestors. In contrast, progressive HCT increase with shore height among mid- and high-shore nonlittorinoideans suggests that they evolved from low-shore ancestors. Among all taxa, thermal tolerance increased progressively across rocky shores, mangroves and salt marshes. Littorinoidean snails had greater HCT than nonlittorinoidean snails among all three habitats, suggesting that mangrove and salt-marsh littorinoideans retain the elevated thermal tolerance of their high rocky shore ancestors. In contrast, HCT of nonlittorinoidean snails increased across rocky shore, mangrove and salt marsh habitats suggesting that they pose ditlerent thermal selection pressures on species evolving into them from the low-shore. KEY WORDS: Gastropoda, heat coma temperature, intertidal gastropods, Littorinoidea, thermal adaptation, thermal tolerance, /onation INTRODUCTION A plethora of studies describe the thermal tolerance of intertidal invertebrates relative to vertical zonation. Sessile species, such as barnacles, present a fairly clear pattern of increased thermal tol- erance with increasing zonation level, elevated thermal tolerance of higher zoned species considered an adaptation to the elevated ambient temperatures associated with insolation during prolonged tidal emersion (reviewed by Newell 1979). The situation for mo- bile intertidal gastropods is more problematic because they seek temperature microrefugia (McMahon & Yipp 1992. Chapman & Underwood 1996) and are active only during periods of reduced themial stress (Ruwa & Jaccarini 1988). Thus, their acute upper thermal tolerance limits may not consistently correlate with their vertical distributions (for a review see McMahon 1990). Further, intertidal gastropod mobility results in species having broad, sym- patric vertical distributions with extensive species overlap (Broekhuysen 1940, McMahon 1990, Lang et al. 1998) and species vertical distributions changing with season (Vaughn & Fisher 1992). Thus, assignment of gastropods to narrowly defined vertical zones is difficult at best and often impossible. In a review of their resistance and capacity adaptations. Mc- Mahon (1990) hypothesized that intertidal gastropods can be di- vided into three major /onational groups characterized by unique suites of physiological adaptions. Sublittoral to lower eulittoral species (referred to hereafter as low-shore species) are primarily aquatic, generally not ranging above the mean low tide mark. E-mail: r.mcmahonCS' ula.edu where they are submerged most of the time (>86'7(). Thus, low- shore species appear to be adapted to an aquatic existence and usually unable to sustain aerial gas exchange when emerged. Their low shore habitat is more temperature-stable than periodically em- ersed eulittoral and eulittoral fringe-maritime zone habitats (Mc- Mahon 1988. 1990). Eulittoral gastropod species (referred to hereafter as mid-shore species) are restricted to areas between high and low tide marks where they experience only short-term tidal emersion (<12 h), spending 14-86% of the time in air depending on shore height. Thus, mid-shore species lead an amphibious existence, their ad- aptations hypothesized to center on remaining active during emersion and immersion. Thus, their aerial and aquatic O^ uptake rates are generally equivalent even though aerial gas exchange results in extensive evaporative water loss (McMahon. 1990). When emersed in direct sunlight, mid-shore gastropods may evaporatively cool to prevent lethal tissue warming, rapidly rehy- dratitig during subsequent tidal inundation (McMahon 1990, Mc- Mahon & Yipp 1992. Lang et al. 1998). Upper eulittoral fringe-maritime zone species (referred to here- after as high-shore species) are restricted to regions above the high tide mark where they experience prolonged emersion (8-10 days) between successive spring tides (emersed for 86-98% of the time). Adaptations in this group are hypothesized to center on tolerance of prolonged emersion including elevated temperature tolerance and reduction in O, uptake rates during emersion which conserve organic energy stores and minimize evaporative water loss. Thus, some studies indicate that high-shore gastropods have higher acute upper thermal limits than mid-shore species (Broekhuysen 1940, Cleland & McMahon 1990. McMahon 1990. Britton 1992). hy- 459 460 McMahon pothesized to allow avoidance of water loss due to evaporative cooling during prolonged emergence (McMahon 1990). There is little synipatry among gastropod species occupying low-, mid- and high-shore habitats, but much sympatry within them (McMahon 1990). suggesting that adaptation to the physical conditions of any one zone restricts species from occupying others. Thus, it is hypothesized that low temperature tolerance in low- shore species prevents them from surviving the elevated tempera- tures experienced on mid- or high shores during isolation. Con- versely, modification of the respiratory system for aerial as well as aquatic gas exchange is hypothesized to be a liability to mid- and high-shore species in the almost purely aquatic low-shore habitat. Mid-shore snails greatly increase foraging time by remaining ac- tive when tidally emersed and appear to have evolved mechanisms, including evaporative cooling and large body size, that allow regu- lation of body temperature at metabolically efficient levels during insolation. While their acute upper thermal limits are generally greater than those of low-shore species, mid-shore gastropods' capacity for evaporative cooling is hypothesized to have prevented their evolution of the extremely elevated temperature tolerances characteristic of high-shore species (McMahon 1990). All of the above hypotheses regarding physiological correlates with species zonation patterns hinge on the assumption that ther- mal tolerance increases with height of occupation on the shore. This assumption is based on a series of studies on the relationship between vertical zonation and thermal tolerance in intertidal gas- tropods limited to a single shore. On a single shore, the number of species available for comparison is limited, being maximally 6 to 14 species (Broekhuysen 1940. Evans 1948. Fraenkel 1966. 1968. McMahon & Britton 1985. Cleland & McMahon 1990. McMahon 1990. Britton. 1992). Low species numbers on any one shore leads to difficulties in data interpretation as anomalously high or low thermal tolerance values for one or two species make zonational thermal tolerance patterns difficult to discern (for examples see Broekhuysen. 1940. Evans 1948, Fraenkel 1966. 1968. McMahon & Britton 1985. Britton 1992, reviewed by McMahon 1990). Analysis of data for a restricted number of species from a single or just a few shores also prevents elucidation of latitudinal influ- ences on thermal tolerance and empirical comparisons among taxonomic groups. Only two investigations of intertidal gastropod thermal tolerance have involved two or more geographically sepa- rated shores. Fraenkel ( 1 968) studied gastropod thermal tolerance on shores in Bimini. Bahamas. Ocean .Springs. Mississippi, and Woods Hole, Massachusetts and Stirling (1982). on shores in Hong Kong and Dar es Salaam. Tanzania. In both studies, differ- ences in the thermal tolerance of the gastropod fauna were re- corded between shores, but too few shores were investigated to reveal latitudinal trends. Further, cross-comparison of gastropod thermal tolerance data from studies by different authors on geo- graphicallv separate shores is also problematic because of incom- patible methodologies (Fraenkel 1968. Stirling 1982). In order to empirically assess the effects of zonation level, latitude, habitat, and taxonomic position on the thermal tolerance of intertidal snails. I investigated acute upper thermal limits in 60 gastropod species from ten geographically separate shores over a latitude range of 12.5 - 54.61". encompassing three different in- tertidal habitats (i.e., rocky shore, mangrove and salt marsh) and seven different superfamilies. These data were statistically evalu- ated for differences in upper thermal limit among tested species relative to latitude, and. with latitude as a covariant. among zona- tion levels, habitat types and taxa on the supertamilial level. MATERIAL AND METHODS Species Collection Thermal tolerance determinations were carried out on samples of 60 intertidal gastropod species (Table 1) collected at 10 geo- graphically separate sites ranging from 54.61" N to 35.00" S lati- tude including; Robin Hoods Bay, England (54.6 1°N. number of species in) = 6); Filey Head. England (54.21°N. n = 2): Forth Towyn. Wales (53.04°C. ;; = 1): Woods Hole. Massachusetts (41.54°N, 11 = I): Gulf Shores. Alabama (30.35°N. n = 1); Port Aransas. Texas (28. 84'=N. /; = 2); Hong Kong (22.37°N. « = 16): Jamaica (18.46''N.;i = 13); Darwin. Australia (12.50°S. ;; = 12); and Albany. Australia (35.00°S. /; = 6). Forty species were from rocky shores. 17 from mangroves and three from salt marshes. Of rocky shore species. 18 occurted on the high shore. 13 on the mid-shore and nine on the low shore. Of 17 tested mangrove species, six were high-shore, nine were mid-shore and two were low-shore species. Of three tested salt marsh species, one was a high-shore and two were low-shore species (Table 1 ). The tested species included seven superfamilies: 25 species of Littorinoidea, 13 species of Cerithoidea, nine species of Neritoidea; four species of Trochacea; four species of Muricoidea; four species of Ellobio- idea; and one species of Onchidoidea (Table 1 ). Determination of Heat Coma Temperature Upper thermal tolerance limit was determined as "heat coma temperature" (HCT) using a method (Cleland & McMahon 1990. McMahon & Britton 1985, McMahon & Russell-Hunter 1981. McMahon 1990) modified from that previously used for intertidal gastropods (Gowanloch & Hayes 1926. Broekhuysen 1940. Evans 1948. Southward 1958). At heat coma temperature, gastropods lose normal nervous function leading to muscle paralysis mani- fested by cessation of locomotion, loss of substratum attachment, ventral-medial lateral curling of the foot: and cessation of move- ment including tentacles and radular activity (Cleland & McMa- hon 1990. McMahon 1976. 1990. McMahon & Payne 1980. Mc- Mahon & Britton 1985. McMahon & Cleland 1990. McMahon & Russell-Hunter 1981). Adult specimens of each species for which HCT was deter- mined were collected at low tide and returned immediately to the laboratory. A randomly chosen subsample was randomly divided into five to six subsamples of five to six individuals each which were placed into separate, 50 ml test tubes with 40 ml of water from the collection site. Thus, total sample size was 30-36 indi- viduals for most species. Sample size ranged from 17 {Merita chamaeloii) to 71 {Littoriim lillorea) (Table 1). Sample sizes greater than 36 individuals resulted when HCT determinations were repeated and their results combined (Table 1: Batillaria sor- diclu. Bemhicium vittatum. Littorina littorea. Littorina saxatilis. Monodonia labia. Morula tnusiva). Porous foam plugs blocked tube openings at the tube water surface, keeping individuals im- mersed. Subsample shell length range was representative of natural adult populations. Subsamples of larger gastropod species were placed in larger test containers (up to 500 ml), allowing freedom of movement, avoiding over-crowding and preventing severe O, depletion during determinations, otherwise all other procedures were identical. These larger containers were glass jars with plastic screw caps. After receiving snails, the plastic caps of larger vol- ume jars were screwed on while the jar was submerged in water Thi;rmal Tolerance in Intertidal S NAILS 461 TABLE 1. Mean heat coma temperatures (±standard deviatiiin of the mean! for 60 species of intertidal gastropod collected from different rock\ shore. man!>ro\e and salt marsh localities at different latitudes and from different shore heijjhts i.e.. conation levels. Heat coma Superfamily Species Site Latitude Habitat Zonation level n temp. Cerithioidea Baiillarui inininui Jamaica I8.46"N Rocky Shore (TP) Eulittoral Fringe 3(1 43.5'^C±0.97 Cerilhioidea Balilluna mulliformis Hong Kong 22.37°N Mangrove Lower Eulittoral 30 41.3X±0.79 Cerithioidea Balillario sordida Hong Kong 22.37°N Rocky Shore Lower Eulittoral 42 40.5'*C± 1.63 Cerithioidea Ctrilliidea onuita Hong Kong 22.37°N Mangrove Eulittoral Fringe 30 43.3°C±2.17 Cerithioidea Ccriihidea rhizophoramm Hong Kong 22.37'^N Mangrove Eulittoral 38 45.5X± LI I Cerithioidea CenthideopsUhi djadjariensis Hong Kong 22.37°N Mangrove Lower Eulittoral 38 40.9X± 1.26 Cerithioidea Ceriiliium aniicipata Darwin. Australia 12.50^S Mangrove Eulittoral Fringe 28 44.rC±0.31 Cerithioidea Clypeomoms sp. Darwin. Australia 12.50"S Rocky Shore Lower Eulittoral 20 40.rC± 1.07 Cerithioidea Plunaxis nucleus Jamaica 18.46°N Rocky Shore Lower Eulittoral 25 38.rC±0.71 Cerithioidea Planaxis sukatus Hong Kong 22.37^N Rocky Shore Eulittoral 36 40.8°C± 1.13 Cerithioidea Telescopium lelescoptum Darwin. Australia 12.50'^S Mangrove Eulittoral 27 44.4"C± 1.97 Cerithioidea Terehralia palustris Darwin, Australia 12.50=^5 Mangrove Eulittoral 30 44.7X ± 0.84 Cerithioidea Terehralia semistrtala Darwin, Australia 12.50^S Mangrove Eulittoral 25 43.0°C ± 2.22 Cerithioidea Terehralia .sulcaia Hong Kong 22.37°N Mangrove Eulittoral 30 43.6°C± 1.59 Ellobioidea Cassidula angulifera Darwin. Australia 12.50"S Mangrove Eulittoral 20 37.9°C ± 3.39 Ellobioidea Cussidula rugaia Darwin, Australia 12,50^S Mangrove Eulittoral 21 39.3X±L52 Ellobioidea Ellobium auriskoduc Darwin. .A,ustralia 12.5U^S Mangrove Eulittoral 30 39.2X±0.86 Ellobioidea Melampus bidentalus Woods Hole. Mass. 4I.54^N Salt Marsh Eulittoral Fringe 25 38.8X±I.!2 Littorinoidea Bemhicium villatum Albany, Australia 35.00^S Rocky Shore Eulittoral Fringe 42 40.0°C± 1.64 LiUorinoidea Cenchriiis muricatus Jamaica 18.46^N Rocky Shore Eulittoral Fringe 28 42.rC±2.13 Littorinoidea Liltoraria sp. Darwin, Australia 12.50°S Mangrove Eulillora! Fringe 23 43.9X±0.87 LiUorinoidea Uiloraria fdosa Darwin. Australia 12.50^S Mangrove Eulittoral Fringe 30 42.9"C± 1.24 LiUorinoidea Litiorariu irrorata Port Aransas. Texas 28.84^N Salt Marsh Eulittoral Fringe 33 43.7°C±1.38 Littorinoidea Lilloraria melanosioma Hong Kong 22.37"N Mangrove Eulittoral Fringe 27 44.3°C±1.20 Littorinoidea Liltoraria scabra Hung Kong 22.37°N Mangrove Eulittoral Fringe 30 43.3°C±L88 Littorinoidea Lillorina arcana Robin Hoods Bay, Eng. 54,6 PN Rocky Shore Eulittoral Fringe 35 32.8X±0.94 Littonnoidea Utiorina hrevicula Hong Kong 22.37='N Rocky Shore Eulittoral Fringe 30 40.rC±1.03 Littorinoidea Liuorimi compressa Forth Towyn. Wales 53.04^N Rocky Shore Eulittoral Fringe 36 27.6X ± 0.76 Littorinoidea Lillorina Hllorea Robin Hoods Bay. Eng. 54.6rN Rocky Shore Eulittoral 71 32.0°C±0.76 Littorinoidea Lillorina neglecta Robin Hoods Bay. Eng. 54.6 rN Rocky Shore Eulittoral 35 3L8X±1.32 Littorinoidea Lillorina oblusata Robin Hoods Bay, Eng. 54.6 rN Rocky Shore Lower Eulittoral 26 28.3X±2.78 Littorinoidea Lillorina mariae Rohm Hoods Bay, Eng. 54.6 1 °N Rocky Shore Lower Eulittoral 21 2bAC ± 2.60 Littorinoidea Lillorina siuatih's Robin Hoods Ba>. Eng. 54.6 rN Rocky Shore Eulittoral Fringe 70 32.8°C±0.82 Littorinoidea Melaraplie neritoides Filey Head, Eng. 54.2 rN Rocky Shore Eulittoral Fringe 21 34.8=' ±1.40 Littonnoidea Nodilinorina dikilata Jamaica I8.46°N Rocky Shore Eulittoral Fringe 30 45.4X±1.59 Littorinoidea Nodilillorina e.xigua Hong Kong 22.37°N Rocky Shore Eulittoral Fringe 36 44.8°C ± 1 .46 Littorinoidea Nodilinorina inierrupta Port Aransas. Texas 28.84"N Rocky Shore Eulittoral Fringe 30 44.4X±1.36 Littorinoidea Nodilinorina mespdlum Jamaica I8.46°N Rocky Shore (TP) Eulittoral Fringe 30 45.6X±2.I9 Littorinoidea Nodilinorina pyramidalis Hong Kong 22.37°N Rocky Shore Eulittoral Fringe 35 46.3°C±0.71 Littorinoidea Nodilillorina riisei Jamaica 18.46°N Rocky Shore Eulittoral 30 45.3X± 1.68 Littorinoidea Nodilinorina unifasciata Albany. Ausiralia 35.00°S Rocky Shore Eulittoral Fringe 30 41.rC± 1.24 Littorinoidea Nodilinorina ziczac Jamaica 18.46°N Rocky Shore Eulittoral Fringe 30 46.9^C ± 2.68 Littorinoidea Teciarius antonii Jamaica 1 8.46°N Rocky Shore Eulittoral Fringe 27 43.4°C±L65 Muricoidea Lepsiella vinosa Albany. Australia 35.00°S Rocky Shore Eulittoral 30 39.2X ± 0.96 Murieoidea Morula nnisi\'a Hong Kong 22.37^N Rocky Shore Lower Eulittoral 60 39.6X ± 0.64 Muricoidea Nncella lapillus Filey Head, Eng. 54.2 rN Rocky Shore Lower Eulittoral 36 29.6 ±1.38 Neritoidea Neriia airamenlosa Albany. Australia 35.0O°S Rocky Shore Eulittoral 30 38.8X±2.I4 Neriloidea Nerita balteaia Darwin. Australia 12.50°S Mangro\e Eulittoral 28 41.rC±L08 Neritoidea Neriia chamaeleon Hong Kong 22.37°N Rocky Shore Eulittoral Fnngc 17 43.3X±0.69 Neritoidea Neriia peloronta Jamaica I8.46"N Rocky Shore Eulittoral 29 40.8°C± 1.90 Neritoidea Nerita tesselata Jamaica 18.46'^N Rocky Shore Eulittoral 30 40.9X± 1.20 Neritoidea Neriia versicolor Jamaica I8.46''N Rocky Shore Eulittoral 60 42.3'^C± 1.39 Neritoidea Neritina reclivala Gulf Shores. AL 30.35°N Salt Marsh Lower Eulittoral 30 41.3 ±0.70 Neritoidea Neritina virginea Jamaica 18.46"N Rocky Shore Eulittoral 30 45.4X±1.16 Neritoidea Piiperita pupa Jamaica 18.46'^N Rocky Shore (TP) Eulittoral Fringe 30 44.6°C± 1.81 Onchidioidea Oncluditim sp. Darwin. Australia 12.50°S Rocky Shore Lower Eulittoral 30 37.0X±1.20 Trochacea Auslrocochlea constricta Albany, Australia 35.(H)°S Rocky Shore Eulittoral 30 37.5°C±I.17 Trochacea Austrocochlea concamerala Albany. Australia 35.00''S Rocky Shore Eulittoral 30 35.6X±0.82 Trochacea Lunella coronala Hong Kong 22.37°N Rocky Shore Lower Eulittoral 30 38.8X±1.02 Trochacea Monodonta lahin Hong Kong 22.37°N Rocky Shore Eulittoral 60 37.4X± 1.92 TP Rocky shore species restricted to tide pool habitais 462 McMahon from the collection site in order to exclude all air bubbles and keep individuals immersed. Containers holding snail subsaniples were placed in a water bath to a depth at \\ hich bath water lexel exceeded test tube water level or was equivalent to that in the glass jars. Uniform bath water temperature was maintained by vigorous aeration or a circulating pump. The bath heater switch was manually controlled to raise bath water temperature and. therefore, container water tempera- ture, at a rate of approximately I °C 5 min"'. This rate of increase makes lag between container water and snail tissue temperatures negligible (Broekhuysen. 1940). Container water temperature was monitored with a thermistor probe extended through the foam test tube plug or a small hole in the jar cap and connected to a YSI model 43-TD tele-thermometer or an eomega 866 Thermometer. Bath water temperature was monitored with a laboratory-grade mercury thermometer or separate thermistor probe. Each HCT determination began at room temperature ( 19-26°C, dependent on laboratory ambient temperature). Prior to HCT de- termination, snails were allowed to attach to container walls and actively locomote. Individuals remaining inactive (i.e.. not pedally attaching) during determinations where not included in HCT com- putations, occasionally reducing sample size below the initial 30 or 36 individuals placed in chambers (Table 1 ). Number of individu- als displaying heat coma symptoms in each container was recorded at every 1°C increase in chamber water temperature. Container water temperature was increased at 1°C 5 min^' until all previ- ously active individuals in all subsamples displayed heat coma symptoms. Thereafter, containers were removed from the bath, partially decanted, and allowed to cool to room temperature. Snails remained in the containers and their recovery observed after a 1-12 h post-treatment return to room temperature (i.e.. observed for re-attachment and/or active locomotion). Included in the data set were HCT determinations for seven species of Jamaican littorinoidean snails {Cenchntis inuruatus. Tectariiis antonii. Nodilitlorina dilatata, Nodililtorina :icrac. Nodilittorina angiistior. Nodililtorina riisei and Noditittorina mespillwn. Table 1) determined by Brilton (1992) utilizing the same methodology described above. Across all 60 tested species, HCT was determined for a total of 1,945 individuals. RESULTS Correlation of Species ' Heat Comma Temperature with iMlitude A high proportion of individuals (> 90-95%) in all tested spe- cies recovered from heat coma on cooling to room temperature. For some species, individual subsamples were removed from the bath at sequentially higher temperatures subsequent to entering heat coma. Among these species, lethal acute temperature expo- sures occurred approximately 3-7°C above HCT. indicating that heat coma is a reversible, nonlethal condition in intertidal gastro- pods as it is in freshwater gastropods {McMahon 1976. McMahon & Payne 1980). Mean HCT values among the 60 tested species ranged from a low of 26.4°C for the low-shore littorinoidean. Littorina muriae. from Robin Hoods Bay. England (54.61 °N) to a high value of 46.9°C for the high-shore littorinoidean. Nodilittorina ziczac, from Jamaica (18.4''N) (Table 1). Across all tested individuals. HCT was negatively conelated with latitude (Fig.lA). The relationship between individual HCT and latitude as the controlled vaiiable was best modeled by a quadratic equation: HCT C^C) = 39.83 + 0.29 CLat) - 0.0083 ("Lat") (n= 1.945. d.f. = 2. r- = 0.681. F = 207 1.0.P< 0.00001). whose r value indicated that correlation with latitude explained 68'7rofobserved HCT variation (Fig. lA). Fitting of HCT data for all tested individuals in the superfamily. Littorinoidea (25 species) to a quadratic equation against latitude as the controlled variable yielded the following result: HCT C'C) = 41.99 + 0.30 (°Lat) - 0.0091 CLat") (n = 836.d.l l.r = 0.855, F = 2455.4. P< 0.00001). whose r" value indicated that correlation with latitude explained 867f of HCT variation (Fig. 1 A). Fitting of HCT data for all tested individuals in superfamilies other than the Littorinoidea (35 spe- cies) to a quadratic equation against latitude as the controlled variable yielded the following result: HCT (°C) = 39.59 + 0.25 C'Lat) - 0.0078 ("Lat") (n= 1109. d.f. = 2. r- = 0.409. F = 382.75. P< 0.00001). whose r" value indicated that correlation with latitude explained 419^ of HCT variation (Fig. 1). Mean HCT values for lit- torinoidean species and nonlittorinoidean species fell close to that predicted by quadratic fits of HCT to latitude (Fig. lA). Differences in the HCT among Taxonomic Groups The elevated intercept and slope values of quadratic equations relating HCT to latitude among individuals of littorinoidean and nonlittorinoidean species suggested that, over tested latitudes, lit- torinoidean species had generally higher HCT than nonlit- torinoidean species. This hypothesis was tested by Analysis of Covariance (ANCOVA) of HCT with taxonomic groups as treat- ments (i.e.. Superfamily Littorinoidea versus nonlittorinoidean species) and both latitude of collection and latitude of collection squared as covariates. Use of latitude and latitude squared as co- variates in this analysis removed the influence of the negative quadratic relationship with latitude on treatment mean HCT values (see above), allowing the direct effects of the treatments to be statistically assessed. The analysis yielded species" mean HCT values that were "adjusted" to eliminate latitude effects, hereafter referred to as "adjusted mean HCT" values. ANCOVA revealed that the adjusted mean HCT of littorinoidean snails (±95% confi- dence lunits of the mean) (41.r"C ± 0.446. /; = 836) was signifi- cantly greater than that of nonlittorinoidean species (38.43"C ± 0.209. ;i = 1.109) (« = 1945. d.f. = l.f = 401.7. P < 0.00001 ) (Fig. IB). ANCOVA with latitude and latitude" as covariates also re- vealed differences in HCT among the seven superfamilies repre- sented in the data set {n = 1945. d.f. = 6, f =169.6. P < 0.00001 ). A post hoc Scheffe pair-wise comparison test revealed significant differences [P < 0.05) among superfamily adjusted mean HCT values (Fig. 2). The Littoriniiidea had a higher adjusted mean HCT (42.2"C ± 0.446. n = 836) than the other six super- families represented. Adjusted mean HCT of the Cerithoidea (40.8X ± 0.237. n = 423) and Neritoidea (40.6"C ± 0.277. n = 284) where not different, but were significantly greater than those of the Muricoidea (39.0°C ± 0.808. /; = 126). Ellobioidea (38.6°C ± 0.386. ;; = 96). Trochacea (36.6°C ± 0.287. n = 150) and Thermal Tolerance in Intertidal Snails 463 • 1 ^45 — ^ 1 • M 1 o ■ •^-■•^-. t> 40 o o 0 C 0) o o o 1- O^xx m 3b - c - ^"^*i> o -^ ^sx • • Supe rfamily Littorinoidea "- > Si 30 X 0 Nonlittroinoidean Superfamilies c • (t: Q> S 25 1 I 1 1 1 t 1 1 1 1 10 15 20 25 30 35 40 45 50 Degrees of Latitude (North or South) 55 42 i41 v40 o ^39 i38 - B - A B (35-1109) A (25-836) Littorinoidea Nonlittoriniodeans Group Figure I. A. Rtlationslilp between heat coma temperature (HCT) (vertical axis) and degrees of latitude of collection (horizontal axis) among subsaniples of 60 species of intertidal snails. Solid circles rep- resent the mean HCT values for species in the superfamilv, Lit- torinoidea, while open circles are those of species in nimlillorinoidean superfamilies. The solid line represents the best quadratic regression fit of HCT to latitude of collection for all tested individuals, the dot- dashed line, the best fit for littorinoidean snails and the dashed line, the best fit for nonlittorinoidean snails (See Results for regression parameters). B) Adjusted mean heat coma temperatures (HCT) (ver- tical axis) for littorinoidean and nonlittorinoidean snails. Different letters inside or above histograms indicate significant difference iP < 0.051 between means. Numbers within or above the histograms repre- sent the number of species represented in the sample (left of dash) and number of individuals (right of dash) on which means were based. Error bars are MS'/r confidence limits of the mean. Onchidioidea (35.5''C ± 0.457, n = 30). The adjusled moan HCT of the Muricoidea and Ellobioidea were not ditterent. but were significantly less iP < 0.05) than those of the Littorinoidea, Cer- ithoidea and Neritoidea and greater than those of the Trochacea and Onchidioidea (P < 0.5|. The adjusted mean HCT of the Tro- chacea and Onchidioidea did not differ and were significantly lower {P < 0.05) than those of all fi\e other supeifaniilies (Fig. 2). Differences in HCT among Zonation Levels and Habitat Types ANCOVA with latitude and latitude" as covariates revealed significant differences in adjusted mean HCT of individuals rela- tive to zonation level (;; = 1945. d.f. = 2, f = 340.3, P < 0.00001 ). A post hoc Scheffee pair-wise comparison test indicated that the adjusted mean HCT of low-shore (37.5°C ± 0.515, /; = 43 42 41 40 39 38 36 - A ■ (25-836) 1 B (14-423) B (9-284) c '■ s A (3-126) C -■ ■ ■ __ (4-96) 1 1 1 1 1 D (4-150) '■ ■ ■ ■ ■ D -I I I I I ^ (1-30) O \ 'i. % % \ o (9 o \ o. q- Superfamilies of Intertidal Gastropods Figure 2, Adjusted mean heat coma temperatures (vertical axis) for snails representing seven different superfamilies of intertidal gastro- pods (horizontal axis). Vertical bars above histograms represent 95% confidence limits of the mean. Different letters above the histograms indicate significanl differences between means i.P < t).05l. Numbers above the histograms represent the number of species represented in the sample (left of dash) and number of individuals (right of dash) on which means were based. 380), mid-shore (39.4°C ± 0.320. /; = 750) and high-shore snails (41.4°C ± 0.362. n = 815) differed significantly from each other (P < 0.05). thermal tolerance increasing with zonation level (Fig. 3A). ANCOVA with latitude and latitude" as covariates also re- vealed significant differences in adjusted mean HCT of individuals relative to habitat (n = 1945, d.f. = 2, f = 23.9. P < 0.00001). A post hoc Scheffee pair-wise comparison test indicated that the adjusted mean HCT of rocky shore (40.2°C ± 0.292, n = 1.370), mangrove (41. r'C± 0.228, ;i = 479) and salt marsh snails (41. 7°C ± 0.486, /; = 88) all differed significantly {P < 0.05), rocky-shore snails being least thermally tolerant and salt-marsh snails, most tolerant (Fig. 3B). A multiple factor ANCOVA with latilude and latitude" as co- variates and zonation level and taxonomic group (i.e., lit- torinoidean versus nonlittorinoidean snails) as treatments, revealed significant differences in the adjusted mean HCT of snails by zonation level (d.f. = 2. F = 154.1. P < 0.00001 ) and ta.xonomic group (d.f. = 1. F = 23.5. P < 0.00001) with significant inter- action between treatments (d.f. = 2. F = 48.9. P < 0.00001). Adjusted mean HCT values and 95% confidence limits for each treatment combination were: low-shore littorinoideans. 35.1°C ±0.832, n = 47; mid-shore littorinoideans, 39.8"C ±0.963, n = 136: high-shore littorinoideans, 39.4-'C ±0.463, /) = 653: low- shore nonlittorinoideans. 35.5°C ±0.389. n = 333: mid-shore non- littorinoideans. 36.7°C ±0.259, /; = 614: and high-shore nonlit- torinoideans. 39.2°C ±0.362. ;; = 162 (Fig. 4A). Post hoc Scheffe pair-wise comparisons indicated that adjusted mean HCT for all snails significantly increased (P < 0.05) with zonation level (i.e.. across low-shore, mid-shore and high-shore zones) (Fig. 4A). Ad- justed mean HCT did not differ among littorinoidean and nonlit- torinoidean snails in either the low-shore or high-shore zones, but 464 McMahon 43 42 41 40 39 38 37 36 35 A B c (26-815) . ^ - A (22-750) PI ^^^H (12-380) H hi 1 - 1 ^^H ■ ^^^^^^^1 ^^^^^^^^H ■ ' ' ■ - 1 ^^^H ^^^^H i- '-^1 - 1 ^^H ^^^1 ■Mil Lower Eulittoral Eulittoral Level of Zonation Eulittoral Fringe 43 E A (40-1378) B (17-479) C (3-88) ^ 41 40 ^^^^H 39 - ^^H B 38 - ^^H Mi 37 - ^^H 1*1 36 - pi^J i 1 35 - ■■■■ ■H Rocky Shore Mangrove Habitat Type Salt Marsh Figure 3. (A) Adjusted mean heat coma temperatures (vertical axis) for intertidai snails relative to zonation level (horizontal axis). (B) Adjusted mean heat coma temperatures (vertical axis) for intertidai snails relative to habitat type (horizontal axis). In both figures, vertical bars above histograms represent ')5'^c confidence limits of the mean. Different letters above the histograms indicate significant differences between means iP < 0.05). Numbers above the histograms represent the number of species represented in the sample (left of dash) and number of individuals (right of dash) on which means were based. was significantly greater {P < 0.05) among littorinoi(dean versus nonlittorinoidean mid-shore snails. Within the Littorinoidea. ad- justed mean HCT was significantly lower in low-shore snails [P < 0.05j relative to mid-shore or high-shore species among which adjusted mean HCT did not differ. Among nonlittorinoidean snails, adjusted HCT increased significantly (P < 0.05) across all zonation levels (Fig. 4A). A multiple factor ANCOVA with latitude and latitude" as co- variates and habitat type and taxonomic group (i.e.. Littorinoidea versus nonlittorinoideans) as treatments revealed significant dif- ferences in the adjusted mean HCT of snails among different habi- tats (d.f. = 2, F = 2\A. P < 0.00001) and ta.xonomic groups (d.f. = 1. F = 89.2. P < 0.00001) with significant treatment interaction (d.f. = 2. F = 29.9. P < 0.00001). Adjusted mean HCT values and 95% confidence limits for each treatment com- bination were: littorinoideans on rocky shores (42.2"C ± 0.508, ;; = 693). littorinoideans in mangroves (41.7'C ± 0.277. it = I 10). littorinoideans in salt marshes (42.8°C ± 0.495 /( = ?0.977 in both cases. A plot of lip length against major axis length (see CANl ) separated clearly the coastal samples from the others; for any given major axis length the former always had a longer lip (i.e. greater jugosity). The East Fleet shells were confirmed as having lip lengths intermediate between those of coastal and the other lagoonal samples (Fig. 5a). Plotting aperture width against shell width (see CAN2) did not separate the samples entirely along coastal versus lagoonal lines. Robin Hood"s Bay shells were more similar to those of lagoonal animals, except for lagoonal Golam shells that were grouped with those from East Fleet, coastal Golam, and Wells (Fig. 5b). The shells from Holkham, Snettisham, Golam (lagoonal) and Robin Hood's Bay had the most obtuse spires, the angle decreasing with increase in Whorl Width 2 (see CAN3). Those from East Fleet were noticeably the most pointed but the angle increased with increase in Whorl Width 2. Thus the separation was again not strictly coastal versus lagoonal (Fig. 5c). CAN I expressed most clearly the separation of coastal and lagoonal animals and this is seen when the three canonical variates a 12 E . Wells - RHB ^^'Golam (coastal) ^Cley ^^- -Alderton East F'^s'/^'^^e^-^'''^ ^''^\,-f'\ Holkham ^^"'y' Snettisham Golam (lagoonal) 6 8 10 12 14 16 18 Major axis length (mm) :r 06 O) 5 04 02 00 * Lagoonal t. = 0 980 18 2 4 6 8 10 12 14 16 Major axis length (mm) Figure 4. (a) Shell weight (mg"') plotted against major axis length (mm). Regression lines for each of the nine populations. RHB, Robin Hood's Bay. (b) Shell weight (nig'") plotted against major axis length (mm). Plots of individual measurements and regression lines for the pooled data. Open circles, coastal sites -i- East Fleet; filled circles, other lagoonal sites. are plotted against each other (Fig. 6a). The shells from East Fleet, however, fall into the "coastal" group and indeed there is consid- erable overlap between them and those from Golam (coastal) and Wells. The coastal shells separate out along the CAN2 and CAN3 axes in the order East Fleet, coastal Golam, Wells and Robin Hood's Bay. The Robin Hood's Bay shells separate out almost completely from the above three samples; the box surrounding the cluster of 50 shells from this site contains only one other shell (from Wells) (Fig. 6a). Amongst the lagoonal shells, the only group that could be separated by rotation of the axes was that from Golam (lagoonal), along the CAN2 and CAN3 axes (Fig. 6b); the box surrounding the cluster of 54 shells contains only one other shell (from Alderton). Finally a Discriminant Analysis was carried out on the size- adjusted data (Table 3). As expected from the above, the most consistent samples were those from Robin Hood's Bay and Golam (lagoonal), where 98% and 94% of the shells respectively were self classifying. Except for East Fleet, no misclassifications occurred between lagoonal and coastal shells. The shells from Cley showed the greatest degree of misclassification, with 20% classifying to Alderton, 12% to Holkham and 8% to Snettisham. Snettisham shells also had a high proportion of misclassifications. with 19% LiTTORiNA Shell Shape 473 a 3 ^ 2 - E E, £ Snett Golam coastal Hoik' - Golam tagoonal 8 10 12 Major axis length (mm) E E. £ 4 :s 2 3 Cley /^B Gobm coastal ■ Aldeiton^ /" ^^^^^NaWs. SneHisham ^^'^^^^^^ ^**^ ^.'^ HolVham ^^■^^^'^ ^^J^EasI Fleet X^^^Golani lagoonal 3 4 5 6 Shell width (mm) C 110 100 I 8" 90 J5 80 ^ 60 HoMiam Snetllsham Golam lagoonal East Fleet Whorl width 2 (mm) Figure 5. la) Lip length Imnil pli)tted against major axis length (mm) with ri'gri'ssii)n lines for each of the nine populations; Holt>. Holkham: RHB, Kohin Hood's Bay: Snett, Snettisham. (hi Aperture width (mm) plotted against shell width (mm I with regression lines for each of the nine populations: RHB, Rohin Hood's Bay. Ic) Apical angle (degrees) plotted against whorl width 2 (mm) with regression lines for each of the nine populations: RHB, Rohin Hood=s Bay. b 1 1 ■ \i Figure 6. Plot of the data from the nine sites on Canonical Variables 1 (CANl), 2 (CAN2) and 3 (CAN3). Canonical Variahle axes at the mid point. +, Alderton; *, Cley; ▲. East Fleet: T, (iolam coastal: ♦, Holkham; ■. Robin Hood's Bay and Colam lagoonal; •, Snettisham; *, Wells, (a) The shells on the left are the coastal + East Fleet, and are separated from the lagoonal shells along the CANl axis. The square encloses the 50 Robin Hood's Bay shells plus one from Wells, (b) The square encloses the 54 lagoonal (Jolam shells plus one from Alderton. classifying to Holkham. 99f to Cley an(j 3% to each of Alderton and East Fleet. Amongst the coastal samples, the highest propor- tion of misclassifications were from the coastal Golam shells, with 247f misclassifying to Wells and 11% to East Fleet. Twelve per- cent of Wells shells also misclassified to East Fleet. Of the East Fleet shells, 1 1% misclassified to coastal Golam and 27f to Cley. In most cases, using the raw data (i.e. not excluding size) reduced the proportion of misclassifications. DISCUSSION The Littorina .sa\iiiilis complex includes a wide variety of shell morphs and the taxon is thmighl to he undergoing differentiation 474 Mill et al. TABLE 3. Discriminant analysis of the data after removing the effects of size by standardizing using the geometric mean. From Alderton Cley East Fleet Golam (lagoon) Golam (coastal) Holkham RHB Snettisham Wells Errors N Alderton 89 9 0 0 0 0 0 0 0 11 0 0 i: u 2 n 0 0 0 0 n 0 8 0 0 0 10 0 0 0 0 0 24 0 -) (1 11 40 13 6 34 17 34 20 ?3 Cley 0 0 0 0 0 3 0 60 50 East Fleet 2 4 0 7 0 9 0 87 53 Golam (lagoon) 0 11 0 0 3 8 94 54 Golam (coastal) 0 0 0 0 0 66 38 Holkham 0 0 0 12 83 72 RHB 0 19 (1 98 50 Snettisham 0 0 66 32 Wells 0 80 50 Total 452 Values are percentages: shading Indicales self classification. (Fretter 1980. Ward & Warwick 1980) which might have arrived at. or in the future reach, species status for one or more of the morphs. This process is thought to be aided by direct development (Van Marion 1981; Janson 1982; Grahame & Mill 1989) and hence poor dispersal ability (Ward & Warwick 1980; Janson 1983; Janson & Ward 1984; Faller-Fritsch & Emson 1985). However. L. sa.xatilis is a rapid colonizer of offshore islands (Johannesson & Johannesson 1995). Furthermore, two other, closely related taxa. L arcana and L. compressa. both show comparatively little varia- tion in shell morphology and yet are direct developers. The main reproductive difference between L. sa.xatilis on the one hand and the other two species on the other, is that the former is ovovivipa- rous whereas the latter are oviparous. It is clear from the data that tlie shells in this study can be separated into coastal + the Fleet and lagoonal. and CANl pro- vides an axis for this separation. Lagoonal animals have a lighter shell than correspondingly sized coastal animals; also lagoonal shells lack the jugosity found in coastal populations. The position of the Fleet animals is not surprising as this lagoon is tidal and the sample was taken within a few hundred meters of the lagoon entrance. Of the other lagoonal animals, those from Golam (la- goonal) and Holkham satisfy the strict criteria of Muus ( 1967) for Liuarina tenebrosa. However, although the shells from Golam (lagoonal) are clearly separable from those of other lagoonal samples, those from Holkham are not. This is somewhat surprising in view of the subjective impression of the shells (Fig. .3). How- ever, this may be due to the (apparent) intermediate shape of the shells from Snettisham between those from Alderton and Cley on the one hand and those from Holkham on the other. Barnes ( 1993) has concluded that there is currently insufficient evidence to accept a species status for L. teiieliinsa (seiisii Muus 1967) but that it is clearly distinguishable both in shell character- istics and habitat from L. saxalilis s.s. He suggested the varietal name L. saxatilis var. lagimae for the former. Although, in the current study, the shells from Golam (la- goonal) separate out from those that came from other lagoons, there appears to be a gradation in shape, size and habitat in the lagoonal populations. It would not be surprising if we are witness- ing different degrees of divergence in different lagoonal popula- tions. This might be related to the age of the lagoon and hence to the period of their separation from coastal animals. Only those from Golam and Holkham fulfil all of the criteria for the status of L. saxatilis var. lagiiiiae but others may have changed partially along this route, particularly those from Snettisham. At Snettisham the animals were found apparently permanently submerged but occurred on the substrate (rocks) rather than on macrophytes. At Cley they were on both the substrate and on floating mats of Chaetomoipha. and at Alderton were found on the substrate around the edge of the lagoon. Furthermore, in the last two sites, the animals were not permanently submerged and they reached a si/e similar to that achieved by the coastal animals. The gradation is retlected in the Canonical Discriminant Analysis, where the two taxa do not separate on the CAN2 axis; indeed the Holkham and Golam (lagoonal) shells fall into different groups when aperture width is plotted against shell width. Similarly, they do not separate on the CAN3 axis. Thus, when apical angle is plotted against whorl width 2. the Snettisham shells align with those from Holkham and Golam (lagoonal). but those from Cley and Alderton are more similar to the Wells and coastal Golam shells respec- tively. The current view of speciation is generally that of Mayr ( 1942) in which a geographical barrier develops between populations, isolating them reproductively from each other. Following this, divergence occurs between the populations, even if the separated habitats are identical, and separate species ultimately evolve. However, there is another possible route for speciation. ecological speciation. Although the idea is not new. it has been brought into focus recently that ecological barriers rather than geo- graphical ones may also be important for speciation (Morell 1999). Thus, ecological pressures could favor changes that even- tually cause populations to become reproductively isolated in the absence of geographical barriers. It may be expected that popula- tions that are ecologically separated but genetically similar to each other would be more likely to interbreed than comparable ones that have been separated geographically. However, this is not neces- sarily the case and size differences between genetically similar populations may be sufficient to produce reproductive isolation (Morell 1999). It is entirely possible that this is the case with L saxatilis var. lagunae and L. saxatilis s.s.. since the former has clearly devel- oped sexual maturity at a size much smaller than occurs in the latter. Indeed Barnes (1993) has suggested that L sa.xatilis var. LmoRiNA Shell Shape 475 kigiinae may have a paedomorphic origin, as Raflaelli ( 1979) sug- gested for another taxon within the L. saxatilis complex, i.e. L. neglecta. and that, in the case of the former, small size is a re- quirement of living on submerged macrophytes such as Cluwto- inorpha. It follows from the above that L. saxatilis s.s. should interbreed with North American L. saxatilis but not with L. saxa- tilis var. lagimae occurring in the same lagoon. However, caution is required until the appropriate breeding experiments have been attempted. Furthermore, if as seems to be the case, parallel e\o- lution is occurring in two or more lagoons, and if the individuals in these populations can interbreed with each other but not with adjacent L. saxatilis s.s., then it follows that any resulting "spe- cies" will have a polyphyletic origin. In only one lagoon (Golam) were both L. saxatilis var. lagunae and L. sa.xatilis s.s. recorded and they were separated by some 100 m. the former occurring at the landward end of the lagoon, the latter adjacent to the rocky shore, h seems highly likely that the two populations are isolated reproductively (Wilson et al. 1999). From a conservation point of view it is irrele\ ant as to whether or not we are dealing with two separate species or two morphs. It could be argued that, if priorities have to be decided, for example because of costs, it is more important to conserve at the species level. However, maximum biodiversity must be preserved so that evolutionary processes are allowed to continue. Brackish lagoons are a nationally rare habitat and have been accorded a "priority habitat type" under Annex 1 of the EU Habitats and Species Di- rective (Bamber 199S). It may be that many lagoons are very teneral. lasting only tens, or at best hundreds, of years (Bamber 1998). but others may be sufficiently permanent to allow coinplete separation of species to occur. Hence, it is vital that these lagoonal habitats be conserved. ACKNOWLEDGMENTS This research was supported by the MAST-3 program of the European Commission under Contract MAS3-CT95-0042 (AM- BIOS). We thank Prof R. McMahon and an anonymous referee for their constructive comments on the manuscript. LITERATURE CITED Bamber 1998. Conservation of brackish lagoonal faunas. J. Conch. Special Publication y.li<^-110. Barnes. R.S.K. 1987. A survey of the coastal lagoons of North Wales (R, Clwyd - Aberystwyth). September 1987. Nature Conservancy Council Report CSD 1 \2y. Peterborough: Nature Conservancy Council. Barnes, R. S. K. 194.^. On the nature of the coastal lagoon winkles attrib- uted to Littoriihi tcnchi'osa and Litlorina sti.\atili\. Call. Riol. Mar 34:477-f9.S. Faller-Fritseh. R. J. & R. H. Emson. 1985. Causes and patterns of mortality in Litlorina nulls (Maton) in relation to intraspecific variation: a re- view. In: P.O. Moore & R. Seed, editors. The Ecology of Rocky Coasts. London: Hodder & Stoughton. pp. 157-177. Fretter, V. 1980. Observations on the gross anatomy of the female genital duct of British Liitoriiia spp. / Moll. Stud. 46:148-153. Gosling. E. M.. I. F. Wilson & J. Andrews. 1998. A preliminary study on genetic differentiation in Litlorina sa.mlili.s from Galway Bay. Ireland: Litlorina tenebrosa Montagu - a valid species or ecotype? Hyilrobio- logia 378:21-25. Grahame, J. & P.J. Mill 1989. Shell shape variation in Litlorina sa.xatilis and L. arcana: a case of character displacement? 7. Mar. biol. Ass. U.K. 69:837-855. Grahame. J., P, J. Mill. S. Hull & K. J. Caley, 1995. Litlorina ncKlccia Bean: ecotype or species? J. Nat. Hist. 29:887-899. Hull, S., J. Grahame & P. J. Mill. 1996. Morphological divergence and evidence for reproductive isolation in Litlorina sasatilis (Olivil in Nonheast England. J. Moll. Stud. 62:89-99. Hull, S. L.. J. Grahame & P. J. Mill. 1999. Heat stability and activity levels of aspartate aminotransferase and alanine aminotransferase in British Littorinidae. J. Exp. Mar. Biol. Ecol. 237:255-270. Janson, K. 1982. Phenotypic differentiation in Litlorina \a.\atilis Olivi (Mollusca. Prosobranchia) in a small area on the Swedish west coast. J. Moll. Stud. 48:167-173. Janson. K. 1983. Selection and migration in two distinct phenotypes of Litlorina scLxatilis in Sweden. Oecologia. Berl. 59:58-61. Janson, K. & R. D. Ward. 1984. Microgeographic variation in allozyme and shell characters in Litlorina sa.xatilis Olivi (Prosobranchia: Lit- torinidae). Biol. J. Linn. Soc. U.K. 22:289-307. Johannesson, K. & B. Johannesson. 1995. Dispersal and population ex- pansion in a direct developing marine snail [Littorina sa.xatilis) follow- ing a severe population bottleneck. Hydrohiologia. 309:173-180. Morell. V. 1999. Ecology returns to speciation studies. Science 284:2106- 2108. Muus. B.J. 1967. The fauna of Danish estuaries and lagoons. Medd. Damn. Fislc.-og Havunders iNx Ser.l 5:1-316. Raffaelli. D. 1979. The taxonomy of the Littorina sa.xatilis species- complex, with particular reference to the systematic status of Littorina patula Jeffreys. Zool. J. Linn. Soc. U.K. 65:219-232. Reid. D. G. 1996. Systematics and Evolution of Littorina. Andover, En- gland: The Ray Society. 463 pp. Seaward, D. R. 1980. The marine molluscs of the Fleet. Dorset. Proc. Dorset Nat. Hist. Archaeol. Soc. 100:100-108. Smith, S. M. 1982. A review of the genus Littorina in British and Atlantic waters (Gastropoda: Prosobranchia). Malacologia 22:535-539. Van Marion. P. 1981. Intrapopulation variation of the shell of Littorina riidis (Maton) (Mollusca: Prosobranchia). J. Moll. Stud. 47:99-107. Ward, R. D. & T, Warwick. 1980. Genetic differentiation in the moUuscan species Littorina rudis and Littorina arcana (Prosobranchia: Littorin- idae). Biol. J. Linn. Soc. U.K. 14:417^28. Wilson. I.E.. E.M. Gosling & W. Tapper 1999. The systematic status of the lagoon periwinkle, Littorina tenebrosa. J. Mar. Biol. Ass. U.K. 79:653- 660. Jotimal oj Shrllfiih Research. Vol. 20. No. 1. 477^S3. 201)1. IS LIFE ON A TROPICAL SHORE REALLY SO HARD?: THE ROLE OF ABIOTIC FACTORS IN STRUCTURING A SUPRALITTORAL MOLLUSCAN ASSEMBLAGE DWAYNE MINTON* AND DEBORAH J. GOCHFELDf Hdfstni University Marine Liilyonitory. Priory. Si. Ann. Jamaica. Wc.'it Indies AKSTR.WT Interactions between biotic and abiotic factors are considered to be the principal mechanisms controlling the dynamics of rock) shore communities. Unfortunately, little research has e.xamined how these factors affect community structure of tropical rocky shore assemblages. We examined the effects of wave action and desiccation on a rocky shore moUuscan assemblage on the north shore of Jamaica. This assemblage exists entirely above mean high water (MHW). where physical factors were expected to be more important than biological factors. We compared the molluscan assemblage along ten vertical transects exposed to different levels of wave action and desiccation potential. In all. nineteen species iif mollusk were observed, thirteen of which occurred on >50% of our transects. We found no differences in species number, individual densities, or the vertical distribution of the species between transects with differing levels of wave action or desiccation potential. Coirespondence analysis revealed differences in assemblage structure, but the differences were not associated with wave action or desiccation, suggesting that these physical factors are not operating at the spatial scale studied. However, gastropod mollusks preferentially occupied pit and crevice microhabitats, which are believed to mediate physical stresses. The distribution of these rocky shore mollusks may be the result of the availability of. and competition for. these sheltered micro- habitats. KEY WORDS: Community structure, rocky shore, mollusks. Littorinidae. Ccnchiili.'i. Nmlilitloriiui. Ti-cUiniis INTRODUCTION The supralittoral fringe, the region on the shore above mean high water (MHW). is often characterized as an especially harsh environment in which species are regulated by physical, instead of biological, factors (Underwood 1979, Lang et al. 1998). High tem- peratures, desiccation, salinity extremes, UV exposure and me- chanical wave stress are all piitential abiotic factors that can struc- ture the supralittoral community (Garrity 1984. Lubchenco et al. 1984). These physical factors are considered more extreme along tropical than temperate shores (Moore 1972), and may be a stron- ger regulating mechanism on these tropical communities (Menge et al. 1986. Menge and Sutherland 1987). Unfortunately, few stud- ies have examined community-level dynamics of the supralittoral fringe habitat in the tropics. The supralittoral fringe in the Caribbean supports a diverse and conspicuous molluscan assemblage. These species exist in over- lapping vertical zones, beginning at or near MHW. and experience varying degrees of wave action, but are seldom, if ever, immersed. Desiccation, heat stress, and mechanical wave action have all been shown to affect the behavior of supralittoral gastropods, resulting in limited cyclic activity (Garrity 1984), selective use of micro- habitats (Garrity 1984. Peckol et al. 1989). evaporative cooling (Vermeij 1971 ). formation of multilayer aggregations (Garrity and Levings 1984), and the establishment of shore-level size gradients (Vermeij 1972). How these physical factors affect the structure of the entire molluscan assemblage, however, is poorly understood. This paper examines the effects of wave action and desiccation on the distribution of molluscan species on an exposed rocky shore on the north shore of Jamaica. Although several species of this assemblage have been the focus of previous research (Britton *Corresponding author. Present address: Department of Zoology. Univer- sity of Hawaii. Honolulu. HI 96822 tPresent address; Department of Pharmacognosy. University of Missis- sippi, University, MS 38677 1992, Lang et al. 1998, Gochfeld and Minton 2001), no work has examined community-level dynamics of the entire assemblage. MATERIALS AND METHODS The exposed rocky shore near the Hofstra University Marine Laboratory. Priory, .St. Ann, Jamaica, is a micro-karsted limestone platform rising approximately three meters above MHW and ex- tending landward from 10 to 25 m. It is often exposed to high swell, but is seldom inundated, remaining, for the most part, dry throughout the year. We selected and permanently marked ten transects (Fig. 1). Transects ran from MHW landward to the position of first veg- etation (usually Raehiecdlis americana (Kuntze, 1891), Coccoloba iivifera (Linnaeus, 1758), or Borrichia arborescens (Linnaeus, 1758)). MHW was easily identified as the lower boundary of a yellow band of microscopic algae (Brattstrom 1980). Transects varied in length from 190 to 570 cm, were spaced at least 2 m apart, and extended over approximately 100 m of shoreline. Stud- ies on movement of supralittoral gastropods (Gendron 1977, Peckol et al. 1989. Williams 1995, Williams and Morritt 1995, Gochfeld and Minton 2001 ) suggest that our transects were suffi- ciently far apart that individual snails would be unlikely to move between them. We determined the level of wave energy at each transect line by counting wave impacts over five minute intervals on four different days. If a wave impacted above the MHW. it was counted as a strike. If a wave or its resulting splash landed above the yellow (splash) zone on the rock, it was counted as a severe strike. We conducted all ten surveys on a given day within 30 minutes of each other by randomly dividing the transects between two observers. A preliminary investigation showed no significant bias between the observers. Based on the average number of wave strikes, we di- vided the transects lines into high energy and low energy sites. We compared the number of strikes for the two groups using a one- tailed, two-sample /-test with different sample variances. 477 478 MlNTON AND GOCHFELD TABLE 1. ANOVA on change in sponge weight after 1 hour ol exposure on the rocky shore. [D] (D © Figure I. Locations of 10 permanent transects on the exposed shore near the Hofstra llniversity Marine Laboratory. Priory, St. Ann. .Ja- maica. Squared letters are high energy transects: circled letters are low energy transects. Desiccation potential on eacln transect line was measured as the amount of water evaporated from wet sponges after one hour on the shore. Two pre-weighed. 15x15 cm sponges, were placed on each transect and wetted with 20 g of water. One sponge was placed at the uppermost end of the transect; the second was placed at the point corresponding to the mid-point of the splash zone. Where necessary, sponges were held in place by monofilament line tied to the rock. After one hour, sponges were placed in labeled re-sealable plastic bags, stored in a cooler and returned to the lab where they were re-weighed using an electronic balance. Prior to use in the field, sponges were dried at 60°C for at least 12 hours. Because we expected wave action to have an effect on desiccation potential, the experiment was repeated on seven calm and seven rough days. Data were analyzed using a three-way ANOVA with transect type (high and low wave energy), vertical shore height (splash zone and above splash zone), and sea condition (calm and rough) as factors. The slope of the substratum was determined at 15 cm intervals along each transect using a hand-made inclinometer consisting of a t-square. carpenter's level, and pivoting protractor. The device performed adequately, and we were able to reproduce transect profiles on graph paper and estimate the \ertical height above MHW at any point along the transect line. Additionally, air and rock temperatures at the uppermost end of each transect line were made using a hand-held thermometer. Rock temperatures were measured by placing the thermometer bulb directly on the rock surface. To estimate the percent cover of the available microhabitats on our transect lines, we divided a 25 x 25 cm quadrat into a grid with 16 points using string and used the point intercept method to estimate the percent cover of three microhabitat types: ( 1 1 exposed surfaces; (2) pits, defined as depressions in the rock large enough to hold a single adult snail of approximately 1 cm length; and {?>] crevices, defined as indentations in the rock of sufficient size to hold multiple adult individuals. Quadrants were placed at every meter along each line, and the microhabitat beneath each of the 16 points was determined. All mollusks within 0.5 m on either side of each transect line were identified to species and their position along the transect line recorded. Several small (<.^ mm) individuals of striped littorinids (genus Nodilittorinu) were found on the shore, and because of Factor df SS F P Transect Type (TT) 151.6 8.73 0.003 Vertical Height (VH) 8416.7 484.7 <0.00l Sea Condition (SO 1299.1 74.81 <0.001 TTx VH 176.6 10.17 0.002 TTx SC 21.4 1.23 0.268 VH X SC llfi4.8 67.08 <0.001 TT X VH X SC 13.2 0.76 0.384 Error 231 4011.3 Total 2.3S difficulties identifying these individuals to species, we designated them as Nodilitlniina spp. For all coiled gastropods (i.e. littorinids. neritids. etc.). we also recorded the type of microhabitat which they occupied. For each transect line, we determined the position of the splash zone. Using the graphical transect profiles, we con- verted the positions on the transect lines to vertical heights above MHW. Transects were surveyed over five consecutive days in January 1998 with one randomly selected high and low energy transect surveyed each day. We tested for differences in species densities, vertical position on the shore, and the vertical breadth of a species' distribution (i.e. range) using MANOVA. Because of limited replication, we re- duced the number of dependent variables in the analysis and ex- amined only the five species that occuiTcd on the greatest number of transect lines. Significant results were followed up with multiple univariate ANOVA. Cominunity structure was compared between transects using correspondence analysis on the species count data, as recom- mended by Jackson (1997). Only species comprising at least 1% of the total assemblage were included in the analysis. Data for microhabitat use by transect type was obtained by pooling individuals from the five high and five low energy transects and computing the percent of snails occupying each of the three microhabitats. We examined snail microhabitat use data for differences between high and low energy sites for each of the three microhabitats using r-tests. B 14i Q) D5 C TO .C O ^ 0 0) -8 14i -8 + High Low High Low Efiergy Energy Energy Energy (a) (b) Figure 2. Mean (±SE) change in weight of sponges on high and low energy transects. Sponges were placed at two vertical heights on the shore: la) in the splash zone: (h) above the splash zone. Solid bars represent replicates conducted on rough days (n = 7). Open bars rep- resent replicates conducted on calm days (n = 7). Abiotic Factors in Assemblage Structure 479 RESULTS High energy transects experienced a significantly greater num- ber of wave strikes (47.9 ± 21 .4) compared to low energy transect lines (11.5 ± 8.59; one-tail, two-sample f-test. t = 7.06, p < 0.0001. d.f. = 24). We observed considerable variation in the number of wave strikes from day to day. but regardless of the ocean conditions on any given day. the five high energy transects always experienced greater wave exposure than the low energy transects. High energy transects also experienced a significantly greater number of severe wave strikes (2.35 + 3.88) compared to low energy transects (0.5 ± 0.827; one-tail, two sample /-test. / = 2.08, p = 0.025. d.f. = 20). Day to day variation in the number of severe wave strikes was high; on one day. we recorded no severe wave strikes at any of our transects lines, but on another day. we recorded as manv as 15 severe wave strikes in a five minute interval at a high energy transect compared to a maximum of one strike at our low energy transects. Desiccation potential depended upon transect type (high energy or low energy), shore height (in or above the splash zone), and sea conditions (rough or calm) (Table 1). Sponges placed above the splash zone lost on average 5.92 ± 1.62 g of water, regardless of transect type or surf conditions. Change in weight for sponges placed in the splash zone varied with surf conditions and transect tvpe. On rough days, sponges in the splash zone gained an average of 10.49 ± 5.5 g of water regardless of transect type (Fig. 2). On calm days, no significant change in sponge weight was observed ( 1 .45 ± 6.27 g), but sponges at high energy transects often gained water while sponges at low energy transects usually lost water. This result, however, was not manifested in a significant three-way interaction in the ANOVA (Table 1 ). Nineteen species of mollusk were recorded during this study (Table 2). All were gastropods, except for two species of polypla- TABLE 2. Mean height on the shore (±sd) and density (±sd) for all mollusl* species on a Jamaican shore. Species Transects occupied Mean height on shore (cm) Mean density (lndiv./m") Total individuals NoililiUDriiHi tlihitiiki (d'Orbigny, 1842) Ccih'liiius iiiiini'tini.s (Linnaeus l7.'iS) NodiliHiinnti ((/h,'/(a//(ic (Morch, 1876) Nddiliunniui riisfi (Miircli, 1876) Noclilittorina -ic-ac (Gmelin. 1791) Ncnia iccvK !)/<)/ (Gmelin. 1741) NiHlililloniui spp. Acuiilluiplciini f^niiiiiltiui iGmelin. 1791 1 Nokiaciiuu'ci imlilliinim (Sowerhy. 1831) Techiriiis uiitonii (Philippi. 1846) Neriui icsm-IUiIh (Gniclln. 1791) Chiton iiihcrciilciiiLs (L,innaeus, 1759) Plaiui.\i.s lim-alu.s (da Costa. 1778) Nodililtonna nicspilluin (Miihlfield. 1842) Neriiu pcloniiua (Linnaeus. 1758) Fis.sutcllti iioildsd (Boin. 1778) Fis.stiiclUi haihadi'iisi'. (Gmelin. 1791) Diiiildiii ccixciicn.sis (Lamarck. 1822) Tliais ruslka (Lamarck. 1822) Thais sp. high 5 low 5 hi>jh 5 low 5 high 4 low 4 hJEh 3 low 3 high 5 kiw 4 high 5 low 3 high 3 low 2 high 5 low 4 high 4 low 5 high 4 low 4 high 4 low 2 high 3 low 3 high 3 low 3 high 1 low 1 high 1 low 2 high 2 low 0 high 0 low 1 high 1 low 0 high 1 low 0 high 1 low 0 69.2(11.6) 73.7(9.6) 112.5(19.5) 118.8(8.7.5) 72.9(20.1) 81.0(18.56 39.3(9.3) 45.7 (8.7) 55.9(25.1) 68.6(20.1) 50.5 (20.8) 65.8(16.1) 44.1 (24.7) 36.2 (254) 15.20(4.7) 10.55(9.1) 17.1 (10.6) 14.3(11.4) 85.9(15.8) 82.0(20.1) 36.2 (22.0) 40.7(41.5) 14.3(11.2) 20.9(18.9) 40.2(0.7) 48.5 (94) 32 76 88 87.5 (24.7) 12.3 (2.9) 74.5 14 0 8.5 0.311 (0.274) 495 0.524 (0.758) 641 0.117(0.033) 207 0.155(0.058) 230 0.176(0.159) 222 0.157(0.256) 151 0.241 (0.297) 229 0.135(0.150) 113 0.046(0.029) 79 0.047(0.031) 71 0.034 (0.024) 64 0.063 (0.037) 61 0.077(0.113) 70 0.075 (0.078) 37 0.051 (0.039) 83 0.016(0,012) 13 0.033 (0.022) 42 0.033(0.019) 42 0.017(0.008) 23 0.050 (0.074) 48 0.024 (0.029) 40 0.017(0.019) 7 0.006(0.002) 8 0.018(0.012) 20 0.013(0.018) 12 0.006 (0.004) 6 0.031 9 0.009 5 0.003 1 0.009 (0.008) 5 0.002 (0.000) 2 0.002 1 ().()()3 1 ().()()4 1 0.002 1 480 MiNTON AND GOCHFELD cophorans. The number of species did not differ significantly be- tween higli energy (1 1.40 ± 1.67) and low energy (9.80 ± 3.63) transects (two sample /-test, t = 0.89, p = 0.41, d.f. = 5). NnJil- ittohna dilatata (d'Orbigny, 1842) and Cenchritis muricatus (Lin- naeus, 1758) were the most abundant snails at both high and low energy transects, forming >30'7f of each assemblage. The majority of species present formed less than 5% of the assemblage. Few differences were found in the species composition be- tween high and low energy transects. Thirteen species occurred on >50'7r of the transect lines. Three species (Nodilittorina ziczcic (Gmelin. 1791). Acanlhoplewa grunulata (Gmelin, 1791), and Notoaanaea aiitillarmn (Sowerby, 1831 )) occurred on 909c of the transect lines and two species, Nodilittorina dilatata and C. muri- catus, were present on every transect line. Four species [Thais rustica (Lamarck, 1822), Tluds sp., Diodora cayenensis (Lamarck, 1822), and Fissurella barhadensis (Gmelin, 1791)) were each rep- resented by a single individual and another, F. nodosa (Bom, 1778) was represented by two individuals. With the exception of F. barhadensis, these rare species were found on high energy transects. The first two components from the correspondence analysis explained 70.8% of the variation (Table 3). A scatter plot of the components (Fig. 3) showed six of the ten transects clustering near the origin. Transects D. E, and F fell outside the cluster as a result of a low component 1 value. Cencliritis nu{ricalus contributed most heavily to component 1 (Table 3) and formed a large pro- portion of the community on these transects (68.8, 77.1 and 41.7%, respectively). Transect 1 is also distinguishable from the cluster due to a low component 2 value. Component 2 receives a large contribution from Nodilittorina riisci (Morch. 1876), and to a slightly lesser extent from Nodilittorina spp. and N. dilatata (Table 3). The assemblage at Transect I is differentiated from other transects by its relatively low percentage oi N. dilatata (32.5%) and its high percentage of M riisei (25.1%). Total mollusk densities did not differ significantly between high energy and low energy transects (Mann-Whitney test; W = 21; p = 1.000). Densities of the five most common species {Acan- thopleura granulata, Cenchritis muricatus, Nodilittorina dilatata. N. ziczac, and Notoacmaea antillaruni] also were not significantly different (MANOVA, A = 0.353, p = 0.367, d.f. = 5,4). Un- TABLE 3. The coordinate (Coor) and contribution (Cont) values for the first two components of the correspondence analysis. Component 1 Component 2 Species Coor Cont Coor Cont Nodititliiniia diliitata 0.228 0.074 0.198 0. 1 59 Cenchritis ntiiricatus -1.083 0.636 -0.114 0.020 N. angustior 0.297 0.041 0.130 0.023 N. riisei 0.480 0.098 -0.546 0.366 N. ziczac -0.259 0.013 0.337 0.061 Nerita versicolor -0.366 0.021 0.214 0.021 Nodilinoriua spp. 0.633 0.053 -0.583 0.130 AcanthopUiira granulata -0.064 0.000 -0.403 0.056 Notoacmaea anlillarum -0.741 0.057 -0.430 0.056 Tecrarius anicmii 0.198 0.003 0.564 0.081 Nerita lessellata 0.270 0.003 -0.453 0.027 Proportion explained .5 259 .1822 Cumulative .5 259 .7081 C^J C 707f of available microhabitats at all levels of the shore (Fig. 5). The availability of microhabitats did not vary between high and low energy transects. Use of exposed surfaces by gastropods was significantly higher Abiotic Factors in Assemblage Structure 481 160 - 140 i 120f ^ 100 CD I 80 CD "i> 60f o X 40 20 - 0 s D T A.g. N.a C.t. N.t. Nt. P.I. N.v, N.z. N.d. N.an. T.a. Cm. Figure 4. Distribution of the 12 most ubundaiit specits of supralittoral mollusks on the ixpostd shore near the Hofstra University Marine Laboratory, Priory, St. ,\nn, Jamaica. Dashed line represents the approximate upper limit of the splash zone. Bar in the upper left corner is equivalent to 10 indi\iduals per m". A.g. = Acaiithopleiiru granulala; N.a. = Nutoacmaea antillariim; C.t. = Chiloii tiiberculatiis: N.t. = Nerita tessellata; N.r. = Sudilittorina riisei: P.I. = Plaiiaxis liinaliis; N.v. = Xerila versicolor; N.z. = Nodilittorina ziczac; N.d. = Nodilittorina dilatala; N.an. = Nodilitlorina angustior; T.a. = Tectarius antonii; Cm. = Cenchritis muricatus. on high energy transects than tin low energy transects (Mest, / = 2.73, d = %.p = 0.022). No significant differences were observed for pit or crevice use between high and low energy transects. DISCUSSION Wave action and desiccation stress have been identified as important abiotic stresses structuring intertidal communities (Gar- rity 1984, Menge et al. 1986), particularly for rocky shore gastro- TABLE 4. Percent of snails occupying pits, crevices, and exposed surfaces. Species Pits Crevices (%) (%) Exposed surfaces (%) Cenchritis mtiricattis his;h 44.1 39.5 low 44.7 49.4 Nodilitlorina riisei high 80.4 12.9 low 70.6 20.2 Nodilittorinti uni^iisticr high 76.5 15.0 low 72.7 24.0 NodilittoriiHi :ic-ac high 60.? 25.9 low 69.0 21.1 Nodlltttnrinu ditattttu high 73.0 20.3 low 155 21.9 Tectarius antcitii high 51.2 21.7 low 19.2 12.5 Nodilittorina spp.t hiuh 78.9 14 low SS.2 2.9 Nerita tessellata high 62.2 2.7 low 57.1 14.3 Nerita vcrsicitlor high 27.0 54.0 low 60.9 31.3 Plaitcviis lineatiis high 66.7 25.0 low 100.0 0 11.4 6.0 6.7 9.2 8.5 3.3 13.6 9.9 6.7 2.6 2.6 8.3 19.7 8.8 35.1 28.6 19.0 7.8 8.3 0 100 0 100 CD c T r-i-n a i r^ r^ ^ J3 CO 0 o 100 o E c: Q) () CD Q. 0 100 b 1 , 1 ^ — i— 1 , 1-^— ] c 1 — i — , a t Includes all Nodilittorina individuals <2 mm Pits Crevices Surfaces Figure 5. Percent microhabitat availability and use by supralittoral mollusks at four heights above .MHW. Data are means (±SE) for 10 transect lines. Solid bars represent the relative abundance of pits, crevices, and exposed surfaces available to snails and open bars rep- resent the percent of snails occupying pits, crevices, and exposed sur- faces, (a) 4 m above .MHW; (b) 3 m above MHW: (c) 2 ni above MHW; (d) 1 m above MHW. 482 MlNTON AND GOCHFELD pods occupying the supralittoral fringe. On exposed shorelines, the splash zone directly above MHW is subject to intense mechanical stress from wave action. The ability to adhere to the substrate when under severe impact from waves should be a major factor control- ling the distribution of species (Rafaelli and Hughes 1978), as individuals knocked from the substrate are likely to be swept away, crushed, or fall victim to predation (Levings and Garrity 1983. Garrily and Levings 1984). Many supralittoral gastropod species remain eniersed. which may reduce predation from subtidal predators (Levings and Garrity 1983. Garrity and Levings 1984). Emersion also increases their risk of desiccation and heat coma from exposure to high tempera- tures. Temperatures on exposed rocky shores in the tropics are often sufficient to cause mortality in supralittoral gastropods (Gar- rity 1984. Garrity and Levings 1984). Laboratory studies on the heat coma temperatures (HCT) of five species of Caribbean lit- torinids from Jamaica, indicate that daily temperatures at the study site can exceed the lethal limits for these species (Britton 1992). During this study, ambient rock temperatures of 45°C exceeded the lethal levels for two of the five litlorinid species. Our results suggest that neither wave action nor desiccation are important factors structuring the moUuscan assemblage at this lo- cation. Species numbers, densities, assemblage composition, and the range of distribution of individual species were constant across the physical factors investigated. Correspondence analysis re- vealed that four transects were different from the others (Fig. 2). but three of these transects (D, E, and F) were related more by geographical proximity to one another than by transect type, and this suggests that relatively small scale, localized events are acting on the assemblage. Other studies have correlated habitat heterogeneity with the distribution of rocky shore mollusks (Peckol et al. 1989). and we observed a significant preference among the supralittoral gastro- pods for the less abundant pit and crevice microhabitats over ex- posed surfaces. Because our transects did not vary in microhabitat availability, the absence of differences in the molluscan assem- blage between high and low energy transects could be explained if microhabitat availability is the primary factor structuring the com- munity. Unfortunately, microhabitat availability alone is not an adequate mechanism: microhabitats should act indirectly on com- munity structure by altering mortality levels caused by physical or biological mechanisms. Pit and crevice microhabitats may be cooler and more humid than exposed surfaces, resulting in lower desiccation stress and ambient temperature (Levings and Canity 1983, Garrity 1984). These microhabitats may also provide protection from mechanical stress (Raffaelli and Hughes 1978). reducing the likelihood of individuals being swept from the substrate when struck by waves. Patterns of microhabitat use between high energy and low en- ergy transects suggest that pits and crevices may be important in alleviating desiccation stress. On high energy transects, nearly all species were found more frequently on exposed suifaces (Table 3). If pits and crevices were important in reducing the risk of being swept from the shore by wave action, we would expect fewer snails on exposed surfaces at high-energy sites. As this was not the case, we believe wave action to have an insignificant direct effect on the distributions of these species. Indirectly, high wave action may lower desiccation potential and rock temperature compared to low energy sites and therefore, individuals on exposed surfaces would be less likely to suffer mortality from desiccation along high energy transects. In contrast to our findings. Lang et al. ( 1998) found little pref- erence for pits and crevices over exposed surfaces for Nodilittorina angustior. N. riisei. and N. antonii in research conducted on nearly the same area of the Jamaican shoreline. However, rock tempera- tures during this earlier study were approximately b'C lower than during our study and were well below the measured HCT for these species: this may have allowed the species to distribute themselves in a more random manner over the rock surface. McMahon (1490) has suggested that mollusks occupying the supralittoral fringe have sufficient physiological tolerances to high temperature and desiccation that these factors play little role in determining zonation patterns. Britton (1992) found only a weak relationship between desiccation resistance and HCT and the rela- tive positions on the shore for seven species of Caribbean littiirin- ids. The littorinid species examined possessed high desiccation resistance and were able to survive from 12 to 30 days of emersion (Britton 1992). Interestingly, the lowest HCT was for C. muricatus (42.I°C). the littorinid occupying the highest position on the shore. Unfortunately, methods used to determine HCT temperatures entail submerging the animal in a heated water bath. In nature, these species are seldom, if ever, submerged. This casts doubt on the ecological relevance of this type of study (Garrity 1984). Mor- tality of supralittoral gastropods in the tropics has been linked to elevated temperatures (Levings and Garrity 1983). and even if desiccation and temperature are not directly responsible for mor- tality, they may indirectly lead to mortality via some other factor (Menge and Sutherland 1987). Therefore, any means of amelio- rating temperature stress would be advantageous to the individual. We do agree, however, with McMahon's ( 1990) contention that the distributional patterns observed for tropical rocky shore mol- lusks may be the result of a suite of factors, including the physical, biological, and physiological environment. Because pits and crev- ices account for <30% of the available habitat, competition for these microhabitats may be an important factor controlling species distributions and ultimately, community structure. The interaction of biological (i.e. competition and predation) and physical (i.e. desiccation and temperature stress) factors may be responsible for the distribution of supralittoral gastropods on exposed tropical shorelines. ACKNOWLEDGMENTS We thank Joseph Britton and Robert McMahon for useful dis- cussion that improved this research. Eugene Kaplan. Sandra Walters. Sandra Ross and the staff of the Hofstra University Ma- rine Laboratory provided facilities and encouragement, and Dirk assisted in the field. Joseph Britton and an anonymous reviewer made helpful comments on the manuscript. This research was supported by a grant to D.M. from the Conchologists of America. LITERATURE CITED BraUstrcim. H. lySO. Rocky shore zonation in the Sant;i Marta area. Co- lumbia. Sarsia 65:163-226. Britton. J. C. 1992. Evaporative water loss, behavior during emersion, and upper thermal tolerance limits in seven species of eulittoral-liinge Lit- torinidae (Mollusca: Gastropoda) from Jamaica. In: J. Grahame. P.J. Mill cS: D.G. Reid. editors. Proceedings of the Third International Sym- posium on Littorinid Biology. London: The Malacological Society of London, pp. 69-83. Abiotic Factors in Assemblage Structure 483 Garrity. S. D. 1984. Some adaptations of gastropods to physical stress on a tropical rocky shore. Ecology 65:559-574. Garrity, S. D. & S. C. Levings. 1984. Aggregation in a tropical neritid, Veliaer 27:\-6. Gendron, R. P. 1977. Habitat selection and migratory behaviour of the intertidal gastropod Lilloriiw lirrorea (L.). J. Aiiim. Ecol. 46:79-92. Gochfeld. D. J. & D. Minton. 2001. When to move and where to go: movement behavior of the Caribbean littorinid Ccnchiiiis miiruiiius (Linnaeus. 1758). / Shellfish Res. 20:447^52. Jackson, D. A. 1997. Compositional data in community ecology: the para- digm or peril of proportions? Ecology 78:929-940. Lang, R. C, J. C. Britton & T. Metz. 1998. What to do when there is nothing to do: the ecology of Jamaican intertidal Littoriniidae (Gas- tropoda: Prosobranchia) in repose. Hydrohiologia 378:161-185. Levings, S. C. & S. D. Garrity. 1983. Diel and tidal movement of two co-occurring neritid snails: differences in grazing patterns on a tropical rocky shore. / £v;j. Mar. Biol. Ecol 67:261-278. Lubchenco, J., B. A. Menge. S. D. Garrity. P. Lubchenco, L. R. Ashenkas, S. D. Gaines, R. Emiet, J. Lucas & S. Strauss. 1984. Structure, persis- tence, and role of consumers in a tropical rocky intertidal community (Taboguilla Island, Bay of Panama). J. Exp. Mar. Biol. Ecol. 78:23-73. McMahon, R. F. 1990. Thermal tolerance, evaporative water loss, air-water oxygen consumption and zonation of intertidal prosobranchs: a new synthesis. Hydrohiologia 193:241-260. Menge. B.A.. J. Lubchenco, S.D. Gaines & L.R. Ashkenas 1986. A test of the Menge-Sutherland model of community organization in a tropical rocky intertidal food web. Oecologia iBerliii) 71:75-89. Menge. B. A. & J. P. Sutherland. 1987. Community regulation; variation in disturbance, competition, and predation in relation to environmental stress and recruitment. Am. Nul. 130:730-757. Moore, H. B, 1972. Aspects of stress in the tropical marine environment. Adv. Mar Biol. 10:217-269. Peckol, P.. S. Guarnagia & M. Fisher. 1989. Zonation and behavioral patterns in the intertidal gastropods Nodilillorina (Tecuninus) antoni (Philippi. 1846) and Nerita versicolor Gmelin, 1791. in the Bahamas. Veliger 32:8-\5. Raffaelli, D. G. & R. N. Hughes. 1978. Effects of crevice size and avail- ability on populations of Lillorina riulis and Litlorina neriliiides. J. Anim. Ecol. 47:71-83. Underwood, A. J. 1979. The ecology of intertidal gastropods. Adv. Mar. Biol. 16:111-210. Vermeij, G. J. 1971. Temperature relationships of some tropical Pacific intertidal gastropods. Mar. Biol. 10:308-314. Vermeij. G. J. 1972. Intraspecific shore-level gradients in intertidal mol- lusks. Ecology 53:693-7fKJ. Williams. G. A. 1995. Maintenance of zonation patterns in two species of flat periwinkle. Liltorinu ohiiisaui and /.. iiniruie. Hydrohiologia 309: 143-150. Williams. G. A. & D. Mornlt. 1995. Habitat partitioning and thermal tolerance in a tropical limpet. CelUma grata. Mar. Ecol. Prog. Ser. 124:89-103. Jimnuil iifSlwllfish Research. Vol. 20. No. I. 4S5-;SS. 2001. EFFECTS OF TEMPERATURE AND DESICCATION ON TISSUE URIC ACID DYNAMICS IN LITTORINA SAXATILIS (OLIVI) DELMONT C. SMITH DepartweiU of Biological Sciences. State University of New York. College at Brockport. Brockport. New York 14420 ABSTRACT Previous studies have shown (Smith & Smith 1998) that the tissue concentration of uric acid in Littoriim sa.xaiilis varies considerably on a seasonal basis. In order to investigate possible physical parameters that might account for this variation, periwinkles were kept in controlled temperature rooms at 25°C or IO°C, and either immersed in seawater or dry (emersed). The combination of 25°C and emersion caused the highest concentration of uric acid, and lO'C and immersion the lowest. Returning emersed snails to seawater caused the tissue uric acid concentration to decline. This re-immersion also produced an increase in ammonia excretion. It appears that uric acid accumulates during emersion, which would mitigate water loss, and that this uric acid is re-converted to aminonia for excretion during subsequent immersion. The time course for these changes is such as to accommodate the natural cycle of immersion/emersion. A7;>' WORDS: Liiroriiia .•^axatilis. uric acid, ainmonia. temperature, desiccation INTRODUCTION Research on the influence of environmental factors on uric acid concentration in inteilidal gastropods began with the classic papers by Needham (1935, 1938) with substantial contributions later by Potts (1967). These workers demonstrated that species located high in the intertidal zone contain a relatively greater amount of uric acid than those from lower. Previous reports indicate (Smith et al.l995) that shore location results in differences in uric acid levels within species {Liltoriiia saxatilis and L incciini). with crev- ice-dwelling animals having a lower uric acid concentration than animals collected from the surface of boulders. The principal end product of nitrogen metabolism has long been recognized to be related to the amount of water that is avail- able for its excretion. Aquatic animals are able to excrete toxic ammonia since water is plentiful and can readily flush ammonia from the body. Terrestrial species of gastropods (and other ani- mals), however, are ureotelic or uricotelic (Delauney. 1931, Needham, 1935). since urea and uric acid are relatively non-toxic and can therefore be permitted to accumulate and ultimately be excreted with a minimal volume of water. The disadvantage of urea and uric acid is that they are metabolically more costly to synthesize than ammonia. Uric acid is the most costly, but is also the most conservative of water. Since it is highly insoluble, it may be stored in tissues without detrimental osmotic effect, and may. in those animals that excrete uric acid, be excreted in crystalline form with very little water loss. Whether in fact periwinkles actually excrete uric acid as such is questionable; Heil and Eichelberg (1983) reported that the only excreted metabolite they were able to detect in L liuorea was ammonia. Littoral gastropods may be thought of as being intermediate between aquatic and terrestrial snails since they are subjected to alternating periods of immersion with plentiful seawater, and of emersion with the possibility of dehydration. Under such condi- tions it is to be expected that seasonal changes would have a significant effect on the accumulation of uric acid since the pre- sumed factors that stimulate uric acid accumulation, namely in- creased temperature and desiccation, will be increased in summer and reduced in winter. This has been shown in land snails (Jeze- wska et al. 1963; Laziradou-Dimitriadou & Kaloyianni. 1989), and. in an earlier paper (Smith & Smith 1998) we were able to show that there is indeed a much greater concentration of uric acid in winkles in summer than in winter. Although it seems reasonable to assume that the physical parameters that lead to accumulation of uric acid in the summer are temperature and desiccation, it is necessary to test the effects of these factors under controlled labo- ratory conditions. For this reason, winkles were brought to the laboratory to be maintained in controlled conditions of temperature and desiccation in order to independently test these factors on the animal's uric acid content. Further, the ultimate fate of this uric acid, conversion to ammonia, was investisated. MATERIALS AND METHODS Collections E-mail: dsmith@brockport.edu Samples of Littorina saxatilis were collected from a boulder field at Filey Brigg. North Yorkshire, on the northeast coast of England (British National Grid Reference TA 133814; this loca- tion is labeled "B" in Smith et al. 1995). All animals used in this study were collected between February 10 and April 19.1999. Treatments The animals were transported to the laboratory where they were maintained in temperature controlled rooms either at 10° (desig- nated here as "C") or 25° (designated as "H"). In each of these rooms animals were separated into either a dry aquarium ("'D""l or an aquarium where they were kept submerged in seawater ("W") by means of a nylon net attached to the walls of the aquaria at the surface of the water. With each collection of winkles, a number of algae-covered stones were taken from the same shore and were placed on the bottom of the aquaria (Warwick 1983). This made it possible to keep the winkles in good condition well beyond the duration of the experiment. The winkles were maintained in these conditions for at least 48 hours before being assayed for uric acid. In order to determine whether and how much animals that had been adapted to dry conditions would reduce their uric acid con- 485 486 Smith centration upon re-immersion, a sample of 45 C/D animals was transferred to CAV conditions. An equal sized sample of H/D was transferred to H/W and forty-eight hours after transfer, each sample was assayed for uric acid. Most determinations were done after 48 hours in the chosen temperature/moisture conditions. To determine whether the rale of uric acid accumulation during desiccation would occur within a time frame that corresponds to typical emersion cycles in nature, winkles were placed in H/D and then uric acid concentration in samples of 15 animals immediately measured, at 3 hours, 6 hours. 1 8 hours, 28 hours, and 48 hours. Uric Acid Delermination This method is the same as that reported in earlier papers (Smith el al 1995, Smith & Smith 1998). The shells were cracked and the body was removed under a dissecting microscope. The body was weighed and then homogenized in 100 |jl1 of phosphate buffer (pH 7.8) at room temperature. After centrifugation the su- pernatant was removed and frozen for later analysis for uric acid. The supernatant was assayed using a modification of the Sigma Diagnostics procedure no. 685. This procedure essentially follows the method of Duerr (1967) and utilizes uricase to promote the oxidation of uric acid to allantoin, CO, and H^O,. The H,0, then reacts, in the presence of pero.xidase, with 4-aminoantipyrine and .3,5-dichloro-2-hydroxybenzenesulphonate to form a quinoneimine dye. The intensity of the color formed is proportional to the con- centration of uric acid and is read at 520 nm. Concentration is reported as p.g of uric acid per gram of wet tissue weight (ixg g"' ). Body Weight The wet weight of four samples of 20 winkles was obtained by weighing each animal individually after blotting external water with filter paper. Each group of snails was then placed in one of the four-temperature/ moisture conditions for 5 days, at which time the animals were individually weighed once more. This was to deter- mine possible changes in body weight, especially in those animals that had been subjected to desiccating conditions. Ammonia Production Fony-five animals were subjected to H/D for 48 hours: another 45 animals were placed in H/W for the same period of time. They were then placed, in groups of 15 animals, in 30 ml of seawater at 25''C. Their rate of ammonia production was determined by mea- suring the concentration of ammonia in the seawater over a 4 h period using an Omega ammonia colorimeter. The rate of produc- tion was expressed as jjlM NH, g~' h"'. Statistics Non-parametric methods were used for all statistics reported in this paper: two-way analysis of variance using the method of Tukey, comparison of individual means by the Mann-Whitney rank sum test, and paired tests by the Wilcoxan method. RESULTS Temperature and Desiccation Effects The effects of the four combinations of temperature and mois- ture treatments: 10°C and immersed (C/W), 25°C and immersed (H/W), 10°C and dry (C/D), 25°C and dry (H/D) are illustrated in Figure 1 . The bar labeled "pre"" is the mean uric acid concentration 250 » 200 150 50 _I1 X X Pre CAV C/D HAV H/D Figure 1. Mean (± SEM) wel tissue uric acid concentration ((ij; g"') in Littorina saxatilis maintained at the indicated conditions of tempera- ture and moisture for 48 h. Pre = control fresh from the shore (N = 42); CAV = 10 C immersed (N = 39); C/D = 10 C emersed (N = 45); H/W = 25 C immersed (N = 38); H/D = 25'C emersed (N = 36). of 42 periwinkles measured immediately upon arrival at the labo- ratory. The fact that this is higher than the next bar, which is the mean of 39 animals 48 hours after immersion in seawater at 10"C, is probably because the animals had already been emersed for some hours by the time they arrived at the laboratory, since the collections were taken at about the time of low water. The C/W w inkles showed the lowest concentration of uric acid, and the H/D animals the highest. Table I shows the results of statistical testing of the data by the Tukey non-parametric method. Clearly, both temperature and des- iccation have highly significant effects on uric acid concentration, but there is no interaction between the two factors. Testing be- tween individual means (Mann-Whitney) reveals highly signifi- cant differences (P < 0.001 ) when comparing C/W vs C/D, H/W vs C/W. and C/W vs H/D: significant differences {P = 0.022) with H/D vs H/W and H/D vs C/D: and no significant difference be- tween H/W and C/D. Effect of Re-liydration It was clear that desiccation leads to an increase in uric acid concentration. What if the animals are now re-immersed in sea- water'.' Samples of winkles (;; = 45) were kept in H/D and C/D conditions for 48 hours to raise their uric acid concentration. After 48 hours half of the surviving animals in each sample were assayed for uric acid and the remaining animals in each sample were im- mersed in seawater at the same temperature and assayed for uric acid 48 hours later. The results are shown in Figure 2. ANOVA of the data of Figure 2 is highly significant (P < 0.001). Testing of T.\BLE 1. Anova table of the effects of temperature and desiccation on uric acid concentration in Littorina saxatilis. Source of variation MS F P Temperature Desiccation Interaction 284284 177169 20116 46.2 28.8 3.3 <0.00l* <0.001* n.s. Desiccation Effect on Uric Acid in Winkles 487 Dn to Wei Figure 2. Klftcl of transftr of Littorina saxatilis after 48 h desiccation ("dry", eitlier Harm |25 C] or cold |1() C|l to 48 h immersion in sea- water ("dry to wet", either warm or coldt on mean wet tissue uric acid concentration (jig g''). N = 19 (C7D), 24 (t/I) to CAV), 18 |H/I)|. 15 (H/1) to HA\ 1. individual sainplL- responses to re-hydralion shows ii highly sig- nificant {P < 0.00 1 ) reduction in uric acid when winkles are moved from H/D to HAV, but the response to a move from C/D to C/W is considerably less iP = 0.17); not statistically significant at a = 0.05. Weight Loss during Desiccation Upon desiccation, animals will almost certainly have some re- duction in weight as a result of water loss. To determine whether such loss of weight is sufficient to account for the concomitant increase in uric acid concentration, four groups of 45 snails were initially weighed, then each group put into one of the four tem- perature/moisture conditions for 5 days. Table 2 shows that the animals that had been kept submerged had a very slight (not sig- nificant) gain in body weight. Desiccated animals had a significant (P < 0.001) weight loss. However, if the % change in weight is compared with the 9r change in uric acid as a result of desiccation (data from Fig. I ) it is clear that the loss of body weight cannot account for the large increase in uric acid concentration. Animals kept C/D had a 6.2Vf weight loss, but a 43.0% increase in uric acid concentration; those animals maintained H/D had a 6.0'/f weight loss but a 79.8% gain in uric acid. Kale of Increase in L'rie Acid during Desiccation The determinations described in the preceding sections were done following 48 h of exposure to the chosen temperature/ moisture conditions. Winkles were placed in H/D and then samples of 15 snails were immediately assayed, at 3 h, 6 h, 18 h, 28 h, and 48 h. This was done to determine whether the uric acid accumu- lation that is due to desiccation would occur within the time con- straints of a typical immersion/emersion cycle in nature. The re- sults are illustrated in Table 3. and show that there is indeed a significant increase in uric acid by 6 h, within the emersion time of a normal tide cycle for these animals. Ammonia Production Forty-five periwinkles were conditioned to H/D for 48 h to increase their uric acid concentration. The control was an equally sized sample held at the same temperature but kept immersed. After 48 h, groups of 15 animals were placed in 30-ml seawater and their ammonia production was measured over a 4-h period. The 45 animals that had been H/D had an ammonia production rate of 0.41 p,M NH, g"' h"'. while the H/W controls produced at a rate of 0.25 (xM NH, g^' h"'. These rates are consistent with those reported by Uglow and Williams (2001 ) for Hong Kong littorines. Statistical testing between these means was not done, since, al- though 45 animals were in each sample, they were combined into groups of 15, and 3 experimental and 3 control samples are insuf- ficient for statistical testing. DI.SCUSSION Shore location and seasonal effects suggest that the determin- ing physical causes of increased uric acid concentration are prob- ably temperature and desiccation. This hypothesis is borne out by the experiments reported here using temperature-controlled rooms to carefully control the temperature of periwinkles that were either completely submerged in seawater or kept in dry conditions. Both increased temperature and desiccation cau.sed significant gains in uric acid burden of L. sciMililis. While no interaction between these two physical factors was demonstrated, one would expect that the effect of temperature would be manifested by an enhancement of the drying effect. However, the actual loss of body weight due to desiccation was in fact rather slight, and was no greater at 25°C than it was at 10°C. Possibly, then, the temperature effect is to metabolically increase the rate of synthesis of uric acid. Heil and Eichelberg (1983), and Eichelberg and Heil (1989) have suggested that uric acid serves as a temporary storage form of nitrogenous wastes in Ultoriiia lilliireci during times of water lack and low salinity. They were unable to detect any excreted form of nitrogen other than ammonia. If the uric acid were a temporary T.VBLK 2. The effects on whole body weight (g) of 5 days of treatment under various temperature/moisture conditions. Initial weight Final weight A %A Treatment N (mean ± SEM) (mean ± SEM) Weight Weight P C/W 20 0.407 ± 0.032 0.410 + 0.031 0.003 1.1 0.381 C/D 20 0.474 ± 0.040 0.45 1 ± 0.038 -0.28 -6.2 <0.001* H/W 20 0.387 ± 0.038 0.389 ± 0.037 0.005 1.4 0.064 H/D 20 0.528 ± 0.060 0.498 ± 0.056 -0.27 -6.0 <0.00I* C/W = I0°C, immersed; C/D = 10 C emersed; H/W = 25°C immersed; H/D = 25 C emersed. 488 Smith TABLE 3. Mean tissue uric acid concentration (pg g"' fresh body Height) at timed intervals following emersion at 25°C. Uric acid Hours A' (mean ± SEM) F P 0 14 1214 ± 14.9 — — 3 15 167.(1+ 19.8 3.30 0.080* 6 15 197.0 ± 17.9 10.34 0.003* 18 14 223.1 ±20.3 16.24 <0.001* 28 14 234.2 ±21.4 18.62 <0.001* 48 14 279.5 ± 20.8 38.05 <0.001* storage form, then one would e.xpect it to be depleted once water becomes plentiful again. I have found this to be the case; re- immersion in seawater causes significant reduction in uric acid concentration. The fact that animals that were warm as well as dry had the highest concentration of uric acid, and then had the lowest concentration after immersion in warm water, further bears out the metabolic role of temperature. Once the metabolic pathways that degrade uric acid become active, they appear to proceed faster at a higher temperature than at a lower temperature. The temporary storage of uric acid avoids the loss of water that would accompany e.xcretion of nitrogenous wastes in any form, but especially if the nitrogen is excreted in the form of ammonia. Subjecting animals to desiccation has the possible effect of causing water loss, which would cause a relative rise in concen- tration of all materials in the body. Not only did the experiments reported here rule out any such effect on uric acid concentration (since the concentration of uric acid increased so much more than the body weight decreased), but the desiccation had a remarkably small effect on body weight. Even after 5 days in dry conditions, the animals only lost about 6% of their original weight, indicating the effectiveness of the opercular seal as well as the water sparing effect of uric acid storage. Littorina sa.xatilis lives rather high in the intertidal zone. There- fore, these animals spend a great deal of time out of the water. Many individuals may he submerged only during spring tides. The time course for uric acid change during desiccation was shown to be sufficiently rapid to account for the changes in uric acid that probably occur during natural immersion/emersion cycles, espe- cially during neap tides when many individuals may not be im- mersed for days, since the range in concentration I obtained within 48 hours approximates the range reported on a seasonal basis (Smith & Smith 1998). It is interesting to consider what might serve as the signal to convert nitrogenous wastes to uric acid. Since the actual water loss was rather slight (about 6'7f after 5 days), and the increase in uric acid was significant within hours of einersion. it is unlikely that water loss can be the signal for the metabolic change. Winkles withdraw into their shell w ithin minutes of water being withdraw n. and remain so until re-imniersed. Possibly the same signal that stimulates withdrawal also serves to alter nitrogen metabolism. A possible stimulus for uric acid production could be a lowering of pH in emersed animals. When emersed individuals withdraw into the shell they become partially anaerobic in their metabolism. Anaerobic metabolism leads to a decrease in metabolism and pei- haps stimulates uric acid production. My results support the thesis of Eichelberg and Heil ( 1 989) that winkles are not truly uricotelic if uricotelic is taken to mean being an excretor of uric acid. Winkles do synthesize uric acid as a water conserving means during times of emersion, but they apparently convert this uric acid back to ammonia in order to be excreted, and this only when water is again plentiful. Such enzymatic conversion of uric acid to ammonia should be possible, since Daguzan ( 1967) delected all the enzymes required for uricolysis in L. littorea. and it seems reasonable to e.xpect that the same enzymes are present in L. sa.xatilis as well. ACKNOWLEDGMENTS I wish to thank the School of Biology. Leeds University, for the opportunity to spend a sabbatical leave there from January to May 1999. Also thanks to Drs. P.J. Mill and J. Grahame for their hos- pitality and advice. LITERATI! Daguzan. J. 1967. Recherches sur les Littorinidae. These Doct. Sci. Nat. Rennes. France. Delaunay. H. 1931. L'excretion azotee des invertebres. Bial. Rev. 6:265- 301. Duerr. F. G. 1967. The uric acid content of several species of marine prosobranch snails. Coinp. Biochem. Physiol. 22:333-340. Eichelberg, D. & K. P. Heil. 1989. Studies on the significance of the uric acid formation in the periwinkle Littorina littorea (L. 1758). Medio Amhiente 10:69-76. Heil. K. P. & D. Eichelberg. 1983. Unlersuchungen zuni Harnsaureme- tabolismus von Littorina littorea (Gastropoda). Helgohiiuler /ii/.v.';. Meeresunters. 35:465^72. Jezewska. M. M.. B. Gorzkowski & J. Heller. 1463. Seasonal changes in the excretion of nitrogen wastes in Heli.x poinatia. Aeta Bioehiin. Pol. 10:309-314. Laziradou-Dimitriadou. M. & M. Kaloyianni. 1989. .'\ge. seasonal and diel variations of the uricolytic enzymatic system in the snails Xeropieta arenosa (Ziegler) and Cenuiella virgala (DaCosta). / Conip. Physiol. B 159:115-121. Needham. J. 1935. Problems of nitrogen cataholism in invertebrates. 11. RE CITED Correlation between uricotelic metabolism and habital in the Phylum Mollusca. Bioehem. J. 29:238-251. Needham. J. 1938. Contributions of chemical physiology to the problem of reversibility in evolution. Biol. Revs. Cambridi;e Phil. Soc. 13:225- 251. Potts. W. T, 1967. E.xcretion in the molluscs. Biol. Rev. 42:1— fl. Smith. D. C. P. J. Mill & J. Grahame. 1995. Environmentally induced variation in uric acid concentration and xanthine dehydrogenase activ- ity in Littorina .sa.xatilis (Olivi) and L. arcana (Hannaford Ellis). Hy- drobiologia 309:111-116. Smith. D. C. & T. P. Smith. 1998. Seasonal vanation in soluble uric acid concentration in Littorina sa.xatilis (Olivi). Hydrohiolos;ia 378:187- 191. Uglow. R. F. & G. Williams. 2001. Variation in ammonia efflux rates with emersion of three Hong Kong Nodilittorina species. J. Shellfish Res. 20:489-493. Warwick. T. 1983. A method of maintaining and breeding memhers of the Littorina sa.xatilis (Olivi) species complex. J. Molhiscan Studies 48: 368-370. Joiinuil ofSlu'llfish Research. Vol. 20. No. I. 489-^93, 2001. THE EFFECTS OF EMERSION ON AMMONIA EFFLUX OF THREE HONG KONG NODILITTORINA SPECIES R. F. UGLOW' * AND GRAY A. WILLIAMS" ^ Department of Bloloi;leal Sciences. University of Hull. Hull HU6 7RX. United Kingdom: -Department of Ecology A Biodiversity and The Swire Institute of Marine Science. The University of Hong Kong. Hong Kong ABSTRACT On moJerateh -exposed HotiL' Kong shores, the nodihtlonnids. Nuditittonna iiocliuides and N. nuliahi occur in the hisih-splash /one and N. vidmi shghtly helow this level, extending into the eulittoral. These species experience long emersion times and high rock and air temperatures. After exposure on the natural rock for 0 min (awash and active, control animals) or 1 . 4 or 22 h emersed. groups of animals (n = 6) were re-immersed and their ammonia effluxes (weight-normali/ed for fresh tissue mass) measured after 0.5. 1.0 and 2.0 h. Re-immersed groups had initial (30 min) effluxes of 3.91, 6.01 and 3.53 (jimoles NHj x g'' x h"' for N. imchoides. N. radiaia and N. vidua respectively, which were high compared with the final values of 1.66. 2.02 and 0.32 jimoles NHj x g"' x h"' for the same species after 2 h of re-immersion. There were clear inter-specific efflux differences and the handling procedure evoked enhanced ammonia excretion rates possibly as a stress response. Re-immersed animals had lower effluxes than control groups and, generally, such rates were negatively related to the duration of the preceding emersion period. Effluxes of M vidua, measured 30 min after re-immersion were always higher than those measured at 1 or 2 h but this occurred with the other species only after they had been emersed for 1 h. Rates at all three sampling times following 4 or 22 h of emersion were very similar. The more eulittoral N. vidua may continue to produce ammonia during emersion periods of >1 h but the high-shore species do not. The energetic cost implications of this difference in post-emersion effluxes may be that emersion tolerance, hence vertical height on the shore, is limited for N. vidua. No evidence of a switch from ammonotely to uricotely was found for any species, but the production of non-excreted uric acid is not precluded. KEY WORDS: Ammonia, effluxes, emersion. NoditinnnNu spp., tropical rocky shore INTRODUCTION Three common Nodilittorina species occur high on Hong Kong rocky shores (Oghaki 1985, Reid 1992, Williams 1994, Mak 1996) and have littoral distributions that result in lengthy periods of emersion and. during neap tide cycles, individuals may be con- tinuously emersed for some days (Williams 1994). Summer day- time temperatures in Hong Kong can be high and rock surface temperatures >50'C are not exceptional; conditions which enhance the risk of desiccation (see Williams and Morritt 1995. Williams and McMahon 1998). Each of the Nodilittorina species exhibit measures which are known to aid the maintenance of a balanced heat budget, including the development of a dried mucus "anchor" moving to forage while awash and ceasing activity shortly after becoming emersed (see Garrity 1984. Britton 1992). All three species are principally ammonotelic. a mode of producing and removing nitrogenous metabolic waste that is inexpensive ener- getically but which entails a loss of water that would be intolerable for a non-aquatic species. The fact that these nodilittorinids are numerous and dominate high shore regions in the tropics is clear evidence that they have evolved means to tolerate protracted pe- riods when nitrogen metabolism continues but normal means of removing the ammonia pioduced are probably impractical because of the entailed water losses. As ammonia is toxic, it is of interest to know how these animals deal with their nitrogenous wastes at *Corresponding author. E-mail: R.F.Uglow@biosci.hull.ac.uk (R.F. Uglow); hrsbwga(a>hkucc.hku.hk (G.A. Williams) such times and a closer examination of the interspecific and in- traspecific means used may improve our understanding of small- scale distributional differences shown by these, and similar, high shore invertebrates. Several studies on the effects of emersion on bivalves have been made (e.g. Widdows et al. 1985. Schick et al. 1988. Sadok et al. 1999) which indicate considerable metabolic and behavioral variability can occur in these animals. Literature is limited on nitrogen metabolite fluxes of intertidal gastropods in general and no information on such fluxes in Nodilittorina spp. could be found. Sv\ itching biochemical pathways so that an increased relative pro- portion of other forms of metabolic waste is produced is energeti- cally an expensive option, were a species able to do this, but would effectively reduce water loss, hence desiccation stress, and thereby decrease the risks of associated toxicity or tonicity. Consequently, much of the emphasis of the study of nitrogenous waste removal in gastropods in general, and littoral species in particular, has con- centrated on the relative rates of production, storage and release of uric acid in relation to ecological preferences and water con.serva- tion (Delauny 1931. Needhani 1935. 1938). Terrestrial gastropods are principally uricotelic and it is logical to speculate that high- shore gastropods may have developed .so that the proportions of ammonia and uric acid produced may be intermediate between the terrestrial and aquatic representatives. Smith and Smith (1998) found that summer levels of soluble uric acid concentrations in Littohna saxatilis were significantly higher than winter levels, partly because of higher desiccation levels, but they suggest, mainly because of increased protein catabolism rates in summer. 489 490 Uglow and Williams Whichever nitrogenous metabolic end-products are produced, their rates of production are reflections of metabohc rate. Comparisons of nitrogen tluxes between emersed and immersed httoral animals are complicated because emersed animals eventually cease loco- motory movements resulting in general levels of slow metabolism. These preliminary studies on three NodHiriorinci spp. were made to determine the weight-specific ammonia eftlu-X rates of normal (immersed) animals and to examine whether post-emersion ammonia effluxes are influenced by the emersion duration. It is not known whether brief, but inevitable, handling procedures consti- tute a stress sufficient to alter normal effluxes or what the magni- tude and duration of these changes may be. Consequently, the ammonia effluxes of groups of each of the three species were assessed in terms of time elapsed since being handled and of preceding emersion duration. As ammonia is the product of re- spiring tissue and because a large proportion of the total wet weight of each animal comprises the non-respiring shell, the rela- tive weights of shell and flesh were determined using a represen- tative sub-sample of each species. All effluxes given here are thus corrected to a tissue mass basis. MATERIALS AND METHODS NdililiiliiriiHi tiochoides (Gray 18.'(9), N. radiata (Eydoux & Souleyet 1852) and N. vidua (Gould 1859) were collected while awash on the falling tide at Cape d'Aguilar. Hong Kong (22°N. 114°E) in July 1996. Eighteen individuals of each species were individually transferred to Eppendorf lubes containing 1.4 ml of the ambient seawater as a control group (awash) and the time was recorded. Fifty-four individuals of each species were also collected and placed on rock above the high water mark immediately in front of The Swire Institute of Marine Science. After 1. 4 and 22 h of emersion, three sets of six individuals of each species, one animal per tube, were transferred to Eppendorf tubes containing 1 .4 ml of ambient seawater and the time recorded. For each set of animals, after 30. 60 and 1 20 mm. six mdi\ idu- als of each species were removed from their tubes, damp dried, weighed to the nearest 0.1 mg and replaced on the shore at the appropriate tidal level. The labeled Eppendorf tubes with the water samples were frozen (-20°C) until required for analysis. Separately, a sample (n = 54) of freshly caught animals of the three species were damp-dried, weighed (± 0.1 nig), killed in boil- ing water and the flesh removed from the shells which were re- weighed (± 0.1 mg). A fresh tissue weight vs. total wet weight relationship was determined for each species. Water samples were analyzed for dissolved total ainmonia (NHj = NH, -t- NH4*) and uric acid (as urate). Ammonia con- centrations (as total volatile base, TVB) were quantified using a flow injection/gas diffusion technique (Hunter and Uglow 199.^) using a 50()-ml sample loop. Calculated ammonia concentrations were transformed to weight-specific effluxes (ixmoles NH4 x g flesh weight"' x h"'). Uric acid concentrations were determined as urate using Sigma diagnostic kit No 685. Data were analyzed using Two Way Analysis of Variance (ANOVA). When significant differences were detected, the Tukey multiple range test was applied to compare means. All statistical analyses were performed at the 5% significance level. RESULTS Tissue weight vs. total weight relationships for the three species show that in each case, the relationship was linear and thus allowed the estimation of flesh weight from a simple wet weight determi- nation (Table 1 ). The weight-specific ammonia efflux data of the freshly-caught, non-emersed (control) animals decreased progressively (Fig. I), particularly during the first 30 minutes, such that the 120 min values were significantly less than those measured at 30 min (P < 0.05 in each case). Presumably, such decreases illustrate a pro- gressive recovery from the brief, initial handling shock and that the 60 and 120-min values better represent the normal, emersed ef- fluxes of the species. The animals were crawling when captured and the ammonia concentration in any extra-corporeal water that they may have had would have been similar to that of the ambient water. The apparently high initial effluxes shown by N. nidiula are probably not a reflection of the general activity of this species v\ hen first caught as it was noticeable that A", vidua was much more active than the other two species. The 120-min effluxes also indi- cate that those of the most eulittoral N. vidua are significantly less iP < 0.05) than tho.se of the two higher shore species. The data were uncorrected for losses that may have occurred due to bacte- rial consumption. Presumably, some losses will have occurred in this way in the 1 and 2 h samples, but such losses were negligible in seawater blanks and it is assumed, therefore, that such losses would not alter the main findings substantially. The two highest zoned species, N. trochoides and N. vadiata had ammonia excretion rates following 1 h of emersion which were either similar or slightly less (P > 0.05) than those of their control groups (Fig. 2a-b). Longer periods of emersion evoked re-immersion effluxes that were significantly lower than those of the control groups iP < 0.05 in all cases). Quite a different pattern of post-emersion ammonia eftlux rate was found to occur in the species with the lowest vertical distri- bution. N. vidua (Fig. 2c). The control value for this species was significantly (P < 0.05) less than those of the other two species and, after each of the three emersion periods, the reimmersed efflux rates were significantly higher than the control group in all instances except for the 120 min post 4 h emersion group. The I and 4 h emersed groups showed a pattern of progressively dimin- ishing post-emersion efflux rates but the 22 h emersed group main- tained a reasonably constant, high efflux (c.f. the control group). None of the animals excreted measurable amounts of urate during these experiments. The 120-min eftlux data were analyzed using Two way ANOVA which revealed no significant differences due to species or emersion duration, but a significant species x emersion time interaction, suggesting that the rates of re-immersion adjustment differed between species. Tukey tests revealed that the mean eftlux of freshly-captured N. radiata was significantly different from that of freshly-captured N. vidua and that the mean effluxes of the 22 h emersed N. vidua were significantly higher than those of 22 h emersed N. trachaides or N. radiaia. TABLE L Relationships betueen total wet weight (X) and fresh tissue weight (Y) for 3 Hong Kong \odililtoriiui spp. Species Regression details Noclilinoriiui vidua Nodilinoriiiu tntclioides NitdilitUirina radiata Y = 0.1 27X + 0.004 r = 0.9298 Y = 0.159X + 0.001 r = 0.8500 Y = 0.093X -I- 0.002 r = 0.7300 n 54 in all cases. Ammonia Efflux in Hong Kong Nodilittor/na 491 J E a. 10 7 - 6 - • N trochoides ■ N radiata * N. vidua 20 40 60 80 100 120 140 Post handling time (mln) Figure 1. Weight-specific ammonia efflux of 3 species of Nodilitlorina at various times following a brief handling procedure. Values are pre- sented as means ± S.E. for separate groups of (n = 61 animals in each case. (T = 29.4 C, S = 30). DISCUSSION These preliminary findings reveal that considerable interspe- cific and intraspecific variability of ammonia efflux occurs amongst these littoral gastropod species and that some of the vari- ability can be attributed to the positions that the animals are found on the shore. Brief handling and emersion evoked high ammonia effluxes after being handled and emersed for a very short period of time. These "shock"" response rates persisted for at least 30 minutes and possibly, 120-min rates better represent the "normal" im- mersed, active effluxes of these species (Table 2). The control rates used here were those measured at 120 min following the brief handling procedures and were 1.66 ± 0.29. 2.02 ± 0.68 and 0.32 ± E ■3, i>-i\ N. radiata • control ■ 1 h emersion * 4 ti emersion ▼ 22 h emersion k >- l^--^#r__-r__^ -20 0 20 40 60 80 100 120 140 Time after re-immersion (min) Figure 2b. Weight-specific ammonia effiux of Xodilittorina radiata at various post reimmersion times following emersion of Ih, 4h or 22h. Values are presented as means ± S.E. for separate groups of (n = 6) animals in each case. IT = 29.4'C. S = 30). 0.08 (imoles NHj x g~' flesh weight x h"' respectively for N. trochoides. N. radiata and N. vidua. Ammonia has an oxycalorific value of 0.0689 cal x (xmole"' (Brafield and Solomon 1972) and, on the broad assumption that N. trochoides. N. radiata and N. vidua were immersed for 4, 5 and 12 h x d~' respectively at the collection site, the daily calorific loss through this route when immersed would amount to 0.46, 0.70 and 0.26 cal x g"' x d"' for the three species. The equivalent ammonia losses would be 6.64, 10. 10 and 3.84 pimoles NH4 x g flesh weight"' x d"'. Duerr (1968) found no urea produced but 0.3-6.0 ixmoles NH4 x g wet tissue"' X d~' produced by seven species of marine prosobranchs and Crisp E 3 o 1 H N. trochoides • control ■ 1 h emersion * 4t\ emersion ▼ 22 ti emersion -20 0 20 40 60 80 100 120 140 en E 3. \ ■^ N. vidua • control ■ 1 h emersion * 4 h emersion ▼ 22 h emersion \ ^ -20 20 40 60 80 100 120 140 Time after re-immersion (min) Figure 2a. Weight-specific ammonia effiux of .\iidilillorina trochoides at various post reimmersion times following emersion of I h. 4h or 22h. Values are presented as means ± S.E. for separate groups of (n = 61 animals in each case. (T = 29.4 C, S = 30). Time after re-immersion (min) Figure 2c. Weight-specific ammonia efflux rales o( .Xodilittorina vidua at various post reimmersion times following emersion of Ih. 4h or 22b. \alues are presented as means ± S.E. for separate groups of (n = 6) animals in each case. IT = 29.4 C, S = 30). 492 Uglow and Williams TABLE 2. Ammonia efflux (fimoies NHj • g"' • h ') of 3 Nodilittorina species collected in silu at high tide whilst awash and given no emersion. Measurement period A". Irochoides N. radiata N. vidua (mini Mean ± S.E. Mean ± S.E. Mean ± S.E. 30 3.91 ±0.48 6.01 ±0.69 3.52 ± 0.42 60 2.21 ±0.44 2.25 ± 0.55 0.83 ±0.17 120 1.66 ±0.29 2.02 ± 0.68 0.32±O.OS Groups of n = 6 animals were used in each ease (T = 29.4^C, salinity = 30). et al. ( 1981 ) found that control group animals of the carnivorous gastropod, Nassiiriiis reticiilatiis excreted ca. 0.76 (xmoles NHj x g wet tissue"' x h"' ( = 18.20. d~'). Further studies will refine these estimates and will include estimates of the volatilised am- monia excreted during the emersion periods and the variability corrected for actual duration of the emersion periods experienced during the various tidal cycles at the collection site. Clearly, the ammonia excretion rates following emersion vary according to species and the duration of the emersion period. After 1 h of emersion, the subsequent efflux rates for the first 30 min were as high (N. radiata) or higher {N. vidua) than the relevant control values. This suggests that ammonia production in these species had not only continued over this emersion period, though possibly at a lessened rate, but that it had been stored in a form and a site such that it was rapidly available for excretion as ammonia. During the 30 min following re-immersion after all the emersion times tested, N. vidua, excreted ammonia at rates which exceeded their control values but, in N. radiata showed this behavior after the 1 h emersion period only and N. trochaides did not show this behavior at all. These interspecific differences may relate to nor- mal tidal distributions as the more eulittoral N. vidua will spend relatively much less time emersed than the high-shore N. iro- choides. Consequently, the trade- off between the need to conserve water and to tolerate supra-normal ammonia concentrations is less critical in eulittoral species than for high shore species. Some of the post-emersion ammonia efflux that occurred in N. vidua and N. radiata may have been attributed to the flushing of ammonia- enriched mantle cavity water — but this was unlikely to have accounted for the entire amount. Similar occurrences have oc- curred with virtually all the crustacean and mollusc species we have examined and the source of this rapidly expelled ammonia is still under investigation. They are distinct from the effluxes de- rived from the general metabolic increase derived from handling. The facility to reduce or cease ammonia production during periods of anaerobiosis is probably a prerequisite to successful colonisation of the high littoral levels where, at and above mean high water level at Spring tide (MHWST). prolonged emersion events are frequent. Additional studies will be needed to examine more critically the interspecific variability of emersion- dependance of volatilised ammonia losses and of uric acid pro- duction. No uric acid was found to be excreted in the 22 h emer- sions used in these tests but the production and storage of uric acid during emersion periods of this length are not precluded and such storage has been shown to vary seasonally in Littorina saxatilis (Smith and Smith 1998). The freshly-caught specimens of all three species showed an enhanced ammonia efflux following brief han- dling but this response is abolished if the handling is preceded by emersion for 4 h or more in N. radiata or for 1 h or more in N. irochoides but persists in N. vidua. In the last example, the rapidly excreted ammonia could represent the release of that produced during emersion and .stored at some site such as the mantle cavity reservoir. In this instance, such reservoirs act less as a means of ensuring a modicum of aerobiosis during emersion, than as a la- trine to which potentially toxic wastes can be consigned until they are flushed away on re-immersion. The techniques u.sed appear to be suitably sensitive to allow accurate estimates of nitrogenous effluxes to be measured using the small volumes of water, or other fluids, which are often all that are available for analyses. They thus lend themselves to better estimations of energy fluxes in these types of animals, whether emersed or immersed, and to comple- ment oxygen consumption studies which, often, are technically difficult, expensive or not feasible to use for certain types of en- ergy flow studies. ACKNOWLEDGMENTS We thank Del Smith for his helpful comments on this paper. Attendance at the conference was funded by a CRCG grant from The University of Hong Kong to GAW. LITERATURE CITED Brafield, A. E. & D. J. Solomon. 1972. Oxy-calorific coefficients for animals respiring nitrogenous substrates. Cump. Biochem. Physiol. 43A;837-841. Britton, J. C. 1992. Evaporative water loss, behaviour during emersion, and upper thermal tolerance limits in seven species of eulittoral fringe Littorinidae (Mollusca: Gastropoda) from Jamaica. In: J. Grahame, P. Mill and D. Reid, editors. Proceedings of the third international sym- posium on littorinid biology. London: The Malacological Society of London, pp. 1-15. Crisp, M., C. W. Gill & M. C. Thompson. 1981. Ammonia e.\crelion hy Nassarius reticulatus and Buccinium uiuiatiini (Gastropoda: Prosobran- chiata) during starvation and after feeding. / Mar. Biol. Ass. U.K. 61:381-390. Delauny, H. 1931. L'e.\cretion azotee des invertebres. Biol. Kcv. Ciiiii- bridge Phil. Soc. 13:225-251. Duerr, F. G. 1968. Excretion of ammonia and urea in seven species of marine prosobranch snails. Coinp. Biochem. Physiol. 26:1051-1059. Garrity, S. D. 1984. Some adaptations of gastropods to physical stress on tropical rocky shores. Ecology 65:559-574. Hunter, D.A. & R.F.Uglow. 1993. A technique for the measurement of total ammonia in small volumes of seawater and haemolymph. Opiuliu 37:41-.';0. Mak. Y, M, 1996. The ecology of Ihe high-zoned littorinids. Nochlillonihi trochoidcs. N. nuliuut and N. vidua, on rocky shores in Hong Kong. Unpublished Ph.D. thesis. Hong Kong: The University of Hong Kong. Needham, J. 1935. Problems of nitrogen catabolism in invertebrates II. Correlation between uricotelic metabolism and habitat in the phylum Mollusca. Biochem. J. 29:238-251. Needham, .1. 1938. Contributions of chemical physiology to the problem of reversibility in evolution. Biol. Rev. Cambridge Phil. Soc. 13:225-251. Ammonia Efflux in Hong Kong Nodiuttohina 493 Oghaki. S. I. I9S5. Di^tribution of the family Littorinidac (Gastropoda) on Hong Kong rocky shores. In: B. Morton and D. Dudgeon, editors. The nialacofauna of Hong Kong and southern China II. Proceedings of the second International workshop on the malacofauna of Hong Kong and southern China ( 1983). Hong Kong: Hong Kong University Press, pp. 457-464. Reid, D. G. 1492. The gastropod family Litlonnidae ni Hong Kong. pp. In: B. Morton, editor. The marina flora and fauna of Hong Kong and Southern China HI. Proceedings of the fourth international marine bi- ology workshop: the marine flora and fauna of Hong Kong and south- ern China (1989). Hong Kong: Hong Kong University Press, pp. 187- 210. Sadok, S.. R. F. Uglow & S. J. Haswell 1999. Some aspects of nitrogen metabolism in Myliliix ediilis: effects of aerial exposure. Mar. Biol. tBcrl.) 135:297-305. Shick. J. M.. J. Widdows & E. Gnaiger. 1988. Calorimelric studies of behaviour, metabolism and energetics of sessile intertidal animals. Am. Zuol. 28:161-181. Smith. D. C. & T. P. Smith. 1998. Seasonal variation in soluble uric acid concentration in Liiiniiiui \ii.\aulUs (Olivi). Hytln)hiol(ii;ici 378:187- 191. Widdows. J & J.M. Shick 1985. Physiological responses o( Mylilit.^ cdulis and Cardium editle to aerial exposure. Mar. Biol. (Berl.) 85:217-232. Williams. G. A. 1994. Grazing by high-shore littorinids on a moderately exposed tropical shore. In: B. Morton, editor. The malacofauna of Hong Kong and southern China III. Proceedings of the third Interna- tional workshop on the malacofauna of Hong Kong and southern China, 1992. Hong Kong: Hong Kong University Press. Williams. G. A. & D. Morritt. 1995. Habitat partitioning and thermal tolerance in a tropical limpet, Cellana ^ntui. Mar. Ecul. Prog. Ser. 124:89-103. Williams, G. A. & B. R. McMahon. 1998. Haemolymph pH and oxygen levels in a naturally stressed tropical limpet. Cellana grata. In: B. Morton, editor. The marine biology of the South China Sea. Proceed- ings of the third international conference on the marine biology of the South China Sea, 1996. Hong Kong: Hong Kong University Press. i Journal of Shcllthli Research. Vol. 20, No. 1. 495-499, 2001. TRANS-ZONAL MOVEMENTS IN WINKLES, LITTORINA LITTOREA (L.): REASONS AND CONSEQUENCES G. F. WARNER* Centre for Maiiue Sciences. The University of tlie West Indies. Mono. Kint^ston 7. .lamaica. West Indies ABSTRACT Two years of monthly samples of L. littorea at three levels on an estuarine shore at Southampton. U.K. have provided evidence of long-term movements up and down the shore. Increased population densities at the middle level of winkles in their second year of growth indicate movements of young winkles toward the center of their zonation range. Increases in population densities of older winkles at the higher and lower levels indicate later dispersion away from the center of the zonation range. Evidence of movements between the middle level and the lower level is provided by the presence or absence in the shell of bore-holes made by the polychaete fohdora ciliaui (Johnston) which recruits mainly on the lower shore. At the upper level, fluctuating population densities and observations of feeding fronts in the spring indicate seasonal migrations, down in winter and up in spring. Ad\antages and disadvantages of living at upper or lower levels are discussed and related to differences in growth rates and predation risk. KEY WORDS: Littorina litiorea. movement, migration, zonation. age INTRODUCTION Littorina littorea is one of the best known members of the Littorinidae (Fretter & Graham 1994, Reid 1996, McQuaid 1996a, b). It is a widespread and common species both regionally and locally, occurring on both sides of the North Atlantic on a variety of substrata. Its preferred habitat is toward the middle and lower zones of sheltered rocky shores but the precise zonation varies somewhat from place to place (Underwood 1973) and extends from the upper regions of some shores into the sub-littoral of others. L littorea also occurs on muddy and estuarine shores where there are pebbles and shells to cling to {Moore 1937, Warner 1997). There have been many studies on populations of L littorea that have produced a variety of conclusions concerning settlement and subsetiuent movements. Some authors observed settlement within the zone occupied by the adults (Moore 1937. Williams 1964) while others have reported sub-littoral settlement followed by up- shore migration of juveniles during their first year of life (Smith & Newell 1955. Lambert & Farley 1968). Seasonal migrations up the shore in the spring and down in the winter have also been observed (Lambert & Farley 1968. Williams & Ellis 1975). In apparent contrast, short-term maintenance of zonation by '"homing" of in- dividual winkles has been reported (Newell 1958, Gendron 1977) as well as short-term "random" movements (Petraitis 1982). In some cases, differences in zonation of the various size-classes have been observed in which larger winkles occurred more commonly on the lower shore (Moore 1940, Smith & Newell 1955. Williams 1964, Lambert & Farley 1968. Fish 1972. Gendion 1977. Warner 1997). Little explanation has been offered other than that the lower shore may be "optimal" for L littorea (Williams 1964). However if this were so, one would expect more winkles of all sizes to live there. In the work reported here, individual ages of L. littorea were estimated by counting annual growth rings on the shells (Black 1973). Growth check marks may be caused by a variety of factors (Ekaratne & Crisp 1982). but at this site annual growth rings are prominent and appear to result from the temporary slowing of *E-mail: gfwarner@uwimona.edu.jm growth from late winter to May (Warner, in prep). Using these age estimates, age-specific zonation patterns were investigated. These patterns, and additional observations, are used here to infer orien- tated movements and as a basis for discussion of the balance of costs and benefits of living in any particular zone. METHODS The study site was at Netley on the northern shore of Southampton Water, which is an elongated, sheltered inlet on the south coast of England, receiving input from several rivers. The surrounding land is urbanized and industrial with considerable shipping activity. The tidal range is up to 5 ni and the salinity is normally about 30%f. Wide mud and gravel Hats are exposed at low spring tide, the distance from high to low water being about 300 m. Three sampling sites were established (Warner 1997). The upper site was close to mean tide level at 2.7 m above Chart Datum, the habitat was damp stony gravel with sparse Entcroinor- pha sp. and Fiieiis sp.; this site was chosen as the highest site where a sample of winkles could reliably be collected at all times of year. The middle site was 80 m further down the shore at 1 .4 m above CD where the substratum was wet muddy sand with gravel on which grew a mixed algal community including Choiulnis cris- piis Stackhouse. The lower site, close to MLWS, was a further 80 m down the shore at 0.7 m above CD where the substratum was wet mud with stones and shells and a sparse algal community similar to that at the middle site. Samples of L. littorea were collected at approximately monthly intervals at each site from November 1995 to March 1998. Samples were collected within 50 x 50 cm quadrats, all winkle shells being collected from 2 or nwre quadrats at each site until more than 30 live indixiduals had been collected. 3.218 live winkles were collected in total (there was no indication that this regular removal of snails had any effect on the population on this extensive beach). All dead winkle shells were also collected from within the quadrats from .lanuary 1996. Winkles were killed by boiling (ca 30 sec) soon after collection and were extracted from their shells. The shells were dried and scored for the occurrence of epifauna. especially for the presence of Polydora eiliata (Warner 1997). Epifauna were scraped off and the length of the shells 495 496 Warner measured with calipers to an accuracy of 0.1 mm. Annual growth rings (Black 1973) were recognized as prominent growth check marks on the body whorl separated by several mm of shell growth. New annual growth rings were added in May/June of each year as new rapid growth recommenced following spawning (Warner, in prep). Winkle shells at this site become progressively eroded (Warner 1997). making younger and older sections of shell easily distinguishable, and accentuating the ring in between. Annual growth rings were counted to estimate age at 2+ years and above; 0+ and 1+ shells showed no annual growth check but these cohorts were distinguishable by non-overlapping si/e ranges. It was as- sumed that each new year of life started at the end of May (the end of the spawning season — Williams 1964. Grahame 1975: per- sonal observation at this site). Additional information on winkle dispersion and movements in the upper part of the zonation range was collected just above the upper site in the vicinity of a line of aggregated vsinkles. parallel to the shoreline, which was observed in May and June of each year. This line coincided with the lower edges of wide patches of the green alga Enleminoiplui growing on the pebbles. In 1997. winkle numbers were counted on a 4 m long transect of 16 con- tiguous 25 X 25 cm quadrats positioned normal to this line. In 1998. the position of the line of winkles was marked with 4 pegs 10 m apart and revisited twice at 14-day intervals; on each visit the position of the line in relation to each peg was measured. Aspects of mortality were studied by collecting dead shells as described previously. These were measured and examined for damage re- sembling that caused by crab predation (large, irregular holes in the body whorl or spire — verified by personal observation of green shore crab. Caninus imieiuis (L.). predation on winkles in aquaria) and for bore holes of P. ciliata. RESULTS Relative Movements of Different Age Classes The distributions of the various ages at the three levels on the shore, summed over the entire sampling period and expressed as percentages, are shown in Figure 1 . These age distributions are very significantly different from each other (contingency test. X' = 295, d.f = 14, /) = 0.000). showing an association between Upper Middle Site on the shore Figure 1. The percent frequency of different age classes of L. littorea at three sites on the shore, showing an association between level and age (contingency test on raw data, \' = 295, d.f. = 14, p = (1.000). .At the middle site, there were relatively higher proportions of younger winkles (age 1+ years), and lower proportions of older w inkles (ages 3+ and 4+ years). level and age. The differences include proportionally fewer age 1 -i- w inkles at the upper and lower sites, and more \+ winkles at the middle site. There were also proportionally fewer ?+. 4+ and 5-1- winkles at the middle site, more 2+ and 3-i- at the upper and more 3-I-. 4-I-. and 5-i- winkles at the lower site. These data could be explained by movement of l-i- winkles toward the middle site, mainly from lower locations, and dispersal of older winkles from the middle site both up and down the shore. Total winkles col- lected at any one age increased from O-i- to 2-I-, then declined, but winkles estimated at more than seven years old were found at all sites. Winkle density was much higher at the middle site than at the upper and lower sites (Table 1 ). Indirect evidence for movement of juvenile wnikles toward the middle of their zonation range, and of older winkles away from the middle, was also provided by a biological marker: the bore holes of P. ciliata. Examination of live winkles showed P. ciliata holes in a proportion of shells at all levels (Table I ). However, infection was high at the lower site and much lower elsewhere, leading to the conclusion (Warner 1997) that settlement of P. ciliata is mainly restricted to the lower shore. The pattern of P. ciliata infection of winkles is that the worms prefer to settle on older areas of shell which lack periostracum (Warner 1997). In O-i- winkles, no P. ciliata borings were found at any site. In l-i- winkles, the only area of shell which is ""old" and thus suitable for settlement is the relatively small area of the spire, since the whole body whorl is young and smooth. This leads to a characteristic pattern of infec- tion in juveniles on the lower shore in which borings are often found only in the spire. Figure 2 shows the percent distribution of P. ciliata infection in winkles of different ages from the lower and middle sites (the upper site had too few bored winkles for feasible analysis). In both cases, there is a highly significant association between age and bore-hole distribution (contingency tests, lower: X" = 203. d.f. = 6.p = 0.000. middle: x" = 79. d.f = 6, /? = 0.000. data for >4-i- snails contained a zero so was excluded). At the lower site, it is clear that borings only in the spire were com- mon in I -I- winkles. Figure 2 shows that this pattern is gradually eliminated later in life by borings distributed elsewhere on the shell. At the middle site, l-i- winkles with P. ciliata borings only on the spire were not common, but a small proportion of 2-1- winkles showed this pattern. These may have been individuals that were lower down the shore at an earlier age and moved up at about l-l- years old. In contrast, the relative increase in proportions of un- bored 2+ and 3-i- winkles at the lower site, shown in Figure 2, provide evidence of downshore movements in older snails. TABLE 1. Mean densities at the three sampling sites of live winkles and dead winkle shells, and the percentages of each of these that were bored liy P. ciliata. .\lso included are the percentages containing bore-holes of those dead shells that appeared to have been cracked by crabs. Upper site Middle site Lower site Live per m" ± SD 54.1 ±28.7 116,7 ±25.1 41.1 ± 10.8 Dead per ill" ± SD 8.3 ±5.7 17.y± 13.7 4.8 ±2.6 % live -1- bore-holes 4.7 10.2 63.5 % dead + bore-holes 8 35.5 84 "/r cracked + bore-holes 35 76 97 Calculated from all samples (n and 3/98. 30 at each level) collected between 1/96 Movements of LirroRiNA littorea 497 100% 80% 60% 40% 20% 3 1 1 A ■ P widespread D P in spire D Un bored age in years Figure 2. Distribution of bore-holes of the pohchaete Piilydora ciliala on the shells of different age-classes of live L. littorea at two levels on the shore, show ing that the incidence of boring increases with age, and that on the lower shore, boring in the spire precedes more widespread boring (contingency tests on raw data, lower: \~ = 2(13, d.f , = 6. p = ().»(»(), middle: x" = 7**. df = 6. p = (1.0(10: data for >4+ snails contained a zero so were excluded). A = middle site, B = lower site; P = Pulydura. Seasonal Migrations In Muy and June, just above the upper site, a line of aggregated winkles parallel to the shoreline was observed to coincide with the lower edges of wide patches of the green alga Entenmitn-pha growing on the pebbles. Figure 3 shows the results of the 1997 transect across this line. In 1998. the line was found to move upshore by a mean distance of 103 cm in 14 days (n = 8, range = 45-169 cm), the winkles having apparently grazed the Entero- morpha leaving hare clean pebbles behind. By July, these "feeding fronts" had dispersed but winkles were still present above the upper site, often clustering at low tide beneath sparse clumps of Fiicits. By November, winkles became rare above the upper site. These observations indicate upshore migration occurring in spring and summer. Since there were no clear seasonal changes in num- bers of dead shells, downshore movements presumably occurred in autumn. Fluctuations in monthly population density measurements at the upper site were much greater than at the other two sites (Table I ) and appeared to correlate roughly with the seasons. The lowest densities were recorded between October and March (mean 35.3 m"'. SD = 21.1, range = 7-65, n = 14) and the highest between April and September (mean = 65.6 m"", SD = 29.7, range = 31-148, n = 16). These means are significantly different (two-sample f-test. t = -3.25; d.f. = 27; P < 0.01). Some Costs and Benefits of Life on the Lower Shore Data on growth rates collected during this study show that growth was faster at the lower site than at the other two sites. Winkles of the same age were larger at the lower site, and .showed greater annual growth increments at the shell margin than those from higher up the shore. These data will be published elsewhere. The mean percent of dead shells relative to total (live + dead) shells from the different sites are shown in Figure 4A and are not significantly different. These data show that dead shells occur in proportion to the live population, and suggest that mortality rate is similar throughout the shore, despite differences in age distribution (it was assumed that dead shells "survived" to be sampled equally 70 60 50 40 30 20 green algae bare gravel 1 n nlln nnn^ 7 8 9 10 11 quadrat number 12 13 14 15 16 Figure i. Results of a 4-m long transect, normal to the shoreline, of contiguous 25 x 25 cm quadrats on the upper part of the shore in May 1997, from gravel covered with green algae (higher) to bare gravel (lower, total drop ca. 5 cm), showing a high density of L. littorea forming a feeding front at the boundary of the algal growth. at the three sites). Figure 4B shows the mean percent of dead shells at the three levels that appeared to have been cracked by crabs. These proportions are significantly different (Single factor ANOVA, F = 11.9, d.f. = 68./? = 0.000): cracked shells were far more common at the lower site than at higher sites. Table I shows the incidence of P. cilicita bore-holes in the samples of dead and live shells from the 3 sites. It may be seen that while the incidence of bore-holes in both live and dead winkles decreased up the shore, a consistently higher proportion of dead shells at all sites contained bore-holes, and this proportion was higher still in those dead shells presumed cracked by crabs. The chi-square contin- gency tests on the raw data showed significant, positive association between being dead and the incidence of bore-holes at both the middle and lower sites (x^ = 100.7 and 17.6 respectively, d.f. = l.p = 0.000). Similarly, among dead shells at all three sites there was significant, positive association between the incidence of bore-holes and being cracked (x^ = 6.6-23.8. d.f. = 1. /x 0.01 ). DISCUSSION Some previous work on L. litlnrca has shown that larvae settle on the shore throughout the zone of the adults (e.g. Williams 1964, Upper Middle site on shore Figure 4. The mean frequencies (±SE) of dead (as percent of live + dead) /_. littorea at three sites on the shore: not significantly different. B. The mean frequencies (±SE) of cracked (as percent of total deadlL. littorea at three sites on (he shore, showing a greater proportion cracked at the lower site (single factor .\NOVA. F = 1 1.9, d.f. = 68, p = 0.000). 498 Warner Underwood 1973). but other authors (e.g. Smith & Newell 1955; Gardner & Thomas 1987) concluded that settlement is mainly sublitioral and is followed by upshore migration. Part of the dis- agreement appears to be due to the difficulty of finding very small winkles. This problem was evident in the present work, with rela- tively few 0+ winkles being found at any site (Fig. 1). However, small (5-7 mm) 0+ winkles appeared at about the same time in autumn in the samples at all three sites suggesting that they had been there, unobserved, since settlement. The mixed nature of the substratum at Netley provided many hiding places for tiny winkles (between pebbles, amongst algae, etc). .Settlement at Southampton is therefore probably littoral rather than sublittoral. but the data for 0+ winkles (Fig. 1) suggest that recruitment was more frequent in the middle of the zonation range (middle site) than at the upper site at mean tide level. In older winkles, changes in population density at a site must be due to movements of snails between zones or to differential mor- tality of snails. Smith & Newell (1955) and Williams ( 1964) found evidence of an upshore movement of young winkles (first and second year respectively), and Smith and Newell (1955) specu- lated that a downshore movement of older winkles could account for the presence of larger ones on the lower shore. The demon- stration here of an excess of l-t- winkles at the middle site (Fig. I ) suggests active movement of young winkles toward the middle of the zonation range from both lower and higher levels. Smith and Newell (1955) ruled out the alternative of differential mortality of young snails on the grounds that dead shells were found distributed in proportion to the living, and this was also found here (assuming that dead shells were not unequally moved, fragmented or buried at the different levels). Similarly, subsequent dispersal seems the only plausible explanation of the increasing proportions and den- sities of older age-classes at the upper and lower sites. The data from P. ciliata borings (Fig. 2) provided evidence of both the movement upwards of juveniles from the lower shore and the movement down of age 2+ and 3+ winkles from the middle of the zonation range. However, it should be clear from these data that the full extent of trans-zonal movements were not performed by all individuals; winkles of all age groups remained present at all sites, indicating that some moved less than others. Age-class specific movements may be thought to conflict with movements that have been described as maintaining the particular zones of individuals (Fretter & Graham 1994). These include small-scale circular movements (Newell 1958) and "home" orien- tated movements following displacement (Gendron 1977). How- ever, the movements described here did not involve all winkles and occurred over a time scale of months to years, thus they can coexist with short-term maintenance of individual shore level, and with the short-term "random" movements described by Petraitis (1982). They also coexist with the winter downshore and spring upshore "migrations"of winkles (Lambert & Farley 1968. Williams & Ellis 1975) which have been observed only in the upper part of the zonation range. The question remains: What functions do these movements perform in the life of the winkle? Movements in early life toward the middle of the zonation range may serve a "habitat location" function. Movement of older adults away from the center may then be a response to high population density, serving to reduce com- petition. However, movement up has very different consequences from mo\ement down. Movement up gives access to the spring growth of green algae but also entails an annual retreat from upper levels as temperatures fall in the winter. Movement down brings the benefit of faster growth, probably due to longer immersion allowing longer feeding periods. Since faster growth leads to larger size, it may also lead to greater fecundity in lower shore individu- als. However, life on the lower shore also entails increased risk of predation from crabs (Fig. 4) and increases the variety of epibiota. including P. ciliata. which settle on the shell (Warner 1997). Al- most all dead, cracked (i.e. crab-predated) shells on the lower shore were bored by P. ciliara (Table 1 ) suggesting that borings may increase predation risk by weakening the shell (Buckley & Ebersole 1994). L. littorea has few defences against epibiota, some of which have adverse effects through increasing drag on the snails (Wahl & Sonnichsen 1992, Wahl 1996). Thus, there are several reasons for not regarding the lower shore as "optimal" for this species (Williams 1964). despite the faster growth that can be achieved there. ACKNOWLEDGMENTS 1 thank the numerous Reading University undergraduate stu- dents who counted winkles for me on various occasions, both for the useful data and for their enthusiastic support. LITERATURE CITED Black. R. 1973. Growth rates of intertidal molluscs as indicators of unex- pected incidents of pollution. / Fish. Res. Bd. Can. .30:1385-1.388. Buckley, W. J. & J. P. Ebersole. 1994. Symbiotic organisms increase the vulnerability of a hermit crab to predation. J. e.xp. Mar. Biol. Ecol. 182:49-64. Ekaratne. S. U. K. Si. D. J. Crisp. 1982. Tidal micro-growth bands in intertidal gastropod shells, with an evaluation of band-dating tech- niques. Proc. R. Soc. Land. B 214:305-323. Fish, J. D. 1972. The breeding cycle and growth of open coast and estua- rine populations of Littorina liltarca. J. Mar. Biol. Ass. U.K. 52:101 1- 1019. Fretter. V. & A. Graliam. 1994. British prosohranch molluscs. 2'"' edition. London: The Ray Society. 820 pp. Gardner. J. P. A. & M. L. H. Thomas. 1987. Growth and production of a Linorina litlorea(L.) population in the Bay of Fundy. Ophelia 27:181- 195. Gendron. R. P. 1977. Habitat selection and migratorv behaviour of the intertidal gastropod Littorina littorea (L.). ./. Anini. Ecol. 46:79-92. Grahame. J. 1975. Spawning of Littorina littorea (L.) (Gastropoda: Proso- branchiata). / E.xp. Mar. Biol. Ecol. 18:185-196. Lambert. T. C. & J. Farley. 1968. The effect of parasitism by the trematode Ciyptocotyle lingua on zonation and winter migration of the common periwinkle Littorina littorea. Can. J. Zool. 46:1 139-1 146. McQuaid. C. D. 1996a. Biology of the gastropod family Littorinidae. I. Evolutionary aspects. Oceanogr. Mar. Biol. 34:233-262. McQuaid. C. D. l'596b. Biology of the gastropod family Littorinidae. II. Role in the ecology of intertidal and shallow marine ecosystems. Oceano.i^r. Mar. Biol. 34:263-302. Moore. H. B. 1937. The biology of Littorina littorea. Part I. Growth of the shell and tissues, spawning, length of life and mortality. J. Mar. Biol. Ass. t/. A" 21:721-742. Moore. H. B. 1940. The biology of Littorina littorea. Part II. Zonation in relation to other gastropods on stony and muddy shores. ./. Mar Biol. Am. U.K 24: 227-237. Newell. G. E. 1958. The behaviour of Littorina littorea (L.) under natural conditions and its relation to position on the shore. / Mar Biol. Ass. U.K. 37:229-239. Movements of Littorina uttorea 499 Pelriiitis, P. S. 1982. Occurrence of random and directional movements in Wahl. M. & H. Sonnichsen. 1992. Marine epihiosis. IV. The periwinkle the periwinkle, Z-mon/m /morcfl (L.). y. £v/). Mm: Biol. Ecoi 59:207- Linorina lillorea lacks typical antifouling defences : why are some 217. populations so little fouled? Mar. Ecol Prog. Ser. 88:225-235. Reid, D. G. 1996. Systematics and evolution oi Linnriiw. London: The Ray vVamer. G. F. 1997. Occurrence of epifauna on the periwinkle. Linorina Society. 463 pp. linorea (L.), and interactions with the polvchaete Polvilora cilialu Smith. J. E. & G. E. Newell. 1955. The dynamics of the zonation of the (Johnston). Hxdrobiologia. 355:41-17. common periwinkle Linorina linorea (L.) on a stony beach. /. Anim. F- I ■'4-35-5fi Williams. E. E. 1964. The growth and distribution ot Z./rfon/w mm/ca (L.) Underwoo'd,' A. "j. ' 1973. Studies on zonation of intertida! prosobranch ™ ^ '^''^^ ^hore in Wales. J. Auun. Ecol. 33:413-432. molluscs in the Plymouth region. J Anim. Ecol. 42:353-372. Williams, I. C. & C. Ellis. 1975. Movements of the common periwinkle Wahl, M. 1996. Fouled snails in tlow: potential of epibionts on Linorina Linorina linorea (L.), on the Yorkshire coast in winter and the influ- linorea to increase drag and reduce snail growth rates. Mar. Ecol. Prog. ence of infection with larval Digenea. / &/;. Mar. Biol. Ecol. 17:47- Ser. 138:157-168. 58. Journal ot Shellfish Research. Vol. 20. No. 1. 50 1 -508. 2001. CORRELATION OF MORPHOLOGICAL DIVERSITY WITH MOLECULAR MARKER DIVERSITY IN THE ROUGH PERIWINKLE LITTORINA SAXATILIS (OLIVI) CRAIG S. WILDING,* JOHN GRAHAME, AND PETER J. MILL Centre for Biodiversity and Conservaiion. School of Biology, University of Leeds. Leeds. West Yorkshire, LS2 9JT, United Kingdom ABSTRACT Two morphological varieties of Liiinnna sumiuHs. widespread around the United Kingdom, are a thin-shelled, high- shore niorph (L saxalilis H) and a thick-shelled, mid-shore animal (L. saxatilis M). Mitochondrial DNA (mtDNA) analysis by PCR-RFLP was used to test whether gene flow between these morphs is restricted. At Galloway. Scotland, replicated sampling (different years and different transects over a distance of 800 m) has been undertaken. One mtDNA haplotype is predominant in H and a different haplotype common in the M animals. Repeatability in space and time suggests a real H/M differentiation. A similar pattern of mtDNA haplotype variation is seen in a single sample of the Swedish E and S morphs of Littorimi saxatilis. However, this pattern is not evident everywhere. Variation in the mtDNA and four nuclear DNA loci was examined within and between L saxalilis H and M morphs from the south coast of England. Because shape variation in this region additionally separates into three shape groupings (regions identified from inultivariate morphometric analysis where shape is more homogeneous within, than between groups), genetic variability was examined within and between these groupings as well as between H and M. On the south coast, an apparent association of shape and mtDNA haplotype is identified, but AMOVA analysis shows no support for the association being with shape grouping or H and M morphs. Although the nature of this shape-genotype association is unknown, a mtDNA haplotype and an allele at the nuclear CAL-2 locus are confined mainly to one shape group. Analysis of association of mtDNA haplotype with H and M morphology suggests a strong correlation can be found in some areas (Galloway, Mumbles (south Wales), and between similar morphs in Sweden) yet no association is seen at others (Ravenscar, UK, Ballynuhown, Ireland, and the south coast of England), Thus, unravelling the basis of the H and M forms will require more detailed studies, with replication, as at Galloway, and also with additional molecular markers, KEY WORD.S: AMOVA, Littorimi saxalilis. mitochondrial DNA. morphometries. PCR, RFLP, shape INTRODUCTION Shell shape in gastropod mollusks is known to be affected by environmental factors. Selection as a result of crab predation or stone damage can produce heavy shells with small apertures, while animals in exposed conditions where crab damage is minimal, but risk of dislodgement by waves is high, tend to have thin shells with wide apertures (Boulding 1990). In Litrdhna .S(«((///(.s (Olivi) two forms typical of these morphological extremes are widespread. Hull et al. (1996) referred to the thin-shelled, patulous form, found in high-shore areas as L. sci.xatilis H, and the thick-shelled variant that occurs in the mid-shore as L .';axalilis M. Although the basis of this shell polymorphism undoubtedly has a strong environmen- tal component, evidence from embryological characters (Hull et al. 1996) and detection of assortative mating for these morphotypes (Hull 1998. Pickles & Grahame 1999) suggests that there may be a genetic component to the differentiation of L. .w.wlili.t H and M. However, the only inolecular level study of genetic differentiation between these forms was limited to two sites on the east coast of Yorkshire. England. Random amplified polymorphic DNA (RAPD) analysis showed differing degrees of differentiation at these two sites (Wilding et al. 1998). No wide scale geographic studies have yet been undertaken. This is important since such evidence is necessary to determine the extent of gene flow between these morphs, which is critical for the understanding of their re- lationships. Mitochondrial DNA (mtDNA) is a useful molecular marker for such a study because of its high rate of evolution (A vise 1994), and a substantial portion of the L. saxatilis mitochondrial genome has been sequenced (Wilding et al. 1999), thus permitting the targeting *Corresponding author. E-mail: bgycswCsUeeds.ac.uk of primers. Population genetic studies of gastropods have often relied on mitochondrial markers (Kirby et al. 1997, Hellberg 1998. Kyle & Boulding 1998) and Wilding et al. (2000a) have recently examined intDNA cytochrome oxidase 1 (Coll variation in UK and Irish Littorimi. We also have developed primers for the amplifi- cation of four variable nuclear DNA (nDNA) loci in Littorina sa.xatilis (Wilding et al. 2000b). Here shape variation and genetic \ ariation are examined at four levels. ( 1 ) Between L. saxcitilis H and M in replicated samples from southwest Scotland. (2) Between L. .m.xotilis H and M on the south coast (where nDNA variation is also examined). (3) Between L. sa.xatilis H and M at other locations, (4) Along the south coast of the United Kingdom where three 'shape groupings" are recog- nized. It is shown that both shape and mitochondrial DNA do vary geographically, but that there is only a limited correlation of the two. There is however evidence for differentiation at the mtDNA level between L. sa.xatilis H and M in some areas of the UK. MATERIALS AND METHODS Animals were collected from 17 sites around the UK (Table 1), Images of the shells were captured with a digital camera and used for morphometric measurements. Measuring of shell images was performed using Sigma Scan to measure eight shell variables (Fig, 1 ), These shell variables were standardized for size using the geo- metric mean transfonn (Reist I98.'i) and then submitted to a ca- nonical variate analysis in SAS v 6,0 (SAS Inc. 1990). The shell was subsequently crushed and a portion of the head-foot removed for DNA extraction using the single-fly DNA extraction protocol of Ashburner (1989). DNA concentration was measured by fluo- rometry and adjusted to 10 ng.ijtp'. A 967 bp segment of mito- chondrial DNA spanning the cytochrome oxidase 1-cytochrome 301 502 Wilding et al. TABLE 1. Collection sites for samples of L. saxatilis H and M. South Coast samples are grouped in the three shape zones used in some analyses. British National Collection site Grid Reference A'lH) N m) South coast Kent St Margaret's at Clitte TR3ftS444 10 10 Folkestone TR24?369(H) & TR244373(M) 10 9 South-central St Alhan's Head SY959754 10 Portland Bill SY675683 9 East Fleet SY799635 10 The Fleet (gravel) SY75S665 10 Golden Cap SY4079I8 20 Pinhay Bay SYS 18907 5 South-west Cargreen SX436627 10 The Lizard SW6991I4 10 Trevaunance SW725519 10 3 Other locations Galloway St. Ninian's Cave NW4 17364 48t 49t Back Bay NW368394 7 10 South Wales Mumbles SS632873 10 10 North East England Ravenscar NZ9K402 1 10 10 Ireland Ballynahdwn IGR: L 992202 10 10 Sweden Ursholmen 58'5O"I0^N 0'59"4'E 9* 8* N = sample size. t replicated samples taken at Galloway, see Fig. 4. * niorphs in Sweden are E and S, not H and M — .see text. IGR = Irish Grid Reference. oxidase II (CoI-CoII) gene boundary (Wilding et al. 1999) was then amplified using the primers saxCoI and saxCoII (Wilding et al. 2000a). PCR was performed in 50p.l volutnes containing 50 mM KCl. 10 mM Tris-HCl pH 9.0, 1.5 mM MgCK, 0.1% Triton X-IOO, 0.01% gelatin, 200 |j.M each dNTP, 25 pmol each primer. 25 ng DNA and lU Taq {Supertaq, HT Biotechnologies). PCR Figure 1. The eight shell variables measured in Sigma Scan, aa = apical angle, al = aperture length, aw = aperture width, cl = columellar length. II = lip length, ww 0-2 = whorl width 0-2. conditions of 1 x 94°C, 5 min; 35 x (94°C. 1 min, 55°C, 30 sees, 72''C. 1 min): 1 x 72"C, 5 min were employed. In an initial study. 400 bp of mtDNA were sequenced in 50 individuals from 18 populations (Wilding et al. 2000a). From the resultant sequences variable restriction sites were deduced. Variation was then as- sessed by RFLP using three restriction enzymes — Ddel, Dial and Hindlll {Dial is not variable in L. sa.\atilis but is in other rough periwinkles). Restriction enzyme digestion was undertaken in 15 10.1 volumes containing 3-5U enzyme in I.X buffer (supplied with enzyme) and 5 jxl PCR product. Following digestion, restriction fragments were separated on 2% agarose gels. Variant patterns generated by each enzyme were labeled A, B, C etc. and a three letter composite haplotype describes the variation in each animal. An analysis of molecular variance (AMOVA: Excoffier et al. 1992) was used to examine genetic variation from animals along the south coast of England. Littorina saxalilis in this region is known to exhibit substantial shape variation which has been shown to partition into three groups (Grahame and Mill 1992, Mill and Grahame 1 995 ), a south-west, a south-central, and a Kent grouping (Fig. 2), Upon this broad-scale variation there is superimposed the H and M differentiation. AMOVA was used to quantify the par- titioning of mtDNA variation into a 'shape grouping' component and a 'H and M' cotiiponent. If substantial variability is accounted for by either of these then the expectation is that these shape differences are indicative of underlying genetic differentiation. If little genetic variability is partitioned into either of these conipo- Trevaunance Lizard \ South-central South-west Figure 2. Delineation of shape groupings as described by Grahame and Mill (1992) and Mill and Grahame (1995). Genotypk-Phenotype Associations in Littorina 503 nents then there is no detectable genetic differentiation at the mi- tochondrial DNA level between these groups. AMOVA was per- formed in ARLEQUIN vl.l (Schneider et al. 1997). In addition to the mtDNA analysis, patterns of nuclear DNA (nDNA) polymorphism have been examined in these animals (Wilding et al. 2000b). Four nuclear loci comprised of two cal- modulin introns (CAL-I and CAL-2), and two anonymous loci (X80 and DELETION) were examined for variation. All nDNA PCR reactions were undertaken in 25 p.1 of 50 niM KCl, 10 mM Tris-HCl pH 9.0. 1.5 mM MgCK. 0.1% Triton .X-100. 0.01% gelatin. 200 |jlM each dNTP, 12.5 pmol each primer. 12.5 ng DNA and 0.5 U Taq (Super/oi?, HT Biotechnologies). PCR cycling con- ditions were I x 94°C. 5 min: 35 x (94°C, 1 min, 55°C. 30 sees. 72 C. 1 min): 1 x 72°C, 5 min. although for CAL-1 the annealing temperature was increased to 57°C, and DELETION in which annealing was at 57°C and the extension time was reduced to 30 seconds. CAL-1 and CAL-2 polymorphism was examined by Taq\ and Dcle\ digestion respectively and in .\-80 with MsjA. Restriction digestion was performed at the appropriate temperature for 2 hours and products were separated on 2% agarose gels. For the DELE- TION locus, alleles differed due to length variation and this was recognized after PCR products v\ere run on 3% agarose gels. These nDNA data were not suitable for AMOVA analysis (because al- though the genetic distance between genotypes was calculable for those loci examined by RFLP, the mutational distance between alleles at the length variable DELETION locus was not. Thus the relationships among multi-locus genotypes, necessary for AMOVA analysis, were unclear). Correlation of shape variation with molecular variation was therefore examined using a Mantel test employed in NTSYS through comparison of the population matrix of genetic distance (different genetic distance measures were tested in case of subtle differences due to method employed) with the Mahalanobis distance matrix. RESULTS Mitochondrial DNA \ urialion A neighbor-joining tree based on all sequences of CoI-CoII from Littorina shows two groupings of L. .sa.\atilis (Fig. 3). Al- though there are 13 separate L. sa.xcitilis sequences, these cannot all be distinguished using RFLP since there are many variable posi- tions at this locus where there is no restriction enzyme with a recognition sequence spanning the site. However, three enzymes iHindlU. Dra\ and Dcle\) do recognize variable positions in this sequence within the rough periwinkles (Wilding et al. 2000a). although Dra\ is not variable in L. saxutilis. In L saxatilis. HiniilW produced two patterns and Dtli'\ four patterns. The four composite haplotypes encountered in L. saxatilis were designated AAA, ADA. BBA and BCA. The most pronounced evidence for association of mitochon- 29 37 r Ml 73 p A,C 91 I C C S A,C, fifil 66 D Scale: each - » distance of 0.000695 Figure 3. Phylogeny of rough periwinkle Col sequences (adapted from Wilding et al. 2000a) using neighbor-joining of .lukes-Cantor distances. S = /.,. saxatilis. A = L. arcana. C = L. compressa, F = L. fahalis and () = /,. ohtusata. The two divisions of /„ saxatilis sequences are denoted, note that some L. arcana and L. compressa sequences are shared with /.. saxatilis or cluster in these "L. saxatilis" groups. Flats = sequences from the flat periwinkles L. fabalis and L. obtusata. The outgroup is /,. sitkana. Numbers on the nodes are bootstrap values from 100 pseodoreplicates. 504 Wilding et al. 4°30' W "^^ , i Back Bay i^ N ^OVA/ ^^ V. 1-3(1 Om apart) ^ ^^^-,^ St Ninian's Cave 0 1 2K.m AAA- V^ 54°40' N Burrow Head 12 10 2 0 4 .1 M H M I ■ BCA DBBA HAAA H M H M H M H M H M Back Bay 3 1997 Figure 4. Distribution of nitDNA haplotvpes in separate transects, undertaken in 1997 and 1998, at Galloway, southwest Scotland. Sites 1-3 were spaced 10 m apart and are separated from site 4 by approximately 800 m of shingle beach. The 1997 sample was taken from the same site as site 3 of 1998. drial haplotype with morphology is found at Galloway in south- western Scotland. Here one collection in 1997 and five collections in 1998 (from different transects) were typed. The AAA haplotype was predominant in L. .'^axulilis H and the BCA haplotype in L. suMirilis M (Fig. 4). Additional data from Mumbles (South Wales) also suggests differences between H and M. but at Ravenscar (east coast of England) and Ballynahown (western Ireland) no differ- ences were detectable (Table 2). At Ursholmen, Sweden, differ- ences were detected between the E and S niorph of L sciMirilis. The E (exposed) morph, like L suMitilis H. is wide apertured and thin-shelled and the S (sheltered) morph thick-shelled. However. unlike L. saxatilis H and M. where animals are separated vertically on the shore. E and S are separated locally, by habitat along the shore, being found in exposed and sheltered localities respectively. For the small sample examined, all nine E had the AAA haplotype, and 7 of 8 S had the BCA haplotype. These are the same haplo- types separating the H and M at Galloway. Soulh Coast of England — Morphological LittoriiHi sii.yatilis are shown to cluster into two groups on the basis of shell variation analyzed by canonical variate analysis (Fig. 5). When canonical variate means are plotted, the two groupings are associated with L saxatilis H and M. South Coast of England — Molecular The distribution of haplotypes along the south coast within L. sa.xatilis H and M is shown in Table 2 and Fig. 6. There is sub- Genotype-Phenotype Associations in Littohina 505 TABLE 2. Observed number of the 4 nitDNA haplolypes in the studied populations, and allele frequeneies at the 4 nDNA loci for the south coast samples. Collection site MtDNA haplotvpe CAL-1 CAL-2 X-80 Deletion AAA ADA BBA BCA A B A B C A BCA B A' South coast South-east St Margaret's at Cliffe— H St Margaret's at Cliffe— M Folkestone — H 1 10 8 3 1 7 0.4.5 0.1. "S 0.55 0.45 0.85 1 1 1 0,2 0,1 0.05 0.8 0.9 0.95 0.3 0.6 0.25 0.7 0.4 0.75 10 10 10 Folkestone — M 1 T 6 0.167 0.833 1 1 0.222 0.778 9 South-central St Alban's Head 1 9 0.6 0.4 0.4 0.6 0.15 0.85 0.35 0.65 10 Portland Bill 6 3 0.333 0.667 0.06 0.944 1 0.389 0.611 9 East Fleet T 8 0.2.'i 0.75 1 0.1 0.9 0.45 0.55 10 The Fleet (gravel) Golden Cap Pinhav Bay 20 5 10 0.4 0.325 0.8 0.6 0.675 0.2 0.1 0.5 0.8 0.9 0.5 0.2 0.275 1 0.725 1 0.3 0.1 0.4 0.7 0.9 0.6 10 20 5 South-west Cargreen 5 5 0.85 0.15 0.95 0.(.)5 0.1 0.9 0.25 0.75 10 The Lizard 6 4 0.25 0.75 0.1 0.9 0.55 0.45 1 10 Trevaunace — H Tre\aunace — M 1 1 6 1 1 0.4 1 0.6 1 I 0.2 0.667 0,75 0.0,5 0.333 0.167 1 0.833 10 3 Other locations Galloway St. Ninian's Cave- -H 39 2 1 48 St. Ninian's Cave- -M 11 -) 36 49 Back Bay— H Back Bay— M 5 5 2 5 7 10 South Wales The Minnhles — H 3 5 1 10 The Munihles — M 7 3 10 North East England Ravenscar — H 10 10 Ravenscar — M 10 10 Ireland Ballynahown — H 1 9 10 Ballynahown — M 10 10 Sweden Ursholmen — E 9 9 Ursholmen — S 1 7 8 sample size. stantial variation within both H and M ftirnis of Z,. snxntilis but no obvious difference in haplotypes between the two groups. How- ever, there is some evidence for association of haplotype with shape grouping since the haplotype ADA is found mainly on the south coast in the south central group of shape variation. To in- vestigate if haplotype partitioned with shape, haplotypes were plot- ted onto a canonical variate analysis of shape. The resultant, ap- parently random, distribution of haplotypes throughout the mor- phospace defined by the first three canonical variates shows little association (Fig. 7). Nevertheless, discriminant analysis with cross validation shows that the best classification on the basis of shape variables is back to the parent haplotype group (Table 3) indicating some underlying association of shape and haplotype. However. AMOVA analysis (Table 4) suggests that neither shape groupings nor H and M are the covarying factor. Variation was also high in the nDNA dataset (Table 2: G^y = 0.1601. Ht = 0..3447. H^ = 0.289.5). Mantel comparisons of the morphological Mahalanobis distance (D") with Nei's (unbiased) distance and the Prevosti distance from the nuclear-DNA RFLP data showed that. iiTCspective of the genetic distance matrix em- ployed there is no evidence of association of shape and genotype; p [random z < obs. z] = 0.337 and 0.345 for association with the Nei and Prevosti distance respectively. However, as was found for the intDNA dataset there is an association between nDNA geno- type and shape grouping in that the allele CAL-2'^ is found at the highest frequency in the south-central region where the haplotype ADA is encountered. DISCUSSION It is apparent that around the coastline of the UK and Ireland there is a separation of L. saxciiiUs into a thin-shelled, patulous, hich-shore form {L. saxutitis H). and a thick-shelled mid-shore 506 Wilding et al. Canonical means by site C AN 1 2.43 0.78 C A N2" -2.43 Figure 5. Population means from canonical variate analysis of south coast L. saxatilis. form {L. saxatilis M). In this study we examined whether mito- chondrial DNA variation provided evidence for genetic differen- tiation of these morphological varieties. At Galloway. Scotland, where repeated sampling on the shore has been undertaken, defi- nite genetic differences are found between the morphs. with con- sistently different haplotype frequencies between H and M over both separate years and over multiple transects. Such repeatability shows the pattern is neither a temporal phenomenon, nor a simple distance effect. We found similar genetic differentiation between the Swedish E and S morphs of Janson (1982) which have similar morphologies to H and M. Interestingly, the differentiation was due to differing frequencies of the same haplotypes (high fre- quency of AAA in H/E and high frequency of BCA in M/S). Differentiation of nitDNA is also suggested at Mumbles (South Wales) where the frequency of AAA in L. saxatilis H is 0..\ compared to 0 in M — once again, it is the AAA haplotype which is involved. In contrast, no differentiation is indicated at Ravenscar or Ballynahown, but there is little mtDNA haplotype variation to partition at either of these sites. On the south coast of England there are two obvious patterns to variation in Litlorina saxatilis shell shape. On a broad scale, there are three "shape groupings" within which shape, analyzed by mul- tivariate canonical variate analysis, is typically homogeneous but among which shape differs (Grahame & Mill 1992, Mill & Gra- hame 1995). Superimposed upon this, is the morphological dis- tinction into H and M. Both mtDNA and nuclear DNA variation have been analyzed in samples from along the south coast to examine whether there is evidence that mtDNA differentiates H and M as at Galloway, and additionally if there is any evidence for restriction to gene How between the shape groupings. Discriminant analysis of shape variation with mtDNA haplotypes as groups suggests a correlation between mtDNA haplotype and some aspect of shape on the south coast. However, there is no detectable as- sociation of mtDNA or nDNA with either shape groupings or H and M using AMOVA. Thus although a shape-mtDNA correlation has been detected it is not due to either of the a priim groupings considered here. In contrast to this, both an mtDNA haplotype and an nDNA allele at the CAL-2 locus are mainly limited to one of the shape groupings: the south central grouping of shape variation. It is likely that the AMOVA analysis of shape groupings does not detect this as a significant association due to the high variability of mtDNA within the shape groupings, effectively masking the 'be- tween group" component. Nevertheless, the correlation of genetic differences at both the mitochondrial and nuclear DNA level with a known shape-group suggests that there is a real population dif- ference. Thus the pattern of genetic variation in L. saxatilis H and M is not simple. MtDNA evidence does not suggest that L saxatilis H are consistently different from L saxatilis M. although this can be the case on a particular shore. Given that these forms have differ- ent embryological characteristics (Hull et al. 1996) and display assortative mating (Hull 1998, Pickles and Grahame 1999) and that, in Littoriiui. there is a known substantial genetic component to shell shape (Grahame and Mill 1993), it is perhaps surprising M Pinhay Bay Golden Cap St Margarets at Cliffe Cargreen Trevaunance \-{ ^^ — ^ Lizard ADA "^ — ^ bha Figure 6. Distribution of the four mtDNA composite haplotypes along the south coast L. saxatilis M are shown abo\ e the map and L. saxatilis H below. Genotype-Phenotype Associations in Littokina 507 C A N 1 5 7 2 0 - 0 . 1 B 2.55 2.41 A O = AAA ^ = BBA O = ADA * = BCA 0.65 C Figure 7. Plotting of haplot>pf onlo canonical variate plot of shape measurements on an Individual basis. that clearer genetic differences have not been detected. However, the rough periwinkle group is itself young (Reid 1996) and sorting of mtDNA in recognized species has not gone to completion (Wilding et al. 2000a). Therefore the pattern within currently rec- ognized species is likely to be complex, as noted here. Is there a general framework for these observations? Appar- ently an aspect of the polymorphism of L. saxutilis (and perhaps of direct-developing intertidal snails in general) is the repeated ap- pearance of similar phenotypes in different habitat regions of the shore, in response to similar selective pressures imposed by these habitats. Thus in the Galician region of Spain there are found a high-shore, ridged, banded and large morph (RB), and a low-shore, smooth, unhanded and small morph (SU) of L. saxutilis. These morphs show a variety of differences, considered to be at least partly genetically controlled, and are partially reproductively iso- lated (Johannesson et al. 1993). The likely selective factors are considered to be for small and slow growing snails in the lower shore, with larger and faster growing snails in the upper shore (Johannesson, Rolan-Alvarez and Erlandsson 1997). In the British Isles there are completely different morphs. referred to here as H and M, and it may be supposed that an important selective pressure is that of crab predation in the lower shore. Thus. L. saxutilis M closely resembles the common morph of L. compressa Jeffreys a low shore rough periwinkle, while L. suxatilis H is very like the typically higher shore L. airuini Hannaford Ellis. These H and M forms are like those in Sweden referred to as E and S (Janson 1982), but while in Britain there is evidence for partial reproduc- tive isolation between H and M (Hull et al. 1996, Hull 1998, Pickles & Grahame 1999), this has not been reported previously for the Swedish animals (Erlandsson & Rolan-Alvarez 1998). Our TABLE 3. Discriminant analysis with cross validation on canonical variance »ith haplotypes as predictors. Original haplolype group Destination haplotype group AAA BBA ADA BCA AAA BBA ADA BCA 62.50'7c \2.507c 12.5()V; 12.50% 11.70% 41.70% 23.30% 23.30% 22.00%^ 22.00% 41.50% 14.70% 38.50%- 19.20% 25.90%. 30.80%r observation of some difference in mtDNA haplotype frequency between them suggests that the E and S situation should be exam- ined further. The repeated nature of such phenotypic differences over large spatial scales (-1,000 km) together with the evidence of nascent reproductive barriers, yet superimposed on this undoubted evi- dence of gene flow between H and M (and E/S in Sweden, RB/SU in Spain), suggests analogy with the parallel speciation scenario proposed by Schluter and Nagel ( 1995) for sticklebacks. Here the proposed scale becomes important: it may be easier to envisage gene flow in Littoriiiu populations along the British coast than between stickleback populations in isolated lakes, but gene flow between snail populations in Spain and Britain may well be very small. In addition the parallel speciation scenario may be appli- cable even over smaller scales (within Britain, within Galicia). Evidence to date from Galician L saxutilis (Johannesson et al. 1983) and H and M around Britain for neutral loci (Wilding et al. in press) supports the interpretation that populations are more closely related at a site than between sites and yet display the same pattern of morphological differentiation. The application of additional markers with replicated samples will aid in uncovering the genetic basis underlying L. saxutilis H and M. Further repeated sampling, as for that implemented at Galloway, is also needed to test if the patterns at Mumbles and Ursholmen, Sweden are robust to repeated sampling or simply an artifact brought about by the small sample sizes. It is clear that much more needs to be done on the biogoegraphy, behavior, and TABLE 4. Nested analysis of molecular variance (AMOVA) of RFLP variation in Littorina. Source of variation % of total Group = shape zones among groups 6.78 among populations within groups 52.31 within populations 40.90 Group = 'H' and "M' among groups -6.70 among populations within grcnips 63.48 within populations 43.22 See Table 1 for details of placement of samples into particular groupings. 508 Wilding et al. genetic consitution of the forms of L. saxatilis as an interesting example of speciation in progress. ACKNOWLEDGMENTS This research was supported by the MAST-3 program of the European Commission under contract number MAS3-CT95-0042 (AMBIOS). CSW is grateful for a grant from the Genetical Society of Great Britain toward attendance at the 6th International Lit- torinid conference. We wish to thank David Reid of the Natural History Museum for the samples of Swedish Lirtoiiiui .sa.xalilis. Paul Chippindale and one anonymous referee provided helpful comments on this manuscript. LITERATURE CITED Ashbiirner. M. 1989. Drosaplula: a laboratory manual. Cold Spring Harbor Laboratory Press. Avise. J. C. 1994. Molecular Markers. Natural History, and Exolulion. London: Chapman & Hall. Boulding. E. G. 1990. Are the opposing selection pressures on exposed and protected shores sufficient to maintain genetic differentiation between gastropod populations with high intermigration rates? Hydrobiologia 193:41-52. Erlandsson. J. & E. Rolan-Alvarez. 1998. Sexual selection and assortative mating by size and their roles in the maintenance of a polymorphism in Swedish Litlorina saxatilis populations. Hyilrobiologia 378:59-69. E.xcoffier. L.. P. E. Smouse & J. M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: ap- plication to human mitochondrial DNA restriction data. Genetics. 131: 479-491. Grahame. J. & P. J. Mill. 1992. Local and regional variation in shell shape of rough periwinkles in southern Britain, hi: J. Grahame. P.J. Mill and D.G. Reid, editors. Proceedings of the Third International Symposium on Littorinid Biology. London: Malacological Society of London. 99- 106 pp. Grahame, J. & P. J. Mill. 1993. Shell shape variation in rough periwinkles: genotypic and phenotypic effects. In: J.C. Aldrich, editor. Quantified phenotypic responses in morphology and physiology. Ashford: JAPAGA. pp 25-30. Hellberg, M. E. 1998. Synipatric sea shells along the sea's shore: the geography of speciation in the marine gastropod. Tegiila. Evolution. 52:1311-1324. Hull. S. L., J. Grahame & P. J. Mill. 1996. Morphological divergence and evidence for reproductive isolation in Litlorina sa.xatilis (Olivi) in nonheast England. J. Moll. Stud. 62:89-99. Hull, S. L. 1998. Assortative mating between two distinct micro-allopatric populations o( Littorina saxatilis (Olivi) on the northeast coast of En- gland. Hydmbiologia. 378:79-88. Janson. K. 1982. Phenotypic differentiation in Littorina sa.xatilis Olivi (Mollusca, Prosobranchia) in a small area on the Swedish west coast. J. Moll. Stud. 48:167-173. Johannesson, K., B. Johannesson & R. Rolan-Alvarez. 1993. Morphologi- cal differentiation and genetic cohesiveness over a microenvironmental gradient in the marine snail Litlorina .m.xatilis. Evolution 47:1770- 1787. Johannesson, K.. E. RoUin-Alvarez & J. Eriandsson. 1997. Growth rate differences between upper and lower shore ecotypes of the marine snail Littorina sa.xatilis (Olivi) (Gastropoda). Biol. J. Linn. Soc 61:267-279. Kirby. R. R., R. J. Berry & D. A. Powers. 1997. Variation in mitochondrial DNA in a cline of allele frequencies and shell phenotype in the dog- whelk Nucella lapillus (L.). Biol. J. Linn. Soc. 62:299-312. Kyle. C. J. & E. G. Boulding. 1998. Molecular genetic evidence for parallel evolution in a marine gastropod. Littmina suhroliiiidala. Proc. R. Soc. Land. B. 265:303-308. Mill, P. J. & J. Grahame. 1995. Shape variation in the rough periwinkle Littorina sa.xalilis on the west and south coasts of Britain, Hydrobio- logia 309:61-71. Pickles, A. R. & J. Grahame. 1999. Mate choice in divergent morphs of the gastropod mollusc Litlorina sa.xatilis (Olivi): speciation in action? Anim. BWinv 58:181-184. Reid, D. G. 1996. Systematics and evolution of Littorina. Andover, En- gland: The Ray Society, 463 pp, Reist, J. D, 1985. An empirical evaluation of several univariate methods that adjust for size variation in morphometric data. Can. J. Zool. 63: 1429-1439. SAS Inc. S.I. (1990) S.AS/ST.Ar^ Users Guide. Version 6, founh edition, volumes 1 and 2. , North Carolina. SAS Institute. 1739 pp. Schluter, D, & L. M. Nagel. 1995. Parallel speciation by natural-selection. Amer Nat. 146:292-301. Schneider. S., J, M, Kueffer, D. Roessli & L, Excoffier, 1997. Arlequin version 1.0 Available at http://anthropologie.unige.ch/arlequin. Wilding, C. S,. J, Grahame & P, J. Mill. 1998. Rough periwinkle poly- morphism on the east coast of Yorkshire: comparison of RAPD-DNA data with morphotype. Hydrobiologia. 378:71-78. Wilding, C. S., P. J. Mill & J. Grahame. 1999. Partial sequence of the mitochondrial genome of Littorina sa.xatilis: relevance to gastropod phylogenetics. / Mol. Evol. 48:348-359. Wilding, C. S., J. Grahame & P. J, Mill, 2000a Mitochondrial DNA Col haplotype variation in sibling species of rough periwinkles. Heredity 85:62-74. Wilding. C. S.. J. Grahame & P. J. Mill. 2000b. Nuclear DNA restriction site polymorphisms and the phytogeny and population structure of an intertidal snail species complex (Littorina). Herediias 133:9-18, Wilding. C. S.. R. K. Butlin & J. Grahame. In press. Differential gene ex- change between parapatric morphs of Littorina sa.xatilis detected using AFLP markers. / Evol. Biol. In press. Jimrmil ,tf Shellfish Research. Vol. 20. No. I, 509-530. 2001. ABSTRACTS OF TECHNICAL PAPERS Presented at The 21st Annual Meeting MILFORD AQUACULTURE SEMINAR Milford. Connecticut February 26-28. 2001 509 i Milt'ord AquacLiltiire Scniiiiar. Miltord. Connecticut Abstracts. Febmai^' 200! 311 CONTENTS Walter J. Blogoslawski Over\ icw, 2 1 "' Milford Aquaculture Seminar 513 Bethaiin Balazsi and Gary H. Wikfors Water Quality and Nutritional Value of Green Water vs. Clear Water Culturing of the Marine Fish. Tautogolabrus adspersus 513 David Berlinsky, Rachel Howell, Jessica Henderson, Mark Watson and Terence Bradley Effect of Sal HI It V on .Sur\ i\al and Growth of Early Life Stages of Black Sea Bass 513 Diane J. Brousseaii and Amy Filipowicz A Theoretical Estimate of the Potential Impact of Asian Shore Crab Predation on Bay Scallops 514 Joseph Choromanski and Sheila Stiles Prclmmiary ln\ estigations of Crab Predation on Bay Scallops 514 Todd Corayer Report on Design of Photovoltaic-Powered Floating Upweller System for Growth of Crassostrea virginica and Merccnariii inercenaria 514 Yvonne Coursey and Mary Kimble De\elopment of the Digestixe Caecum During the P' Larval Stage in the Horseshoe Crab. IJinuhts polrphemus 515 Carmela Cuomo and Paid R. Bartholomew Horseshoe Crab Aquaculture: Results from Initial Spawning Studies 515 John J. Curtis, Sherry W. Lonergan, Paul J. Trupp, Jose Zertuche, Rachel Carmona and Charles Yarish A Cooperative Study on the Culture of Clumilnis crispiis (Rhodophyta) for Possible Commercial and Bioremediation Applications in Long Island Sound 516 Maureen Davidson The Effects of Stocking Density in Pearl Nets on the Survival, Growth, and Reproductive Potential of the Bay Scallop, Argopecten irradians irnulians 517 Christopher V. Davis Design and Operation of an Airlift Driven Floating LIpweller System 517 Jasmine DeCrescenzo, Joseph DeCrescenzo and Inke Sunila Growth and Mortality of the Eastern Oyster {Criissostrcii virgiiiicn) and the Quahog iMeneiuirici inercenaria var, notata) in a Taylor Float in Milford Harbor 517 Mark S. Dixon, Gary H. Wikfors and Bethann Balazsi Rotifer Production on Microalgal Diets: First Steps Toward Process Engineering 518 Shuyun Dong and Sylvain De Guise Development of Assays to Evaluate Immune Functions of American Lobsters (Hoiuarns anicricaniis) and their Use in Field Studies 518 Gef Flimlin Importation of Chinese Clams Causes Problems in Local Markets 518 Thommai A. Francis Comparative Efficiency of Hormones on the Maturation and Breeding of Indian Catfish. Hcicrnpnciistcs fossitis 519 Eric Ganger and Marta Gomez-Chiarri Characterization of Vibrio carchariae as a Pathogen of Summer Flounder (Paraliclnliys Jenfiilus) 519 Michael J. Goedken and Sylvain De Guise Flow Cytometry as a Tool to Quantify Oyster Defense Mechanisms 519 David W. Grunden Comparison of two Trap Types Used in a Municipal Green Crab Predator Control Program 519 Rachel Howell, David Berlinsky and Terence Bradley Effects of Photoperiod Manipulation on Reproduction in the Black Sea Bass. Cenlroprislis striata 520 Richard C. Karney. John C. Blake and Thomas E. Berry Aqua-tourism. Turbo-tidal LIpwellers, and Other Developments in the Shellfish Industry on Martha's Vineyard in the Year 2000 520 Nick King Sinner Flounder (Paralichtlns dcnialus) Production at GreatBay Aquafarms. Inc 521 Hauke L. Kite-Powell and Porter Hoagland The Econoimcs of New England Finfish Growout: Aquaculture at an Offshore Site 521 512 Abslnicis. February 2001 Milford Aquaculture Seminar. Milford, Connecticut Jill LaBaiica and hike Siinila The Role of Apoptosis in the Pathogenesis of theEastem Oyster, Crassostrea virginica 521 Kenneth J. LaValley Effects of Nursery Culture Technique on the Morphology and Borrowing Capability of the Softshell Clam. Mya areiuiiia 52- Dale Leavitl and William Burt The Razor Clam (Ensis JirectKs) as a Candidate for Culture in the Northeast: An Introduction 522 Jennie M. Mandeville, Mark T. Watson, and Brandy M. Moran Preliminary Studies on Optimal Rotifer Culture Diets and Alternative Replacement Diets and Enrichments for Larval Black Sea Bass ( Centropiistis stricila ) 52.^ Karen Mareiro, Josefa Dongal. Meggan Dwyer, Kenneth Leonard III, Marta Gomez-Chiarri, and Arthur Ganz Prevalence of Perkinsus nuirimis in the Eastern Oyster. Crassostrea virgiuica. in Rhode Island 523 Christopher Martin Is Sinilpiiliiim zoophthorwn the Animal Eater its Name Suggests? New Evidence of Parasitism 524 Brandy M. Moran, Mark T. Watson, and Cliff A. Goudey Potential Effects of Contaminant Exposure on Cultured Tautog. Tauloga onitis 524 Dana L. Morse Bottoms Up! An Industry-led Project: Bringing an Aquaculture Technique to the Inshore Scallop Fishery in Maine — 525 Jennifer Mugg, Michael A. Rice, and Monique Perron Elects of Filter-Feeding Oysters on Sedimentation Rates and Phytoplankton Species Composition: Preliminary Results of Mesocosm Experiments 525 Dean M. Perry, David A. Nelson, and Robin S. Katersky Laboratory Culture of Larv al Tautog: Recent Updates and Changes 525 Edwin Rhodes Department of Commerce Aquaculture: An Update of Policies. Plans and Programs 525 Gregg Rivara, Timothy Caufield, and Walter Smith Keeping Gentrification at Bay: How a Defunct Shellfish Facility was Kept from the Developers 526 Karen Rivara Development of the Noank Aquaculture Cooperative in Connecticut 526 John Roy, Amber L. Beitler, and Kathryn R. Markey A Comparison of the Growth and Mortality of Juvenile Argopecten irradiaiis in the Various Culture Methods Employed by the Students at the Sound School Regional Aqacluture Center 526 John Sarkes Federal Crop Insurance for Quohog Farms 527 Kim W. Tetrault, Robert M. Patricio, Rory MacNish, and Jonathan P. Polistena An Introduction to the S.P.A.T. (Special Programs in Aquaculture Training Initiaitive 527 Bethany A. Walton and Brian F. Beal Broodstock Conditioning of the Sea Scallop. Placopecten imigelUmiciis. with Various Microalgal Diets 528 William C. Walton Removal of Blue Mussels. Mytihis fdulis. From Mudflats: Does it Improve Softshell Clam Mya arenaria. Habitat?.... 528 Scott Weston, Joseph Buttner, and Mark Fregeau Nurturing a Softshell Clam Priv ate/Public Initiative on Massachusetts' North Shore 528 James C. Widman Jr. and David Veilleux Growth and Survival of Bay Scallops. Argopecteii irradians irradians. Fed Tetraselmis chiii by Two Methods 529 Gary H. Wikfors, Jennifer H. Alix, Mark S. Dixon, and Barry C. Smith Effect of Feeding Ration and Regime upon Growth and Food-Convention of Juvenile Quahogs. Mercenaria mercenaria. and Comparison w ith Bay Scallops and Eastern Oysters 529 Mlllord Aqiiaculture Seminar. Milford, CoiinccliLUt Abstracts, February 2001 513 OVERVIEW. 21' MILFORD AQUACULTURE SEMINAR. Walter J. Blogoslawski, U.S. Department of Commerce. National Oceanic & Atmospheric Administration. National Marine Fisher- ies Service. Northeast Fisheries Science Center, Miltord Labora- tory. 212 Rogers Ave.. Milford. CT 06460 With over 170 registrants representing industry, research, and academic interests, the 2r' Annual Milford Aquaculture Seminar had Ihe highest attendance ever for this series of meetings. This blend of both the theoretical and practical aspects of aqua- culture ensured a meeting that permitted attendees to be exposed to areas of the practice of aquaculture outside their own expertise and provided a forum where the latest innovations were introduced and discussed. The formal papers and posters were presented by persons from 10 U.S. coastal states and the countries of India and Canada. They represented members of three vocational aquaculture high schools. 14 universities, tlve marine labs, and several state and federal employees involved in shellfish and finfish aquaculture. Topics included aquaculture policy, crop insurance, education, disease, nutrition, and culture techniques. Studies provided information and experiences from laboratory research as well as full-scale com- mercial installations. The Seminar has developed a tradition of offering the latest information available in the field in an informal atmosphere. This has succeeded in promoting a free exchange among all with an interest in the success and future of aquaculture. This Seminar continued that approach, allowing all attendees to enjoy and learn from the formal presentations, while also affording informal op- portunities to discuss the latest developments pertinent to this im- portant, expanding field. The meeting was sponsored by the US Department of Com- merce. NMFS Milford Laboratory. Milford. CT; abstract printing was courtesy of the US Department of Agriculture. Northeastern Regional Aquaculture Center. N. Dartmouth. MA. Their support is greatly appreciated. WATER QUALITY AND NUTRITIONAL VALUE OF GREEN WATER VS. CLEARWATER CULTURING OF THE MARINE FISH, TAUTOGOLABRUS ADSPERSUS. Bethann Balazsi, Southampton College of Long Island Univer- sity. Southampton. NY 11968. and Gary H. Wikfors, USDOC. NCAA, National Marine Fisheries Service. Northeast Fisheries Science Center. Milford Laboratory. Milford. CT 06460 Experiments were designed to test the relative importance of nutritional and water quality factors in green vs. clear water cul- turing of marine finfish larvae. Two microalgal strains, Nan- nochloropsis sp. (UTEX2341) and Isoihrysis sp. (T-ISO), were used. Algae were added to fish tanks either in suspension or im- mobilized in beads or in a dialysis cassette. Two fish. Tautogola- bnis cidspersus (canners). were placed into each tank. Artemia was the food source. Artemia counts were done every day. as were per cent transmittance readings on algal samples from treatments with suspended algae. This was done to determine feeding rates of fish on artemia and of artemia on algae. Water samples were taken periodically throughout the experiment and analyzed for ammonia. The volume displacement was taken as the cunners were added to the tanks, and also at the termination of the experiment. Immobilized algae in alginate beads was found to be unstable in seawater. Dialysis cassettes were found to be successful for eight days before the dialysis membranes losl integrity. In the dialysis cassette experiment, artemia were replenished lOO^f when resources were abundant. Algae were replaced each day. It was found that the anemia were cleared within 10 minutes to a few hours, depending on the treatment. The percentage of algae to replace declined in respect to the rate of clearing artemia. Volume displacement measurements did not show any significant growth of the cunner. Algae immobilized in dialysis cassettes removed ammonia from solution, but not as effectively as algae in free suspension. Nevertheless, these findings show dialysis cassettes are a useful tool in understanding how green water cultures work. Funding provided by a Northeast Fisheries Science Center. Southampton College of Long Island University Cooperative Ma- rine Education and Research (CMER) grant. EFFECT OF SALINITY ON SURVIVAL AND GROWTH OF EARLY LIFE STAGES OF BLACK SEA BASS. David Berlinsky. Rachel Howell, Jessica Henderson, Mark Watson and Terence Bradley, University of Rhode Island, Fisheries. Ani- mal, and Veterinary Science. Kingston. RI 02881 The black sea bass. Centropristis striata, is currently being investigated as a candidate for commercial aquaculture. One ob- stacle to production is high mortality rates during early life stages. Environmental salinity has been implicated as an important factor influencing larval survival in numerous species of finfish. The present study was conducted to determine the effects of a range of salinities on the growth and survival of black sea bass at several larval stages. Experiment 1 examined larval survival through yolk- sac absorption (3 days post hatch; dph) at salinities of 0. 5. 10. 15. 20. 25. 30 or 35 parts per thousand (ppt). Ten fertilized eggs were cultured in triplicate at each of the salinities in 500 ml glass con- tainers. At 3 dph. the number of live larvae for each treatment was determined. In experiment 2. 380 actively feeding larvae (8 dph) were cultured in 40 L glass tanks at salinities of 1 5. 20. 25 or 30 ppt. Each treatment was replicated in triplicate. At 15 dph. the number of surviving larvae was enumerated for each treatment and a sample of 12 larvae from each tank was preserved in 10% neutral buffered formalin for measurement of length. In experiment 3. 200 actively feeding larvae (23 dph) were cultured in 40 L glass tanks at salinities of 15, 20. 25 or 30 ppt. Each treatment was replicated in triplicate. At 32 dph. the number of surviving individuals was enumerated for each treatment and a sample of 12 larvae from each 514 Abstracts. February 2001 Milford Aquaculture Seminar, Milford, Conneclicut tank was preserved in 107r neutral buffered formalin for measure of length. Results from experiment 1 indicate that salinities of at least 10 ppt are required for survival of black sea bass larvae. Survival improved from 42% at 10 ppt to 77% at 15 ppt to >90% at 20-35 ppt. In experiment 2, survival (67%) was higher at a salinity of 30 ppt than at 25 (41%), 20 (47%), or 15 (41%) ppt. No significant differences in growth were observed among treatments. In experi- ment 3, no significant differences in survival or growth of larvae were observed among the treatments. These studies indicate that black sea bass larvae require at least 10 ppt salinity and become more tolerant to lower environmental salinities with de\elopment. A THEORETICAL ESTIMATE OF THE POTENTIAL IM- PACT OF ASIAN SHORE CRAB PREDATION ON MUS- SEL SETTLEMENT. Diane J. Brousseau and .Amy Filipowicz, Fairfield University, Biology Department. Fairfield, CT 06430 The Asian shore crab, Hemigrapsus sanguineus, readily con- sumes juvenile blue mussels. Mytilus edulis. In laboratory feeding experiments {n = 59: sexes pooled) each Hemigrapsus ate an av- erage of 12.7 ± 1 1.6 mussels (60 mm) caught by recreational fishermen in the towns that have such policies. Field studies of bay scallops have suggested a variety of causes for population fluctua- tions including habitat loss, genetic inbreeding depression, and predation. It is generally known that crab predation can be a major REPORT ON DESIGN OF PHOTOVOLTAIC-POWERED FLOATING UPWELLER SYSTEM FOR GROWTH OF CRASSO.STREA MRGINICA AND MERCENARIA MERCE- NARIA. Todd Cora.ver, Block Island Shellfish Farm. Ocean Ave.. P.O. Box 1342. Block Island. RI 02807 The intent of this experiment was to design, construct, and evaluate the performance of a floating shellfish upweller which would offer the same growth enhancement opportunities of a tra- ditionally powered unit, yet be completely independent of shore- side power. The unit would be adaptable to salt pond shellfish farms where traditional power was unavailable or cost prohibitive. The concept and construction was partially funded through an R&D Partnership Award from the Slater Center for Ocean Tech- nology at the University of Rhode Island and through an allcrna- tive energy grant from the United States Department of Energy. Block Island Shellfish Farm (BISF) formed a partnership with Chris Warfel PE. owner of Entech Engineering, specialists in solar Milt'did AqiiacLilture Seminar. Milford. Connecticut Ahsimcls. Februar> 200 1 515 and wind energy systems, to facilitate the electrical engineering of the upweller. The upweller consists of single and double-tiered PVC silos with submersible pumps positioned atop. The shellfish were di- vided between the silos with a test case being established without the aid of a water pump to identify any ancillary growth assistance from the structural design of the upweller. Another test case was located in the abutting lease of Block Island Shellfish Farm (BISF). These test case animals were placed in 1 mm ADPI grow- out bags and maintained along with BISF's inventory of Cnissos- trea virginica and Merceiuiria mercenaria. Animals in both sites were measured for volume and size changes regularly, and col- lected data were evaluated against the traditional farm-raised test cage. Provisions were incorporated into the design to prevent the silos from coming into contact with the pond bottom in situations of extreme low water events. Constructed of pressure treated lum- ber, the unit was also designed to survive any severe weather which a coastal pond might endure. An insignificant amount of heavy metals leaching was detected as a result of systematic water testing inside the upweller as well as up and down stream. At this point in our research, regular measurement recorduigs indicate a successful mating of the photovoltaic cells and the spe- cially designed upweller silos. Essential to the success were ani- mal-to-screen densities and the ability to regulate water flow rates. Experiments are ongoing to refine the structural elements of the unit and the silos to decrease maintenance requirements. DEVELOPMENT OF THE DIGESTIVE CAECUM DURING THE FIRST LARVAL STAGE IN THE HORSESHOE CRAB, UMULUS POLYPHEMUS. Yvonne Coursey and Mary kinible. Department of Biology. University of South Florida, Tampa, FL 33620 The internal development of Linnihis polypliemus is not com- plete at the time of hatching from their embryonic membranes. A 1st instar, an animal that has just hatched from its egg casing, does not have a complete digestive system. It lacks the digestive cecum, which branches off from the intestine, and thus can't absorb nu- trients. Instead, the 1st instar relies on the yolk present in coni- partmented areas where the gut will arise. The foreaut. which includes the esophagus, crop, and giz/ard. is fully formed in the 1st instar. In addition, the hindgut is also present. The intcsline and the digestive cecum de\elop as the animal progresses from a 1st to a 2nd instar, approximately 1 1-14 days. In this time period the mid- gut and digestive cecum develop, becoming contiguous with the foregut and the hindgut. Juvenile horseshoe crabs. 1st and 2nd mslars (a 2nd instar is a 1st instar that has molted.), were fixed in Bouin's tliiid or a 3'/r formaldehyde/seawater solution. They were dehydrated through an alcohol gradient series, and embedded in Unicryf" or Spurr"s resin. Sections were cut at 3-4|j.ni for light micro.scopy and 60-90 nni for electron microscopy. This study determined the internal morphology of P' and 2'"' instars as a precursor to the identification of the site(s) of hemo- poeisis in Liiiiiihis pohphciinis. Horseshoe crab blood cells are used to produce Liimdus amebocyte lysate (LAL), which is used to test for sterility of intravenous solutions. A long-term goal of our research is to establish methods for culturing horseshoe crab cells /;; vitro. HORSESHOE CRAB AQUACULTURE: RESULTS FROM INITIAL SPAWNING STUDIES. Carmela Cuomo, Yale Uni versity. Department of Geology & Geophysics, New Haven, CT 06520, and Paul R. Bartholomew. SUPERB Technical & Envi- ronmental, Hamden. CT 065 1 7 The horseshoe crab, Linndus polypliciniis. is considered by many to be a living fossil, more for its primitive appearance than for its actual ancestry. The actual species dates back only about 20 million years, although its relatives can be traced back over 200 million years in the fossil record. At the present moment in time it IS. perhaps, more important to the biological community than to the geological one. Sea turtles and at least 20 species of Arctic- bound shorebirds feed upon Linudiis eggs deposited in the shallow tidal flats all along the eastern seaboard of the United States during their spring migrations. Several critical "staging areas"" for these birds, including Milford Point. CT. are also known to be "critical habitats" for Liimdus spawning. Fuilhermore, Linndus adults are harvested for "bleeding"" by several biomedical companies. Liinu- lus blood is a copper-based blood that coagulates in the presence of gram-negative bacteria. Several companies have developed a biomedical assay, known as Liimilus amoebocyte lysate (LAL), which is used to test for human diseases (e.g. spinal meningitis and gonorrhea), the purity of pharmaceutical products, and clinical laboratory water purity. The demand for LAL doubled between the 1980s and the 1990s, and doubled again between 1990 and 2000. It is predicted to keep rising well into the 21st century. Coincident with the increased demand for Limuhis has been a drastic decline in the Limulus populations along the central and northeastern Atlantic Coast of the United States. This decline has been attributed to many factors including over-fishing, declining spav\ning grounds, deliberate destruction, increased harxesting for LAL, and pollution. In 199S. the Atlantic Marine Fisheries Com- mission prepared a report on the status of the horseshoe crab populations along the east coast of the United States and a man- agement plan for the fishery. The central goal of this plan was to conserve and pi'otect the horseshoe crab resource in order to main- tain sustainable levels of standing stock to ensure both its impor- tant role in natural ecosystems, as well as its role in commercial enterprises. Although this is a good initial step, it does not repre- sent the solution to the declining horseshoe crab populations. Breeding grounds are likely to continue to decline due to increased 516 Abstmcls. February 2001 Milford Aquaculture Seminar. Milford. Connecticut shoreline development, increased bii)niedical demands on horse- shoe crab, and inability to maintain a steady-state. More fishermen are turning to commercial eel and conch fishing as other fishing stocks decline. The solution to the problem lies in the ability to produce a reliable stock that can be maintained even as natural stocks undergo population fluctuations. To this end. in the summers of 1999 and 2000. preliminary studies were undertaken at the NMFS laboratory in Milford, CT to test whether or not it was possible to induce Linnilus adults to spawn in captivity. Various factors were tested for their impor- tance, including sediment grain size, tidal height, light, and tidal regime. The results from this research have proved promising 1 pairs of adult Limuhis successfully mated, resulting in the produc- tion of thousands of eggs. The eggs were allowed to develop under normal conditions and had a hatching rate of approximately 609f over a two month period. Experiments evaluating the long-term survival and growth of the larvae are presently underway. It is our goal not only to rear the eggs successfully through hatching, but also to rear the hatchlings successfully through at least the first seven molts. It is anticipated that the results from these experi- ments will translate directly to industry as they will provide horse- shoe crabs to companies that need them for years to come. A COOPERATIVE STUDY ON THE CULTURE OF CHON- DRUS CRISPUS (RHODOPHYTA) FOR POSSIBLE COM- MERCIAL AND BIOREMEDIATION APPLICATIONS IN LONG ISLAND SOUND. John J. Curtis. Sherry W. Lonergan and Paul J. Trupp. Bridgeport Regional Vocational Aquaculture School. 60 St. Stephens Rd.. Bridgeport, CT 06605; Jose Zer- tuche. University of Connecticut. Stamford. CT 06901, and Uni- versidad Autonoma De Baja California, Ensenada. B.C.. Mexico; Rachel Carmona and Charles Yarish. University of Connecticut, Stamford, CT 06901 As part of an agreement with the University of Connecticut at Stamford, students and stalf at the Bridgeport Regional Vocational Aquaculture School are participating in a cooperative study on the feasibility of growing Chondrus crispiis (Irish Moss) in Long Is- land Sound for commercial puiposes, and as a possible candidate for bioremediation. Dr. C. Yarish. Professor of Ecology and Evo- lutionary Biology, and Dr. J. Zertuche, a visiting scientist and Professor from Universidad Autonoma De Baja California, for- mally introduced Chondrus crispus culture techniques used on Prince Edward Island. Canada and Baja. California to our school as a project proposal. Since the initial introduction, these scientists and other UConn faculty and staff have visited iiur school to teach the students the life history and biology of Clumdnis. and also the design and construction of the equipment necessary for its experi- mental commercial culture The first task was to assemble the components necessary to maintain and nurture the Chdiidnts in the laboratorv settina. An available .61 x 1.20 x 2.44 m (height x width x length) gel-civited fiberglass tank with a gray interior was selected. To provide aera- tion and the required water turbulence, a 1 horsepower regenera- tive blower was used to inject air, through the use of a manifold, into four (4) 1.27 cm PVC pipes. Each 1.27 cm pipe was drilled, in a single horizontal row, with 1.6 mm diameter holes spaced 2.54 cm apart. The drilled pipes were placed in the bottom of the tank with the holes facing upward and running parallel to the 2.44 m of the tank. They are equally spaced from each other and leveled to provide for an equal amount of air discharge for a consistent tum- bling action. Suspended above the tank is a 1000-watt metal-halide light fixture delivering 400 |jl mol photons ni~" s^'. which is on a timed cycle 10:14. L;D cycle. The water is obtained from Long Island Sound and maintained in the tank at a salinity of 27 ppt and at 15 C through the use of a chiller. The water depth is maintained at .36 m to allow for the correct tumbling action and with consis- tent tissue-to-water volume ratio (500 p,m in diameter were in- duced to spawn by implantation of 25-50 |jLg leutinizing hormone- releasing hormone analog (LHRH,). Females reared under the compressed cycle attained follicles 500 |j.m in diameter and were induced to spawn in March, approximately 2 mo in advance of individuals reared under the simulated natural photoperiod. Males on both cycles began spermiating in late January and continued through the duration of the experiment. Measurement of plasma levels of 173-estradiol. testosterone, and 1 1-ketotestosterone are ongoing and will be presented. AQUA-TOURISM, TURBO-TIDAL UPWELLERS, AND OTHER DEVELOPMENTS IN THE SHELLFISH INDUS- TRY ON MARTHAS VINEYARD IN THE YEAR 2000. Richard C. Karney. Martha's Vineyard Shellfish Group. Inc.. Box 1552. Oak Bluffs, MA 02557; John C. Blake, Sweet Neck Farm. Box 1468, Edgartown, MA 02539; and Thomas E. Berry, Martha's Vineyard Shellfish, Box 1660, Edgartown, MA 02539 When polled, shellfish growers in Massachusetts have consis- tently identified the general public's perception of aquaculture as one of the most important factors affecting the growth of aquacul- ture in the Bay State. The strong and enduring tradition of active government participation by the Massachusetts citizenry makes a public education project which stresses the benefits of shellfish aquaculture crucial to the development of an industry that relies on the use of the marine common lands. Under grant funding from the Massachusetts Department of Food and Agriculture, the Martha's Vineyard Shellfish Group conducted a collaborative, niultifaceted public relations project with the goals of showcasing local public and private aquaculture operations and extolling the environmen- tal, economic, and social benefits of the shellfish aquaculture in- dustry. We employed a multitude of communication outlets to spread the gospel of aquaculture. The project included aqua-tours of private shellfish farms, web site development, tastings of cul- tured shellfish, public access cable programs, posters, and the dis- tribution of printed information. With financial support from the National Fish and Wildlife Foundation, five growers cultured over a million seed oysters in tidal upweller nurseries. Some of the nurseries were modified with the addition of side vents. Although these "turbo-charged" nurs- eries experienced an improved flow rate, there appeared to be no improvement in oyster growth rate compared to nurseries without the added vents. Both public and private oyster stocks were tested for disease. Private growers in Katama Bay were relieved to tlnd that their cultured oysters, which were believed to be infected with MSX {Haplosporidiiim iwlsonil upon PCR analysis, were found to be infected with the less serious SSO {Haplosporidiiim costale). Oys- ters sampled for Dermo (Perkinsiis marimis) from public beds in Tisbury Great Pond showed an increased infection from 8'7r in 1999 to over 50% in 2000. We report the discovery of an unusual flattened form of Mer- Milt'ord Aquacultiire Seminar. Miltbrd. Connecticut Abslracl.\. February 2001 521 cenaria mercenaria from a sand bottom in Katama Bay. It mea- sured 57.8 mm in length. 49.3 mm in height, and 19.5 mm m width. SUMMER FLOUNDER (PARALICHTHYS DENTATUS) PRODUCTION AT GREATBAY AQUAFARMS, INC. Nick King, GreatBay Aquafarms. Inc.. 153 Gosling Road. Portsmouth. NH 03801 Since 1996. GreatBay Aquafarms (GBA) has been producing summer flounder at its hatchery in Portsmouth. NH. Recently. GBA has sent its first major crop of flounder to market from its demonstration growout facility that utilizes re-circulation technol- ogy developed by the company. Survival offish through the hatch- ery phase (6 mo) tends to be less than 30'7f and varies with pro- duction lot. Growth of individuals differentiates following meta- morphosis, and by six months considerable variation in body size occurs around the mean size of 10 g. Size grading begins at 70 days (ph) and is necessary through the growing cycle to avoid nipping and to ensure proper feed management. Currently, fish reach an average size of 650 grams in two years from hatching, and our studies have shown that growth rates differ between the sexes during commercial ongrowing. Great strides toward improving production standards and improving the growth curve for summer flounder can be obtained through advances in husbandry, nutrition, fish health, and selective breeding. THE ECONOMICS OF NEW ENGLAND FINFISH GROW- OUT: AQUACULTURE AT AN OFFSHORE SITE. Hauke L. Kite-Powell and Porter Hoagland. Marine Policy Center. Woods Hole Oceanographic Institution. Woods Hole. MA 02543 We have developed a bioeconomic model of an open ocean finfish growout operation. The model optimizes stocking and har- vesting schedules, and projects financial flows. It allows for com- parison of alternate growout sites based on their physical charac- teristics (distance from shore, water temperature, water depth, etc.). The model takes into account seasonal variability in the price of fish landings as well as the effect oi water temperature on fish growth rates. We have illustrated the use of the model by applying it to hypothetical growout operations for cod. salmon, and flounder off the coast of New England. The model's optimization procedure assumes that the growout operation is to produce a fixed amount of fish (v,,. by weight) each month (or in specified months only). The model determines the optimal stocking time and number of fish for each harvest month. It also calculates expected financial flows and summary values such as project NPV and the amount of up-front investment re- quired. For each harvest niniith h (h= 1.2.3 12). the model uses a species-specific growth function to calculate the weight at harvest of an individual fish (f,,(m)) stocked as a fingerling in month m (m = h-23. h-22. h-21. ...). The model then calculates the number of fish at harvest n,,(m) = V|,/f|,(m). and works backward, using the mortality function, to calculate the number of fingerlings to be stocked. For each harvest month h. the model then identifies the stock- ing month m that results in the maximum net revenue (discounted difference between revenue and variable cost). Revenue is the product of harvest weight and price, which varies with fish size and time of year; v^,*p(h.f,,(in)l. (Stocking months that result in sub-market-size fish result in zero revenue and are not considered.) Variable costs include the cost of fingerlings. feed and medication, and harvesting, including associated vessel costs. The maximum net revenue determines the optimal stocking month (and length of growout). as well as the number of fingerlings. Once stocking decisions have been optimized, the model cal- culates the financial performance of the growout operation month- by-month over 1 5 years to determine projected cash flows, project NPV. and investment capital needed, as well as operational pa- rameters such as vessel utilization and feed volume. The table below summarizes model results for three species. TABLE \. Cod Salmon Flounder number of cages 72 72 6 fingerlings/year 1 .78S.()()t) 532.000 261.000 tons of feed/year 5.416 5.281 474 boat days/year 81 77 14 average harvest weight, g 1.466 4.093 1.025 harvest, tons/month 175 151 125 NPV at 5% discount rate. $m 1 .50 14.16 1.07 investment capital required $m 8.61 9.24 1.17 THE ROLE OF APOPTOSIS IN THE PATHOGENESIS OF THE EASTERN OYSTER, CRASSOSTREA VIRGINICA. Jill LaBanca. Fairfield University. Fairfield. CT 06430; Inke Sunlla. State of Connecticut. Department of Agriculture. Bureau of Aqua- culture. P.O. Box 97. Milford. CT 06460 Apoptosis. programmed cell death, is an essential part of cell renewal and embryonic development. During apoptosis. cells un- dergo shrinkage and zeiosis. or blebbing. Endonucleases digest DNA into 200 base pair fragments. Later, the cell breaks into apoptotic bodies that are ingested by phagocytes. Apoptosis. or the lack of apoptosis. is also pivotal in the pathogenesis of several different diseases. Peikinsus muiinus (Dermo). Haplosporidiwn nehoni (MSX). and Juvenile Oyster Disease (JOD) are common diseases of the eastern oyster (Crassiistrea virginica) along the New England shoreline. The complete pathogenesis of these dis- eases is still unknown. We studied the effects of apoptosis on the progression of these diseases. Archived oyster tissues from the State of Connecticut's Bureau of Aquaculture were classified according to the type of infection 522 Abstnicls, February 2001 Milford Aquaculture Seminar. Milford. Connecticut and fixe categories were created: oysters infected with only Dermo. oysters infected with only MSX, oysters infected with only JOD. mature oysters with no infections, and juvenile oysters with no infections. Ten oysters from each group underwent in situ hy- bridization to detect DNA fragments by end labeling. A 6 (jim section from each oyster was deparaffinized and digested with proteinase K. A digoxigenin nucleotide probe was used to anneal to and label the 200 base pair fragment yielded by apoptosis. The probe was detected immunohistochemically using an antidigoxi- genin peroxidase conjugate. The complex was stained with a per- oxidase substrate and the slide was counter-stained with methyl green. Apoptotic hemocytes were found to be present in the gill and stomach epithelia and surrounding connective tissue at a regular rate in both sets of healthy oysters. The oysters infected with Dermo had a reduced number of apoptotic hemocytes present in their tissues. These oysters had Dermo weighted prevalences from 0.5 to 4. The prevention of hemocyte apoptosis yields a greater number of hemocytes in which Dermo houses itself. Large num- bers of Dermo cells in some infected oysters were eliminated via apoptosis in the stomach epithelia, disabling the spread of infec- tious particles through sea water. The oysters infected with MS.X also had reduced numbers of apoptotic hemocytes. MSX appears to prevent hemocyte apoptosis. In some infected oysters large numbers of MSX-plasmodiu were eliminated via apoptosis. Part of the vesicular connective tissue cells of MSX-infected oysters were apoptotic. Oysters infected with JOD had the presence of large, unidentified, apoptotic cells in the stomach epithelia whose role has not yet been determined. Apoptosis enhances progression and prevents transmission of oyster diseases. The ability of oysters to eliminate parasites via apoptosis may establish the genetic basis of disease resistance. EFFECTS OF NURSERY CULTURE TECHNIQUE ON THE MORPHOLOGY AND BURROWING CAPABILITY OF THE SOFTSHELL CLAM, MYA ARENARIA. Kenneth J. La- Valley, Spinney Creek Shellfish. Inc.. Eliot. ME 03903 In August of 1999 trials began to evaluate the benefits of a raceway nursery system compared to a Floating Upweller System (FLUPSY). Raceways are shallow rectangular trays into which water is introduced at one end, flows over the seed clams, and exits at the other end through a drain. During normal operation, silt from the estuarine waters and pseudofeces from shellfish produce a substrate. This substrate would be detrimental to other bivalve shellfish (oysters, quahogs) and would represent an intense main- tenance problem. However, to the juvenile softshell clam the sub- strate represents an artificial mud flat. It was initially hypothesized that the ability to burrow into a substrate would allow enhanced growth by reducing the physi- ological stress experienced while in an upweller system. The an- terior and posterior adductor muscles, having evolved to suit a habitat several inches beneath the sediment, are unable to maintain a "normal" gape or closed position outside of sediment. This con- dition results in a physiological stress, which incurs abnormal growth. During initial FLUPSY trials a significant stress (reduced growth, shell deformities) was observed after juveniles had reached 15 mm. To determine the effect of initial size and planting density, two size ranges of softshell clam seed, 6 and 9 mm, were planted at three densities (4, 8, and 16 clams/square inch) within the raceway. In order to compare the performance between the two nursery techniques both size ranges were stocked in the floating upweller system at a density of 8 clams per square inch. The weekly growth rates were consistently higher at the raceway than the floating upweller. The raceway clams grew at a rale of approximately 2 mm per week, while the upweller clams did so at a range of .75 to 1 .0 mm per week. Stocking density did not significantly affect growth rate of raceway clams. Following growth and density comparisons, culture observa- tions were statistically evaluated by comparing "wild" softshell clam seed to cultured animals, as well as by comparing the mor- phology of market size "wild" softshell clams to cultured seed. We also determined and compared the burrowing capability of "wild" and cultured seed. Softshell clams produced in the FLUPSY significantly differed in their morphology compared to clams cultured in the raceway or "wild" seed. Clams cultured in the FLUPSY also demonstrated a significant decrease in their ability to burrow compared to raceway or "wild" seed. The percentage of clams that successfully bur- rowed in a two hour period was proportionally lower in seed produced in the FLUPSY (65% compared to 80%), Raceway seed did not significantly differ from "wild" seed or market size animals in their morphology or ability to burrow. In summary, nursery technique can significantly affect the morphol- ogy and behavior of cultured shellfish. These differences may significantly impact initial survival of softshell clam .seed post- planting. For softshell clams, a sediment-filled culture environ- ment appears to be the optimal nursery strategy. THE RAZOR CLAM [ENSIS DIRECTUS) AS A CANDI- DATE FOR CULTURE IN THE NORTHEAST: AN INTRO- DUCTION. Dale Leavitt and William Burt, Southeastern Mas- sachusetts Aquaculture Center. Hurley Library, Mass. Maritime Academy, 101 Academy Dr.. Buzzards Bay, MA 02532 There is an urgent need for the shellfish culture industry in the northeastern United States to expand their list of candidate species for culture and to diversify their crop. Reliance on two species of bivalve mollusk, given the historic and cuiTent prevalence of de- bilitating diseases, may result in lost opportunities to use sites for farming shellfish. By expanding the selection of candidate species, growers will have better success in conducting their business by providing alternate crops that may be appropriate for their specific growout situation. The razor clam represents one candidate species that has a high Milford Ac|iiaculture Seminar. Milt'ord. ConnectiLiit Abslracl.s. February 2(K)1 523 potential for commercialization. Mari^et demand for the razor clam seems to be constant and can be expanded given the relative lack of awareness on the part of consumers regarding the acceptability of the product. Landed value is currently at a level that makes farming the razor clam econonucally attractive. The constraints at this point are focused on developing the appropriate technology for growing the razor clam to a market size. The razor clam exhibits biological properties that suggest it is a reasonable candidate for culture. It undergoes a rotiline bivalve larval cycle that should adapt to hatchery conditions readily. It grows relatively quickly and has been successfully cultured as a by-product under quahog anti-predator netting, due to wild recruit- ment onto the site. It naturally inhabits intertidal and subtidal areas that are currently being used for quahog culture. The razor clam, however, will present some challenges to de- \elopment of it as a commercially viable farmed bivalve. These include: 1. Mobility 2. Over-winter survival 3. Predators & disease 4. Shelf-life 5. Overall lack of knowledge about the species. To aid in diversifying the industry, the Northeast Regional Aquaculture Center (NRAC) has funded us to begin work on de- veloping technology for razor clam growout. The overall objective of this project is to provide an opportunity for the current shellfish culture industry to investigate, develop, and optimize the growout technology for a cultured razor clam. An overview of the projected work and a solicitation for industry involvement is presented. PRELIMINARY STUDIES ON OPTIMAL ROTIFER CUL- TURE DIETS AND ALTERNATIVE REPLACEMENT DI- ETS AND ENRICHMENTS FOR LARVAL BLACK SEA BASS {CENTROPRISTIS STRIATA). Jennie M. Mandeville, Middlebury College. MC Box 3365, Middlebury, VT 05753: Mark T. Watson and Brandy M. Moran. Massachusetts Institute of Technology Sea Grant College Program, MIT Bldg. E38-300. 292 Main St., Cambridge, MA 02139 Rotifers {Bnwhionus plicatilisj are excellent first-feed food for cultured larval fish as they are one of the fishes" naturally occur- ring food sources, small in size, mobile, and easily cultured. It is important that the rotifers are fed a healthy diet that promotes rapid growth in order to maintain a constant food source for the larvae. Several rotifer diets have been tested in the past; this study exam- ines three new, as well as previously tested, diets for their effect on rotifer population growth. The diets were Instant Algae (Reed Mariculture, CA), consisting of Namiochloropsis ociilata (Nanno). Isocluysis galbana (T-Iso), and Pavlova sp., bakers yeast and Al- gamac 2000 (Aquafauna Bio-Marine, Inc., Hawthorne, CA). Fif- teen combinations of these diets were evaluated for their perfor- mance. Nanno supported the highest maximum population growth rate, and was significantly hicher than 8 other diets (P < 0.05). However, three other combinations of algae and yeast significantly (P<().()5) enhanced population growth rates as well. Food combi- nations including Algamac resulted in significantly (P < 0.05) poorer growth rates. While preliminary, these findings can be used to guide aquaculturists in their choice of rotifer diet, and also as a foundation on which to base future studies exploring the nutri- tional value that these diets confer to larval fish. Advancements in culture methods of alternative marine finfish species, especially early life stages, are necessary in order to en- courage commercial production. Alternative diets and more effi- cient live feed enrichments are two ways to enhance growth and survival of cultured fish. Black sea bass (Centropristis .striata) eggs, spawned in captivity at the University of Rhode Island (Kingston, RI), were transported to the Massachusetts Institute of Technology Sea Grant College Marine Finfish Hatchery (Charles- town, MA). The eggs were incubated in filtered (1 p.. with UV) harbor water (22 "C, 28 ppt salinity), and 24 h pholoperiod) until hatch and then transfeired into nine 40 L tanks. In preliminary studies we tested a live-feed replacement. Revo- lution (Salt Creek Inc., UT) liquid fish diet, and compared it to rotifers enriched with Instant Algae (Nanno and T-Iso). Newly hatched larval sea bass were fed three combinations of diets: (1) lOOVf Revolution diet, (2) 50<7f Revolution and 50<>'f Rotifers, and (3) lOO'^f rotifers. Results of growth and survival showed that there was no difference between diets #2 & #3; diet #1, however, re- sulted in lOO'/r mortality after 10 days, demonstrating that the Revolution diet can be used as a supplement in larval fish cultures, but not as a substitute for rotifers. Recent studies have shown that larvae allowed to feed for several days on live feed prior to re- placement diets show higher survival. This is thought to permit proper digestive tract development. A second series of studies compared various enrichment media for rotifers and examined their effect on growth and survival of newly hatched black sea bass larvae. Three popular, commercially available enrichments were tested: Instant Algae (Nanno and T- Iso); Culture Selco dnve Aquaculture Inc., Grantsville, UT): and Algamac 2000. Preliminary results showed that Algamac 2000 provided the highest growth and the most consistent survival (50%) as an enrichment. Research is continuing to compare the economic benefits of the various enrichments. PREVALENCE OF PERKINSUS MARINUS IN THE EAST- ERN OYSTER. CRASSOSTREA VIRGINICA, IN RHODE IS- LAND. Karen Mareiro, .losefa Duugal, Meggan Dwyer. Ken- neth Leonard HI and Marta Gomez-Chiarri. University of Rhode Island, Fisheries, Animal and Veterinary Science, Kings- ton. RI 02881: Arthur Ganz, Rhode Island Department of Envi- ronmental Management, Coastal Fisheries Laboratory, Wakefield, RI 02879 Populations of Eastern oyster, Crassustrea virgiiiica. along the East Coast of the United States have been severely decimated by \ arious diseases in recent years. One of the most common of those 524 Abstmcts. February 2001 Milford Aquaculture Seminar. Milford, Connecticut diseases is Dermo disease, which is caused by the protozoan para- site Perkinsus maniuis. Since 1998. our laboratory has been track- ing the prevalence of this disease in oysters from several locations in Rhode Island. Thirty oysters were collected from 12 sites throughout Rhode Island waters, including aquaculture lease sites, in August and November of 1998-2000. The presence and inten- sity of Dermo disease was tested using the Ray's Fluid Thiogly- collate Media (RFTM) method. Dermo infections were detected at most sites, with the highest intensity of infections being observed in Barrington River and Point Judith Pond (wild). Areas where prevalence of Dermo was low include Block Island and Prudence Island. In addition. Dermo prevalence was also low in oysters collected from aquaculture lease sites. Differences in prevalence of Dermo disease in these areas could be due to lower exposure to parasite, resistance to parasite, food a\ ailability. or abiotic factors including flushing rates, salinity, and temperature variability. Fur- thermore, in 1998 and 1999. intensity of infections was higher in August than in November. However, in 2000 the reverse was true. It is possible that the late peak in prevalence and intensity could be due to the cold weather of early 2000 winter. This research has been funded bv the Rhode Island Department of Environmental Management. Division of Fish and Wildlife. IS SIROLPIDIVM ZOOPHTHORUM THE ANIMAL EATER ITS NAME SUGGESTS? NEW EVIDENCE OF PARASIT- ISM. Christopher Martin, USDOC. NOAA. National Marine Fisheries Service. Northeast Fisheries Science Center. Milford Laboratory. Milford. CT ()(i46() A phycomycetous fungus has been observed repeatedly in lar- vae of the bay scallop. Arfiopectcii irnulicms. at the Milford Labo- ratory. This microorganism has been tentatively identified as Sirol- pidiitm zoophthoiiim Vishniac. first observed at this laboratory by V.L. Loosanoff almost fifty years ago. The morphology and de- velopment of the fungus have been previously described. While apparently enzootic in scallop cultures at our laboratory. S. zooph- thonim has not been tied directly to mass mortalities of this spe- cies. Loosanoff regarded it as parasitic in the bay scallop and in the larvae and juveniles of other bivalves. However, his evidence was largely circumstantial, i.e.. conspicuous presence of "infected" lar- vae in cultures suffering high mortality. Using pure cultures of S. zoophthorwn, freshly isolated from affected scallop larvae, it has been possible to demonstrate that this fungus is one likely cause of observed mortality. Exposure of 72 h scallop larvae to suspensions of recently emerged zoospores re- sulted in approximately 70% mortality in 4 days. Fungal thalli were detected in up to 88% of dead larvae. Untreated controls remained unaffected. S. zonphlhoriiiii was successfully re-isolated from parasitized larvae, thus satisfying Koch's Postulates. POTENTIAL EFFECTS OF CONTAMINANT EXPOSURE ON CULTURED TAUTOG. TAUTOGA ONITIS. Brandy M. Moran, Mark T, Watson, and Cliff A. Goudey. Massachusetts Institute of Technology Sea Grant College Program. MIT Bldg. E38-300. 292 Main St., Cambridge, MA 02139 MIT Sea Grant College Program has been demonstrating aqua- culture in recirculating systems in the Boston Harbor since May 1998. The hatchery is located in the inner harbor within the Charlestown Navy Yard, Charlestown. MA. The Navy Yard was in full commission from 1799 to 1975. Due to the shipbuilding, equipment facility, and cordage activities conducted at the yard and the overall industrial character of the port, the surrounding seatloor sediments are contaminated by waste products such as heavy metals. The contaminated sediments and other pollution sources in the harbor may influence the water quality used in the recirculating systems. Because there was a concern about con- taminant accumulation in the fish reared at the hatchery, a moni- toring program began. To establish an economical and commercial hatchery in Boston Harbor, the possibility of cultured fish accu- mulating harmful pollutants needed to be investigated. Tautog exclusively raised in a recirculating system from egg. spawned in July 1998. are cunently 2 Vi years. Collaborating with the Massachusetts Water Resource Authority, the muscle tissues of these fish were analyzed for heavy metals and contaminants in February 1999 (7 mo). November 1999 ( 16 mo) and June 2000 (23 1110). The data reveal traces of heavy metals and other contami- nants in fish tissue, but all levels are below FDA regulations, action levels, and recommendation levels. This demonstrates that tlsh being held for long periods of time in Boston Harbor water are safe for human consumption under current regulations. Additional unique findings in the reared tautog were observed during this study. In February 2000. when the tautog were 18 mo, fertilized and unfertilized eggs were found in the recirculating system. This spawning occurrence is early according to literature on tautog reproductive strategies. Tautog are known to mature sexually between 2 and 3 y. By June 2000. some of the tautog showed signs of an enlarged abdomen and unbalanced swimming behavior. The fish were sent for a full diagnostic analysis. The tautog with swollen abdomens were all females that were not successfully reabsorbing unspawned eggs. The tautog also had gill defonnities consisting of migrating chloride cells to the tip of the lamellae, fusion of lamellae, and enlarged mucus cells. The water quality of the system was stable over the 2 '/: year time span except for the ammonia, which fluctuated between 0 ppm-5 ppni. These abnormalities are potentially associated with water qual- ity and/or low-level contaminants. The impact of the fluctuating ammonia may have caused a stress-induced acclimation to the gills to be able to tolerate its surroundings. It is also possible that exposure to low-level contaminants for an extended period of time could cause these gill deformations. Both the stress from the am- monia and contaminants may affect the female tautog ability to reabsorb eggs that were not spent. Although the tautog are edible MMford Aqiuiciilture Seminar, Milford. Coiinci.'tii.'iit Abstracts, February 2001 525 under the current FDA regulations, the exposure to these contami- nants could be affecting the overall health of the fish. Further research is needed to determine tautog's exposure tolerance to metals, ammonia, and other critical environmental parameters. This information would aid in the siting and success of a commer- cial hatchery in industrial harbors. BOTTOMS UP! AN INDUSTRY-LED PROJECT: BRING- ING AN AQUACULTURE TECHNIQUE TO THE IN- SHORE SCALLOP FISHERY IN MAINE. Dana L. Morse, University of Maine, Darling Marine Center, 193 Clark's Cove Rd.. Walpole, ME 04573 In May of 1999, a delegation from Maine tra\eled to Japan to learn about the scallop culture industry. Focusing on Aomori Pre- fecture, the group quickly realized that the Japanese industry relied on a successful spat collection program, and that elements of this program could yield benefits in Maine's wild fishery. The late summer and fall of 1999 saw many fishermen up and down the coast experimenting with spat collection gear, particularly in the port of Stonington. Since the start, industry has had the leading voice in imple- menting and continuing this work, with strong collaborations with a wide variety of agencies and organizations. Catches of seed from the 1999 year class were spotty but encouraging, and preliminary assessments from the year 2000 set indicate that catches will be good. Many obstacles exist for the continued development and ac- ceptance of the program. However, the effort has yielded promis- ing results thus far, and industry support has remained strong. Future work will rely on this continued support, and from the productive working relationships with other groups in the state. EFFECTS OF FILTER-FEEDING OYSTERS ON SEDI- MENTATION RATES AND PHYTOPLANKTON SPECIES COMPOSITION: PRELIMINARY RESULTS OF MESO- COSM EXPERIMENTS. Jennifer Mugg, Michael A. Rice, and Monique Perron, University of Rhode Island, Fisheries, Animal, and Veterinary Science, Kingston, RI 02881 Eutrophication is occurring in many coastal estuaries. A pos- sible solution to this problem is to raise aquaculture oysters to improve water clarity and to help remove excess nitrogen. In order to determine what effects aquaculture oysters have on the envi- ronment, a niesocosm study was performed at the Marine Ecosys- tem Research Laboratory (MERL) from June to October 2000. The MERL facility is located adjacent to Narragansett Bay with thir- teen 13,000 L niesocosm tanks that simulate the environmental conditions of the Bay. Two hundred oysters (=35 mm in valve length; nominally filtering about 48 L da"' ind^' ) were placed into three mesocosms, and three mesocosms were maintained without oysters as controls. Experiments were run with varying rates of water exchange in the tanks ranging from 0% to 100% per day (0-13.000 L da"'). Several parameters were measured and com- pared between the two treatments, which included chlorophyll-a, particulate organic and inorganic matter, sedimentation rates, ni- trate, ammonia, phytoplankton analysis and growth rates. Preliminary results show that oysters have an effect on species composition of phytoplankton in the water column and induce increased rates of sedimentation to the benthos. Diatoms of the genus Nitzschia were predominant in mesocosms with oysters, and in the control tanks Skeletonema were dominant. Tanks with oys- ters consistently showed rates of sedimentation greater than twice the control tanks. We speculate that this increased organic sedi- mentation by actively filter feeding oysters may contribute to in- creased rates of sediment deposition leading to increased denitri- fication in natural systems. This is work from RI-AES Project H-886, and is publication number 3857 of the College of the Environment and Life Sciences, University of Rhode Island. LABORATORY CULTURE OF LARVAL TAUTOG: RE- CENT UPDATES AND CHANGES. Dean M. Perry, David A. Nelson, and Robin S. Katersky, USDOC, NCAA, National Ma- rine Fisheries Service, Northeast Fisheries Science Center, Milford Laboratory. Milford, CT 06460 Adult field-captured tautog, Tautoga onitis. were spawned in the laboratory. Larvae were cultured according to standard labo- ratory procedures developed at the Milford laboratory using a re- circulating system containing six 1 140 L conical rearing tanks. During the summer of 2000, the following changes were made to our protocol which increased larval survival. These changes in- cluded increasing live feed density in the rearing system, extending the duration of rotifer feeding, and decreasing the initial larval stocking density. Larvae were fed an average of 9 x 10^' rotifers per day, which is double the amount fed in previous years. They were also fed rotifers for a 25 d period compared to 14 d in prior years. With this increased larval survival we are cun-ently monitoring growth rates of these juveniles under laboratory conditions. From December 14, 2000 to February 2, 2001, total length (mm) and biomass were measured and recorded biweekly. Specific growth rate was 0.30 mm d"' and biomass increased at an average of 0.04 gd-'. DEPARTMENT OF COMMERCE AQUACULTURE— AN UPDATE OF POLICIES, PLANS AND PROGRAMS. Edwin Rhodes. Aquaculture Coordinator, National Oceanic and Atmo- spheric Administration, National Marine Fisheries Service, 1315 East-West Highway, Silver Spring, MD 20910 The Department of Commerce Aquaculture Policy, signed in 1999. has created a renewed interest in aquaculture within the Department. Guidance to Commerce agencies concerning aqua- 526 Abstracts, February 2001 Milford Aquaciilture Seminar, Milford. Connecticut culture has been developed that should insure that aquaculture is considered in all relexant Department programs. The chiet funding increase for aquaculture came through an additional S3 million appropriation to the Office of Oceanic and Atmospheric Research that will result in a competitive distribution of $5 million in this fiscal year, and a similar amount next year. Funding from this program in the last cycle went primarily toward management and regulatory issues, but this year's priorities include a strong em- phasis on research and encourages industry, government, and aca- demic partnerships that will result in commercialization of aqua- culture, and which also will contribute to the Commerce goal of a fivefold increase in U..S. aquaculture production by 2025. Other NOAA programs that support the de\elopnient of envi- ronmentally sound aquaculture include the in-house aquaculture R&D programs at our Northeast and Northwest Science Centers, and the Fisheries Finance Program that made significant loans to the aquaculture industry in 2000. NOAA has also drafted offshore aquaculture legislation that would authorize long term leases in the U.S. FEZ. Additionally, NOAA, along with other Federal partners and with stakeholder input, is developing a code of conduct for responsible aquaculture in the U.S. exclusive economic zone. NOAA is also providing some funding for offshore aquaculture development. KEEPING GENTRIFICATION AT BAY: HOW A DE- FUNCT SHELLFISH FACILITY WAS KEPT FROM THE DEVELOPERS. Gregg Rivara. Cornell Cooperative Extension of Suffolk County, Marine Environmental Learning Center. Southold, NY. 11971: Timothy Caufield. Peconic Land Trust, P.O. Box 1776, Southampton, N"^' 11969; and Walter Smith. P. O. Bo.x 395, Orient, NY 1 1957 The loss of waterfront real estate to non-water dependent uses in the Northeast and elsewhere has made siting new shellfish mari- culture operations expensive, if possible at all. The Shelter Island Oyster Company (SIOC) was formed in 1924 by Anna and John Plock Sr. By the 1960s the company had offices in Manhattan and Greenport, and a nursery /grow out facility in Southold, New York. John Plock has been called a marketing genius — many types of SIOC marketing devices can today be found in online auctions. As a result of Plock's business sense, the company was well known, and plans to sell its "greenhouse" oysters by mail were featured in National Fisherman in the 1970"s. A number of unfortunate events led to the demise of the com- pany in 1984. What remained of SIOC's holdings was the Southold location, a 22-acre site of buildings, ailificial lagoons, tide gates, and decaying culture gear — all in all. a facility that could not be built with today's environmental regulations. After a series of failed development plans that would have precluded mari- culture at the site, the Peconic Land Trust, a not-for-piofit land conservation organization, and Cornell Cooperative Extension worked on a plan with the Plock family to save most of the area used for shellfish culture v\hile limiting development to four two- acre lots. Dedicated in 1996, the "Shellfisher Preserve" consists of 14 acres with structures and water access that are offered by the Trust as an incubator for shellfish mariculture firms. Cunently there are two such tenants, both with hatchery and nursery operations in place or planned for 2001. DEVELOPMENT OF THE NOANK AQUACULTURE CO- OPERATIVE IN CONNECTICUT. Karen Rivara. The Noank Aquaculture Cooperative, 100 Main St.. Noank, CT 06340 The Noank Aquaculture Cooperative (NAC) is now a legal entity which will operate a commercial shellfish hatchery and nursery system. It will provide shellfish seed to our members to grow to market size. The market shellfish will then be sold through the Cooperative. In addition to operating a commercial hatchery, the Cooperative will have wet storage and a processing facility at the Noank site. The Noank Aquaculture Cooperative (NAC) will coordinate with the Long Island Sound Regional Vocational Aquaculture School to provide hands-on learning experiences for their students. The Groton Shellfish Commission will be involved in our educa- tion outreach program. We will be working with researchers in Connecticut, both fmm the universities and the Bureau of Aqua- culture. to develop disease resistant oysters. Future work with lobsters, finfish, and seaweeds is possible. The Groton Shellfish Commission (GSC) will be managing a portion of the shellfish seed provided by the Cooperative to the Town of Groton. This seed will be used for Recreational Shellfish Programs and for their Shellfish Restoration Programs. The GSC will most likely be help- ing the NAC with our public education outreach programs. The Cooperative has received a lot of help and support from the Groton Shellfish Commission, the Noank Fire District, the Con- necticut Department of Agriculture, Bureau of Aquaculture, and the USDA. With this support we should become an important asset to the Southeastern Connecticut Maritime Community and to the Connecticut Shellfish industrv in seneral. A COMPARISON OF THE GROWTH AND MORTALITY OF JUVENILE ARGOPECTEN IRRADIANS IN THE VARI- OUS CULTURE METHODS EMPLOYED BY THE STU- DENTS AT THE SOUND SCHOOL REGIONAL AQUA- CULTURE CENTER. John Roy. Amber L. Beitler. and Kath- ryn R. Marl0.5 m in diameter), adjacent to a clump, or in an open area without mussels (>l m from any mussels). Lastly, half of the pots were meshed to protect them from predation with flexible 6 mm meshing. All pots were collected on September 28"", 2000. and surviving clams were counted and measured. Survival was significantly lower in unprotected pots (P = 0.0001) than protected pots, and was also low underneath mussels. More interestingly, among un- protected clams not under mussels, survival tended to be better in the mussel bed than on the flat (adjacent, P = 0.0801; open, P = 0.0380). In terms of growth, final shell length was lowest under- neath clumps of mussels; among clams not underneath mussels (where survival was very low), growth tended to be better on the mudtlat than in the mussel bed. This result was greater for unprotected clams (P < 0. 1437) than for protected clams (P > 0.2803), suggesting a possible interaction with predation. Thus, the presence of mussels has a mixture of effects upon clams, which need to be considered when contem- plating removal. NURTURING A SOFTSHELL CLAM PRIVATE/PUBLIC INITIATIVE ON MASSACHUSETTS' NORTH SHORE. Scott Weston. Joseph Buttner, and Mark Fregeau, Northeastern Massachusetts Aquaculture Center and Department of Biology. Salem Stale College, Salem, MA 01070 Commercial aquaculture is in its infancy on the North Shore of Massachusetts. Interest and support for aquaculture continue to grow as evidenced by an increasing number of restoration and enhancement initiatives that target the "Ipswich clam" (softshell clam. Mya arenaria). Many North Shore towns (e.g.. Ipswich, Rowley, Gloucester and Essex) have sponsored public/private ef- forts involving shellfish wardens, advisory boards, shellfishers and the Merrimack Valley Planning Commission in innovative, volun- teer projects to identify technologically effective and socially ac- ceptable methods of seeding and managing local clam flats. The Department of Marine Fisheries and Department of Food and Ag- riculture (DFA) have provided permits and/or funding. Acknowl- edging the social and economic importance of a viable aquaculture industry on the North Shore and throughout Massachusetts, three regional centers have been created by legislative act and supported through the DFA. The Northeastern Massachusetts Aquaculture Center (NEMAC), housed at Salem State College, serves as a catalyst to forge the knowledge and experience gained through Milt'ord AqiiaciiltLire Seminar. Milt'ord. Connecticut Abslmcls. Fehruarv 2001 529 independent efforts by North Shore comnuinities into a regional, collaborative softshell clani initiative. Since its official opening in April 1999, NEMAC"s Cat Cove Marine Laboratory has made courses, reference materials, and out- reach (extension) assistance available to aspiring/practicing aquac- ulturists. Perhaps the most significant contribution has been filling a void as a local supplier of clam seed. In its first year of operation, 30,000 softshell clams (3 mm) were acquired from the Beals Island Regional Shellfish Hatchery (ME) in October 1999, successfully raised by laboratory personnel and released as 10 mm spat in Ipswich in October 2000 by town officials, volunteers, and people from the MVPC and NEMAC. Currently, an estimated 500,000 softshell clams reside at NEMAC's Cat Cove Marine Laboratory. Most clams originated from the Beals Island Regional Shellfish Hatchery (300.000) or were collected from mud Hats in Rowley by the shellfish warden and volunteer shellfishers (200,000). A mod- est number of clams (1,000-2,000) were spawned from local broodstock maintained in the Cat Cove Marine Laboratory (July 20(J0). As these clams attain an appropriate size they will be re- leased onto the flats of lpsv\ich. Rowley. Gloucester, and Essex in the early summer and/or early fall 2001 . Production and seeding of clams are expected to continue increasing as laboratory personnel gain experience and the facility becomes fully functional. Through the input of varied stakeholders an economically fea- sible and environmentally sustainable aquaculture industry is gradually developing on Massachusetts' s North Shore at a rate and in a manner consistent with the expectations of all collaborators, specifically through supplementing the traditional wild harvest with increasing numbers of cultured clams. Aquaculture is grow- ing because local shellfishers and towns are supportive, technical expertise and an increasingly dependable source of seed exists, and state agencies are making significant financial and logistical com- mitments. The private/public approach to establish a sustainable aquaculture industry on Massachusetts' North Shore is working and could prove applicable elsewhere. GROWTH AND SURVIVAL OF BAY SCALLOPS, AR- GOPECTEN IRRADIANS IRRADIANS, FED TETRASELMIS CHUI BY TWO METHODS. James C. Widnian Jr. and David Veilleux, USDOC, NCAA, National Marine Fisheries Service, Northeast Fisheries Science Center, Milt'ord Laboratory, Milford, CT 06460 Tetraselmis cJiiii was grown using two different methods, car- boy vs. GRAMPS (Greenhouse Algal Mass Production System), and fed to bay scallops, Ari^npeclen inad'uins inadians. Growth and survival were monitored. Algae grown by the traditional Mil- ford carboy method utilize artificial light and artificial seawater with the addition of various nutrients, trace elements, and vitamins. GRAMPS-grown algae utilize sunlight and natural seawater from Milford CT harbor, which is UV treated and then enriched with a commercially available F/2 media. Current results indicate a slight growth advantage when using algae grown in the carboys. In- creases in mean shell height ranged from 2.6-3.1 mm for scallops fed with GRAMPS-grown algae, while scallops fed carboy-grown algae increased 3.6—4.1 mm. Most surprising was the decrease in survi\al when scallops were fed Tetraselmis grown using the GRAMPS method. Survival of scallops on the GRAMPS-grown Tctrasi'lmis averaged 64%, and scallops fed carboy Tetraselmis averaged 89%. Additional research is needed to determine the cause of the growth and survival discrepancies observed between the two growing methods. EFFECT OF FEEDING RATION AND REGIME UPON GROWTH AND FOOD-CONVERSION OF JUVENILE QUAHOGS, MERCENARIA MERCENARIA, AND COM- PARISON WITH BAY SCALLOPS AND EASTERN OYS- TERS. Gary H. Wikfors, Jennifer H. Alix, Mark S. Dixon, and Barry C. Smith. USDOC, NCAA, National Marine Fisheries Ser- vice. Northeast Fisheries Science Center, Milford Laboratory, Mil- ford, CT 06460 Our previous studies have demonstrated effects of both feeding ration (how much) and regime (how often) upon growth and feed- conversion efficiency of juvenile bay scallops, Argopecten irnuli- ans, and eastern oysters, Crassostreci virginicii. As this informa- tion is useful in design and scaling of land-based hatchery and nursery processes, we conducted an experiment identical to those done previously with bay scallops and oysters, this time with ju- venile (3mm) clams, Mercenaria mercenaria (notata variety, gen- erously provided by the F.M. Flower Co., Oyster Bay, NY). The feeding experiment was a factorial design in which ration, at 1,2, 5, or 10% of clam live weight in dry weight-equivalent algal feed per day, and regime, with the daily ration di\ ided equally into 2, 4, or 16 daily feedings, were co-varied. Our computer-controlled molluscan rearing chambers were used to accomplish the multiple daily feedings of a 50:50 mix of two high-lipid Tetraselmis strains, PLY429 and PLAT-P, that were cultured free of bacteria in semi- continuous carboy assemblies. The 50 clams held in each chamber were removed weekly, and growth was monitored as live weight, volume displacement, and shell size. Rations were adjusted weekly to remain consistent with live-weight increases, and the same clams were returned to the chambers. Organic weight growth (ash- free) of clams was determined after six weeks of feeding and compared with organic weight of algae fed to calculate conversion efficiency (CE) as the percentage of algal mass converted to clam tissue mass. A multifactor Analysis of Variance (ANOVA) was used to analyze main and interactive effects of the independent variables ration and regime upon the dependent variables growth and CE. Clam survival was near 100% during the experiment. Growth of clams was significantly affected by ration, but not regime. Growth was not significantly greater than zero on 1 and 2% ra- tions, and increased exponentially at 5 and 10% rations. CE was 530 Abstracts. February 2001 Milford AquacLilture Seminar. Milford. Connecticut affected significantly by both ration and regime, as well as by the interaction of these two variables. CE was highest at the \0% ration given as 16 daily feedings, but was only slightly over 2%. i.e.. it took 50 g of algal dry weight to produce 1 g of clam tlesh. These results contrast sharply both qualitatively and quantitatively with findings for scallops and oysters, in which maximal growth occurred on 5% rations, and CE values were in the range of 20- 25%. Possible reasons for this contrast mav include lower nutri- tional value ot the Tetniseli\iis strains for clams than for the other bivalves, differences in initial size of individuals in experiments with different .shellfish, or the fact that ash-free dry weight was used for clams, whereas dry tissue weight was used previously. Nevertheless, it appears likely that northern quahogs are far less efficient at using low levels of algal food, and require more food than other bivalves and, thus, are poorer suited to land-ba.sed, controlled iiurserv culture. Jcmnial of Slicllfnh Rcseurch. Vol. 20. No. I, 531-562. 2001. ABSTRACTS OF TECHNICAL PAPERS Presented at AQUACULTURE 01* Triennial Meeting of: WORLD AQUACULTURE SOCIETY NATIONAL SHELLFISHERIES ASSOCIATION FISH CULTURE SECTION OF THE AMERICAN FISHERIES SOCIETY Orlando, Florida January 21-25. 2001 *Editor"s Note: Since this meeting was topic-oriented rather than society-oriented, papers presented here represent those whose authorts) signified membership in the National Shellfisheries Association. 531 National Shellfisheries Association. Orlando. Florida Ahsiratts. January 2001 533 CONTENTS Normal and altered gametogenesis in the green sea urchin: implications for aquaciiltiue 537 William D. Anderson and Guy M. Yianopoulos Using GLS. GP.S. and digital photography in shellfish resource management 537 Kathleen Apakiipakul, Antonio Villalba, Sandra M. Casus and Kimberly S. Reece Molecular analyses of a Pcikinsus species in the European flat oyster Ostrca ediilis 537 Gwynne I). Brown and Kimberly S. Reece Variation in serine protease genets) among Pcrl fully differentiated spermatozoa in the testis); or (2) somatic cells called nutritive phagocytes (NP) and present in both sexes. In an effort to help others better understand normal game- togenesis in the green sea urchin, we present an accompanying web page, which provides a detailed catalog of the appearance and functions of these two populations of cells during the following stages: 1. intergamelogenesis and NP phagocytosis; 2. pregametogenesis and NP renewal; 3. gametogenesis and NP utilization; and 4. end of gametogenesis, NP exhaustion, and spawnmg. This web page: http://zoology.unh.edu/faculty/walker/urchin/ gametogenesis.html also covers gamete structure, fertilization, and very early sea urchin development, gives an extensive bibliogra- phy, and will be amended as new information becomes available. We also encourage others with more extensive information than our own to help us to develop this web page. Intimate knowledge of these topics in the sea urchin is useful in aquaculture, if it does no more than enhance our abilities to manipulate reproduction and development. High-quality sea urchin gonads (often called roe) are distinguished by superior size, taste, color, texture, and firmness (= extended shelf life). Ovaries and testes containing fewer ga- metes relative to somatic cells (NP) are prefeiTed in most cultures that eat sea urchin gonads. Gametogenesis usually has a negative influence on the quality of sea urchin gonads. This is particularly true near the conclusion of gametogenesis, when gonads are very fragile and contain mainly gametes. Rough handling of sea urchins during harvesting results in disintegration of the gonads. In the future, a more thorough understanding of the cell and molecular biology of sea urchin gametogenesis will permit us to manipulate the process at various stages to accomplish particular marketing goals. For example, it would be advantageous for aquaculturists to understand: ( 1 ) how to suppress gametogenesis to produce high- quality gonads for consumption; and (2) how to promote gameto- genesis for the increased production of seed stock. The concept of producing "designer"' urchins with gametogenesis customized to particular market needs is not beyond reach. Based on our current understanding of the normal process of gametogenesis in the green sea urchin, we present methods that can accomplish both of these very different goals within a land-based facilitv. Supported by the University of New Hampshire Center for Marine Biology and by Sea Grant (R/FMD-146) to CWW and MPL of UNH. DOC SBIR to MD and CWW and Hatch Grant #333 to CWW and Dr. Larry G. Harris. We thank Elia Heath- Antonelli, Erin Zook, and Paula Rodgers for their efforts in pre- paring tissues for this study and Tanya Hakala for preparing the web site. USING GIS, GPS. AND DIGITAL PHOTOGRAPHY IN SHELLFISH RESOURCE MANAGEMENT. William D. Anderson and Guy M. Vianopoulos, South Carolina Department of Natural Resources. Marine Resources Center, P.O. Box 12559. Charleston. SC 29422. USA. E-mail:andersonb@mrd.dnr.state. sc.us Multiple data layers, consisting of intertidal oyster iCrassu- strca viii^iiiica) spatial distributions, management categories, wa- ter quality, and boundary markers, all linked to attribute databases and digital photography, are used to characterize bivalve habitats and manage shellfish resources in coastal South Carolina. The three-dimensional intertidal oyster habitat, which serves as a criti- cal nursery to resident and transient fish populations, is character- ized as a keystone species in a geographic area devoid of sub- merged aquatic vegetation. Environmental perturbations, such as coastal development, boat wakes, nonpoint source runoff, and silt- ation. complicate resource conservation and management for a species chronically exploited by recreational and commercial har- vesting. This presentation describes how intertidal oyster resource as- sessments are conducted utilizing GIS, GPS, laser rangefinders. and digital photography. Color ArcView® maps are produced to delineate harvest areas for commercial fishermen; disseminated to recreational harvesters, and made a\ailable (in easily readable .PDF format) on the SC Department of Natural Resources' web page (www.dnr.state.sc.us/marine/regs/stateshell). Mariculture and wild stock shellfish permit interface with water bodies condi- tionally approved, restricted, and closed to harvest by coliform bacteria — dynamic and complex environment — its management facilitated by multiple GIS layers and linked attribute data. Digital maps depict oyster populations, water bodies closed by coliform bacteria, and shellfish management areas, all linked to commercial harvest and resource assessment data. MOLECULAR ANALYSES OF A PERKINSUS SPECIES IN THE EUROPEAN FLAT OYSTER OSTREA EDULIS. Kath- leen Apakupakul, Antonio Villalba, Sandra M. Ca.sas, and Kimberly S. Reeee, Virginia Institute of Marine Science. College of William and Mary. Gloucester Point. VA 23062. USA. E-mail: kathleen (a' vims.edu Members of the genus PerkiiisKs are protozoan parasites that have a subtropical distribution and infect a variety of shellfish host species. Perkinsus species have caused high mortality among the world's shellfish and are. therefore, the subject of intensive study to develop effective management strategies for the health of com- mercial shellfish fisheries. Currently, there are five recognized 538 Ahslnicts. January 2001 National Shellfisheries Association. Orlando. Fkirida species of Perkinsus. An infection of Pcrkinsus has been discov- ered in the tissues of the European flat oyster Ostrea ediilis. a previously unreported host species. To determine the species of parasites infecting the flat oyster, portions of the internal transcribed spacer (ITS) region and the small subunit ribosomal (SSU) RNA gene were sequenced and compared to the sequences of other Perkinsus species as well as to those of other protozoans. Based on these data, phylogenetic analyses showed that the parasites are a distinct sister clade to the Perkinsus atlanlicus/olseni group. Whether or not these parasites comprise an entirely new species of Perkinsus or a strain of P. atlanlicus/olseni has yet to be determined. Perkinsus species are known to have rather wide host selectivity. The sequences we obtained are different from any of those deposited in GenBank or published to date. If morphology data corroborate the differences seen in the molecular data, these parasites may be considered a new species of Perkinsus. A Perkinsus genus-specific DNA probe was designed based on the alignments of the SSU rRNA sequences of all known Perkin- sus species. In in situ hybridization protocols, this probe has been effective in labeling P. ntarinus. P. atlanticus. P. olseni. and Per- kinsus species from Chania pacificus, Ruditapes philappinaruni. Macoma balthica. and Mya arenaria. In situ hybridization was performed on histologically prepared infected Ostrea edulis tissue sections. We found that this probe also hybridizes to the parasites infectine O. edulis. VARIATION IN SERINE PROTEASE GENE(S) AMONG PERKINSUS MARINUS ISOLATES. Gwynne D. Brown and Kimberly S. Reece, Virginia Institute of Marine Science. College of William and Mary. Gloucester Point. VA 2.'^062. USA. E-mail: gbrown@vims.edu Perkinsus nuiiinus. the causative agent of dermo. is a major pathogen of the eastern oyster. Crassostrea virginica along the U.S. Adantic and Gulf coasts. Although .scientists have studied the organism for fifty years, little is known regarding the pathogenic mechanisms of the protozoan. Serine proteases present in cell-free supernatants of P. marinus cultures //; vitro are currently under investigation as a putative virulence factor. Serine proteases have been found to play key roles in pathogenesis of several parasitic protozoans. In this study, we identified and sequenced a serine protease gene from P. marinus. We also examined variations in the identified gene among different isolates. We first amplified a 475 bp subtilisin-like serine protease gene fragment from P. marinus using "universal" degenerate primers. The 475 bp fragment was then labeled with digoxigenin and used to screen a P. marinus \ phage genomic library. DNA from hy- bridizing phage was isolated and subjected to Southern blot analy- sis. Two different recombinant clones were identified. Clones were subcloned and sequenced using internal primers. Within one sub- clone, a 1,254 bp open reading frame was identified containing the probe .sequence. BLAST analysis confirmed the similarity to sub- tilisin proteases, reporting an E value of 7e-40. Genomic Southern blot analysis using the 475 bp digoxigenin- labeled probe showed that variations exist among different isolates of P. ntarinus. Two clonal isolates from Louisiana. LA 5-2 and LA 10-1. produced similar hybridization patterns with Nci I and Sal I; whereas, additional bands were detected in the isolates P-1 (York River. VA) and MA 2-1 1 (MA). These results suggest that either the P- 1 and MA 2- 1 1 genomes contain a second gene or that they possess an alternate allele of the same gene. GENE TRANSFER IN CELL CULTURE OF THE EAST- ERN OYSTER CR.ASSOSTRE.\ VIRGINICA. John T. Bucha- nan, Ta-Chi Cheng, S. Reyna Gonsalves, Jerome F. La Peyre, Richard K. Cooper, and Terrence R. Tiersch, Department of Veterinary Science. Louisiana State University Agricultural Cen- ter. Louisiana Agricultural Experiment Station, Baton Rouge. LA 70803. USA. E-mail: jbuchanan(a'agctr.lsu.edu The eastern oyster Crassostrea virginica supports a valuable industry along the East and Gulf coasts of the United States. Re- cently the oyster industry has been injured by disease problems, especially infections of the protozoa Perkinsus marinus (causing dermo) and Haplosporidium nelsoni (causing MSX). Both proto- zoa have been significant factors in the decline of the oyster fish- ery in the Chesapeake Bay. Along with techniques such as selec- ti\e breeding and ploidy manipulation, gene transfer may be an effective tool to combat these disease problems. Cell culture serves as an effective model for in vivo work, and gene transfer in cell culture would be useful for screening promot- ers and genes for use //) vivo. Because a cell line from bivalve mollusks has not been established, it is necessary to work with primary cultures of oyster cells. To this end, ventricle cells from C. virginica were isolated, and conditions for optimal gene transfer and gene expression were investigated. Cells were cultured in 96-well plates in 100 |jlL of medium LA-2 at 4 x \0^ cells per well. The transfection reagent Effectene* (Qiagen Inc.. Valencia. CA) was used to deliver DNA to cells, and expression of the reporter gene luciferase was used to evaluate promoters suitable for use in oyster cells. One of the promoters tested, the heat shock promoter (HSP70) from the snail Biomphalaria glabraia was found to be inducible in oyster cells, with significantly higher levels of luciferase detected after transfection and heat shock (f < 0.0000 1 ). The optimal conditions for heat shock were 40°C for 1 h{P < 0.05). Gene expression varied over time, with highest levels of expression observed 16 to 24 h after heat shock (P = 0.0001). Various conditions for transfection with Effectene® were tested. and optimal conditions per well were 0.3 |Jig DNA mixed with 3 {X.L of Effectene® {P < 0.05). Although gene expression was de- National Shcllfishcrics Association. Orlando. Florida Abstracts. January 2001 539 tected. we found significant toxicity associated with transfection with Effectene* (P < 0.05). Other transfection reagents may pro- vide adequate levels of gene expression in oyster cell culture with less toxicity. We have also observed luciferase expression using the following promoters: C. gigas actin. Drosophila melanogaslcr actin, D. ineUmogaster heat shock, and human cytomegalovirus (CMV) immediate early promoter. Future work will involve char- acterization of gene expression with these promoters and identifi- cation of promoters suitable for constitutive and inducible gene expression in vivo. THE INFLUENCE OF SUBSTRATES AND CULTURE ME- DLA FORMULATIONS ON THE ATTACHMENT AND SPREADING OF EASTERN OYSTER CELLS IN PRIMARY CULTURES. John T. Buchanan. Y. Li, and Jerome F. La Peyre, Department of Oceanography and Coastal Sciences. Loui- siana State University, Baton Rouge, Louisiana 70803, USA. E-mail: jbuchanan@agctr.lsu.edu The utility of cell culture in research has long been recognized. However, cell lines from any bivalve moUusk have yet to be es- tablished. In addition, a lack of standard protocols for primary cell culture has hindered bivalve research. Improved media have been recently developed in our laboratory. Enhanced cell attachment and spreading in primary culture would improve protocols and may prove essential for the development of a bivalve cell line. To this end. ventricle cells from Crassostrea virginica were isolated and the attachment factors collagen I, collagen IV. fibronectin. laminin. and poly-D-lysine were tested for their ability to promote cell attachment and spreading in the media JL-ODRP-4. LA-2, and LA-5. Also tested were two commercially available uncoated tis- sue culture plates (Falcon. Inc. and Costar. Inc.). A spectrophoto- metric assay (MTS assay) based on the reduction of tetra/oliuni salts to a formazan dye by metabolically active cells was employed to estimate metabolic activity and cell numbers remaining in a well after washing. A fluorescent assay (CyQuant assay) measuring the amount of nucleic acid in a well was used to estimate cell number in wells after washing, cell survival, and RNA production on the various substrates. Cell response to each attachment factor was measured at 1 .5 d. 3.5 d. and 6.5 d. Measured with the MTS assay, cell attachment was signifi- cantly affected by length of culture, media, and substrate {P < 0.0001). Interaction between media and substrate was significant iP < 0.0001). At day 1.5, significantly greater attachment was detected in JL-ODRP-4 on poly-D-lysine (F < 0.05), in LA-2 on fibronectin. poly-D-lysine, and uncoated Falcon substrates (P < 0.05). and in LA-5 on fibronectin, poly-D-lysine, laminin. and uncoated Costar substrates. At day 3.5, significantly greater at- tachment was detected in JL-ODRP-4 on all substrates other than laminin {P < 0.05), in LA-2 on all substrates other than laminin (P < 0.05), and in LA-5 on all substrates other than collagen 1. At day 6.5, significantly greater attachment was detected in JL- ODRP-4 on the uncoated Costar substrate, in LA-2 on poly-D- lysine. uncoated Falcon, and uncoated Costar substrate, and in LA-5 on the uncoated Costar substrate. Measured with the CyQuant assay, cell attachment was signifi- cantly affected by length of culture, media, and substrate {P < 0.0001). Interaction between media and substrate was significant (P < 0.0001 ). Results were similar to results for attachment with the MTS assay. Cell survival and RNA production were signifi- cantly affected by length of culture, media, and substrate (P < 0.0001), with significant interaction between these three factors. In summary, cell attachment, cell survival, and RNA produc- tion were all significantly affected by the media and the substrate cells were cultured with. Furthermore, the response of cells to the various substrates tested varied with the media used indicating potential interaction between media components and substrates. OPTIMIZATION OF GENE DELIVERY FOR IMPROVED OYSTER HEALTH. John T. Buchanan. Jerome F. La Peyre. Terrence R. Tiersch, and Richard K. Cooper. Depailment of Veterinary Science. Louisiana State University Agricultural Cen- ter. Louisiana Agricultural Experiment Station. Baton Rouge. LA 70S03. E-mail: jbuchananCs'agctr. Isu.edu In Louisiana, over 300,000 acres of bottom area are privately leased for production of the eastern oyster Crassoslrea virginica. Recently, disease problems have plagued this industry, particularly the protozoan parasite Perl mean) in all groups. These preliminary observations suggest that H. nelsoni has an adverse effect on protein content and proteolytic activity in sus- ceptible oyster stocks. A MODEL OF WITHIN-HOST POPULATION DYNAMICS OF THE OYSTER PARASITE PERKINSUS MARINUS: SIMULATED EFFECTS OF TEMPERATURE AND SALIN- ITY. Lisa M. Ragone Calvo and Eugene M. Burreson, Virginia Institute of Marine Science. College of William and Mary, Glouc- ester Point, Virginia 23062. E-mail: ragone@vims.cdu A computer model was developed to investigate the population dynamics of the protistan parasite, Perl5 17.44 39.95 27.17 'H myo-inositol 0.10 2.56 1 .9(1 1,23 ' This fraction may also contain unidentified glycolipids This research was funded bv NSF (MCB9728284). ALTERED PERKINSUS MARINUS PROTEASE PROFILES UPON EXPOSURE TO SELECTED OYSTER TISSUE HO- MOGENATES. Alanna Maclntyre and Stephen Kaattari, De- partment of Environmental Sciences. Virginia Institute of Marine Science, College of William and Mary. Gloucester Point. VA, USA. E-mail: alanna@vims.edu Utilizing in vitro culture methodologies for the growth of P. marinus, we have examined the possible role of oyster tissue fac- tors in the modification and/or regulation of parasitic secretory products. Various concentrations of oyster tissue homogenates were used to supplement a defined culture medium. The growth National Shellfisheries Association. Orlando. Florida Abstracts. January 2001 549 rate and protease expression with the various supplements was then determined. It was found that although the growth rate of the F. iiianiiKs cells is relativel_\ unchanged ii[ion exposure to the homogenates of various oyster species, a distinctly different array of proteases, as determined by gelatin-gel electrophoresis were observed with Crassostrea virginicu homogenates. but not with Crassostix'd oriakensis. Typically at least five proteases within the molecular weight range of 65-135 kDa can be observed in bovine serum albumin (BSA)-free medium or BSA-free medium supple- mented with C. ariakensis or homogenates. However addition of C. virginica homogenates reveals the presence of five additional protease bands in the 10-50 kDa molecular weight range. This is of particular interest, because C. virginica is the most susceptible oyster species to P. marimis infection. This research was funded by the NOAA-Virginia Sea Grant Association. Oyster Disease Research Pro<;ram. TEMPORAL VARIATION IN THE GAMETOGENIC CYCLE OF MARINE MUSSELS. MYTILUS EDULIS AND MYTILIS TROSSULUS. IN COBSCOOK BAY, MAINE. Aaron Maloy, Bruce J. Barber, and Paul D. Rawson, University of Maine. School of Marine Science. 5735 Hitchner Hall. Orono. ME 04469-0001. E-inail: aaron.nialoyCS' uniit.maine.edu Culture of the edible blue mussel. Mytihis edulis. is expanding in the state of Maine to meet a growing demand. Current practice involves the dredging of juveniles from natural seed beds for placement on bottom leases or collection of seed on ropes for suspension from rafts. It has recently been discovered that a second species of mussel. Mytihis trossiihis. occurs sympatrically with M. edulis in eastern Maine. Future economic viability of the industry is dependent on the ability of growers to collect the more valuable M. edulis seed preferentially. A lack of natural hybridization be- tween the two species and observations of two distinctive spat settlements in the region suggests there may be a temporal differ- ence in spawning that could be utilized for preferential collecliiin of M. edulis seed. This study was undertaken to compare the gametogenic cycles of M. edulis and M. trossidus froin a single population from Cobscook Bay. Maine. Throughout 2000. one hundred to one hundred thirty mussels were sampled from a low intertidal site at monthly to bimonthl> inter\als. Three-polymerase chain reaction (PCR)-based nuclear markers and one ribosomal marker were used to differentiate be- tween mussels species. Histologically based quantitative assess- ments of gametogenesis were carried out using image analysis software. The gametic volume fraction (GVF) for twenty individu- als of each species was obtained from each sample and used to make assessments concerning gametogenic development and tim- ing of the spawning events. Preliminary analysis indicates that M. triissidus initiates sametoaenesis earlier than M. edulis. MASS MORTALITIES OF SOFT-SHELL CLAMS i.MYA ARENARIA ) IN ATLANTIC CANADA ASSOCIATED WITH UNPRECEDENTED LEVELS OF HEMIC NEOPLASIA. Sharon E. McGladdery.' Gregory S. MacCallum." Neil G. MacNair,"' and JetTrey T. Davidson". 'Department of Fisheries and Oceans. Canada. Gulf Fisheries Centre. Moncton, NB. Canada. E-mail: mcgladderys@dfo-mpo.gc.ca; "Atlantic Veteri- nary College. UPEl. Charlottetown. PEL Canada; and 'PEI De- partment of Fisheries and Environment. Charlottetown. PEL Canada. In July. 1999. a soft-shell clam [Mya areuaria) farmer reported heavy losses that seemed to be increasing in severity. A sample of moribund and neighboring, apparently healthy, clams showed lev- els of hemic neoplasia >95'7f . Low prevalences ( 75 mm) were present within the cultch. Cultched plots also contained a higher number of oyster-associated species than natural reefs and noncultched plots. After two years, densities within the cultched plots for the most abundant species including C. viiginica. Bakinus aiiiplutiite. Paii- opeus herbstii, and Eurypanopeiis depresstis were equivalent to natural reefs. These species were not present within the non- cultched plots. In addition, the fringe of oyster cultch seemed to enhance marsh use by some marsh resident nekton species and those known to have an affinity for oyster substrate, including Cyprinodon variegatiis. Fimduhis heteroclitiis, Gobionellus boleo- soma, Cobiosoma bosci. and Palaemonetes spp. The cultch may. however, serve as a barrier for species that move onto the marsh early in the tidal cycle including Lcinstoimis xauthiiiiis and Calli- necties sapidus. Cultch placement at the lower fringe of created salt marsh can reduce marsh sediment erosion and vegetation loss in areas prone to storm-related wave activity. The addition of cultch to the lower marsh fringe may promote resident marsh nekton use beyond that of noncultched marsh. Furthermore, cultch additions can provide qualitv habitat for oyster reef fauna. IN VITRO STUDIES OF DISSEMINATED NEOPLASMS OF MYTILUS TROSSULUS. James D. Moore. Hiroko Shike. Chi- sato Shiniizu. Ralph A. Elston. and Jane C. Burns, U.C. Davis Bodega Marine Laboratory, 2009 Westside Road, Bodega Bay, CA 94923. USA. E-mail: jimmooreCs'ucdavis.edu Establishment of a cell line requires immortalized or trans- formed cells that divide continuously in culture. Naturally occur- ring neoplasms provide a source of highly concentrated, rapidly proliferating, transformed cells. The only impediment to establish- ing a cell line from this material is to identify culture conditions that will support long-term growth. We are exploring the in vitro culture of disseminated neoplasms of mussels, Myliliis trossulus from Puget Sound, Washington, U.S.A. Approximately 0.5-1% of mussels in these populations contain advanced neoplasms in which more than 90% of the cells in the hemolymph are rapidly cycling, transformed cells with altered DNA content. We assessed the cycling activity of normal and neoplastic cells /;; vitro by tritiated thymidine incorporation. Cells from a mussel with an advanced neoplasm showed different levels of DNA syn- thesis in different media, thus demonstrating the utility of this assay for optimizing culture conditions. To study the transmissibility of the neoplasms, we sought a method to mark the cells genetically. Neoplastic cells were ex- posed in vitro to pantropic retroviral vectors that contain a substi- tuted envelope protein that permits infection of invertebrate cells. The vector, LLRNL (Moloney murine leukemia virus long termi- nal repeat (LTR) (L)-firetly luciferase (L)-Rous sarcoma virus LTR (R)-neomycin phosphotransferase (N)-MoMLV LTR), was used to infect cells. Transduction of cells was stable over a six-day period, was independent of polybrene, and was proportional to number of vector particles. These results suggest that pantropic vectors can be used to mark genetically neoplastic cells ex vivo. which can then be transplanted into healthy mussels to study trans- mission and provide a continuous source of neoplastic cells for in vitro experiments. OPTIMIZING CULTURE CONDITIONS FOR CREATION OF AN OYSTER CELL LINE. James D. Moore. Viviane Boulo. Chisato Shimizu. Hiroko Shike. Carolyn S. Friedman. and Jane C. Burns, U.C. Davis Bodega Marine Laboratory, 2099 Westside Road, Bodega Bay, CA 94923, USA. E-mail: jimmooreCs'ucdavis.edu To aid in the formulation of media to support oyster cells nutritionally in culture, we analyzed the hemolymph components from Crassoslrea virginica and C. gigas. We measured DNA syn- thesis in primary cultures incubated in different media and at dif- ferent temperatures to optimize culture conditions. Natioiuil .Shcllfislieries Association. Orlando. Florida Ahslracls. Januarv 2001 551 Pooled hemolymph (two to three oysters/pool) was obtained by cardiac puncture. Henioeytes were removed by centrifugation (1,000 X g) and the supernatants stored at -70°C. Analysis of hemolymph components included free amino acids, organic acids, carbohydrates, metals, electrolytes. pH. and osmolality. Cultures of heart and embryos were established according to published methods (Boulo et al. 1996. 2000). DNA synthesis in cultured cells was assessed by ""H-thyniidine uptake. Two media formulations based on published data were com- pared: (1) KS medium (based on Kleinschuster and Swink 199.^); L-15 adjusted to 750 mOsm with synthetic sea salts, and supple- mented with amino acids, lipids, carbohydrates, vitamins, 109f fetal calf serum (FCS). and 10% C. virginica hemolymph; (2) 2X L-15 (Boulo et al. 1996) adjusted to 750 mOsm with NaCl plus 10% FCS. Selected components that were lower in the media than in ihe pooled hemolymph samples are shown below (mean \ alues ± SD; NA = not available) C. virginica 2X L-15 Coniponcnl hemolymph KS medium mcdiuni/10%FBS Taurine 195.9 ± I2.lmg/L 51.9 NA Pniline 90.5 ±3.7 34.2 NA Calcium 40.7 ± 6.2 19.8 9.3 Strontium 4.7 ±0.9 3.0 0.1 Boron 4.4 ± 0.7 NA 0.1 Zinc 1.3 ± 1.3 NA 0.1 pH ranged from 6.4-6.9. Preliminary comparison of 'H-thymidine uptake by primary C. gigas heart cells cultured in different media at different tempera- tures revealed that DNA synthesis occurred between 48 and 72 h postinitiation of the cultures under all conditions. In contrast, dis- sociated embryo cells had maximal DNA synthesis during the first 24-h period in culture and 2X L-15 at 23°C supported the highest level of DNA synthesis. Additional comparisons will guide for- mulation of media to nutritionally support primary cultured oyster cells. Supported in part by California Sea Grant NA86RG(107.^ Project No. RYA-l lOB and the California Department of Fish and Game. REVIEW OF GULF OYSTER INDUSTRY PROGRAM GRANT PROJECTS: LOUISIANA OYSTER LEASES VER- SUS COASTAL RESTORATION AND CLEAN-UP OF CON- TAMINATED OYSTER BEDS. Erinn W. Neyrey, Joe Steven- son, and Michelle .Marney, LSU Sea Grant Legal Program. LSU Law Center. Rni. 170, Baton Rouge, LA 70803, USA. E-mail: eneyrey (a' lsu.edu Louisiana oysier leases versus coastal restoration: Louisiana's coastal wetlands account for more than 25% of all wetlands found within the LInited States; however, this important ecosystem is disappearing at an astonishing rate. To counter coastal erosion, Louisiana has developed an extensive plan consisting of numerous restoration projects. One of the major components of the plan is the construction of freshwater diversion projects. The diversion of sediment-rich river water is designed to mimic the natural flood cycle and allow the river's .sediments and nutrients to be deposited in the coastal zone. Oyster farmers in the restoration areas have experienced negative impacts on their over-all productivity caused by the changes in salinity. These impacts have led to lawsuits being filed by the oyster farmers against the state for damages. In response. Louisiana has passed legislation creating the Oyster Lease Relocation Program (OLRP). The OLRP is designed to offer alternatives to the oyster farmers who find themselves in a resto- ration area. Although this program may not provide a cure for all of the problems that may arise, it does provide an extrajudicial means to resolve many of the disputes between oyster fanners and the state resulting from coastal restoration projects. Clean-up of contaminated oyster beds: Louisiana is the top processor of oysters within the Gulf region, which produces nearly 60%> of the nation's oysters, and is home to nine hundred oyster farmers. Successful business planning by members of this signifi- cant industry benefits not only those in the industry, but the state as well. Relocation of leases because of fouled oyster beds effects not only the livelihood of an oyster farmer, but lifestyle as well. Both the Clean Water Act and Louisiana's Water Control Law clearly prohibit the discharge of pollutants into waters; however, these laws do little to address the actual clean-up of polluted water bodies. Therefore, solutions to the water pollution problems are turning increasingly on inteipreti\e regulation and policy of the U.S. Environmental Protection .^gency and the private citizen ac- tions that seek alternatives to the imposition of ci\'il penalties. Supplement environmental projects offer a potential flexibility that traditional enforcement actions do not. A REVIEW OF ORGANISMS ASSOCIATED WITH OYS- TERS CULTURED IN FLOATING SYSTEMS. F. X. O'Beirn, P. G. Ross, and M. W. Luckenbach, Virginia Institute of Marine Science. E-mail: francisCs'vims.edu Given the increase in oyster {Crassostrea virginica) aquacul- ture on the eastern seaboard of the United States, an issue to be considered is the influence of these intensi\e operations on faunal assemblages found adjacent to and among the oysters themselves. In fact, little is known concerning the organisms associated with such aquaculture operations. Using oysters in a typical culture scenario, we assessed the number of macrofaunal species associ- ated with floating culture systems and the number of organisms within broad taxonomic groups. Within each floating structure (2 ft X 8 ft X 1 ft) oy.ster numbers ranged from 488 to 1.381. Overall, forty-four species of macrofauna were recorded from the floating structures. The species richness within each float ranged from twenty-four tothirty-six. Abundances ranged from 12.746 to 92.602 individuals per float. In an attempt to determine if the organisms were influenced directly by live oysters or the structure 552 Abstracts. January 2001 National Shellfisheries Association, Orlando, Florida provided by oyster shell, we evaluated communities associated with live oysters, oyster shell, and empty floats (culture systems). Preliminary results indicate that the live oyster and oyster shell have similar abundance and biomass of individuals, but the com- position of individuals is different. These results suggest that struc- ture of any sort is important to certain species (e.g.. grass shrimp), but the presence of live oysters influences the presence of other species (e.g., anemones, blue mussels). Given the ephemeral na- ture of these communities (mediated by harvest and handling schedules), many organisms within these floating communities may not actually mature to reproduce, and these systems may be sink populations. Certain associated species may actually compete with the oysters for food resources (e.g., blue mussels) and in- crease the organic loading emanating from the culture systems. In addition, altering the habitat type in a particular area may change the constituent organisms of a community and species interactions within a system. For example, fishes nesting or taking refuge in the cages may reduce grazer abundances (e.g.. amphipods): thereby, affecting macroalgal abundances. As the scale of aquaculture op- erations increase, it will be increasingly important to further elu- cidate these and other ecolonical interactions. SPATIAL AND INTERANNUAL OCCURRENCE OF A BROWN SHELL CONDITION IN NEWFOUNDLAND FARMED BLUE MUSSELS {MYTILUS SPP.) G. Jay Parsons. Kelly Moret, Cyr Couturier, and Miranda Pryor, Centre for Aquaculture & Seafood Development. Marine Institute. Memorial University. P. O. Box 4920. St. Johns. NF. Canada AlC 5R3. E-mail: Jay.Parsons@mi.mun.ca The appearance of a brown coloration on the surface of blue mussels in Newfoundland prompted an investigation into the ex- tent, potential impact on mussels, and factors influencing its oc- currence throughout the province. In severe infections, this brown shell condition results in the loss of the periostracum and has been described as the "mycotic periostractal sloughing" (MPS) disorder. The etiological agent is a fungus yet to be identified (T. J. David- son, pers. comm.); however, similar mycotic infections are com- mon in cultured mussels worldwide. The mode of transmission is unknown, but studies on PEI mussels suggest that initial infection occurs in early fall in mussels that are more than a year old (T. J. Davidson, pers. comm.). The objectives of this study were to de- temiine the spatial distribution of the fungal shell infections (FSI) (resulting in the brown shell condition) throughout Newfoundland, determine the interannual variability, and determine if progression is size related. The shells of mussels were examined from twenty-one culture sites throughout the province in 1998 and twenty-two sites in 1999. There were about sixty mussels sampled from each site. Mussels ranged in size from about 35 to 85 mm shell length. Mussels were examined for the occurrence of FSI on the surface of the exterior shell using a dissecting microscope to observe the occurrence of brown foci. On mussels with FSI present, the percentage coverage (to the nearest 5%) was recorded. Data on the occurrence and percent coverage were summarized by farm site and geographic zone. The occurrence of FSI in both years ranged from 0 to 100% of mussels at sites around Newfoundland. At sites where the brown shell condition was recorded, the percentage coverage ranged from 1 to 9y:'f and 1 to 99% for 1998 and 1999, respectively. When the farms were grouped by zones, the most northern sites had the lowest occurrence of FSI and the lowest percentage coverage. Sites in central Newfoundland had a higher rate of occurrence but a low percentage coverage. Sites on the south coast had both a high rate of occurrence and a high percentage coverage of FSI. A similar pattern occurred in both years of the study. There were, however, no consistent differences in the occurrence or percentage coverage of FSI on individual sites from 1998 to 1999. Some sites had a lower rate of occurrence of FSI in 1998 as compared to 1999. but other sites had a higher rate in 1998 as compared to 1999. Per- centage coverage was significantly, but weakly, coirelated to mus- .sel shell length for both 1998 and 1999. There was no apparent decrease in condition or health in mussels with the brown shell condition. Mussels with FSI were found throughout Newfoundland; how- ever, the condition known as MPS was rarely observed even in larger, older mussels. Although there were distinct geographic dif- ferences in the rate of occurrence and percentage coverage, we have not yet been able to relate these differences to specific envi- ronmental parameters or husbandry conditions, except that it seems size/age, temperature, and husbandry practices may influ- ence the extent of coverage. INDUCTION OF TRIPLOIDY IN THE AMERICAN OYS- TER CRASSOSTREA MRGINICA: A RE-EVALUATION OF POLAR BODY I INHIBITION. Stefano Peruzzi and Ximing Guo. Haskin Shellfish Research Laboratory. Institute of Marine and Coastal Sciences, Rutgers University, 6959 Miller Avenue, Port Norris, NJ 08349, USA. E-mail: speruzzi@hsrl.rutgers.edu In mollusks. triploidy can be induced by inhibiting either polar body I (PBl) or polar body 2 (PB2). It has been shown in the Pacific oyster that PB 1 inhibition results in primarily aneuploids and heavy larval mortality. In the American oyster, an early study has shown that PBl triploids grow faster than diploids and PB2 triploids. possibly because PBl triploids are more heterozygous. The objective of this study is to re-evaluate if PBl inhibition is a valid method for triploid induction and if PBl triploids grow sig- nificantly faster than diploids and PB2 triploids in New Jersey. Growth performance of triploids is known to vary among different environments. In this work, three experimental groups were produced: a nor- mal diploid control, a PB 1 triploid group, and a PB2 triploid group. Cytochalasin B (0.75 mg/L) was used as the induction agent. The Natioinil Shellfisheries Association. Orlando. Florida Abstracts, January 2001 553 experiments were replicated four times using a total of twenty-five females and fifteen males oysters. Both methods were effective in generating variable levels of triploidy (41 to lOC/r). Viability of PBI triploids was actually higher than PB2 triploids over the first two months of life. Oysters will be deployed at two sites in NJ. Oysters will be sampled at regular intervals for ploidy determina- tion and growth measurements. Genomic heterozygosity of experi- mental oysters will be determined using allozyme markers. Sam- pling and analyses are currently underway and will be reported at the meeting. REDUCTION OF RED TIDE TOXIN IN CLAMS BY OZONE PURIFICATION AND RELAYING. Richard Pierce, Michael Henry, and Gary E. Rodrick. University t)f Florida. Department of Food Science and Human Nutrition. Gainesville. FL, USA. E-mail: GER@GNV.IFAS.UFL.EDU A study of the accumulation and purification of harmful algal biotoxins in the clam. Mercenaria menenaria was initiated to investigate the use of ozone as a means to enhance the natural purification of toxins. Clams were exposed to viable cells of the Florida red tide organism, Gymnodiniiim breve (3 x 106 cells/ clam/day) for nine days. Following exposure, the clams were di- vided into groups for natural relay versus ozone purification. The concentration of brevetoxins (PbTx-2 and PbTx-3) in the exposure water and clam tissue was monitored by high-power liquid chro- matography-ultraviolet (HPLC-UV) analysis and by receptor- binding assay. The amount of total PbTx-2 and PbTx-3 in the original culture before dilution for clam exposure and brevetoxin concentration in the effluent water after exposure clams show that the clams were exposed to appropriate levels during the study. These results show the presence of brevetoxins in the spiked samples and a definite accumulation of active toxins after ten days of exposure. After three days of ozone purification, the treated clams exhibited a 30* drop in the brevetoxin activity level; whereas, the nonozone-treated clams exhibited a 380* drop in the ten-day exposure level. Relaying the red tide contaminated clams to a clean seawater area for fifteen days removed (lOOVr) toxin activity when compared to nonrelayed controls. OYSTERS REEFS AS HABITAT FOR FISH AND DECA- PODS: SPECIES AND LANDSCAPE CONSIDERATIONS. Martin H. Posey, Troy D. Alphin, and Thomas K. Frazer, Cen- ter for Marine Science and Department of Biological Sciences, University of North Carolina at Wilmington. One Marvin K. Moss Lane. Wilmington. NC 28409. USA. E-mail: Poseym@uncwil.edu Structural habitats have long been recognized as miportant nursery areas for juvenile fish and decapods, as indicated by work in seagrass and marsh habitats. Recent interest has focused on the potential importance of oyster reefs as a refuge habitat. Along much of the southeastern coa.st of the United States and in some areas of the Gulf of Mexico, seagrasses are largely absent, and oyster reefs may provide the primary structural habitat in low to shallow subtidal areas. However, use of these oyster reefs may vary among species and may depend on the occurrence of other habitat types. We have examined abundances of decapods and fish in oyster and marsh edge/sand habitats in two regions of North Carolina — one that has seagrasses (New River) and one that completely lacks seagrass beds (Hewletts Creek). In the absence of seagrass beds, grass shrimp and certain fish were more abundant within oyster beds during peak periods of abundance in winter and spring (Fig. 1 ). However, other decapods may not show this pattern. In the New River, where seagrasses provide an alternative habitat, there were low abundances of all groups sampled within oyster beds (Fig. 2). Abundances were somewhat higher in sand/marsh edge habitats and were much higher for most groups within seagrass beds. In all cases, there are strong seasonal effects on use of oyster beds, including shifts in the major species present. Our results emphasize the |Totential importance of oyster reefs as habitat for certain fish and decapiids. However, the patterns of use varied between locations (especially with the presence of al- ternative habitats), varied between taxa, and varied seasonally. DERMOWATCH: A WEB-BASED APPROACH FOR MAN- AGING PERKINSUS MARINUS DISEASE OF OYSTERS. Sammy M. Ray. Thomas M. Soniat. and Enrique V. Kortright, Department of Biological Sciences. Nicholls State University, Thi- bodaux, LA 70310, USA. E-mail: biol-tms@mail.nich.edu A web site called DermoWatch (www.blueblee.com/dermo) has been established to help manage Perkiiisus iiniriinis ( = Der- mocystkliuin marimim) disease of eastern oysters, Crassostrea vir- i;liilca. The main page provides the most recent data for nine stations in Galveston Bay. Texas. Data include water temperature (T) and salinity (S). weighted incidence (WI) and percentage in- fection, and estimated time lo a critical level of disease dcrJ- Archived historical data for each site is available from the main page and via a map of the bay. With a utility called the Dermo Calculator, the web site is useful wherever dermo disease is found. The Dermo Calculator allows anyone with information on water T and S, oyster length (L). and initial WI of disease to calculate a '(. nr ^ prototype utility has also been developed that uses real-time values for water T and S from a fixed monitoring station. The utility requires input of the time period of interest and L: it returns, in graphical form, a time course of T, S and t^.^,,. More frequent 554 Abstracts. January 2001 National Shellfisheries Association. Orlando, Florida values of T and S permit more frequent estimates of tcr„ and should increase the reliabilitv of the model. INTER- AND INTRA-SPFXIFIC GENETIC VARIATION AMONG PERKINSUS SPECIES: IMPLICATIONS FOR SPECIES IDENTIFICATION AND DEVELOPMENT OF MOLECULAR DIAGNOSTICS. Kimberlv S. Reece, Gwynne D. Brown, Karen L. Hudson, and Kathleen Apakupakul, Vir- ginia Institute of Marine Science, The College of William and Mary, Gloucester Point, VA 23062, USA. Peikinsiis species are protozoan parasites of marine and estua- rine mollusks. Globally, these pathogens have had devastating ef- fects on wild and cultured host populations. There is considerable interest in developing efficient, sensitive, and accurate means of identifying marine pathogens to facilitate monitoring and devel- opment of appropriate management practices. The traditional methods for detection of Perkinsits species, histological analysis, and the fluid thioglycollate medium (FTM) assay, is time consum- ing and does not distinguish among recognized Perkinsiis species. Molecular diagnostic techniques have the potential to improve on the speed and sensitivity of more traditional diagnostic methods. Unfortunately, molecular protocols often prove to be too expensive and technically cumbersome to be practical for routine monitoring purposes. In addition, unless molecular sequence data are available for many strains of a particular species and for closely related species, there is the chance that a molecular diagnostic may give false-negative or false-positive results. To design genus- and species-specific molecular diagnostics appropriately it is best to obtain DNA sequence information from as many different strains of a particular species and as many spe- cies within the genus as possible. We have been examining the DNA sequences for the ribosomal RNA small subunit gene, the internal transcribed spacer (ITS) region, and an anonymous locus from genetic strains of P. marinus and from other species of Per- kinsiis. Examination of intra- and interspecific DNA sequence variation, as well as alignments to DNA sequences of other alveo- lates, has allowed development of Perkinsits probes for in situ hybridization and primers for use in polymerase chain reaction (PCR). We have a genus-specific DNA probe that hybridizes to Per- kinsus species (except P. ifiigwadi) infecting tissues from a variety of hosts. Perkinsus cells have been detected in host tissue samples from Asia. New Zealand. Australia. North America, and Europe. PCR primers have been designed based on sequence analysis of the ITS region that specifically amplify Perkinsus species DNA from host tissue. In addition. DNA sequencing of the amplified ITS region has allowed identification of Perkinsus species. We also have species-specific PCR primers designed from DNA se- quence of other areas of the genome to amplify only P. marinus. We are testing and optimizing a quick, inexpensive, and sensitive PCR assay for detection of Perkinsus species. HSP70 RESPONSE IN OYSTER CRASSOSTREA VIR- GINICA EXPOSED TO CD"* AND PAHs SORBED TO AR- TIFICIAL SEDIMENTS. Luis A. Cruz-Rodriguez and Fu-Lin E. Chu, Virginia Institute of Marine Sciences, School of Marine Sciences, College of William and Mary, Gloucester Point, VA 23062. USA. E-mail: lcruz@vims.edu The induction of Hsp70 has been used as a marker of exposure to pollutants. Ongoing studies are determining if changes in Hsp70 levels could be used as indicators of environmental quality. Our previous studies showed a general increase of total Hsp70 in oys- ters exposed to field-contaminated sediments containing a mixture of PAHs, heavy metals, and PCBs. To analyze further which con- taminants are responsible for the response, we exposed oysters to a heavy metal (Cd*") or PAHs sorbed to artificial sediments. Oys- ters were exposed to 1 5 ppb or 25 ppb Cd*" sorbed to I g or 2 g, respectively, artificial sediments, and to 40 |j.g, 200 p-g or 400 p.g PAHs sorbed to I g, 1.5 g, or 2 g, respectively, artificial sediments daily for 40 days. Concomitantly, to determine the effect of sedi- ments alone, oysters were exposed to 0 g, I g, 1 .5 g. or 2 g sediments for forty days. Oysters exposed to 1 g, 1.5 g, or 2 g sediments showed no statistically significant difference in Hsp70 levels compared to the 0 g oysters after forty days (Fig. lA). Oysters exposed to PAHs- sorbed to the sediments showed a significant increase in the Hsp70 response compared to those exposed to sediments only, but no dose dependency was observed. Oysters exposed to 15 ppb or 25 ppb Cd*" did not show a statistically significant effect as compared to those exposed to sediments only, although the oysters exposed to the highest dose (25 ppb) showed an increase in the Hsp70 levels (Fig. 1 B). In a separate study, however, Cd*" .sorbed to algae induced a statistically significant Hsp70 response as compared to controls (data not shown). PAHs seem to be more available and/or have a greater exchange between the sediments and the oyster gills than the heavy metal. In summary, based on the above results, Hsp70 potentially could be used as an indicator of exposure to contaminants (PAH and Cd*") in aquatic environments. This re- search was funded by the EPA. COMPARISON OF OYSTER (CRASSOSTREA VIRGINICA) CULTURE NURSERY AND GROW-OUT TECHNIQUES. P. G. Ross, F. X. O'Beirn, and M. W. Luckenbach. E-mail: pg@ vims.edu Along the mid-Atlantic coast, decline of wild eastern oyster (Crassaslrea virginica) stocks, combined with collateral declines of other fisheries, has prompted development of alternative sources of income from nearshore waters. As a result, interest in oyster aquaculture in the region has increased. Initially, natural reefs were enhanced to take advantage of natural spawning and subsequent settlement and growth of wild oysters. However, the current trend is toward culture of hatchery-spawned oysters. Currently, most commercial oyster aquaculture focuses on three phases: hatchery. N;ition;il Sliellfisheries Association. Orlando, Florida Abslracl.s. January 2001 555 nursery (~3 to 25 nnn shell height) and grow-out (->25 mm shell height-market size). Different techniques are being utilized in each phase. Much research has focused on the hatchery phase with different nursery and grow-out techniques being utilized based on anecdotal comparisons. To further a technical comparison of post- hatchery culture, we evaluated growth and mortality of two nurs- ery and several grow-out techniques. Two nursery techniques were evaluated and compared: (1) fine mesh bags (1/16-in mesh) sus- pended in the water column via polyvinylchloride (PVC) tloats and 1-in x 1-in wire mesh cages, which relies on natural water Mow; (2) a forced upweller system (tlupsy) utilizing a pump and suspending oysters in the water column via square plastic "silos" with mesh bottoms, which enhances water flow mechanically. Oysters entered both systems at the same size ( x = 3.08 mm shell height, SD = 0.36) and density (-2,500 oysters). After si.xty-three days (April 5 to June 7) oysters had grown to 20.99 mm (SD = 3.02) in the "tlupsy" and 18.87 mm (SD = 4.27) in the bag/tloat system. Low mortality (s 1%) was observed for both culture techniques. Although there was a slight growth differential and similar mortality, the most striking difference between these two techniques is their size variation. Variable sizes at the end of the nursery phase can lead to increased handling and sorting, leading to higher costs, lii addition, two densities of oysters were com- pared within the "tlupsy" system. Oysters were stocked at -2.500 oysters and -120.000 and. after si.xty-three days, had grown to 20.99 mm (SD = 3.02) and 18.51 mm (SD = 2.98). respectively. Mortality was similar at both densities (s I'/r). With similar mor- tality and size variance, the most important density effect within the "tlupsy" system seems to be on growth. Cost comparisons are currently being analyzed for both techniques. Three grow-out tech- niques are currently being evaluated: ( 1 ) mesh bags (3/8-in mesh) in PVC/wire mesh tloats (subtidal system): (2) mesh bags in in- terlidal 1-in x 1-in wire mesh trays with legs that keep oysters -4 to 6 in off the bottom: and (3) mesh bags placed directly on intertidal shell bottom. Growth, mortality, and costs will be com- pared between these techniques. It is expected that the advantages conveyed by subtidal growth will be offset by reduced biofouling intertidally. IMPACT OF WATERFLOW ALTERATION UPON OYS- TER GROWTH AND DISTRIBUTION WITHIN ESTUAR- IES OF SOUTHWEST FLORIDA: IMPLICATIONS FOR MANAGEMENT AND RESTORATION. Michael Savarese and Aswan! K. Volety, Florida Gulf Coast University. 10501 FGCU Boulevard South, Fort Myers, FL 33965, USA. E-mail: msavares@fgcu.edu Water management practices within the fast growing coastal regions of Southwest Florida have drastically altered natural water quality conditions within estuaries. Two of the more profound changes have been to salinity and nutrient influx. Storm water management has caused some estuaries to receive excessive fresh- water: whereas, others receive less because of intenuption of sheet tlow: and runoff from suburban and agricultural lands delivers higher nutrient loads to some estuaries. Many of these estuaries are targeted for restoration; however, our knowledge of the prealter- ation conditions is limited. This research employs the eastern oys- ter, Crassostrea virfiinica. as an indicator species of ecosystem health. The patterns of reef distribution, oyster productivity, and living density compared among pristine and altered estuaries with historically similar hydrologic conditions are used to establish tar- get water quality conditions for restoration efforts. A "spatial ho- mologue" approach is employed, whereby conditions are com- pared at sites with similar geomorphologies and geographic posi- tions along the estuarine axis. In the Faka-Union estuary, a system that receives excessive freshwater during the rainy season, the distribution of reefs, the regions of maximum living density, and the foci of maximum oyster productivity are displaced seaward relative to pristine estu- aries. Henderson Creek, an estuary receiving greater input of nu- trients, has oyster populations w ith higher mean productivities and higher living densities. Two demographic patterns of oyster pro- ductivity are consistently discernable in all estuaries: one of greater variance in biomass. indicating a wider age distribution; and a second skewed heavily toward smaller individuals, indicat- ing greater levels of juvenile mortality. The first pattern persists at sites that experience pulsed release or protracted inundation of freshwater. Faka-Union. because its watershed area has increased significantly because of management practices, exhibits the "fresh- water demographic signal" at all but the most downstream spatial honiologue. Here, freshwater inundates the estuarine bays for five to SIX months of the year. The upstream homologues within the estuary below Henderson Creek's weir, whose simplistic design releases pulses of freshwater, exhibit the same freshwater demo- graphic signal. The alternative, "marine demographic pattern" dominates in natural watertlow settings and at the downstream homologues in Henderson. The persistence of small-sized oysters with few older, large individuals, seen in the marine demographic pattern, may be related to the greater occurrence of the disease- causing protozoan Pcrkinsiis within environments of higher salin- ity. Both Henderson Creek and Faka-Union are scheduled for wa- tershed restoration. These patterns of oyster distribution will help establish target environmental conditions. LABORATORY-BASED TRANSMISSION STUDIES OF QUAHOG PARASITE UNKNOWN (QPX) IN MERCE- NARIA MERCENARIA. Roxanna Smolowltz, Dale Leavitt, Bruce Lancaster, Ernie Marks, Rhea Hanselmann, and Chris- tine Brothers, Marine Biological Laboratory. 7 MBL St., Woods Hole, MA 02543, USA. E-mail: rsnioKs' mbl.edu Disease caused by the protistan QPX has caused severe mor- tality in aquacultured clams stocks in Massachusetts in the past few years. Before 1995, it had only been identified in Canada, but 556 AbMiiuis. January 2001 National Shellfisheries Association. Orlando. Florida had been responsible for mortality in liard clams there. In Massa- chusetts, up to 447f of clams sampled from a group of severely infected submarket sized clams have shown swellings and nodules in the mantles. Often these lesions are located at the mantle edge, close to. or directly adjacent to the siphon or adductor muscle. Mortality is usually the most severe in the spring and summer months and is associated with the appearance of the gross lesions in the mantle. The QPX organism has been identified as an unusual member of the phylum Labyrinthulomycota. probably in the family Thraus- tochytriidae. As QPX proliferates in culture media, it produces abundant mucoid material that binds the individual organisms to- gether. This mucoid material, which is also produced as QPX grows in clam tissues, may prevent phagocytosis and. thus, act as a pathogenic mechanism in the clam tissues. Cultured QPX pro- liferates best, and produces abundant mucus, at 22°C. The method of infection of clams by QPX is unknown. To study the pathogenesis of QPX infection and resultant disease in hard clams, we designed methods for the study of QPX transmis- sion in the laboratory by exposing naive clams to QPX. Methods used involved injecting clams with cultured QPX (with and with- out the associated mucoid material) in the mantle and pericardial sac. exposing clams to cultured QPX added to the water column and exposing clams to infected clams harvested from severely infected leases. In one set of experiments, clams were placed in a raceway and exposed to filtered seawater at two temperatures and fed two different levels of prepared algae preparations. In the second set of experiments, animals were housed in aquaria at room temperature in the laboratory. Clams used in these experiments were approximately one-year-old. Interestingly, results to date shov\ that QP.X is a directly infec- tive organism. Naive clams, exposed over a twelve-month period to infected clams, developed mild focal QPX infections after three months of exposure, which progressed to severe infections and associated mortality at the end of the experiment. In contrast to the identification of the mantle as the primary location in submarket- sized aquacultured clams, the one-year-old clams first developed infections in the open vascular spaces at the base of the siphon. Clam injected with QPX washed of the mucus coat or exposed to washed QP.X in the water column did not develop infections or disease. Studies of clams exposed to mucoid enveloped QPX are ongoing. FATTY ACID COMPOSITION AND S^ NTHESIS OF PER- KINSUS MARIM'S MERONTS AND PREZOOSPORAN- GIA. Philippe Soudant, Fu-Lin E. Chu, Eric Lund, Jerome La Peyre, and Aswini Volety, Laboratoire de Physiologie des Inver- tebres. Centre IFREMER de Brest BP 70. 29280 Plouzane. France. E-mail: Philippe.Soudan@ifremer.fr Although parasitic protozoa generally require exogenous sources of essential fatty acids to support their growth and life cycle completion, some are capable of modifying exogenous fatty acids (FAs). Thus, their fatty acid profiles could differ from their culture media and their hosts. To investigate whether P. mariims maintains its characteristic fatty acid profile when cultured outside its host, meronts of this parasite were cultivated in two media with different lipids: (1) medium with lipids derived from bovine fetal serum (terrestrial characteristic lipids); and (2) medium with cod liver oil extract (marine characteristic lipids). Results revealed that despite the difference in lipid sources in the two media. FA com- position of P. mariniis meronts cultivated in these two media were similar and were dominated by 14:0. 16:0. 18:0. 18:I(h-9). 20:l(«-9), 18:2(H-6), and 2QA{n-6) (Fig. 1 ) Also, the FA quantity in meronts increased significantly as compared with amounts con- tained in their media. This suggests that P. marimis is able to de novo synthesize fatty acids during proliferation. The FA profile of prezoosporangia developed from tissue-associated meronts (PDFTAM) cuhivated in fluid thioglycollate medium (FTM), which is deficient in 20:4(»-6) and low in 20: 5{n-'i) (\A%) and 22:6(;)-3) (0.6%) resembled its host, although it contained a higher weight percentage of 20:4(/i-6) (Fig. 2). Surprisingly, the pr- ezoosporangia developed from tissue-free meronts (PDFTFM) cul- tivated in alternate FTM supplemented with cod liver oil had very high levels of 18:l(«-9) and very low levels of 20:5(/;-3) and 22:6(«-3). This research was funded bv NSF (MCB9728284). DEFENSE-RELATED ACTIVITIES IN "NATURAL DERMO RESISTANT" OYSTER STOCKS. Shawn Stickler, Vincent Encomio, Luttrell Tadlock, Jerome LaPeyre, Standish K. Allen, Jr., and Fu-Lin E. Chu, Virginia Institute of Marine Science, College of William and Mary, P.O. Box 1346, Gloucester Point, VA 23062. USA. E-mail: stick@vims.edu The restoration of eastern oyster {Crassostvea virfiiinca) popu- lations can be accelerated with the development of strains resistant to dermo disease, caused by the protozoan parasite, Perkinsus marinus. To date, this has meant the slow, methodical approach of selectively breeding oysters that have survived repeated exposure to the parasite. The identification of natural ""dermo-resistant" wild stocks and related defense activities can benefit selection efforts. Our goals were to identify wild populations with natural dermo resistance or tolerance (NDR) in a common garden experiment and to identify effective defense activities in surviving individuals or populations for use as breeding markers. We collected and spawned oyster from putative resistant and susceptible control stocks from Louisiana (LOT. LHB. LOB ) and the Chesapeake Bay (CCR. CRB. CTS). and a hatchery strain (XB). The progeny were deployed in December 1999 at two dermo-enzootic bay sites. Kin- sale and Reeent for crow-out. Growth and mortalitv have been National Shellt'isheries Association. Orlando, Florida Abstracis. January 2001 557 recorded monthly and samples analyzed for dermo prevalence and intensity. As of July 2000, the CRB stock has grown significantly faster at both sites. Mortality levels, unrelated to Perkinsus infec- tion, were low in all stocks, and lower in Louisiana (L) than Chesapeake (C) stocks (Fig 1). Infection prevalence and intensity were minimal in all stocks. In July, we collected hemolymph samples to assay hemocyte counts, and we measured plasma pro- tein and lysozyme levels. Protein levels were significantly differ- ent between stocks, but lysozyme levels were not (Fig. 2). Future hemolymph samples will also be assayed for P. mahnus killing by hemocytes. and plasma protease inhibition activity. This research was funded by the NOAA-Virginia Sea Grant-Oyster Disease Re- search Prosram. NATURAL DERMO RESISTANCE AND ITS ROLE IN THE DEVELOPMENT OF HATCHERIES FOR THE GULF OF MEXICO. Shawn M. Stickler. Eric Wagner, Vincent G. En- coniio. Standish K. .Allen. Jr.. and Jerome F. LaPeyre, Aqua- culture Genetics and Breeding Technology Center (ABC). Virginia Institute of Marine Science (VIMS), P.O. Box 1346. Gloucester Point, VA 23062, USA. E-mail: stickCs'vims.edu The value of developing selectively bred dermo-resistant oyster seed for aquaculture and perhaps stock enhancement is obvious. An important step in developing such a value-added product is identifying wild stocks that may have acquired resistance as a result of natural selection. Our goal was to provide clear evidence for the existence of "naturally resistant populations" of eastern oysters Crassostrea viri;iiiica by using putatively dermo-resistant brood stock from both the Gulf of Mexico and the Chesapeake Bay and determining the inherent resistance of their progeny to demio disease in a common garden experiment. We identified and collected putative resistant and susceptible control stocks from both the Gulf (HAB, GTR, OYS) and the Chesapeake Bay (RPP. TSO. COK. XXB) that were spawned at the ABC Gloucester Point hatchery in the summer of 1999. Seed were deployed that fall in Taylor floats at two dermo-enzootic Gulf sites. Grand Isle and Grand Terre. Growth and mortality were recorded monthly, and samples were analyzed for dermo preva- lence and intensity using a modified body burden assay. As of July 2(X10, dermo prevalence was similar for all stocks; whereas intensity le\els in non-Louisiana stock were considerably higher than in Louisiana stocks (Fig. 1 ). Mortality levels correlated with disease prevalence. Louisiana stocks showed little mortality attributed to disease, whereas, imported stocks experienced high mortality levels. Although there are differences between LA and CB stocks, there is little difference between the two putatively resistant populations from LA and their susceptible control HAB (Fig. 2). VALIDATION OF DNA-BASED MOLECULAR DIAGNOS- TICS FOR THE HARD CLAM PARASITE QPX (QUAHOG PARASITE UNKNOWN) AND THE OYSTER PARASITE SSO {HAPLOSPORIDIVM COSTALE). Nancy A. Stokes. Lisa M. Ragone Calvo. Kathleen Apakupakul. Eugene M. Burre- son. Inke Sunila. and Roxanna Smolowitz. Virginia Institute of Marine Science, College of William and Mary, Gloucester Point, VA 23062. USA. E-mail: stokes@vims.edu Traditionally, the diagnosis of parasitic diseases of bivalve mollusks has primarily relied on standard histological techniques. In recent years, DNA-based molecular diagnostics have been de- veloped for a suite of bivalve parasites; however, for the most part, the utility of these novel tools for shellfish health certification, routine disease monitoring, and basic epizootiological research has yet to be determined and is dependent upon validation with stan- dard diagnostic techniques. DNA-based molecular diagnostics. DNA probes, and polymer- ase chain reaction (PCR) assays, were recently developed in our lab for two protistan parasites of bivalve mollusks, QPX (Quahog Parasite Unknown) and Htiplospoililiiiiii cosiale (Seaside Organ- ism or SSO). QPX is a parasite of hard clams, Mercenaria iner- cciuiria. that has been found in cultured hard-clam populations in New Brunswick and Prince Edward Island. Canada and in Mas- sachusetts, New Jersey, and Virginia. U.S.A. The disease has spread south to the United States in the last decade, causing sig- nificant clam mortalities in some areas. SSO is a parasite of eastern oysters, Crassostrea virii'mica. The parasite has been reported in oysters on the Atlantic Coast of the United States from Virginia to New Jersey, and organisms histologically similar to SSO have been noted in the northeast up to Maine. Studies using DNA probes recently confirmed the presence of SSO in Connecticut and Massachusetts. Historically, oyster mortalities attributed to SSO were confined to Virginia and Maryland: however, in 1998, the parasite was associated with oyster mortalities in Massachusetts. The no\el molecular diagnostic tools for QPX and SSO have been shown to be sensitive and specific. We are presently con- ducting studies to validate QPX and SSO PCR assays with the standard histological diagnostic method. Cultured hard clams are being sampled bimonthly from a Virginia coastal embayment and diagnosed for QPX by histological examination and PCR of tissue DNA. Oysters from Virginia, Connecticut, and Massachusetts are being sampled monthly and diagnosed for SSO and Haplospo- ridiuni netsiini (MSX) by histological examination and by PCR of hemolymph DNA and tissue DNA. Both studies began in spring 2000 and will continue at least one year. To date diagnostic results between methods have been comparable for QPX: whereas, PCR has resulted in detection of more SSO infections than histological examination. 558 Abstmcls, January 2001 National Shellfisheries Association, Orlando, Florida LOUISIANA'S DERMO ADVISORY PROGRAM: INCI- DENCE AND PREVALENCE OE PERKINSUS MARINVS ON LOUISIANA'S PUBLIC OYSTER GROUNDS. John Su- pan, Ron Dugas, Tom Soniat, Jerome Lapeyre. Ron Thune, John Hawke, and Al Camus, Office of Sea Grant Development, Louisiana State University, Baton Rouge, LA 70803, USA. E- mail: jsupan@lsu.edu Seed (25 to 75 mm) and market oysters l>75 mm) were col- lected along coastal Louisiana and analyzed for Perkinsus marimis during the summer, 1997 to 1999. The sampling program is funded by the Louisiana Oyster Task Force in response to industry obser- vations of high oyster mortality on the state's public oyster grounds. Sampling locations included Cabbage, Three-Mile Bay, Mozambique Pt., Black Bay, Telegraph Pt., Bay Crabe. Bay Gar- dene. Lonesome Island, Hackberry Bay, Sister Lake, Bay Junope, Vermillion Bay, and Calcasieu Lake oyster reefs. The assay uti- lized rectal tissue in Ray"s Fluid Thioglycollate Media, using the Mackin scale for qualification. The sampling program is the basis for the state's Dermo Advisory, printed annually in the task force's newsletter to assist industry in the management of private oyster leases. State fisheries managers also use the data to manage the public oyster grounds. Perkinsus intensity varied annually at each site and oyster cat- egory and was greater during 1997 than subsequent years. On the prime grounds in the eastern portion of the coast, seed oysters ranged from 0.1-1.9 weighted incidence, with eight out of nine stations >1.0; prevalence ranged from I6-I007r, with six stations >90%. Market oysters ranged from 0.6-2.0 and 59-100%. respec- tively. Vermillion Bay was the lowest site during all three years. Incidence and prevalence has declined during 1998 to 1999 at most stations: whereas. Hackberry Bay. in the central portion of the coast, had the highest infection levels for market oysters during 1999. High market oyster mortality and heavy spatfall have been prevalent during these latter years. Sampling for 2000 is underway and will be incorporated in the annual Dermo Advisory. Preliminary stock assessments indicate greater mortality in market oysters a most stations. Higher levels of Perkinsus are expected because of severe drought. The advisory will soon utilize the new DermoWatch program now online, funded by the Gulf Oyster Industry Program. DISEASE STATUS AND PHYSIOLOGICAL RESPONSES OF OYSTERS AS INDICATORS OF WATERSHED AL- TERATION EFFECTS IN SOUTHWEST FLORIDA ESTU- ARIES. Aswan! K. Volety, Michael Savarese, and S. Gregory Tolley. Florida Gulf Coast University, 10501 FGCU Boulevard South, Fort Myers, FL 33965, USA. E-mail: avolety@fgcu.edu Southwest Florida possesses one of the country's fastest grow- ing populations. Consequently, watersheds are heavily managed to accommodate development. Studies on the effects of watershed alterations involving valued ecosystem components, such as oys- ters, are lacking, but clearly necessary. Using the oyster. Crassos- trea \iri>inica. as an indicator species, we are investigating eco- system-wide health effects of watershed management practices in altered (Faka-Union, Henderson Creek, and Caloosahatchee River) and pristine (Blackwater River) estuaries. Currently, watershed management involves the simple opening and closing of weirs in the estuaries, thus delivering pulses of freshwater during rainy season and decreasing freshwater input into the estuaries during the dry .sea.son. This results in the estuaries being mostly freshwa- ter when the weirs are open and mostly higher salinity water when the weirs are closed. Measurements of disease prevalence of Per- kinsus marinus. condition index, and substrate suitability of healthy oyster reefs as essential fish habitat are underway using a "spatial homolog approach" (comparing conditions among hydro- logically and geomorphically similar points between estuaries) along salinity gradients. Preliminary results indicate that in sum- mer months, mean prevalence of P. marinus infection in oysters varied between 33-73% depending on the location: mean condi- tion index varied between 2.4 to 4.7 during July 2000, although lower disea.se prevalence was noted in oysters from homologs 2 and 3 in Faka-Union and is likely attributable to the greater fresh- water input into the estuary, homolog I (upriver) was devoid of oysters because of freshwater-induced mortality. The lower dis- ease prevalence in homologs 3-5 (down river) in Blackwater River and Henderson Creek may be attributable to the disease-induced mortality of these heavily infected oysters during the summer months. This project represents the first study of watershed alter- ation on oysters in Southwest Florida and will help provide target en\ ironmental conditions for restoration efforts. PARTIAL PURIFICATION AND CHARACTERIZATION OF L\SOZYME-LIKE PROTEINS FROM THE PLASMA OE THE EASTERN OYSTER, CRASSOSTREA MRGINICA. Aswan! K. Volety, Fu-L!n E. Chu, and Lu!s Cruz-Rodriguez. Florida Gulf Coast University, College of Arts and Science, 10501 FGCU Boulevard South, Ft. Myers. FL 33912, USA. Lysozyme(s) (and lysozyme-like proteins) are shown to be in- volved in a broad battery of such defense mechanisms as bacteri- olysis and opsonization in both vertebrates and invertebrates. The action of the enzyme on microorganisms is mediated by the hy- drolysis of p-(l-4)-glycosidic linkages between N-acetylmuramic acid and A'-acetylglucosamine in bacterial cell walls. Our previous work showed that lysozyme activity in oysters, although positively correlated with low temperatures and salinities, was negatively correlated with Perkinsus marinus infections in oysters. Condi- tions of temperature below IO°C and salinity below 10%fi are not favorable to the parasite. This suggests that lysozyme may be involved in imparting a protective role in oysters against P. mari- nus infections. As a first step in elucidating the role of lysozyme in oysters' defense, we have partially purified lysozyme-like en- zyme(s) from the plasma of oysters, Crassostrea virginica. National Shellfisheries Association. Orlando, Florida Ahsinicts. January 2001 559 Lysozyme-like proteins were purified by cation-exchange chro- matography using carboxymethyl cellulose, followed by concen- tration, desalting, and separation of proteins using Centricon 3 and Centricon .^0 filtration. The proteins were eluted using a step gra- dient acetate buffer (0.01 M) with 0-0.8 M NaCl. Lysozyme ac- tivity was measured spectrophotometrically at 450 nm by the lysis of Micrucoccus lysodeikriciis suspension in phosphate buffer. Pre- liminary analysis using SDS-PAGE revealed the presence of two lyzozyme-like proteins with apparent molecular weight of 18.2 KDa and 38.8 Kda. respectively (Fig. 1). The separated proteins based on molecular weight cut-offs show the ability to lyse the bacterium Micrococcus lysiHlcikticiis. We hypothesize the 38 KDa protein to be a dimer of the 19 KDa protein. Currently, we are investigating the effects of the separated proteins on P. marinus and various bacterial species. This research was funded by the NOAA-Virginia Sea Grant-Oyster Disease Research Program. SUSCEPTIBILITY OF CULTURED PERKINSUS MARINUS AND VIBRIO PARAHAEMOLYTICUS CELLS TO HEMO- CYTES OF EASTERN OYSTER. CRASSOSTREA VIR- GINICA. Aswani K. Volety, Florida Gulf Coast University. Col- lege of Arts and Sciences, 10501 FGCU Boulevard South. Fort Myers. FL 33965, USA. Oysters possess a very effective defense system comprising cellular and humoral defenses. Various measurements of the abil- ity of oysters to generate a defense response against pathogenic and nonself particles have been developed. However, the relation- ship of these measurements to the actual defense of the organisms is unclear. Therefore, a colorimetric assay was developed to assess the ability of oyster hemocytes to kill Vibrio paraliacmolyticus. in vitro. This assay was subsequently modified to investigate the ability of oyster hemocytes to the kill cultured Pcrkinsus iiuiriiius isolated from various geographical locations. Hemocytes from Rhode Island and Florida oysters were used to assess their ability to kill P. marinus isolates from Connecticut, Delaware. Maryland. Virginia. Louisiana, and Texas. In addition, seasonal intluences on the ability of hemocytes to kill P. marinus and Vilirio parahaemolyticus was examined by monthly sampling of Florida oysters. Hemocytes from both stocks of oysters were able to reduce viability of P. marinus cells by 25-90*^ depending on the isolate. P. marinus isolate from Virginia was the most susceptible; whereas, isolates from Louisiana. New Jersey, and Connecticut were less susceptible to hemocyte killing. Variation in the results seemed to result from differences in the susceptibility of the isolates rather than the ability of hemocytes from two oyster stocks. Killing of P. marinus by oyster hemocytes was lowest during peak summer months (July-August) and increased through the winter months. This trend contrasted with low winter and high summer bactericidal activity of hemocytes against Vilvio para- liacmolyticus. The differences in the susceptibility may indicate different killing mechanisms for bacteria and P. mariiuis. The assays developed for assessing killing of bacteria and P. iiKinnus are relatively simple, inexpensive, reproducible, and en- able numerous replications owing to the low numbers of hemocytes required to carry out the assays. These techniques could be used in investigating the role of environmental factors, mecha- nisms of action, and contaminant stress on host-parasite interac- tions. EFFECT OF PERKINSOSIS ON THE ENERGETIC PHYSI- OLOGY OF THE CLAM RUDITAPES DECUSSATUS. Anto- nio Villalba and Sandra M. Casas, Centro de Investigacions Ma- rinas. ,\unta de Galicia. Aptdo. 13. Vilanova de Arousa 36620, Spain. E-mail: villalbaca^cimacoron.org Perkinsosis has been considered a threat for Galician clam in- dustry since its detection in late 1980s. A program is being devel- oped to evaluate the potential effect of the disease in clam Rudi- tapes decussatus populations of the Galician coast. One of the approaches is estimating the effects of parasitization by Pcrkinsus atlaiiticus on clam energetic physiology. In a first experiment, fifty clams (40 to 50 mm in length) were taken from a natural bed (Ri'a de Arousa, Galicia, NW Spain) with high prevalence of this para- site. Clams were allowed to acclimatize for two weeks to labora- tory conditions (I5'C and 35 ppt) and fed with cultured phy- topUmkton. Temperature value of I5°C was chosen as a midvalue of the range in Galician Ri'as ( 10 to 20C). Respiration and clear- ance rates of each clam were estimated. Then, each clam was processed for disease diagnosis. The intensity of infection by Pcr- kinsus atlanticus was estimated both through histological tech- niques and incubation of a gill lamella in fiuid thioglycollate me- dium. Both physiological rates showed a decreasing tendency as infection intensity increased. Nevertheless, differences between infection intensity categories were not statistically significant, with either diagnostic method. In a second experiment, sixty clams (42 to 47 mm long) from the same bed were allowed to acclimatize for six days to laboratory conditions. Water (15'C and 33 ppt) was continuously pumped from the Ria de Arousa into the acclimatization tanks. In this occasion, respiration, clearance and excretion rates, and absorption efficiency of each clam were estimated. The total Pcrkinsus atlan- ticus body burden was calculated in addition to the two previous diagnostic methods. No significant effect of either infection inten- sity or P. atlanticus body burden on any of the physiological parameters was found. Those results support that perkinsosis does not broadly affect clam scope for growth, at least at 15°C. Nev- ertheless, in a previous study, a statistically significant decrease (up to 25%) of clam condition index was detected only when infection intensity was heavy. Concurrent results allow rejecting that perkinsosis is respon- sible for clam mortality in the study bed, which is one of the most heavily affected by P. atlanticus in Galicia. Clam mortalities as- sociated with perkinsosis have been described in warmer European 560 Abslracls. January 2001 National Shellfisheries Association. Orlando. Florida areas. New experiments will he performed at 2()"C in September 2000, when seawater temperature in the clam bed is around that value to test if seawater temperature modulates the effect of per- kinsosis on host energetic physiology. GROWTH. GONAD DEVELOPMENT, AND MORTALITY OF GAMMA-IRRADIATED JUVENILE EASTERN OYS- TERS. Eric Wagner, Jerome La Peyre, John Buchanan, John Supan, and Terrence Tiersch, Department of Veterinary Science, Louisiana State University, 1 1 1 Dalrymple Building, Baton Rouge, LA 70803, USA. E-mail: eric.wagner@mindspring.com Gamma irradiation has been shown to cause sterility and to increase growth in a variety of organisms, possibly because of partial or complete blockage of gametogenesis and reallocation of energy to somatic tissues. Sterilization of oysters would be advan- tageous, because meat quality and yield would be improved during spawning season. Moreover, an increase in growth rate may allow oysters to reach market size before succumbing to diseases. The development of a sterilization method will also be essential for use of transgenic oysters. The objective of this study was, therefore, to determine the effect of gamma irradiation on the growth, gonadal development, and mortality of juvenile oysters. Juvenile oysters were exposed to 5 and 10 krad of cobalt-6() gamma rays using a Sheperd irradiator at a rate of 1353 rad/min. Three groups of four hundred oysters were irradiated separately at each dose. Control oysters were handled similarly to the irradiated oysters, except they were not exposed to gamma rays. Each group of oysters were placed in Taylor floats (2-in x 3-in) and grown for one year at Grand Isle, Louisiana. The number of dead oysters in each of the nine floats and the size of fifty oysters per float were determined monthly. The gonadal development of iiradiated and control oysters (fifteen per dose) was also evaluated every month from March to June. The oysters receiving lO-krad of radiation experienced signifi- cantly greater mortality than either the 5-krad or control oysters during nine months of grow-out. Significant differences in growth occurred between the control oysters (63 ± 1 mm). 5-krad oysters (54 ± 5 mm) and the 10-krad oysters (42 ± 1 mm). Significant differences in mean gametic stage among the three treatments were detected for the oysters sampled in April (P = 0.0066) and May (P - 0.0032). In each month, oysters irradiated at 10-krad had significantly lower gametic stages than did control oysters (P < 0.05). No differences among treatments were detected for oysters sampled in June (P = 0.6940). Irradiation affected sex ratio among treatments as well. Control oysters were 68% female and 32% male; 5-krad oysters were 45% female, 45% male, 2%' her- maphrodite, and 6% undifferentiated: 10-krad oysters were 20% female, 60% male, 2% hermaphrodite, and 18% undifferentiated. These data indicate no production advantage to irradiation of ju- venile oysters at 5 or 10 krad, but provides insight into possible effects on sex ratio and gonadal development. EXPERIMENTS IN DETERMINING OPTIMUM SIZE FOR PLANTING HATCHERY PRODUCED OYSTER CRASSOS- TREA VIRGINICA SEED. Richard K. Wallace, David B. Rouse, F. Scott Rikard, Jeffrey C. Howe, Blan A. Page, Donald B. Gruber, and John K. Dunne, Auburn University Marine Ex- tension & Research Center, 4170 Commanders Dr., Mobile, AL 36615, USA. E-mail: rwallace(5>acesag.aubum.edu Hatchery-produced oyster seed has been available in the Gulf of Mexico region for .several years. However, the oyster industry has been reluctant to invest in seed without more information on the relationship between cost of seed and survival. We investigated the survival of different size (<5 mm, 6-10 mm, 1 1-15 mm, and 16-20 mm) remote set (on, shell) and cultchless oyster seed planted in replicated units at two sites (low salinity: 10-30 ppt and high salinity: 19-29 ppt) in Mobile Bay. Alabama. Water quality (temperature, salinity, and oxygen) was monitored continuously at the sites, and laboratory experiments were performed to examine survival under anoxic conditions. Mean survival after thirty-two weeks for remote set seed at the low salinity site ranged from 43% for 1 1-15 mm seed to 51% for 16-20 mm seed (Fig. I). There was no significant difference in survival among the four size groups. Cultchless seed at both sites and remote set seed planted at the high salinity site did not survive the first two-week sampling period. Predation seemed to be the cause of mortality. Mean height of seed after thirty-two weeks varied from 48 mm (range = 23-63 mm) for the <5 mm group to 55 mm (range = 37-75 mm) for the 16-20 mm group. There was no significant difference in height among the three larger seed sizes, but the £5 mm group was significantly smaller than the other three seed sizes. Very low dissolved oxygen (<1 ppm) occurred at the low salinity site with varying durations (0.5-19.5 h). Preliminary laboratory studies indicated that the LT-50s for anoxic conditions (at 30°C and 15 ppt) ranged from 94 to 98 h, and there was no relationship between the LT-50s and size of seed. EVALUATION OF A MARSHLAND UPWELLING SYS- TEM FOR THE TREATMENT OF RAW DOMESTIC WASTEWATER FROM COASTAL DWELLINGS. Robert E. Watson, Jr., Kelly A. Rusch, and Tingzong Guo, Louisiana State University, Department of Civil and Environmental Engi- neering, Engineering Lab Annex Building, Baton Rouge, LA 70808, USA. E-mail: rwatson(s'unix I. sncc.lsu.edu The oyster industry depends on a healthy water environment. Waters with fecal coliform bacteria concentrations in excess of the National Shellfish Sanitation Program standard of 14 MPN ''*^/|oo mL are closed to harvesting. Improperly treated .sewage from coastal dwellings is a major source of bacterial contamination. Conventional wastewater treatment strategies are not viable op- tions for many of these dwellings because of their remote loca- tions, sporadic usage, and proximity to water tables that are typi- N;iti(in;il Shcllfisheries Association. Orlando. Kliirida Abstracts. January 2001 561 TABLE 1. Fecal conform removal under employed injection rates (Q = (1.25, (1.5, 1.0 gpm) Flowrate Influent fecal coliform concentration Effluent fecal coliform concentration Requ Ired vector feet of travel 0.25 gpm 0.5 gpm 1 .0 gpm 326.700 '='^'/,„„„„ 930.400 •-^/,„„„,L 559.200 '^^7,„„„,L 2.14 "V 3.08 "^/.ou^L X s 11 10.3 tt 1 1 .7 ft Exceedance probability (>14'^'7„,„„„l IK'f 5% 23% cally at or above the ground surface. Thus increased efforts are being focused upon de\eloping innovative methods to treat do- mestic wastewater effectively in poorly sewered coastal areas. The primary objective of this research was to evaluate the efficacy of a marshland upwelling system in removing fecal pathogens from raw domestic wastewater. Wastewater was injected intermittently down a 15-foot deep well under three tlowrates. Density gradients created between the wastewater and relatively dense native saline groundwater con- fined the injectant within a limited area and forced it vertically toward the ground surface, allowing the natural soil matrix to lunction as an uptlow sand filter removing fecal pathogens. Con- centrations in the 5-ft monitoring wells were used as a conserva- tive estimate of effluent counts. Fecal coliform removal was as- sumed to follow first-order decay, allowing the required travel distance for influent concentrations to fall below the 14 MPN standard to be calculated. Best-fit probability density functions were determined for effluent fecal coliform concentrations under each flowrate and used to predict the probability of effluent counts exceeding the 14 MPN standard (see Table 1 1. Two types of failure are believed to exist: (1) isolated fail- ure created by accelerated intermittent injection flowrates, that may cause localized pressures to build and create transient epi- sodes of channelization: and (2) catastrophic failure induced by excessive hydraulic loading rates that cause global increases in fecal concentrations because of reduced hydraulic retention times and the corresponding widespread declines in pathogen removal. Although treatment efficacy did vary with injection flowrate. and isolated instances of abnormally high fecal counts were observed under the 1 .0 gpm flowrate, the MUS never experienced a cata- strophic failure during the thirteen-month evaluation period. MHACS: MARINE HABITAT ACOUSTIC CHARACTER- IZATION SYSTEM. A PROGRAM FOR THE ACQUISI- TION AND INTERPRETATION OF DIGITAL ACOUSTICS TO CHARACTERIZE OYSTER HABITAT. Charles A. Wil- son and Harry H. Roberts, Coastal Fisheries Institute, Depart- ment of Oceanography and Coastal Sciences. CCEER. Louisiana State University. Baton Riiuge. LA 70803, USA. Coastal Louisiana, like many deltaic land masses, faces con- tinued landscape alteration from natural processes and anthropo- genic impacts that affect estuarine habitat. Steps are being taken at both state and federal levels to slow/mitigate these changes. Most promising of these strategies is river diversions that introduce freshwater and sediment to river-flanking environments (lakes, bays, and associated marshlands). Two such diversion projects (Caenarvon and Davis Pond), planned by Louisiana and the U.S. Army Corps of Engineers, are designed to nourish marshes with water and sediment as well as to help establish ideal salinities over historic oyster grounds. Critical to the success of these programs is a rapid and accurate means to qualify and quantify changes in oyster habitat. Digital high-resolution acoustic instrumentation linked to state- of-the-art data acquisition, and processing software was used to build a baseline of information for evaluating future changes in shallow and shelf water bottoms with special emphasis on oyster habitats. Application of digital side-scan sonar ( 100 and 500 kHz), a broad-spectrum sub-bottom profiler (4-24 kHz) for rapidly ac- quiring water column, surficial and shallow subsurface data has now been accomplished. In our most recent study, geo-referenced side-scan sonar mosaics of a Louisiana estuary were incorporated into a CIS data base. These datasets "calibrated" with surface sampling, coring, and other "ground truthing" have established that numerically indexed acoustic reflectance intensities correlate closely with surface shell and oyster reef density. With image- processing techniques to analyze mosaic reflectance patterns, we estimated the percentage and total acreage of several bottom types. Results were calibrated with field collected ground truth measure- ments. PRODUCTION AND EVALUATION OF MEIOSIS I AND II TRIPLOIDS IN THE HARD CLAM. MERCENARIA MER- CENARIA. Huiping Yang, Jian Wang, and Ximing Guo, Haskin Shellfish Research Laboratory. Institute of Marine Coastal Sci- ence. Rutgers University. 6959 Miller Avenue, Port Nortis, NJ 08349, USA. E-mail: hyang(a>hsrl.rutgers.edu Triploid mollusks grow significantly faster than normal dip- loids in most species studied so far. Results from clams are mixed. Triploid dwarf-surf clams grow 72% faster than normal diploids; whereas, triploid soft-shell clams had the same or slower growth rate than diploids. In the hard clam. Mercenaria mercenaria, pre- vious studies show that triploids grow faster than diploids at four years of age. but not during the first two years. Triploids may perform differently under different environments, and triploid hard 562 Abstracts, January 2001 National Shellfisheries Association, Orlando, Florida clams have not been evaluated in New Jersey. Different types of triploids, such as meiosis I triploids, meiosis II triploids, and mated triploids, also differ in growth performance. The objective of this study is to produce two types of triploids by inhibiting polar body 1 (PBl) and polar body 2 (PB2). respectively, and evaluate their growth performance in New Jersey. Three experimental groups were produced: a diploid control, a PBl triploid group, and a PB2 triploid group. Cytochalasin B (CB, 1.0 mg/L) was used as the induction agent. Four replicates were produced using different sets of brood stock. Ploidy was deter- mined by counting chromosome number at two- to four-ceil stage and by flow cytometry (FCM) at D-stage and beyond. High levels of triploids were produced with both methods, ranging from 62 to 83% in PBl inhibition groups and from 82 to 100% in PB2 inhi- bition groups. Tetraploids and aneuploids were also produced in PBl groups, but became undetectable at late larval stages by FCM. Survival of larvae in PBl groups was lower than that in PB2 groups. Experimental clams will be deployed at two sites in New- Jersey and sampled at regular intervals for ploidy determination and growth measurements. Data from the first two sampling dates, three and six months of age, are presented. A STUDY OF THE EFFECT OF DERMO DISEASE PER- KINSUS MARINUS ON EASTERN OYSTERS CRASSOS- TREA VIRGINICA IN THE PATUXENT RI\ ER. MARY- LAND WITH THE HELP OF COMMUNITY VOLUN- TEERS. George R. Abbe, Brian W. Albright. Erin S. Wood- row, and Shannon B. Campbell, Academy of Natural Sciences, Estuarine Research Center. 10545 Mackall Road, St. Leonard. Maryland 20685, USA. E-mail: abbe@acnatsci.org In studies of natural oyster populations affected by dermo dis- ease in the Patuxent River Maryland, we have examined live oys- ters and boxes from dredged samples over time to gain some understanding of mortalities associated with disease. Not knowing the accuracy of the box- to-live ratio, we suspended three trays of one hundred oysters each from piers along a 33-km salinity gra- dient in the Patuxent in January 1997 and measured growth and survival monthly during the year. After a year, it was evident that oyster mortalities were highly correlated with salinity and dermo infection. We thought that by increasing the number of sites along this salinity gradient we could gain a better understanding of oyster- salinity-disease interactions but lacked the resources and time to do so. With the help of community volunteers who donated piers and/or time and effort, we were able to conduct a similar study during 1998 at twenty sites along the Patuxent River. Each tray was stocked with one hundred oysters from the nearest natural oyster bed. By January 1999, it was evident that oyster mortalities decreased in an upriver direction with decreasing salinity as did the prevalence and intensity of dermo as measured by an infection index. Dermo levels in tray-held oysters were highly correlated with disease levels on nearby natural bars, and mortalities in trays were also highly correlated with estimated mortalities from dredged samples based on box-to-live oyster ratios. One of the most valuable results of this study was the increased awareness of the participants of environmental issues in their own communities. They learned to collect data year round at the same time each month and how their efforts fit into the larger riverwide picture. Relationship of dermo infection index to oyster mortalities in trays as compared to those salinity in the Patuxent in September 1998. estimated from natural beds during 1998. IMPROVING ACCURACY IN THE DETERMINATION OF MEAT CONDITION INDEX FOR THE EASTERN OYSTER CRASSOSTREA VIRGINICA. Brian W. Albright and George R. Abbe, Academy of Natural Sciences. Estuarine Research Cen- ter. 10545 Mackall Road St., Leonard, MD 20685, USA. E-mail: Albright@acnatsci.org The meat condition index (MCI) of a bivalve is a numerical representation of the quality of its soft tissue. Based on the per- centage of the internal shell volume occupied by an oyster's soft body tissue, enumeration of a quantitative index is possible. Older methods sought to measure shell cavity volume volumetrically; however, this is not practical. In 1982, Lawrence and Scott devel- oped a method to determine MCI gravimetrically. where shell cavity capacity is determined by the difference between whole oyster weight and empty shell weight after drying for 24 h. After using this method for more than fifteen years, we ques- tioned its accuracy. Because any water contained in the shell itself (not between the valves) was included in the whole oyster weight, it seemed this should be included in the weight of the empty shells as well. Drying shells for 24 h could make the shell cavity appear larger than it really was. resulting in a reduced meat condition. To determine the significance of weighing shells at 0 h verses 24 h of drying, meat condition indices were first determined based on true cavity volume using a nioldable material of known density and then compared to the two methods. Weighing shells immedi- ately after processing was determined to more accurately estimate cavity capacity whenever shells lost more than 3% of their weight because of drying. Because oyster shells are highly variable with respect to po- rosity, we set out to determine the average weight loss of the \ alves after 24-h drying time. Over a 3-y period, monthly collec- tions from many oyster bars in the Patuxent River resulted in the examination of 1.749 oysters, of which 74.3% lost more than 3% shell weight. Several other sites in the Chesapeake Bay were also examined yielding similar results. Weighing shells at 0 h increased accuracy for most of the oysters we examined and also saved time and space, because shells do not need to be held for an additional 24 h for weighing. To date, this method has neither been tried for other species nor over the entire range of C. virginica. Differences in shell morphology and fouling communities may influence shell porosity, favoring one technique over the other. Journal of Shellfish Research. Vol. 20. No. 1. 56.i. 2001. ACKNOWLEDGMENT OF REVIEWERS THANK YOU In addition to the editorial board, many indi\iduais have contributed their time and efforts to the review process. Without the continued efforts of such individuals, the Journal af Shellfish Research could not maintain its standards of publication. It is a pleasure to thank the following individuals who have reviewed manuscripts over the past three years: George Abbe Laura Adamkewicz David Aiken Richard Alexander Standish Allen William Ambrose Jay Andrews Klaus Anger Alan Ansell Ron Appeldoorn Robert Armstrong William .Arnold Patrick Baker Shirley Baker Bruce Barber Stephen Bates Andrew Bauder Robert Bayer Brian Bayne Brian Beal Andrew Beaumont Ricardo Beiras Peter Beninger Sandra Blake Norman Blake Juan Blanco Anne Boetlcher Neil Bourne Susan Bower Jeanne Boylan Andrew Brand Monica Bricelj Diane Brousseau Craig Browdy Lindsay Broun Gavin Burnell Eugene Burreson David Bushek Carlos Cacere-Martinez Alan Campbell Dan Campbell James Carlton Melbourne Carriker Michael Castagna Jim Childress Martina Chintala John Christiansen Fu-Lin Chu Loren Coen Tracey Collier Peter Cook Steve Coon Cyr Cotourier Peter Coutteau Simon Cragg Derek Cropp Sarah Culloty Mike Dadswell William Dall Richard Dame Chris Da\is Joth Da\is Megan Daxis-Hodgkins Robert Day Lewis Deaton Allison DeFerry Thomas Deitz Margaret Deksheniaks Waher Diehl Robert Dillon Mike Dredge Robert Elner Ralph Elston Arnold Eversole Richard Fassler Steve Fegley William Fisher Gef Flimlin David Folt/ Susan Ford Richard French Carolyn Friedman Patrick Gaffney Louis Gainey David Garton Rodman Gelchel Paul Geoghegun Sbrenna Giovanni Elizabeth Gosling Ron Goldberg Jon Grant Charles Griffiths P. Grosjean Ximing Guo Nancy Hadley Sherwood Hall Karolyn Hanson Juliana Harding Richard Harlnoll Anthony Hawkins Barbara Hayden Michael Heasman Dennis Hedgecock John Hinimelman Eileen Hofmann Simon Hooker Steve Hopkins Rita Horner Hunt Howell Roger Hughes John W. Hurst. Jr. Juan Illanes Lindsay JoU Malcolm Jones Stephen Jones Steve Jordan Richard Kamey Jeff Kassner Bruce Keafer Kevin Kelly Victor Kennedy Ellen Kenchington Jake Keogh James Kirkley John Kraeuter Maureen Krause Dan Kreeger Chris Langdon Greg Langlois Richard Langton Patrick Lassus Peter Lawton Jeff Levinton Tim Littlewood Junde Lin Bruce MacDonald Sandra MacFarlane Lincoln MacKenzie Kevan Main Steve Malinowski Roger Mann Joan Manuel Dan Marelli Islay Marsden Jennifer Martin Gloria Martinez James Mason George Matthiessen John McDonald David McKee Sharon McGladdery Nature McGinn Paul McShane Steve Morton Enrique Navarro Richard Neves Michael New Carter Newell Roger Newell Francis OBiern Michael Osterling Jay Parsons A.J. Paul Kennedy Paynter Jan Pechenik Frank Perkins Alain Plusquelec Grant Pogson Eric Powell Warren Pryor Michael Quilliam T.A. Rawlings Paul Rawson Robert Rheault Michael Rice Chris Richardson Hans Riisgard Gregg Rivara Steve Robinson Terrance Rowell John Scarpa Timothy Scott William Shaw Scoresby Shepard Thomas Shirley Scott Siddall Neil Sims D. Smith Roxanna Smolowitz Thomas Soniat Patrick Sorgeloos Paul Southgate Melissa Southworth Brian Spencer Martin Sprung Alan Stoner Kenneth Stuck Hal Sunimerson Harry Taylor Steven Tettelbach Ray Thompson Gudrun Thorasinsdotter Terry Tiersch Michelle Tomlinson Jorge Toro John Tremblay George Trevalyan Susan Utting Susan Waddy Richard Wahle Thomas Waller J, David Walker Randall Walker J. Evan Ward Les Watling Elizabeth Wenner David Whitaker Robert Whitlach J.N.C. Whyte John Widdows Gary Wikfors James Winstead Chris Woods H.A. Woods 56?. THE NATIONAL SHELLFISHERIES ASSOCIATION The National Shellfisheries Association (NSA) is an international organization of scientists, manage- ment officials and members of industry that is deeply concerned and dedicated to the formulation of ideas and promotion of knowledge pertinent to the biology, ecology, production, economics and man- agement of shellfish resources. The Association has a membership of more than 1000 from all parts of the USA. Canada and 18 other nations; the Association strongly encourages graduate students" mem- bership and participation. WHAT DOES IT DO? — Sponsors an annual scientific conference. — Publishes the peer-reviewed Journal of Shellfish Research. — Produces a Quarterly Newsletter. — Interacts with other associations and industry. WHAT CAN IT DO FOR YOU? — You will meet kindred scientists, managers and industry officials at annual meetings. — You will get peer review through presentation of papers at the annual meeting. — If you are young, you will benefit from the experience of your elders. — If you are an elder, you will be rejuvenated by the fresh ideas of youth. — If you are a student, you will make useful contacts for your job search. — If you are a potential employer, you will meet promising young people. — You will receive a scientific journal containing important research articles. — You will receive a Quarterly Newsletter providing information on the Association and its activities, a book review section, information on other societies and their meetings, a job placement section, etc. HOW TO JOIN — Fill out and mail a copy of the application blank below. The dues are 65 US $ per year ($35 for students) and that includes the Journal and the Newsletter! Name: Mailing address: NATIONAL SHELLFISHERIES ASSOCIATION— APPLICATION FOR MEMBERSHIP (NEW MEMBERS ONLY) For the calendar year: Date: Institutional affiliation, if any: Sheilfishery interests: Reguku' or student membership: Student members only — advisor's signature REQUIRED: Make checks (MUST be drawn on a US bank), international postal money orders or VISA for $65 ($35 for students with advisor's signature) payable to the National Shellfisheries Association and send to Nancy Lewis, Bookkeeper. PO Box 350. V.I. M.S. Eastern Shore Lab. Wachapreague. VA 23480. USA. INFORMATION FOR CONTRIBUTORS TO THE JOURNAL OF SHELLFISH RESEARCH Original articles dealing with all aspects of shellfish re- search will be considered for publication. Manuscripts will be judged by the editors or other competent reviewers, or both, on the basis of originality, content, merit, clarity of presentation, and interpretations. Each article should be carefully prepared in the style followed in prior issues of the Journal of Shellfish Research before submission to the Editor. Papers published or to be published in other journals are not acceptable. Title, Short Title, Key Words, Abstract: The title of the paper should be kept as short as possible. Please include a '"short running title"" of not more than 48 characters including spaces, and key words. Each manuscript must be accompanied by a concise, informative abstract, giving the main results of the research reported. The abstract will be published at the beginning of the article. No separate suminary should be in- cluded. Text: Manuscripts must be typed double-spaced throughout on one side of the paper, leaving ample margins, with the pages numbered consecutively. Scientific names of species should be underlined or in italics and. when first mentioned in the text, should be followed by the authority. Conmion and scientific names of organisms should be in accordance with American Fisheries Society Special Publications 16 and 17: Cdinmon and Scientific Names of Acjiiatic /iivertehrates from the United States and Canada: Molluslis and CSNAIUSC: Decapod Crus- taceans, or relevant publications for other geographic regions. Abbreviations, Style, Numbers: Authors should follow the style recommended by the si.xth edition ( 1994) of the Council of Biology Editors [CBE] Style Manual, distributed by the American Institute of Biological Sciences. All linear measure- ments, weights, and volumes should be given in metric units. Tables: Tables, numbered in Arabic, should be on separate pages with a concise title at the top. Illustrations: Line drawings should be in black ink or laser print and planned so that important details will be clear after reduction to page size or less. No drawing should be so large that it must be reduced to less than one third of its original size. Photographs and line drawings should be prepared so they can be reduced to a size no greater than 17..^ cm x 22.7 cm. and should be planned either to occupy the full width of 17.3 cm or the width of one column. 8.4 cm. Photographs should be glossy with good contrast and should be prepared so they can be reproduced without reduction. Originals of graphic materials (i.e.. line drawings) are preferred and will be returned to the author. Each illustration should have the author's name, short paper title, and figure number on the back. Figure legends should be typed on separate sheets and numbered in Arabic. Digital Figures: Authors may provide digital figures (they are not required); they must be accompanied by hardcopy fig- ures of equal quality, which the printer will use for comparison and backup. If digital figures are supplied, please note the following instructions: • Each piece of art should be saved as its own file. • Files must be one of the following fomiaLs: TIF. EPS. or JPG. • Each file should be named according to its figure number and format (e.g., ■■fig2b.tif"). • Figures must not be embedded in a word-processor or spreadsheet document: the printer cannot use images stored in Word. WordPerfect, Excel, Powerpoint, etc. • Resolution: line shots: 1000 dpi: lialftone.s/gra\scales: 300 dpi if no lettering, 500 dpi if figure contains lettering. • Color figures: save the files as CMYK-encoded TIF images (prefened) or CMYK-encoded EPS or JPG images. Color figures have the same resolution requirements a B/W, above. Color illustrations will not be accepted unless the author agrees to cover the cost of associated reproduction and printing. Literature Cited: References should be listed alphabeti- cally at the end of the article. Abbreviations in this section should be those recommended in the American Standard for Periodical Title Abbreviations, available through the American National Standard Institute, 1430 Broadway. New York. NY 10018. For appropriate citation format, see examples below: Journal: Watts. R. J.. M. S. Johnson & R. Black. 1990. Effects of re- cruitment on genetic patchiness in the urchin Echinometra mathaei in Western Australia. Mar. Biol. 105:145-151. Book: Claudi. R. & G. L. Mackie. 1994. Practical manual for Zebra Mussel monitoring and control. Boca Raton. FL: CRC Press. 227 pp. Chapter in Edited Bool<: Davio. S. R., J. F. Hewetson & J. E. Beheler. 1985. Progress toward the development of monoclonal antibodies to saxitoxin: antigen preparation and antibody detection. In: D. M. Ander- son. A. W. White & D. G. Baden, editors. Toxic dinoflagel- lates. Amsterdam: Elsevier, pp. 343-348. Page Charges: Authors or their institutions will be charged $100.00 per printed page. All page charges are subject to change without notice. A handling fee of $50 will be charged for all manuscripts accepted for publication. Proofs: Page proofs are sent to the conesponding author and must be corrected and returned within seven days. Alter- ations other than corrections of printer's enors may be charged to the author(s). Reprints: Reprints of published papers are available at cost to the authors. Information regarding ordering reprints will be available from The Sheridan Press at the time of printing. Cover Photographs: Appropriate photographs may be sub- mitted for consideration for use on the cover of the Journal of Shellfish Research. Black and white photographs and color illustrations will be considered. Corresponding: An original and two copies of each manu- script submitted for publication consideration should be sent to the Editor. Dr. Sandra E. Shumway, Natural Science Division, Southampton College. LIU. 239 Montauk Highway. Southampton, NY 11968. Ph. 631-287-8407. FAX 631-287- 8419. E-mail: sshumway@southampton.liu.edu Membership information may be obtained from the Editor or the Treasurer using the form in the Journal. Institutional subscrib- ers should send requests to: Journal of Shellfish Research, P.O. Box 465. Hanover. PA 17331. Contents continued from following page Paul A. Hohenlohe and Elizabeth G. Boulding A molecular assay identities morphological characters useful for distinguishing the sibling species Liuorina sciilulaui and L. plena 453 Robert F. McMahon Acute thermal tolerance in intertidal gastropods relative to latitude, superfamiliy. zonation and habitat with special emphasis on the liltorinoidea 459 P. J. Mill. A. P. Clarke, D. C. Smith. J. Grahame. and C. S. Wilding Lagoonal littorinids: shell shape and speciation 469 Dwayne Miuton and Deborah J. Gochfeld Is life on a tropical shore really so hard.': the role of abiotic factors in structuring a supralittoral molluscan assemblage 477 Delmont C. Smith Effects of temperature and desiccatiini on tissue uric acid dynamics in Litunina saxmilis (Olivi) 485 R. F. Uglow and Gray A. Williams The effects of emersion on ammonia efflux of three Hong Kong Nodilittorina species 489 G. F. Warner Trans-zonal mo\ ements m winkles. Lillcirimi lituirca ( L. ): reasons and consequences 495 Craig S. Wilding. John Grahame. and Peter J. Mill Correlation of morphological diversity with molecular marker diversity in the rough periwinkle Liuorina saxatilis (Olivi) 501 Abstracts of technical papers presented at the 21st Annual Aquaculture Seminar. Milford. Connecticut, February 26-28. 2001 509 Ab.stracts of technical papers presented at the 93rd Annual Meeting of the National Shelllisheries Association. Orlando. Florida. January 22-25, 2001 53 1 Acknowledgment of reviewers 563 COVER PHOTO: Cencluilis muncatus among leaves of the coastal shrub Sesuviuni puitiilacaMniin: photographed at Swan Island in the Caribbean. March. 1973. This snail is one of several species of tropical Caribbean littorinids and occupies the highest position on rocky intertidal shores. (Joseph C. Britton) The Journal of Shellfish Research is indexed in the following: Science Citation Index*, Sci Search*, Research Alert®, Current Contents®/Agriculture, Biology and Environmental Sciences, Biological Abstracts, Chemical Abstracts, Nutrition Abstracts, Current Advances in Ecological Sciences. Deep Sea Research and Oceanographic Literature Review. Environmental Periodicals Bibliography, Aquatic Sciences and Fisheries Abstracts, and Oceanic Abstracts. Contents continued from inside back cover Michael D. Kaplowitz Uncovering economic benefits of ciiivita {Melongena metongemi Linnaeus, 1758 and Melongena corona bispinosa Philippi, 1844) 317 Andong Qiii, Anjing Shi, and Akira Komaru Yellow and brown shell color niorphs of Corbicida fliwunea (Bi\alvia; Corbiculidae) from Sichuan Province. China. are triploids and tetraploids 323 S. J. Nichols, J. Allen, G. Walker. M. Yokoyania, and D. Garling Lack of surface-associated microorganisms in a mixed species community of freshwater unionidae 329 Brian F. Beal and Samuel R. Chapman Methods for mass rearing stages I-IV larvae of the American lobster. Homarus americcmus H. Milne Edwards. 1837. in static systems 337 Guiomar Rotllant, Mireille Charmantier-Daures, Guy Charmantier, Klaus Anger, and Francisco Sarda Effects of diet on Ncphrops nonrgiciis (L.) larval and postlarval development, growth, and elemental composition 347 Eugenia Bogazzi, Oscar Iribarne, Raul Guerrero, and Eduardo Spivak Wind pattern may explain the southern limit of distribution of a soulhw estern Atlantic fiddler crab 353 Junda Lin and Dong Zhang Effect of broodstock diet on reproductive performance of the peppermint shrimp. Lysiinna witrdemawu 361 Ryan Gandy, Tzachi M. Samocha, Edward R. Jones, and David A. McKee The Texas live bait shrimp market 365 A. J. Paul and J. M. Paul The reproductive cycle of golden king crab Lithodes aequispiniis ( Anomura: Lithodidae) 369 A. J. Paul and J. M. Paul Intermolt durations of captive juvenile and adolescent male tanner crabs Chioijoccelt's Ixiinli 373 Gillian McLaughlin and Maeve S. Kelly Effect of artificial diets containing carotenoid-rich microalgae on gonad growth and color in the sea urchin Psainmccliinus iniliciris (Gmelin ) 377 Technical papers presented at the Sixth International Littorinid Symposium 383 Troy C. Addy and Ladd E. Johnson Liuorine foraging behavior and population structure on a wave-exposed shore: non-linear responses across a physical gradient 385 Robert Black and Michael S. Johnson Contrasting life histories and demographies of eight species of littorines at Ningaloo Reef Western Australia 393 Elizabeth G. Boulding, Deborah Pokes, and Stephanie Kamel Predation by the pile perch. Rhacochilus viicca. on aggregations of the gastropod Littorina sitkana 403 Carlos Brito. Paulinha Lourenfo, Roberto Medeiros, Jose F. Rebelo, Hans de Wolf, Kurt Jordaens, and Thierry Backeljau Radular myoglobin as a molecular marker in littorinid systematics (caenogastropoda) 411 Monica Carballo, Carlos Garcia, and Emilio Roldn-Alvarez Heritability of shell traits in wild Litrniiini saxatilis populations: results across a hybrid zone 415 J. T. Christensen, P.-G. Saurian, P. Richard, and P. D. Jensen Diet in mangrove snails: preliminary data on gut contents and stable isotope analysis 423 Hans de Wolf, Ronny Blast, and Thierry Backeljau Shell size \ ariation in Littunna littorcu in the western Scheldt estuary 427 Alluwee Dobson-Moore and Joseph C. Britton Carbonate processing by intertidal gastropoda on Jamaican limestone shores 43 1 Antonio M. de Frias Martins Ellobiidae — lost between land and sea 441 Deborah J. Gochfeld and Dwayne T. Minton When to move and where to go: movement behavior of the tropical littorinid Ccnchritis iimricatiis (Linnaeus. 1758) .. 447 Contents continued on previous page Contents continued from hack cover William Cameron Walton and William Charles Walton Pniblenis. predators, and perception: management of quahog (hardclam), Mercenaria mercenria. stock enhancement programs in southern New England 127 Lara K. Gulmann, Lauren S. Mullineaux, and Heather L. Hunt Effects of caging on retention of posllar\ al soft-slielled clams (A/ra arenaria) 135 Stuart Alan Goong and Kenneth A. Chew Growth of butter clams. Saxidomus giganleus Deshayes. on selected beaches in the .Slate of Washington 143 Diego C. Luzzatto and Pablo E. Penchaszadeh Regeneration of the inhalant siphon of Dnini.x haiilfymuis (Philippi. 1847) (Bivalvia. Donacidae) from Argentina 149 Mehmet Cengiz Deval The shell growth and the biometry of the striped venus Chamelca gallina (L) in the Marmara Sea. Turkey 155 Michael L. Zettler, Regine Bonsch, and Fritz Gosselck Distribution, abundance, and some population characteristics of the ocean quahog. Arclicu iskindica (Linnaeus. 1767), in the Mecklenburg Bight (Baltic Sea) 161 Maria Sparsis, Junda Lin, and Randolph W. Hagood Growth, survivorship, and nutrient uptake of giant clams {Tiidacna) in aquaculture effluent 171 Mi Seon Park, Chang-Keun Kang, and Pil- Yang Lee Reproductive cycle and biochemical composition of the ark shell Scapharca bruiighumii (Schrenck) in a southern coastal bay of Korea 1 77 Katherine A. McGraw, Michael Castagna, and Loveday L. Conquest A study of the arkshell clams. Noelia ponderosa (Say 1822) and Aitadara ovali.s (Bruguiere 1789). in the oceanside lagoons and tidal creeks of Virginia 185 G. F. Smith, K. N. Greenhawk, D. G. Bruce, E. B. Roach, and S. J. Jordan A digital presentation of the Maryland oyster habitat and associated bottom types in the Chesapeake Bay (1974-1983) 197 Nancy A. Stokes and Eugene M. Burreson Differential diagnosis of mixed Haplosparidiiiiii cnslaU' and Haplosporidiuin nelsoni infections in the eastern oyster. Cnissostrea Virginica. using DNA probes 207 John E. Supan and Charles A. Wilson Analyses of gonadal cycling by oyster broodstock. Cnissostrea virginica (Gmelin). in Louisiana 215 Gustavo W. Calvo, Mark W. Luckenbach, Standish K. Allen, Jr., and Eugene M. Burreson A comparative t~ield study of Crassoslrca ariakcnsis (Fujita 1913) and Crassostrea virginica (Gmelin 1791) in relation to salinity in Virginia 221 Lisa M. Ragone Calvo, Richard L. Wetzel, and Eugene M. Burreson Development and verification of a model for the population dynamics of the protistan parasite, Perkinsiis marinus. within its host, the eastern oyster, Crassostrea virginica. in Chesapeake Bay 23 1 Eleanor A. Bochenek, John M. Klinck, Eric N. Powell, and Eileen E. Hofmann A biochemically based model of the growth and development of Crassostrea gigas larvae 243 Angelika Prael, Simon M. Cragg, and Suzanne M. Henderson Behavioral responses of veliger larvae of Crassostrea gigas to leachate from wood treated with copper-chrome-arsenic (CCA): a potential bioassay of sublethal environmental effects of contaminants 267 Boo-Keun Khim Stable isotope profiles of Serripes groenlandicus shells. II. Occurrence in Alaskan coastal water in south St. Lawrence Island, northern Bering Sea 275 G. Bigatti, P. E. Penchaszadeh and G. Mercuri Aspects of the gonadal cycle in the antarctic bivalve Latermda elliplica 283 Pedro J. Baron First description and survey of the egg masses of Ltiligo gahi (d'Orbigny. 1835) and Loligo sanpaulensis ( Brakoniecki. 1 984) from coastal waters of Patagonia 289 Alfredo Enriquez, Maria Teresa Viana, Carlos Vdsquez, and Armando Shimada Digestion of cellulose by stomach homogenates of green abalone (Haliotis fulgens) 297 Peter G. Beninger, Rozenn Cannuel, Jean-Louis Blin, Sebastien Pien, and Olivier Richard Reproductive characteristics of the archaeogastropod Megatluira crenidata 301 Luciano Rodriguez, Giovanni Daneri, Cristidn Torres, Mati'as Leon, and Leonardo Bravo Modeling the growth of the Chilean loco, Concholepas concholepas (Bruguiere, 1789) using a modified Gompeilz-type function 309 Contents continued on previous page JOURNAL OF SHELLFISH RESEARCH Vol. 20, No. 1 JUNE 2001 CONTENTS Gary Rodrick In Memoriam: Thomas Clement Cheng 1 Debra A. Ingrao, Paula M. Mikkehen, and David W. Hicks Another introduced marine nioliusk In the Gulf of Mexico: the Indo-Pacific green mussel. Pema viridi.s. in Tampa Bay, Florida 13 Amy J. Benson, Dan C. Marelli, Marc E. Frischer, Jean M. Danforth, and James D. Williams Establishment of the green mussel, Penia viridis (Linnaeus I751S) (Mollusca: Mytllidae) on the west coast of Florida.. 21 Paul D. Rawson, Susan Hayhurst, and Brook Vanscoyoc Species composition of blue mussel populations in the northeastern Gulf of Maine 31 C. Rodriguez-Jaramillo, A. N. Maeda- Martinez. M. E. Valdez, T. Reynoso-Granados, P. Monsalvo-Spencer, D. Prado-Ancona, F. Cardoza-Velasco, M. Robles-Mungaray, and M. T. Sicard The effect of temperature on the reproductive maturity of the pensheil Atriim maiira (Sowerby, 1835) (Bivalvia: Pinnidae) 39 /. Leyva-Valencia, A. N. Maeda-Martinez, M. T. Sicard, L. Roldan, and M. Robles-Mungaray Halotolerance. upper thermotolerance, and optimum temperature for growth of the pensheil Atrina maura (Sowerby, 1835) (Bivalvia: Pinnidae) 49 Jose Luis Cordova, Adolfo Jamett, Juan Aguayo, Maria Teresa Faure, Orialis Villarroel, and Leonidas Cardenas An /;; vitro assay to detect Paralytic Shellfish Poison 55 Jose Luis Cordova, Juana Bustamante, and Leonidas Cardenas Specific inhibition of endogenous shellfish protein phosphatase that could be used as a direct reporter of diarrhetic shellfish poison 63 Juan Carlos Uribe, Carlos Garcia, Mariella Rivas, and Nestor Lagos First report of diarrhetic shellfish toxins in Magellanic Fjords, southern Chile 69 D. A. Campbell, M. S. Kelly, M. Busman, C. J. Bolch. E. Wiggins, P. D. R. Moeller, S. L. Morton. P. Hess, and S. E. Shumway Amnesic shellfish poisoning in the king scallop, Pecten maxim us, from the west coast of Scotland 75 Sergio A. Gonzalez, Wolfgang B. Stotz, and Marcelo Aguilar Stranding of scallops related to epiphytic seaweeds on the coast of northern Chile 85 Paul A. X. Bologna, Ami E. Wilbur, and Kenneth W. Able Reproduction, population structure, and recruitment limitations in a bay scallop (Argopecten irradians Lamarck) population from New Jersey, USA 89 Ana Farias and Iker Uriarte Effect of microalgae protein on the gonad development and physiological parameters for the scallop Argopecten purpuratus (Lamarck, 1819) 97 Juan C. Perez-Urbiola and Sergio F. Martinez-Diaz Stephanostomiim sp. (Trematoda: Acanthocolpidae), the cause of "pimientilla" disease in catarina scallop Argopecten ventricosus (circiilarisj (Sowerby 11, 1 842) in Baja California Sur, Mexico 107 N. de Vido de Mattio, M. E. Paredi, and M. Crupkin Influence of the gonadal cycle and food availability on postmortem changes in glycogen, adenosine triphosphate. hypoxanthine. and the 260/250 absorbance ratio in adductor muscles from acaWop Aequipeclen teliiielcluis (d'Orbigny. 1 846) Ill Philip Heath and Martin Pyke King scallop {Pecten maximus) depuration trials 117 Bryce D. Beukers-Stewart, Stuart R. Jenkins, and Andy R. Brand The efficiency and selectivity of spring-toothed scallop dredges: a comparison of direct and indirect methods of assessment 121 Contents continued on inside back cover MBL/WHOI UBRARY H lAAN 7 JOURNAL OF SHELLFISH RESEARCH VOLUME 20, NUMBER 2 DECEMBER 2001 The Journal of Shellfish Research (formerly Proceedings of the National Shellfisheries Association) is the official publication of the National Shellfisheries Association Editor Dr. Sandra E. Shumway Department of Marine Sciences University of Connecticut Groton, CT 06340 Dr. Standish K. Allen. Jr. (2002) School of Marine Science Virginia Institute of Marine Science Gloucester Point, VA 23062-11346 Dr. Peter Beninger (2001 ) Laboratoire de Biologic Marine Faculte des Sciences Universite de Nantes BP 92208 44322 Nantes Cedex 3 France Dr. Andrew Boghen (2001) Department of Biology University of Moncton Moncton. New Brunswick Canada ElA 3E9 Dr. Neil Bourne (2001) Fisheries and Oceans Pacific Biological Station Nanaimo, British Columbia Canada V9R 5K6 Dr. Andrew Brand (2001) University of Liveipool Marine Biological Station Port Erin, Isle of Man Dr. Eugene Burreson (2001) Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Dr. Peter Cook (2002) Department of Zoology University of Cape Town Rondebosch 7700 Cape Town, South Africa EDITORIAL BOARD Dr. Simon Cragg (2002) Institute of Marine Sciences University of Portsmouth Ferry Road Portsmouth P04 9LY United Kingdom Dr. Leroy Creswell (2001) Harbor Branch Oceanographic Institute US Highway 1 North Fort Pierce, Florida 34946 Dr. Lou D'Abramo (2002) Mississippi State University Dept of Wildlife and Fisheries Box 9690 Mississippi State, Mississippi 39762 Dr. Ralph Elston (2001) Battelle Northwest Marine Sciences Laboratory 439 West Sequim Bay Road Sequim, Washington 98382 Dr. Susan Ford (2002) Rutgers University Haskin Laboratory for Shellfish Research P.O. Box 687 Port Norris, New Jersey 08349 Dr. Raymond Grizzle (2001) Randall Environmental Studies Center Taylor University Upland, Indiana 46989 Dr. Mark Luckenbach (2001) Virginia Institute of Marine Science Wachapreague, Virginia 23480 Dr. Bruce MacDonald (2002) Department of Biology University of New Biunswick P.O. Box 5050 Saint John, New Brunswick Canada E2L 4L5 Dr. Roger Mann (2002) Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Dr. Islay D. Marsden (2002) Department of Zoology Canterbury University Christchurch, New Zealand Dr. Tom Soniat (2002) Biology Department Nicholls State University Thibodaux, Louisiana 70310 Dr. J. Evan Ward (2002) Dept. of Marine Sciences University of Connecticut Groton, CT 06340-6097 Dr. Gary Wikfors (2002) NOAA/NMFS Rogers Avenue Milford, Connecticut 06460 Journal of Shellfish Research Volume 20, Number 2 ISSN: 0730-8000 December 2001 www.shellfish.org/pubs/jsr.htm PROCEEDINGS 4th International Symposium on Abalone Biology, Fisheries, and Culture University of Cape Town Cape Town. South Africa February 6-11. 2000 Guest Editor: Peter A. Cook Zoology Department University of Cape Town Rondehosch. South Africa Jounuil oj Shcllfisli Research. Vol. 20. No. 2, 565. 2(){)l. PREFACE The 4lli IiULTTUitKinal Symposium on Abalone Biology. Fish- eries, and Culture was held at the Uni\ersity iif Cape Town. South Africa from 6th to 1 1th February. 2000. The eonferenee attracted about 250 delegates from at least 15 different countries, and was successful in bringing together scientists, entrepreneurs, divers, aquaculturists. and government representatives. This Special Issue of the Journal of Shellfish Research contains nian\ ot the contri- butions presented at the conference. The Cape Town Abalone Symposium was the fourth in u series that started in La Paz, Mexico in 1989. The second was held in Hobarl, Tasmania in 1994 and the third in Monterey, California in 1997. It was decided, at the Monterey symposium, that a three-year interval between symposia was appropriate in view of the rapid changes that were occurring worldwide in abalone fisheries and in farming techniques. The large quantity of new information pre- sented at the Cape Town conference proved this decision to be correct. The choice of Cape Town as the venue for the 4th International Abalone symposium was appropriate due to the fact that the South African abalone tlshery is, perhaps, one of the most over-exploited in the world, principally because of illegal exploitation by poach- ers. The occasion of the symposium was used to highlight this problem, and extensive media coverage provided a platform for scientists to sound alarm bells about the state of the fishery. In addition, the south-western region of the country has developed as the center of South Africa's abalone farming industry and confer- ence delegates were able to benefit from \isits to local farming enterprises. The symposium was opened with an interesting and thought- provoking review of the international supply, markets, and pricing of abalone products. The rest of the symposium was divided into a number of sections including abalone aquaculture. nutrition, ge- netics, diseases, stock assessment, and fisheries management. This volume contains a selection of the papers from each section and all papers have been through the usual stringent refereeing and review process applied by this journal. I wish to thank the authors who submitted manuscripts for publication in this volume and. in par- ticular. I also wish to thank the numerous referees who contributed their time to review the papers. The 5th International Abalone Symposium will be held in China in 2003. Peter Cook Editor DEDICATION This volume is dedicated to the memory of Mia Tegner, whose work on abalone contributed enormously to our understanding of the biology of this unique creature. 565 .lounuil ,>l Slwllfisli Research. Vol. 20. No. 2. 567-57U. 2001. WORLD ABALONE SUPPLY, MARKETS AND PRICING: HISTORICAL. CURRENT AND FUTURE H. ROY GORDON' * AND PETPZR A. COOK"" 'Fislitcch Inc. B(i\ 6SS6. San Rafael. California 94903 'Zoologx Departiiieiu. Uiiivcrsirx of Cape Town. Ronclebosch, 7700. South Africa ABSTRACT The world supplies of wild-caughl uiid cullured ahalone are considered with implications of the past, present and future. Much of the data available in recent years, from various government sources, and even FAO, have often been inadvertently misleading. Reporting has often combined a number of dissimilar abalone products (fresh in shell, frozen meat, canned, dried, etc) resulting in misleading results. Surprising numbers result when an effort is made to standardize production and export information for both the commercial catch and cultured product. Factors affecting abalone FOB and CNF prices are discussed. Different market forms (live, fresh, frozen, canned, dried) affect pricing, and price is also influenced by processing and packaging as well as economic conditions. Suggestions are made for value added abalone products and the challenges of sustaining world demand at premium prices are discussed. KEY WOKUS: abalone, abalone market, abalone prices, abalone process and packaging, abalone future INTRODUCTION World abalone supply statistics have not all been accurate over the past 30 years. In a reexamination of the statistics the industry has used for many years, a need for some substantial adjustments in both the fisheries and the cultured sectors has been found. Annual data for some countries has substantial gaps in reporting. Whilst some countries have reported tonnage as "in shell", others report "meat only". This distinction alone can distort the compara- tive total production of an individual country by two to three hundred percent. To compHcate the analysis further, export and import data often combine a number of dissimilar abalone products (fresh in shell, frozen meat, canned, dried etc) into a single nutnber for tonnage and value. In an attempt to standardize the reporting of the world abalone supply, the following definitions are applied: Abalone fisheries: The total allowable annual commercial land- ing quota (country by country) expressed in terms of "in shell" weight. This category would include the planting of seeds in large areas of the sea (e.g. in Japan) wherein the sea bottom has not been prepared with man-placed rocks or structures. This definition does not include the legal sport catch or any illegal catch worldwide. Cultured abalone: (expressed in terms of "in shell" weight) includes both the farming of abalone on land or in the sea — contained in man-tnade tanks, nets or structures (Intensive culture) and sea planting of abalone seeds in artificially arranged substrate or structures, with or without added food (Extensive culture). The illegal catch: Any harvest of abalone beyond the total allowable annual landing quota. To discuss the topic of world supply without consideration of the illegal catch would be grossly misleading (Muiphy 1999). TEN-YEAR FISHERIES COMPARISON In order to compare the abalone fisheries and cultured totals on an "apples with apples" basis, we must apply a standard "in shell" weight. Retroactive reporting (in particular China, but other coun- tries as well) dramatically changes the cultured abalone totals (New 1999). We now know that the FAO Yearbook of Fisheries Statistics had been reporting China's production of several groups of cultured molhisks. includins abalone. as shelled or shucked *Corresponding author. E-mail: rgordon@fishtech.com (H. Roy Gordon). ''Also corresponding author. E-mail: pcook@botzoo.uct.ac.za (P. Cook). weight, instead of their "in shell" live weight equivalent, which should be normal practice in submitting statistics to FAO. This has grossly underestimated both cultured abalone production and aba- lone fisheries. Adjustments must, therefore, be made to previous reporting over this 10-year tittieframe. Figure la shows that the worldwide catch from abalone fish- eries has declined by about SO^r over the past 10 years. In 1989 worldwide abalone catch was reported as 12.995 mt but. based on the assumptions given earlier concerning incorrect reporting, this has been adjusted to 14,8.30 mt. It is estimated that the catch in 1999 was 10.150 mt. From Figure lb it can be seen that Australia is the only country where abalone fisheries have increased over the past 10 years. Significant declines have occurred in countries such as Mexico, the U.S.A.. and Japan. TEN-YEAR COMPARISON OF CULTURED ABALONE The situation with regard to cultured abalone is completely different (Figures 2a and 2b). Over the ten year period from 1989 to 1999. whilst abalone fisheries declined by about 30%. the pro- duction of cultured abalone increased by over 600%. In 1989 cultured abalone production was reported as 689 mt (adjusted to 1.220 mt). whilst in 1999, it is estimated that production was 7.775 mt. Of the total estimated 1998 cultured abalone world production of 7.165 mt, Asia accounted for 5,500 mt or 75% of that total. Within Asia, the vast majority of the cultured abalone production is from China and Taiwan. ABALONE PRICING With regard to the price of abalone products, it is interesting that the F.O.B. and C.N.F. prices for similar species and sizes tend to equate worldwide. Abalone is marketed in many different forms, each of which is priced differently. Figure 3 illustrates, diagrammatically. a range of prices that are quoted for different forms of abalone, but the examples given below will show that, when processing and marketing costs are taken into account, these prices tend to revert back to a similar "in shell" price of about US$32/kilo. Example one: cultured Red abalone in the 90-mm-size range. (The principle is the same with other species when adjusted for the size/weight ratio of that species.) Many of the Asian traditional recipes call for dried abalone as their preparations began thousands 567 568 Gordon and Cook A 16,000 14.000 12,000 10.000 S.000 6.000 4.000 2.000 0 (Metric tons) ,.'-,-wffi^ .j^SiS S80/kilo (L.inned) Abalone Fisheries B 6,000 AJjiaBd 1919 EitaBlrdlf*f (Metric tons) D Japan DNZ O Australia DUSA D Mexico O S. Africa B Others' Figure 1. (A) Worldwide abalone fisherie.s: ten-year comparison: (B) Abalone fisheries country by country: ten-year comparison. of years before refrigeration. There are a few special (often secret) processing methods creating a dried product which sells for a minimum of US$7()0 per kilo and as much as $12,000 per kilo or more. This dried abalone is not to be confused with the "standard" sun dried product that sells at a lower price. The preparation of this specialty-dried abalone is almost ceremonial, often prepared with unique coals and slow cooking in ceramic cooking utensils. Only about 10% of the "in shell" weight remains in the dried product. The preparation process is highly labor-intensive and takes months to complete. If the processor pays US$32/kilo "in shell", and sells the premium product for US$7(X)/kilo, he is making only a rea- sonable profit for the effort. E.xample two: a live 90-mm abalone at the Japanese wholesale A 80OO 7000 <000 5000 4000 3000 2000 1000 0 - (Metric tons) Cultured abaJon« Sl,200/kilo, lJrK\li S32/kilo ("In shell" weight) S700/kilo (dried) $45/kilo llresli. liiel Figure 3. Prices for different market forms of abalone (for explana- tion see text). (auction) markets. There are fisheries pncing services that regu- lady publish an average of 1 0 major Japanese wholesale markets (Sunee 1999). For the sake of comparison, we will use a price of 4,700 yen/kilo. This equates to approxiinately US $45/kilo. Does that mean you might sell to these markets at a C&F level approach- ing these prices? Not by a long shot: there are auction commis- sions, wholesaler commissions, trading company commissions, lo- cal transportation, customs and duty — all to be deducted froin your selling price along with possible mortality deductions. In effect. $45/kilo equates closer to $32/kilo. Example three: weight los.ses during canning (illustrated in Figure 4). If the "in shell" weight is 1425 g (H. nifescens. 8 pes, 100 mm "in shell", for example) shucking will yield about 640 g ot meat which, after cleaning and trimming, will yield approxi- mately 390 g. After cooking the meat, you will find only 252 grams. In effect, after shrinkage and canning cost, a $20 tin of 252 g meat equates to an "in shell" price of as low as only $l2/kilo. Without the continued development of premium product and brand name tins, canned abalone will remain as a commodity with rela- tively lower prices. ALTERNATIVES TO WORLD COMMODITY PRICING As illustrated in Figure 5. the world abalone market is domi- nated by Asian countries. Prices are currently driven by a handful of Asian nations, guided by their historical and changing customs, preparations, populations and economies. An important additional influence is that of Asian populations living elsewhere in the worid. As displayed earlier, regardless of the form to market (live, fresh, frozen, boiled, canned, dried etc) the abalone price "in shell" tends to equate worldwide. Of course, there will be shorter-term price variations as the economies of that handful of Asian buyer countries weakens or strengthens. In recent years, we have seen a substantial weakening of the Japanese economy — now recovering. On an equivalent yen basis, adjusted to USS, $37/kilo for a given B 3500 3000 ■ 2500 • 2000 ' 1500 1000 ■ 500 0 (Metric tons) HE ^ a Japan OChlna O Taiwan DUSA a Mexico ■ S Africa O Chile ■ Australia a Othei^- 1989 1999 Figure 2. (a) Worldwide production of cultured ahalone: ten-year comparison; (b) Cultured abalone production country by country: ten-year comparison. 1425g 640g 390g 252g $20/Tin of just 252g Abalone = only $12/kilo "in shell" Figure 4. Weight loss during canning of abalone. World Abalonl Supply. Markets and Pricing 569 Figure 5. 1999 World abalonc marki'l. size and quality in January 1998, this would equate to $4 1 /kilo in January of 1999 and $45/kilo in January 2000. During this same period. PR China's economy has markedly strengthened. Just a few years ago, the majority of the abaione farming production in China was designed to export this premium product to Japan and elsewhere. Today, virtually all 3.000+ tons of China's abaione production is consumed internally. To break away from world pricing, a company responsible for large quantities of abaione. whether fished or cultured, would have some important choices to consider. These could include: • Continue the selling pattern • Develop premium market for e,\port or withm the exporting country • Brand name product • Multi functional processing and packaging • Sous Vide and "skin packed" fresh and frozen • Direct approaches to DWE's (distributors, wholesalers, major end users) Conliniie ihe Selling Pattern A company could continue along the lines of the patterns men- tioned earlier (prices likened to a commodity — driven by Asian Markets — varying by species and size) which tend to equate to a similar "in shell" price worldwide. Alternatively, a company could begin to break away from the "price equation" through implementing any one, or several, of the following choices. Develop a Premium Market for Export and Within Currently Exporting Countries Exporting nations should not take for granted that premium prices will continue for ever. Considerable effort will be required. You may hear "we had a strong abaione market within our country in the late 60s early 70s, however, that was when prices were very low". If we re-examine that statement on an inflation-adjusted basis, a .S30/kg price today would equate to a $6.79/kg price in 1970 (Antweiler 2000). That "expensive" restaurant abaione meal at $45 today equates to a $10 "expensive" abaione meal in 1970. The premium pricing in the "home market" of the currently ex- porting nation will not remain as a "given". Abaione will never become a low cost food, however, effort and promotion will be needed to maintain the highest market pricing levels. Brand ,\'ame Product Much of the tlsheries and aquaculture world has worked with brand names for many years, but this is rare in the "abaione world". Premium (higher than the most common denominator) pricing is achievable through developing a high quality, branded product and then telling the world about it. Sons Vide and "Skin Packed" Fresh and Frozen We live in a "food .service — ready to eat — fast food" industri- alized world, yet abaione traditionally goes to market in the same basic forms (live, fresh meat, frozen, canned etc.) and is distributed to the same standard markets. Larger producers should consider developing "heat and serve" processed, prepared abaione dishes. This may sound like heresy to the abaione traditionalist, but it is a clear path toward maintaining premium prices in the home markets of abaione exporting nations. Some companies have been promot- ing vacuum pack/skin pack for many years with rewarding results, but they represent only a small fraction of the marketplace. Fast food need not be a low-priced food. A Hong Kong food chain is currently promoting an abaione fast-food concept, with a single outlet generating over US$2,0()0,00() annually. Multi Functional Processing and Packaging Industrialized nations need to develop low cost production pos- sibilities with highly mechanized food service facilities. (This level of processing is quite expensive, but could be accomplished on a co-op basis.) Today, several industrialized nations have such plants operating for bivalves. A modem, low-cost process complex might encompass six multifunctional production areas including raw material supply, raw material pre-treatment, processing/ transformation, preservation, packing, storage. Direct Approaches to DWE's (Distributor, Wholesaler, and major End Users). This method- ology requires considerable effort, yet is rewarding. It should be your market study that selects the DWE, not the reverse. WORLD SUPPLY/DEMAND RELATIONSHIPS: 1977-1999-2004 Figure 6 illustrates the relationship between world supply and demand in 1975, 1999 and 2004 (predicted). In 1975 the supply was about 20,000 mt (on an adjusted basis, that might have been 24,000 mt or more). There was no major shortage of product, and this resulted in a demand/supply balance. The 1999 supply was 18,000 mt (or 13.000 mt after deducting "new market" H. siiper- te.xta which was not a factor in the 1975 demand) whilst the de- mand remained at over 20,000 mt, leaving a potential shortfall of about 7,000 mt. It is estimated that by 2004 the supply will be about 15,000 mt (after deducting "new market" H. supertexia) but 20,000 18.000 16,000 14,000 12,000 10.000 8,000 6,000 4,000 2,000 0 I supply demand 1999 2004 No(* New niarkei,imalJpr, Iflwer vaJie. ruHujTd H juperlpsta U rujltniluried In -simply" (5.000 ml 1998.6,000 no 2004 Figure 6. Abaione .supply and demand In 1975, 1999 and 2004. 570 Gordon and Cook the potential demand is likely to remain over 20.00U mt. This will leave a potential shortfall of at least 5.000 mt. CONCLUSIONS World Ahalone Supply Over the past 10 years, the abalone fisheries of the world have declined by about 30% while the world's cultured abalone pro- duction has increased over 6009^. The trend toward larger, cul- tured, premium species will continue. Ten years ago a 70 mm animal was considered "market size" whereas today market size is closer to 90 mm. Some production at 120 mm or larger will be required in the future. Looking ahead to the year 2004. we anticipate the abalone fisheries to remain fairly flat at the 10,000 to 11,000 mt levels, while cultured abalone farms are anticipating very substantial in- creases. Political, environmental and pathological events will, of course. ha\e some unknown impact on world abalone supply. World Ahalone Demand We have been discussing a strong world demand and potential "shortfalls" but this does not necessarily imply an automatic de- mand at a premium price. Outside of the Asian world, the desire for this "caviar in a shell" priced animal will be directly related to the ability of the abalone industry to compete successfully in the marketplace. Among other things, this should include Sous Vide preparations, brand identification, sophisticated processing facili- ties and unique DWE programs. For the sake of our own and future generations, we must do a better job of protecting the world's abalone fisheries, however, it is the cultured abalone industry that must not only expand produc- tion hut must also increase efforts to assure a continual world premium market. ACKNOWLEDGMENTS Contributions were made by: Nagashisi Uki (Japan). Nie Zong Qing (China). Ray Fields (USA). Roberto Flores and Enriqe Vaza- quez (Mexico). Atilio Ziomi (Chile). Michael Tokley (Australia). Rodney Roberts (New Zealand). Andre Du Plesis (South Africa). Gavin Burnell (Ireland and Europe). Alawi Salim Al-Hafidh (Oman) and Acnar Steinarsson (Iceland). LITERATURE CITED Antweiler. W. 2000. Historical currency exchange. Pacific Exchange Rate Service. Vancouver. BC: University of British Columbia. Accessed at www.oanda.com Asakawa. T. 1998. Abalone Market in Japan, US National Marine Fish- eries Service. E-mail update to H. Roy Gordon. FAG Fish and Fishery Products. World Apparent Consumption Statistics. 199« Senes 821. Murphy, D. E. 1999. S. Africa abalone poachers. Times/Mirror Co. Los Angeles Times July 8. 1999. New. M. B. 1999. Current trends and challenges for the 21st century. GIoIhiI Aijiiticiillure Magazine 30:8-70. Sunee. S. C. 1999. US National Marine Fisheries. Fislwn Market News, Prices. Tokyo Central Wholesale Market, January 2000. Joiinnd of Shellfish Research. Vol. 20, No. 2. 371-5S6, 2001. A REVIEW OF SETTLEMENT CUES FOR LARVAL ABALONE (HALIOTIS SPP.) RODNEY ROBERTS* Cawthroii Insiitiite. Private Bag 2. Nelson, New Zealand ABSTRACT Settlement of abalone larvae involves larval attachment (a reversible behavior) followed by metamorphosis (which involves irreversible physical changes). Coralline algae induce attachment and metamorphosis in all abalone species tested, and there is some evidence of settlement preferences for certain coralline species. The seltletnent-indiicing chemicals from corallines have not been identified. In one case [Hulimis rufescens) a small peptide is implicated, while in another (Halioiis discus hcimnii) halomethanes are thought to be critical. Corallines are generally regarded as unsuitable for use in hatcheries, but their potential use has not been fully evaluated. Many abalone hatcheries rely on biotllms to induce larval settlement. The activity of biofilms increases with their age. Ungrazed films are generally dominated by fast-growing benthic diatoms, and settlement on these films is variable and often low. Few diatom strains are consistently good for settlement, and strains that are excessively mobile, or form 3-dimensional colonies, can prevent successful settlement. The chemistry of settlement induction by biofilms. and the role that bacteria play, are poorly understood. Bacteria can induce settlement, but metamorphosis occurs gradually over several days, if at all. Pregrazing by conspecific abalone improves the settlement-inducing activity of a biofilm. The mucus from the foot of grazers probably contains chemicals (not identified! that trigger attachment or metamorphosis, particularly when combined with another cue (such as dibromomethane or a biofilm I. Various pure chemicals induce attachment and/or metamorphosis of abalone larvae. These may bind to larval receptors (e.g.. ^-aininobutyric acid = GABA) or act "downstream" of. or "parallel to", the receptors (e.g.. compounds that depolanze membranes or alter levels of cyclic AMP or calcium). None of these chemicals is considered to be a natural settlement cue. and only GABA has been used in abalone hatcheries. The timing and end point of the abalone settlement response vary among cues, and among abalone species. Cues can combine synergistically to enhance settlement (e.g., GABA with lysine, biofilm or coralline extract, dibromomethane with mucus). Signal reception pathways are known only for the case of Haliotis rufescens and GABA-mimetic inducers. The response is controlled through a "morphogenetic pathway", which is subject to up- and down-regulation by molecules acting on a separate "regulatorv pathway". L;u-val attachment and metamorphosis are often uncoupled in abalone. Cues for attachment are much more widespread than cues for metamorphosis. KEY WORDS: abalone, attachment, metamorphosis, diatoms, biofilms. coralline algae, settlement cues INTRODUCTION Abalone larvae swim for several days before becoming com- petent to undergo settlement, which transforms them into a crawl- ing, benthic snail. Settlement is triggered by external cues. Knowl- edge of these cues is critical in abalone culture where complete. rapid and predictable settlement is desired but seldom achieved (Searcy-Bemal et al. 1992a, Roberts et al. 1998). Settlement cues also play a role in determining where abalone recruit, and knowl- edge of cues may be useful in site choice for larval reseeding. Several authors have previously reviewed aspects of abalone larval settlement. The extensive work by Daniel Morse and col- leagues on metamorphosis induction in Haliotis rufescens has been reviewed in detail (A. Morse 1991. D. Morse 1992). McShane's (1992. 199.'i) reviews discuss settlement in natural habitats and attempt to draw links between laboratory and field data. Hahn (1989c) presented a detailed summary of the data on abalone settlement induction as of 1985. In the decade or so since these reviews there has been substantial research on abalone larval settlement. The new work has broadened the range of abalone species studied from its early focus on Haliotis rufescens and Haliotis discus hannai, and has also broadened the range of cues that have been addressed, and the depth of investigation for certain key inducers. The body of data now available is adequate to justify a comprehensive review. This study concentrates on abalone larval settlement cues, whether natural or otherwise. Ecological rel- evance is discussed only briefly. This review attempts to empha- size the distinction between the components of the settlement re- sponse, which has received little critical attention in previous re- views. It also attempts to highlight similarities and differences *E-mail: rodney(n'cawthron.org.nz among abalone species. Both of these tasks were made difficult by varying methodology and terminology, which are discussed in detail at the outset in the hope that greater standardization will result. The review concludes by summarizing some major points arising from the review, making some comments on settlement cues for hatcheries, and providing some suggestions about priori- ties for future research in this field. SETTLEMENT TERMINOLOGY AND METHODOLOGY Terms Used to Describe Settlement Stages Various terms have been used interchangeably in publications dealing with abalone larval settlement (Table 1 ) so clarification is required. In this review, the transition from a swimming larva to a crawling, feeding post-larva is divided into three stages. The first is larval "attachment" — the larvae stop swimming, sink to the bottom, and attach to the substratum by their foot. Larvae retain their velum during attachment and some larvae choose to resume swimming — attachment is a reversible behavior. The next stage marks the initiation of metamorphosis and is conveniently indi- cated by the shedding of the larval velum, committing the abalone to benthic life. This stage is here termed "metamorphosis", and involves an array of irreversible physical changes. If the post-larva continues to develop normally for a day or two after initiation of inetamorphosis. then the mouth will open, feeding will begin, and flared peristomal shell will begin to grow. This further metamor- phic developtnent is sometimes separated from the earlier shed- ding of the velum, and is referred to here as "shell growth". The term settlement is used as a general term for the transition from a larva to a post-larva. Data that do not distinguish between attach- ment and metamorphosis are reported here as "settlement" (in quotation marks). 571 572 Roberts TABLE 1. Examples of terms used in published abalone settlement studies and the corresponding term used in this review. Term Used in Original Paper Equivalent Term in this Review Notes Reference Seltlement Settlement Interred settlement "Settlement" Behavioural metamorphosis Attachment + metamorphosis Attachment Unknown Unknown Attachment Settlement defined as being animals that had shed the velum. Attachment without metamorphosis was not quantified. Attachment scored separately from metamorphosis. Calculated from number of larvae still swimming. "Settled" larvae would include those attached or metamorphosed plus any stuck to surfaces. Paper does not define settlement. Daume et al. tmWa. 1994b. 2000) Morse et al. (1979b) Kawamura & Kikuchi (1992) Searcy-Bernal et al. (1992) Slattery (1992). Seki (1997) Moss (1999) Genade et al. (1988) Morse et al. (1980a) Settling abalone larvae often do not progress rapldiv through attachment, nietainorphosis and shell growth. Larval attachment is not a useful end point for an abalone larva, or an abalone farmer, so the distinction between the .stages of settlement is critical in interpreting data. It is difficult to make these distinctions in the hatchery, and even more so in the sea, so laboratory experiments have proven very useful in the study of abalone settlement cues. Abalone Seltlement Assay Methods In static settlement bioassays the use of antibiotics to suppress bacterial growth is essential for some abalone species to ensure consistently high larval survival (Morse et al. 1979b). Different degrees of bacterial suppression led to conflicting results over the activity of 7-aminobutyric acid (GABA) (Morse 1992). Negative controls are crucial to ensure that settlement responses are not falsely attributed to particular cues. Several studies have reported high attachment or "settlement"" in negative controls (Liu et al. 1986. Genade et al. 1988. Slattery 1992. Moss 1999) making it difficult to interpret settlement induction by other treatments. Some problems with antibiotic toxicity and interference have been reported. Streptomycin was toxic to Hallntis cliversicolor at doses as low as 3 [O-g ml"' (Bryan & Qian 1998) and emetine toxicity caused abnormal loss of velum cells in Haliotis nifescens (Fenteany & Morse 199.'^). Anisomycin caused slight attachment induction in Haliotis ntfcscens. but also stimulation of swimming activity that partly antagonized the response to GABA (Fenteany & Morse 199.^). The methods used will depend on the objectives of an experi- ment. If the objective is to determine larval responses in hatchery conditions then full-scale hatchery systems should be used and antibiotics will be inappropriate. However, if the experiment is to determine the settlement response of larvae to particular cues then the assay system should be designed to eliminate all potentially confounding influences, including bacterial interference. If larval mortality or morbidity in assays complicates interpre- tation then this needs to be discussed. For example. Moss ( 1999) recorded high "inferred settlement"' by day five of an assay but then explained that mortalities were very high (88 to 97% by day seven) tiiaking interpretation of "settlement"" data difficult. Larval and post-lar\al abalone have considerable ability to survive with- out particulate food (Searcy-Bemal 1999. Takami et al. 2000. Rob- erts & Lapworth 2001. Roberts et al. 2001 ). This suggests that high mortality within a few days of settlement induction (Daume et al. 1999b. Moss 1999) is likely to be caused by factors other than starvation. Dead or moribund larvae should be excluded from cal- culation of percentage settlement, but they still introduce uncer- tainty. Would those animals have settled if they were healthy ' Is the larval settlement response influenced by the marginal condi- tions ill the assay? Similar attention needs to be paid to toxic responses that can cause velum loss and could be confused with the onset of meta- morphosis. Typical symptoms of toxicity include partial loss of the velum, an inci'easing number of larvae found lying on their sides (rather than attached by the foot or swimming high in the water colurnn) and lack of shell growth (Fenteany & Morse 199,?). Assay counts may need to include abalone on the walls of assay chambers (Genade et al. 1988) or on the surface film of the assay water (Roberts & Nicholson 1997). The difficulty of accurately counting swimming larvae has been circumvented by doing retro- spective destructive counts (e.g. Roberts & Nicholson 1997) or leinoving the settlement substrate to aid counting (e.g.. Moss 1999). While larval attachment is generally rapid, many cues cause only a gradual rise in the peicentage metamorphosis over a week or more (e.g. Kawamura & Kikuchi 1992. Roberts & Nicholson 1997). Thus, the period of observation influences the settlement observed, and this tnust be taken into account when comparing studies. The timing of observations should be tailored to the ques- tion being addressed, and multiple observations will be most in- formative. Larval age can affect settlement responses. Abalone larvae have a pre-competent phase of several days during which they will not respond to settlement cues. Various morphological and behav- ioral features have been used to indicate competence to settle. These include the appearance of the fourth tubule on the cephalic tentacles (Seki & Kan-no 1977. Ebert & Houk 1984). exploratory Review of Abalone Settlement Cues 573 crawling (Seki & Kan-no 1977) and the presence of at least three rows ofchitinized radula teeth (Moss & Tong 1992a). Morse et ai. (1979b) used a 30-120 minute exposure to GABA to test larval competence to attach. An indirect method is to calculate larval stage based on larval age. rearing temperature and biological zero point (Seki & Kan-no 1977. Hahn 1989b I. Larvae of many abalone species reach competence to settle after several days at 15 to 20°C (Hahn 1989b). The ability to attach arises earlier than the ability to metamorphose, and the speed of a settlement response can vary with larval age (e.g.. Morse 1984a. Barlow 1990). Halious diver- sicolor larvae can begin attaching as early as 1 2 hours after fer- tilization (at 24T) but may not metamorphose until several days later (Liu et al. 1986. Bryan & Qian 1998). Positive controls should be included in experiments to confirm larval competence. Data can even be normalized to a positive control to take account of inter-batch variability in larval responsiveness (Baloun & Morse 1984). Other factors that may affect the outcome of abalone larval settlement assays include: ( 1 ) the use of vertical versus horizontal surfaces (Bryan & Qian 1998). (2) the size of the settlement sub- stratum relative to the assay chamber (e.g.. Daume et al. 1999a. Daume et al. 1999b versus Kawamura & Kikuchi 1992). (3) the increased sensitivity of older larvae to settlement cues (Barlow 1990. Moss & Tong 1992a. Yang & Wu 1995). (4) variability between batches of larvae (e.g.. Baloun & Morse 1984. Trapido- Rosenthal & Morse 1986b. Roberts & Nicholson 1997). and (5) changes in light intensity causing attached larvae to resume swim- ming (Moss & Tong 1992a, Moss 1999). CORALLINE ALGAE AS SETTLEMENT CUES The Corallinaceae are divided into articulated ( = geniculate) and non-articulated (= non-geniculate) forms (Johansen 1981). The non-articulated corallines are commonly referred to in the abalone literature as crustose corallines, so this review uses the terms crustose and articulated corallines. The crustose corallines occur in a wide range of growth forms (Woelkerling et al. 1993. Shepherd & Daume 1996). In the wild, juvenile abalone are closely associated with crus- tose coralline algae (e.g.. Morse et al. 1980c, Shepherd & Turner 1985, McShane & Smith 1988. Shepherd & Daume 1996. Day & Branch 2000). Laboratory experiments show that intact crustose coralline algae are among the best inducers of larval attachment and metamoi-phosis for all abalone species tested (Morse et al. 1980c. Morse & Morse 1984a, Moss & Tong 1992a, Roberts & Nicholson 1997, Takami et al. 1997, Daume et al. 1999a, Daume et al. 1999b, Moss 1999). Crustose coralline algae generally induce strong and rapid metamorphosis (>80% within two days) but Hali- olis nihni and Haliotis laevigata sometimes show much weaker responses to three species of crustose corallines (Daume et al. 1999b, Daume et al. 2000). Articulated corallines were much less effective (<\0% attach- ment) than crustose corallines as settlement cues for Haliotis nife- sccns (Table 2, Morse et al. 1980c, Morse & Morse 1984a). The articulated Coiallina officinalis showed moderate to strong activ- ity with Haliotis virginea but was less effective than crustose corallines (Table 2). Settlement of Haliotis cyclobates on sea grasses may be triggered by the crustose corallines that cover 1 3-20% of the seagrass blades, or by other cues such as the sea grass or its surface biofilm (Stevenson & Melville 1999). When given a choice of crustose coralline species larvae of Haliotis laevigata selected Sporolillion chimin in preference to Mesophylliiin eiigelhartii or Hydrolithon nipeslie (Daume et al. 1999a). This is consistent with the disproportionately high number of Haliotis laevigata recruits found on Sporolithon duniin in the wild (Shepherd & Daume 1996). OTHER MACRO-ALGAE AND BLUE-GREEN ALGAE Crustose red algae are the only seaweeds known to strongly induce Haliotis nifcscciis settlement while intact. This observation is based on a relatively small sample of "other" seaweed species (Table 2). Larval attachment was <1% after 24 hours of exposure to intact foliose algae or blue-green algae (Morse et al. 1980c. Morse & Morse 1984a. Morse et al. 1984). Haliotis discus luiniuii seems to be less discriminatory — Seki ( 1997) found that 6 out of 12 foliose macro-algae induced metamorphosis of Haliotis discus luinnai larvae. The active species included brown, red and green algae (Table 2). Ulvella lens is a small, crustose green alga that grows as pros- trate rosettes on hard surfaces. It is common on heavily grazed surfaces such as abalone tanks, and has been used in commercial abalone hatcheries as a settlement cue. Pregrazed films of Ulvella produced >90'7r attachment (Takahashi & Koganezawa 1988. Seki 1997) and metamorphosis (Takahashi & Koganezawa 1988) of Haliotis discus hamuli larvae while ungrazed Ulvella films in- duced 60-70% attachment and metamorphosis (Takahashi & Koganezawa 1988). Germlings of Ulvella lens induced 10-50% of Haliotis rubra larvae to metamoijihose within one to three days (Daume et al. 2000). This was higher than the metamorphosis recorded on films of several benthic diatom species in that study. Prostrate brown algae of the genus Myrioneina are often abun- dant (along with Cocconeis spp. and Ulvella lens) on the pregrazed plates used successfully to settle abalone (Seki 1980. Suzuki et al. 1987). The author is not aware of any data on the settlement- inducing activity of Myrioneina spp. CHEMICALS EXTRACTED FROM ALGAE AND BLUE-GREEN ALGAE Attempts to purify the chemicals that trigger abalone settlement have focused on crustose ct)ralline algae because of the corallines' consistently strong activity and their role in abalone ecology. The extensive work on purifying chemical inducers of Haliotis riife- scens settlement has been reviewed previously (A. Morse 1991. D. Morse 1985, D. Morse 1990, D. Morse 1992). so only a brief account is given here. When intact seaweed speciinens were tested, only crustose red algae induced high levels of larval attachment and metamorphosis of Haliotis rufescens larvae (Table 2). How- ever, testing of crude extracts revealed that blue-green algae and foliose red algae also contained inducers intracellularly. while two brown algae, a green alga and a bacterium did not (Table 2. Morse & Morse 1984a, Morse et al. 1984). Red and blue-green algae contain phycobiliproteins, which the other algae lack (Morse 1985). The activity from coralline Lithothainnium californicum. the foliose red alga Porphyra sp., and the blue-green alga Spirulina plalensis was initially protein-associated, but could be chromato- graphically separated from proteins into a low molecular weight (-650-1250 Mr) form (Morse & Morse 1984a, b, Morse et al. 1984, Morse 1992). The activity of the protein-associated form was enhanced by boiling, proteolytic digestion, and acid hydroly- sis, suggesting that the active product may have been an amino acid or small peptide (Morse et al. 1979a). The peptide-associated 574 Roberts TABLE 2. Settlement of abalone larvae in response to macroalgae and blue-green algae. Haliolis nifescens H. virginea H. discus hannai Attachment & No. of Algal % Attach. % Met. No. of Algal .attachment & No. of Algal Algal Species Metamorphosis Species (Range of means or Mean ± SE) Species Metamorphosis Species Intact algae 2 da\ 2 da> Crustose coralline + 4 84-95 64-88 4 + 1 Articulated coralline -KS8%) 2 89 ± 2 56 ± 5 1 1 Crustose non-coralline red + I Foliose red - 3 +/- (3/2) Foliose brown - 2 +/- (2/4) Foliose green iUlva spp.) - 1 + 1 Crustose green (Ulvella lens) + 1 Blue-green - 1 Crude extract from: Dose of 20 |xg protein niT' 2.5 day 5,5 da\ Crustose coralline -1- 4 56-81 18-65 Articulated coralline + 2 37 ±6 68 ±13 Crustose non-coralline red + 1 38 ± 5 63 ± 8 Foliose red + 3 88 + 9 Toxic Foliose green - 1 10 ±4 4 + 0 Foliose brown - 2 18-38 2-13 Blue-green + 3 21 ±2 6± 1 Filtered seawater control 6 ± 5 0 + 0 0.01 M Tris buffer control 1±0 0±0 References Morse and Morse (1984a) R. D. Roberts unpiibf data'' Seki (1997) Morse et a . (19841 ° To prepare crude extracts, 4 g of fresh alga (or 1.5 g dry Spiniliiici) was disrupted with a Bellentini bead shaker in -25 ml 0.01 M Tris-HCI (pH 7.0) then cenlrifuged at 24.000 x g for 10 min at 2°C. Protein content of the supernatant ( = crude extract) was determined by the Lowry method and extracts were assayed as per Roberts and Nicholson (1997). Data presented are all for a dose of 20 ji.g protein ml"' of assay medium. Greater doses generally gave higher attachment and metamorphosis, but often became mildly toxic to larvae within a few days. Note: "Attach." = attachment. "Met." = metamorphosis. To read Haliotis discus hannai data: "+/- (3/2)" means induction by 3 species of foliose reds and lack of induction by 2. inducer molecule is known to contain several unidentified unusual amino acids but its precise structure is still unknown (Morse 1992). The inducer purified from Litliothamniiim mimics the activity of GABA. The response of Haliotis viii^ineci to algal extracts is complex. At a low dose, the activity of extracts from various algal groups (Table 2) appears similar to that reported for Haliotis nifescens. except that blue-green algae have low activity. At higher doses, the brown algae induced means of 72-76% larval attachment within two days but caused mild toxicity after five days (data not shown). While most extracts induced some degree of larval attachment quite rapidly, metamorphosis occurred only gradually (Table 2) and shell growth was almost completely inhibited. A similar re- sponse to coralline extracts (Fig. 1 ) has been reported previously for both Haliotis virginea and Haliolis iris (Roberts and Nicholson 1997). The lack of shell growth appears to result from toxicity of co-extracted compounds preventing normal post-larval develop- ment and activity subsequent to the shedding of the velum (Rob- erts and Nicholson 1997). Extracts from the coralii]ie alga Ph\- matolitlum repandum contained inducers in both low molecular weight (600-1100 Mr) and relatively high molecular weight OIO.OOO Mr) fractions of polar extracts (Robeds et al. 1994), paralleling the situation with Haliotis nifescens and Lilhotham- niitm californiciim (Morse et al. 1984), The active chemicals have not been identified. Attempts to pui'ify chemical inducers of urchin metamorphosis from coralline algae (Taniguchi et al. 1994) led to the discovery that metamorphosis of Haliolis discus hannai larvae is induced by a combination of dibromomethane and mucus (Seki et al. 1997), Larval attachment and metamorphosis on corallines are hypoth- esized to be independently triggered by trail mucus and dibro- momethane respectively (Seki et al. 1997), Dibromomethane is a volatile chemical naturally released from corallines and many other alsae (Tanicuchi et al. 1994). & &. &. IH2 d intact ■ 2d extract □ 3d extract □ 4 d extract D 5 d extract Attachment Metamorphosis Shell growth Figure 1. Comparison of the settlement response of larval Haliolis iris to intact plants and aqueous extract (440 pg dry extract ml"') of the coralline alga Phyinati)Ulhon repandum. Intact plants induced rapid attachment, metamorphosis and shell groHlh. whereas extracts in- duced rapid attachment, gradual metamorphosis and negligible shell growth. Data are from Roberts and Nicholson (1997). Review of Abalone Settlement Cues 575 BIOFILMS (INCLUDING BENTHK 1)1 ATOMS) B Biofilni Cinnpttsitiitn A biofilm is a layer of microorganisms and organic matter that develops on aquatic surfaces. Biofilms have long been used to induce larval settlement in abalone hatcheries around the world. Biofilms that develop in flowing seawater with light are often dominated by diatoms. The succession that occurs in the diatom community is important in abalone farming. The film is initially comprised largely of fast-growing, prostrate, benthic diatoms (Kawamura 1993). Over time, diatoms that grow as upright colo- nies (often referred to as 3-dimensional. overstorey or filamentous diatoms) grow above the prostrate forms (Kawamura 1995). Graz- ing by juvenile or adult gastropods removes the overstorey diatoms (Types C-H in Fig. 2), and loosely attached prostrate forms (Type A in Fig. 2), so only grazing-resistant. prostrate algae remain (Seki 1980. Suzuki et al. 1987, Kawamura 1995, Matthews and Cook 1995). The grazed (or secondary! film is often dominated by dia- toms of the genus Cocconeis often along with small prostrate macroalgae such as Ulvella lens and Myrioneina spp. (Suzuki et al. 1987, Seki 1980, 1997). Use of Biofilms in Abalone Hatcheries Some early abalone culture cued settlement with natural bio- tlims. without exerting any control o\er algal composition (Grant 1981, Chen 1989). The observation that filamentous diatoms were unsuitable for settlement (Seki 1980) led to the use of pre-grazing to select for tightly attached, prostrate diatoms (Type B in Fig. 2) and other algae (Seki 1980). Pregrazed films became the settle- ment method of choice in Japan. Some other abalone hatcheries chose to select for small, pros- trate diatoms by using filtered water to condition settlement sur- faces, or by filtering diatom slurries prior to inoculation of tanks. An ungrazed film dominated by small, prostrate diatoms (Type A in Fig. 2l was the standard settlement preparation in many coun- tries including Australia (Hone et al. 1997). Korea (Shallow Sea- farming Research In.stitute 1990), New Zealand (Tong et al. 1992), and the USA (Ebert & Houk 1984). Greater control of the diatom species in the film is introduced by isolating particular strains and growing them in monocultures (e.g., Hahn 1989a, Tong & Moss 1992, Kawamura & Kikuchi 1992). However, knowledge of the dominant diatom species still provides limited description of the characteristics of the biofilm. In addition to diatoms, a biofilm will contain many strains of bacteria, fungi and other microbes, as well as a variable amount of extra- cellular secretions, and a wide array of organic molecules. All of these components will change as the film develops. Given this complexity, it is perhaps not surprising that we still do not know which characteristics of a biofilm are responsible for settlement induction. Settlement Versus Biofilm Age and Source-Habitat Older biofilms appear to be better than younger ones for aba- lone larval settlement. The percentage of Haliotis iris larvae that settled on biofilms developed in flowing seawater increased with the age of the film (Moss & Tong 1992b). In this case maximum film age was nine days, and mean settlement was -22% after six hours. A similar pattern was found with Haliotis virginea larvae — films grown for 36 days induced more attachment (77 ± 9%) and metamorphosis (25 ± 12%) within three days than films grown for ' ~^ Figure 2. Schematic representation of eij;hl groHth forms of benthic diatom (Kawamura 1994). Type A (gliding prostrate type). Solitarv ceils with a prostrate form and swift gliding movements. Adhesive strength is very low. Type tl (adhesive prostrate type). Solitary cells with a prostrate form, slow mo\ement. and high adhesive strength. Type C (non-motile upright t.\pe). Non-motile solitary cells or simple, fan-shaped colonies standing upright with a relatively weak adhesion to the substratum. Type I) (belt-shaped colonial type). Long, bell- shaped or zigzag colonies attached strongly to the substratum by a terminal cell of the colony. Type K (mucous thread solitarv type). Solitary cells or short, belt shaped colonies attached strongly to the substratum with a mucous thread. Type F (mucous thread colonial type), .\rborescent colonies connected with mucous threads and at- tached strongly to the substratum by a single mucous thread. Type G (tube-dwelling colonial t.\pe). Colonj cells are enclosed in a filamen- tous or arborescent branched mucous tube. Cells can move within the tube, and the tube is tightly attached to the substratum. Type H (fila- mentous colonial type). Filamentous colonies attached only weakly to the substratum bv a terminal cell of the colonv. 1 . 4. or 12 days before assayed (Roberts et al. 1997b). This pattern with respect to biofilm age is reinforced by data from specific diatom strains reviewed below. The settlement inducing activity of biofilms is not limited to films from suitable abalone habitats. Biofilms grown in unsuitable habitats (rivers and estuaries) induced more attachment (site means ± s.e. of up to 83 ± 5 and 72 ± 5% respectively) than negative controls (6 ± 2%) after three to four days (Robeils et al. 1997b). Diatoms Diatom films have long been used to induce larval settlement in abalone hatcheries around the world. The species composition and other biofilm characteristics are often not controlled or recorded. Only a handful of studies have examined the settlement of abalone larvae in response to specific diatom strains. The picture they give 576 Roberts is one of variability and complexity tiiat liigliliglits the limited knowledge of settlement induction by diatoms. Some of the physi- cal features of diatoms that affect their settlement-inducing activity are known (see later) but these explain relati\ely little of the varia- tion seen among strains. It is likely that details of biofilm chem- istry will explain settlement patterns, so detailed study of biofilm chemistry will be needed to significantly advance our understand- Fealiires Making Diatom Films Favorable or Unfavorable for Settlement Kawamura and Kikuchi (1992) quantified larval attachment, metamorphosis, shell growth and survival over two weeks of ex- posure to various densities of 22 diatom strains. Data representa- tive of the major patterns of response are shown in Fig. 3. All strains induced over 70% attachment within two days of exposure, but few strains induced rapid and complete metamorphosis. Dia- toms that induced relatively rapid metamorphosis were non- colonial, prostrate forms (Types A and B in Fig. 2). but many diatoms of this growth form did not give strong metamorphosis. At high densities, diatoms with three-dimensional growth forms en- tangled larvae and prevented them from metamorphosing. At low densities these diatoms induced gradual attachment and metamor- phosis, as seen with many prostrate diatoms. An English summary of this study was presented in Kawamura (1996). 100r-»— • g 50 -^8 ■-.,. — - ® ^ '/ / y^ r. Cocconeis scutellum var. parva (4B) -1 1 1 1 1 1 1 1 1 1 1 a 50 A' ^ Achnanthes longipes (lA) 6 24 48 168 336 Time (h) Figure 3. Examples of three patterns of response observed for Haliotis discus liannai larvae settling on hentliic diutiims. Data (mean ± s.e.) show changes in attachment ( — • — I, metamorphosis ( — C' — ), shell growth ( — 0 — ) and survival ( — * — ) over 2 weeks of continuous exposure to three diatom treatments, 4B, 2A andl.A (Kawamura and Kikuchi 1992). For prostrate diatoms, there is generally a positive relationship between diatom density and abalone larval settlement. This rela- tionship is generally apparent for a given diatom strain within an experiment (Kawamura & Kikuchi 1992. Daume et al. 1999a) but does not necessarily apply across diatom species or across experi- ments. Larval attachment of Haliotis discus hamuli within six hours was low when diatom density was less than 1 x 10"' cells. cm~". and high mainly for diatom densities above 3x10"* cells. cm"~ (Kawamura and Kikuchi 1992). Given that the size of diatom cells varies considerably, percent cover may be more useful than density as a measure of diatom abundance. The abundance of a diatom film may be important in providing sufficient inducer to trigger larval attachment. As diatom films develop and age. they go through a series of growth phases (O'Meley & Daintith 1993) and many physical and chemical char- acteristics change along the way. It may be one or more of these factors that causes increased settlement rather than diatom abun- dance per se. Diatom Species Favorable and Vnfavorable for Settlement It is not yet possible to list diatom strains that are consistently good or bad for abalone larval settlement. Cocconeis spp. have gained a reputation of being good for settlement because diatoms of this genus are often dominant on the pre-grazed plates that induce settlement in Japanese abalone hatcheries (Akashige et al. 1981. Suzuki et al. 1987, Seki 1997). Testing of pure strains re- veals a more complex picture. Cocconeis scutellum is often effec- tive as a metamorphosis inducer, although metamorphosis some- times occurs gradually over several days (Kawamura & Kikuchi 1992, Roberts & Nicholson 1997). Daume et al. (1999b) reported only 10-30% metamorphosis of Haliotis laevigata larvae in re- sponse to Cocconeis scutellum. However, if Daume et al.'s much shorter observation period (24 hours) and lower diatom densities are taken into account, the results are comparable to those obtained with Haliotis discus liannai (Kawamura & Kikuchi 1992) and Haliotis virginea (Roberts & Nicholson 1997). Ohgai et al. (1991) found that settlement on Cocconeis sp. was 1 2 times higher than on Nitzschia closterium or a mixed biofilm of naturally seeded dia- toms, but only 1.5 times higher than on Navicula ramosissima. Cocconeis strains are not always good for abalone settlement induction. Induction of metamorphosis can be slow (see above) and some strains of Cocconeis have given poor settlement induc- tion (Ishida et al. 1995). Haliotis rubra larvae did not respond within 24 hours to Cocconeis scutellum or any of the four other diatom strains tested (Daume et al. 2000). Mean tnetamorphosis within 24 hours was S6% in all cases. Similarly, less than 2% of Haliotis rubra larvae metamorphosed within three days on four diatom strains (6-day-old film, diatom densities not presented), while up to 25% of larvae settled on the coralline alga Phymaloli- thon repandum (Daume et al. 1999b). In equivalent bioassays. Haliotis laevigata larvae responded strongly to the same four dia- tom strains (Daume et al. 1999b) implying a major difference in the settlement requirements (or speed of response) between these two Australian abalone species. Cylindrotheca closterium (and the physically simWar Nitzschia longissima) are often poor inducers of attachment and metamor- phosis, perhaps because of their high mobility and low attachment strength (Ohgai et al. 1991. Kav\amura & Kikuchi 1992. Roberts et al. 1997a). However. CyUndrotlicca closterium has on occasion strongly induced metamorphosis of Haliotis discus luumai larvae Review of Abalone Settlement Cues 577 (Kawannira & Takaiiii 1995) and Haliotis UwviKiitct (Dannie et al. 1999b). BACTERIA Bacteria are ubiqintoiis ni so-called ■■diatom" films, yet the role that they play in settlement induction is virtually unstudied. Rob- erts et al. (1997a) (bund that metamorphosis of Haliotis iris larvae was reduced by 607f if the inducing diatom {Nitzschia oralis) was grown in the presence of antibiotics. The proportion of larvae that attached was similar between the treatments. No work has been done on settlement induction by axenic diatoms, so it is not certain that diatoms can induce abalone settlement without bacteria present. The fact that many diatom strains are active when assayed in the presence of antibiotics does not preclude (I) bacterial pro- duction of inducers prior to bioassay; or (2) activity resulting from bacteria unaffected by the antibiotic treatment. It is practical to obtain axenic diatom cultures, but keeping them axenic during a larval assay would be far more challenging. Detection of differ- ences resulting from antibiotic treatments (e.g.. Johnson & Sutton 1994) or near-axenic treatments would suggest a role for bacteria. but negative results would be equivocal. The limited work done on abalone settlement induction by bacteria shows that bacteria can induce lar\al attachment and metamorphosis in some abalone species, but that settlement is slow. Haliotis viiginea larvae incubated with a natural bacterial assemblage, took a week to reach 50% metamorphosis, and two weeks to reach 80% metaniorphosis. Parallel assays with antibi- otics retained low bacterial densities and had less than 2% meta- morphosis over the same period (Roberts et al. 1997b). .Similarly. gradual attachment and metamorphosis has been recorded with Haliotis iris exposed to bacteria in static assays without antibiotics (R. D. Roberts unpubl. data). Morse et al. (1979b) discussed gradual and inefficient attachment and metamorphosis oi Haliotis rufescens larvae occurring in parallel with rising mortality and microbial growth. They suggested that such responses may reflect the production of GABA or the less active glutamate by bacterial metabolism. Bryan and Qian (1998) reported 20-50% attachment ot Haliotis diversicolor within 24 hours on three bacterial strains isolated from a diatom film. Assays were conducted without an- tibiotics, and metaniorphosis data were not reported. None of 33 strains of marine bacteria induced settlement in Haliotis rufescens (A. Morse's unpubl. data cited in Morse 1992. methods not pre- sented). PREGRAZED FILMS AND ABALONE MUCUS Algal biofilms grazed by juvenile or adult abalone are widely effective as inducers of abalone larval attachment and metamor- phosis (Takahashi & Koganezawa 1988. Ishidaetal. 1995. Searcy- Bernal et al. 1992b, Slattery 1992, Conroy et al. 1996, Seki 1997. Biyan & Qian 1998, Daume et al. 2000). Where comparisons have been made between grazed and ungrazed films, the grazed films have performed better for both diatom films (Slattery 1992. Searcy-Bernal et al. 1992b. Seki 1997, Bryan & Qian 1998) and Ulvella lens (Takahashi & Koganezawa 1988, Daume et al. 2000). There are varying reports on the activity of abalone trail mucus as a settlement inducer. Conspecific mucus alone was reported to induce only larval attachment in Haliotis discus luinuai (Seki et al. 1997). In contrast. Yang and Wu (1995) found that mucus induced over 90% metamorphosis and shell growth of Haliotis discus han- nai larvae withm 48 hours. In Yang and Wu"s study the mucus was laid down and assayed in the absence of antibiotics. In Haliotis iris mucus alone induced only a low percentage of attachment and inetamorphosis when assayed in the presence of antibiotics (Fig. 4). For Haliotis diversicolor adult mucus alone induced moderate levels of larval attachment within six hours of exposure, but at- tachment subsequently fell as high mortality developed (Bryan & Qian 1998). Metamorphosis data were not presented. Morse et al. (1980c) recorded no settlement of Haliotis rufescens larvae in response to juvenile conspecifics (that would ha\'e produced trails of mucus). The induction of Haliotis discus hcmnai larval attachment by trail mucus appears to be genus-specific. Larval attachment was induced by trail mucus from four Japanese abalone species but not by trail mucus from three other gastropods (Seki & Kan-no 1981b, see Seki & Taniguchi 1996 for a summary in English). The mucus left by abalone during grazing gave higher larval attachment than that left by crawling abalone, while mucus from abalone pressed or rubbed against .settlement plates did not induce larval attachment (Seki & Kan-no 1981b. Seki & Taniguchi 1996). This pattern was hypothesized to result from differences in the physical or chemical properties of mucus produced during different activities (Seki & Kan-no 1981b). The presence of remnants of grazed biofilm in the grazing mucus treatment raises possible alternatives: ( 1 ) release of chemicals by grazing damage to biofilm organisms; (2) interaction between mucus and biofilm chemicals; (3) microbial proliferation on the trail mucus providing settlement cues. The chemical basis of settlement induction by trail mucus or pre-grazed films has not been studied. Rodriguez et al. (1993) suggested that heparin-bindmg growth factors might be the sub- stance responsible for settlement induction by abalone mucus, but there is no evidence to support this hypothesis. PURE CHEMICALS The Morplwgeiietic and Regulatory Signal Transduction Pathways Before discussing pure substances that can trigger larval attach- ment and/or metamorphosis, it is necessary to briefly describe cunent understanding of the way in which inducer molecules bring about a response in larval abalone. This knowledge is based on „ 100 u 90 S 80 +, 70 £ 60 CD "o 50 - CM 53 40 ra 30 S 20 o S 10 .^^^^L_^ DDBM + mucus 1 □ Mucus only | ■ DBM only HFSW Attachment Metamorphosis Shell growth Figure 4, Synergy between dihroniomethane (DB.\I) and mucus in settlement induction of Haliotis iris larvae. One juvenile Haliotis iris (-10 mm shell length) was held overnight in each well of a new mulli- well plate (Falcon 3t)43), then removed before addition of 3 ml of 0.2 (jni-nitered .seawater, antibiotics ( 15(1 ng ml ' penicillin + streptomycin sulfate), dibromomethane (175-70(( ng ml ', no dose-response within this range) and 5(1- KM) larvae. Counts were made after 48 hours of incubation (dark, 17±0.5 C). F.SW = llltcred seawater. R, I). Roberts (unpubl, data). 578 Roberts exiensive research with Haliotis rufescens (previously reviewed by A. Morse 1991, D. Morse 1984b, 1985, 1990, 1992) and the degree to which it applies to other abalone is not known. A simplified working model describes two interacting bio- chemical signal pathways — the morphogenetic and regnlatory pathways. These pathways are well summari/.ed and illustrated by Morse (1991). The morphogenetic pathway describes how an external chemical signal is translated into a nerve impulse. The model hypothesizes that binding of a GABA-mimetic molecule at stereochemically-specific receptors activates adenyl cyclase enzyme that catalyzes the synthesis of cyclic adenosine- monophosphate (AMP) as a secondary messenger. The cyclic AMP activates a protein kinase that phosphorylates an unknown protein. This protein somehow causes opening of an anion channel in the chemosensory cell membrane allowing efflux of anions and causing depolarization of the cell. This depolarization provides an electrochemical signal that can be transmitted by the larval ner- vous system (A. Morse 1991, D. Morse 1992). This nerve impulse can then trigger larval responses such as cessation of swimming, foot exploration behavior, and metamorphic changes in gene ex- pression and morphology (Cariolou & Morse 1988, Barlow 1990. Groppe & Morse 1993, 1994, Degnan & Morse 1995, Degnan et al. 1995, 1997). The larval response to GAB A and GABA-like molecules (via the morphogenetic pathway) is subject to up- and down-regulation via a separate regulatory pathway (Trapido-Rosenthal & Morse 1985, Trapido-Rosenthal & Morse 1986a, Trapido-Rosenthal & Morse 1986b, Baxter cS: Morse 1987, Wodicka & Morse IWl). Larval sensitivity to GABA analogs can be enhanced by facilitat- ing compounds (such as exogenous lysine or certain other diamino acids), or down-regulated by prolonged larval exposure to the inducer during the pre-competent phase. Facilitating compounds are not thenisehcs acting as inducers (Baxter & Morse 1987). Cues That Artificially Trigger the Morphogenetic Pathway Larval attachment and metamorphosis can be induced by sub- stances that bind to a specific larval receptor, or by treatments that artificially trigger the larval response at some other point in the morphogenetic pathway. The latter applies to several pure chemi- cals that are thought to bring about larval attachment or metamor- phosis by affecting elements of the signal transduction pathway such as trans-membrane ion transport, or intracellular levels of Ca-^ and cyclic AMP (Morse et al. 1980a, Jensen et al. 1990, Morse 1992). An interesting example of an "artificial" cue is the potassium ion (K*). A concentration of K* -10 niM above that found in seawater triggered up to 90% of Haliotis rufescens larvae to attach within 72 hours (Baloun & Morse 1984, Yool et al. 1986). Meta- morphosis was also induced if exposure lasted at least 20 hours (Baloun & Morse 1984). Haliotis discus hannai also metamor- phosed in response to excess K* but required a higher concentra- tion (optimally 40 mM excess) (Yang & Wu 1995). Excess K* induces larval attachment, but not metamorphosis, in Haliotis iris. Haliotis virfiiiica. (Roberts & Nicholson 1997) and Haliotis diver- sicolor (Bryan & Qian 1998). This difference does not appear to result from experimental methods. The latter three species were tested with natural seawater (in contrast to the artificial seawater used with Haliotis rufescens) but KCI addition to natural seawater was effective for Haliotis discus liaiuuii (Yang & Wu 1995) and other marine invertebrates (Pearce & Scheiblina 1994). Variation of larval age, assay container volume and shape, antibiotics used, and KCI concentrations all failed to produce significant metamor- phosis in Haliotis rirf^inea (Roberts & Nicholson 1997). Excess K* induces metamorphosis in a wide range of marine invertebrates, including several gastropod species, and it is thought to act by depolarizing excitable membranes and thereby triggering the settlement response (Baloun & Morse 1984, Pearce & Scheib- ling 1994). The induction by K* of attachment without metamor- phosis in some abalone species is interesting because it suggests that there could be separate cues for attachment and metamorpho- sis, and differences in the signal transduction pathways for these cues. The inducing effect of K* is more specific than a simple excess of cations, Haliotis rufescens did not settle in response to e.xcess Na"", Ca-* or Mg"* (Baloun & Morse 1984). Low external K* or high external Ca"* inhibit induction by GABA, apparently by hy- perpolarizing the receptor cells so that the ion transport brought about by the inducer does not cause cell depolarization and a nerve impulse. These artificial cues are useful in understanding the biochemi- cal pathways of settlement induction, and can provide useful tools for industry or research, but compounds that bind to specific larval receptors are much more likely to be ecologically relevant. These are now discussed. Cues that Bind to Specific iMnal Receptors The inducer molecules purified from Lithothainnium californi- cum (see above) mimicked the activity of GABA in binding to GABA receptors purified from mammals (Trapido-Rosenthal & Morse 1986a). Conversely, GABA mimicked the activity of the inducer from Lithophylluin by binding to abalone larval receptors and triggering larval attachment and metamorphosis (Morse et al. 1979a). Several lines of evidence suggest that GABA and GABA- mimetics bind to the same receptors, and that these receptors are involved in settlement induction (Trapido-Rosenthal & Morse 1986a). The location of the receptors is unknown, but they aie probably external on the larval epithelium (Barlow 1990, Morse 1990). Label bound to receptors was shed at the time of metamor- phosis (Trapido-Rosenthal & Morse 1986a) but the receptors are unlikely to be on the larval velum (Barlow 1990). Published literature contains many contradictions regarding the efficacy and effective dose of GABA ( Akashige et al. 1 98 1 , Morse 1992, Table 3). The susceptibility of GABA to degradation by marine microbes (Morse et al. 1979b: Kaspar et al. 1991, Kaspar &. Mountfort 1995) probably explains much of the variation in results (Searcy-Bernal & Anguiano-Beltran 1998, Morse 1992) yet no study has quantified the decline in GABA concentrations dur- ing settlement experiments. The use of antibiotics in settlement bioassays should reduce bacterial interference (Morse et al. 1979b) but assays using recognized antibiotic treatments can still produce quantitatively variable results for the same species (e,g„ see results for Haliotis gigantea and Haliotis virginea in Table 3). Such varia- tion may result from variations in the quantity or composition of the microbial flora (Searcy-Bernal et al. 1992b) and particularly the abundance of antibiotic-resistant microbes that degrade GABA. Variability between batches of larvae may also contribute (Baloun & Morse 1984, Trapido-Rosenthal & Morse 1986b). Another possible source of variation in GABA experiments is the composition of assay water. The concentrations of GABA required to trigger settlement can be altered orders of magnitude Review of Abalone Settlement Cues 579 TABLE 3. Settlement response of abalone larvae to various doses ofGABA. SelllenienI Response (X hours) A Her Percent of Larvae Responding GABA Concentrations of: to Comments Abaioiif Species 1(1" M 10-^ M 10' M lU " M Zero Reference Hatiotis di\risuith>i supertexla All:ichnienl (24 hi Attachmenl (48 hi SI 77 63 95 96 94 6(1 91 58 94 10' M & 10-" M GABA leihal in 72 h. No errors. Liu et al. (1986) H. diversicolor Atlachmcnt (24 hi Shell growth (96 hi H. (tiscus Ihtnnui Shell growth (48 hi H. discus clisctis Metamorphosis (96 h) Metamorphosis (96 h) H. giganlea Metamorphosis (96 h) Metamorphosis (96 h) H. mulac "Settlement" (18 hi H. iris Attachment (48 h) Metamorphosis (96 h) H. virfiinea Attachnient (48 h) H. cormgaia H. rttjescens H. ritfescens H. rufescens H. rufescens Meianiorphosis (48 h) Attachment (48 h) Metamorphosis (48 h) Inferred selllement (24 h) Inferred settlement (48 h) Inferred settlement (96 h) Metamorphosis (24 h) Attachment (24 h) Metamorphosis (48 h) Shell Growth (48 h) Melanmrphosis (24 h) Metamorphosis (24 h) Metamorphosis (24 h) Metamorphosis (24 h) Attachment (IS h) 48 ±12 45 ±13 45 ±14 32 ±10^ 22 ± 9 0 ± U 0 ± (1 14 ± 7 0 ±0 2 ± 2 Toxu 58 ±15 3.^ ±17 59 ±7 5 + 0 83 ± in 21 ± 1 82 ± J 7±3 21 ±7 2±0 80 80 50 50 98 ±2 9±2 411 ± 1(1 0 ± 0 75 ±5 9±3 3±2 0±0 61 ±5 87 ± 1 73 ±1 4±0'' 10 ±3 1 ± 1 1 ± 1 1 ±0 0±0 0 ± 0 73 ±3 48 ± 2 85 ± 2 60 ± 3 98 ± 0.5 96 ± 2 10(l±0 96 0 0 85 96 0 40 92 0 0 88 0 83 12' 86 12' 75 ± 10 76 ±4 8±8'^ Control larvae attached. Used unspecified antibiotics. No antibiotics used."* Was significantly higher than control at 18 and 72 h. 24 hour exposure followed by 24 hours in seawatcr. No antibiotics. Used 150 (jLg/ml penicillin + streptomycin. Data from 2 experiments. Used 150 (JLg/ml penicillin + streptomycin. Data trom 2 experiments. No, eiTors, Used 33 ppm penicillin. Mean ± 959c CI. Used 150 (jLg/ml penicillin + streptomycin. Mean ± SE of 7 experiments. Used 150 |jLg/ml penicillin + streptomycin, Standard error, n = 3.^ But 2.2 X lO-"" M GABA gave 55 ± 7% attachment. Methods as in Roberts and Nicholson (1997) High settlement in controls. Survival after 7 days was y/v. Flow through system without antibiotics. No antibiotics used. No biofilm present. Convincing data but no errors/stats given. Probably used 150 p.g/ml penicillin + streptomycin (Morse et al. 1979b). No difference between antibiotics (tOO-130 (xg/ml penicillin + streptomycin) (Hb9c) and no antibiotics = 83%. No biotllm present. ^ = diatom film without GABA. Biofilm present and no antibiotics used. ** = biofilm without GABA. 10" M GABA did not improve settlement. Biofilms present. No antibiotics used. Bryan and Q'^n (1998) Yang and Wu (1995) Fukazawa et al, (in press) Fukazawa et al. (in press) Genade et al. (1988) Roberts and .Nicholson (1997) Roberts and Nicholson (1997) R. Roberts (unpubl. data) Moss (1999) Searcy-Bemal et al, (1992a) Morse et al. (19S0a) Searcy-Bemal et al. (1992a) Searcy-Bemal et al. (1992b) Slattery (1992) Boldface type indicates settlement significantly greater than negative controls (zero GABA column). Control.^ ""^ Superscript letters refer to explanatory comments. lacked any deliberate cue. except where noted. by dissolved substances acting on the regulatofy pathway (Trapido-Rosenthal & Morse 1986b). This suggests that the use of artificial seawaler might reduce variability among experiments, but caution is required in the choice o\' seawater. Larvae of Haliotis discus hannui exposed to 1 |xM GABA in "ASP-M" artificial seawater (McLachlan 1964) often initiate metamorphosis more slowly than sibling larvae in natural seawater. and always fail to develop a normal post-larval shell (T. Kawamura & H. Takami. 580 Roberts unpubl. data). Similar results have been obtained with Haliotis iris where metamorphosis induced by 2 jjiM GABA after 5 days was -8-fold lower in ASP-M artificial seawater ( 1 1 ± 3%. mean ± 959^ CD than in natural seawater (86 ± 3%) (R. D. Roberts unpubl. data), while larval attachment was unaffected. Interestingly. Hali- otis nifescens does initiate metamorphosis in two artificial seawa- ter recipes (Baloun & Morse 1984. Barlow 1990). Effects of arti- ficial seawater on the initiation o't metamorphosis may be caused by complex interactions between the ionic composition of the wa- ter and membrane depolarization of chemosensory cells (Baloun & Morse 1984). Data on abalone larval responses to GABA are summarized in Table 3 and Figure ."i. Several additional abalone species are re- ported to settle in response to GABA, but data have not been presented. These include Haliotis fulgens (Salas-Garza et al. 1994, Castro-Gal vez & Searcy-Bernal 1997), Haliotis crachcrodii. Hali- otis kuititschatkaiia, Haliotis tuherculata. Haliotis inidae. and Haliotis ndier (Morse 1984a). Most abalone species will attach within 24 hours in response to concentrations between 10"'' M and 10"^ M. but doses of 10"^ M and 10"' M GABA are toxic within a few days (Table 3). Most abalone species initiate metamorphosis at doses of between 10"'' M and 10"'' M GABA (Table 3, Fig. .S) but higher doses inhibit shell growth (10"'' M) or metamorphosis and shell growth ( 10"'' M) (e.g., Morse et al. 1980a, Searcy-Bemal & Anguiano-Beltran 1998). While high concentrations of GABA are toxic within a few days, 10-30 minute exposures to 10"' M or 10""* M GABA have been successfully used to induce metamor- phosis in Haliotis riifescens (Shilling et al. 1996, Buchal et al. 1998) and Haliotis iris (R. D. Roberts, unpubl. data). Haliotis rirginea and Haliotis diversicolor larvae attach in re- sponse to GABA, but few proceed to metamorphose within four days (Table 3. Fig. 5). For Haliotis virgiiiea. fine scale dose- responses (Fig. 5) and several variations of assay methodology (Roberts & Nicholson 1997) have failed to produce significant metamorphosis. Use of GABA as a settlement cue in commercial abalone hatch- Haliotis rufescens ■ Haliotis rufescens -H discus hannai — ♦ — Hatiotis iris - O- Haliotis virginea 4 5 6 7 GABA concentration (pM) 10 Figure 5. Metamorphosis of four abalone species in response to low doses of GABA. All assays used antibiotics and quantified metamor- phosis after 48 bours of continuous exposure. Tbe lines Joininj; points are not fitted curves. .All abalone species showed similar responses at tbe lowest doses with the exception of Hcdiotis yiriiinca. which did not metamorphose at any concentration tested. Metamorphosis and shell growth ()( Haliotis discus hannai (Morse 1992, no errors presented! was lower at 1,3 fiM than at 0.8 - 1.0 |aM, but no other data suggest such a narrow dose response to G.ABA. Data for Haliotis rufescens (Searcy- Bernal and Anguiano-Beltran 1998) are from two separate experi- ments. Methods for Haliotis iris and Haliotis lirginea )R. D, Roberts unpubl, data) were as described in Roberts and Nicholson (1997). eries was initially discouraged by ( 1 ) poor results from GABA in the absence of antibiotics (Morse et al. 1979b. Slattery 1992, Searcy-Bernal 1994); (2) the toxicity associated with prolonged exposure to high concentrations of GABA (Morse et al. 1980a, Akashige et al. 1981, Liu et al. 1986); and (3) the fact that GABA causes larvae to settle on horizontal surfaces (Morse et al. 1980c, Leighton 1989) where conditions unsuitable for post-larval sur- vival can develop (Leighton 1989). However, successful large- scale induction of metamorphosis by GABA without antibiotics and in the presence of biofilms has been demonstrated (Searcy- Benial et al. 1992b, Searcy-Bernal 1994). To help counter occa- sional poor results with GABA, Searcy- Bernal and Anguiano- Beltran (1998) showed that the GABA concentration can be doubled or tripled without compromising metamorphosis or post- lar\ al sur\ i\ al and growth of Haliotis rufescens. GABA causes anest of swimming cilia (Akashige et al. 1981, Hahn 1989c, Barlow 1990) so we expect indiscriminate settlement in the presence of GABA, but strong substrate selection in the absence of G.^BA (Morse et al. 1980c). In contrast, Yang and Wu (1995) found that Haliotis discus hannai larvae preferentially metamorphosed on certain substrates even in the presence of 1 (jiM GABA. These results may be explained by bacterial degradation of GABA in the absence of antibiotics. A number of compounds that are structurally related to GABA also induced attachment of Haliotis rufescens larvae, presumably by binding to the larva's GABA-sensitive receptor (Morse et al. 1979a. 1980a. b). Investigation of stereochemical specificity re- vealed that attachment-inducing activity decreased as the carbon chain length was varied from that of GABA. and that the activity was dependent upon the presence of both the amino and carboxyl terminal groups (or on groups confeiring similar functional activ- ity) (Morse et al. 1980b). Most GABA analogs were tested only at 1 mM, while GABA was active at 1 |jiM (Morse et al. 1980b). Five common amino acids induced larval attachment only at concen- trations three to four orders of magnitude higher than those re- quired with GABA (Morse et al. 1980a). These studies quantified larval attachment, and did not say whether the active compounds also induced metamorphosis. Haliotis virginea larvae show little metamorphosis in response to GABA, yet its attachment response to several GABA-homologs (Berkett et al. 1994) minors that of Haliotis rufescens (Morse et al. 1980b), implying similar stereo- chemical requirements for attachment induction. Other Chemicals whose Mode of Action is Unknown GABA and other neurotransmitter-like molecules have been implicated in settlement induction of larval marine invertebrates other than abalone (Morse 1985). Various neurotransmitters have been tested as abalone settlement inducers but found to lack ac- tivity (Morse et al. 1980a, Akashige et al. 1981, Berkett et al. 1994, Fukazawa et al. in press). Epinephrine and serotonin stimulated the beating of the velar cilia in Haliotis discus haiuuii (Akashige et al. 1981). Dibromomethane induces low levels of metamorphosis in Hali- otis discus hannai (T. Seki, pers. comm.) but negligible attachment or metamorphosis in Haliotis iris within two days (Fig. 4). How- ever, in both species the combination of dibromomethane and trail mucus gives rapid and complete settlement (Seki et al. 1997, Fig. 4). The biochemical basis of this synergy is not known. Mucus provides a substrate for microbial proliferation, and certain bacte- ria are capable of oxidizing dibromomethane (e.g., Goodwin et al. Review of Abalone Settlement Cues 581 1998). This oxidation would liberate bromide anions and (via methanol) carbon dioxide and water. Huliolis nifescens larvae were induced to attach by replacing 25-75% of the chloride in artificial seawater with various anions including bromide (Baloun & Morse I9S4I. but this represents a bromide concentration far in excess of the effective dibromomethane concentration (Fig. 4). Brominated compounds are involved in various chemical reception pathways (e.g.. Wassermann et al. 1979. McKinney & Richelson 1986) but there is no evidence that such compounds are formed in the presence of dibromomethane. or are active in abalone. Organic solvents (e.g. ethanol at >0.75%) can induce abalone larval settle- ment (Fenteany & Morse 1993). However, the methanol generated by oxidation of dibromomethane would be short-lived, so would not reach concentrations expected to induce settlemcnl in abalone larvae. Thyroxine (Tj) at 33-1.^2 p.M induced 85'/r metamorphosis of Haliotis nifescens larvae in 44 hours (Carpi/.o-ltuarte & Rosa- Velez 1993). The thyroid hormones Tj. 3. 3. 3'-triiodothyronine (T-,), 3, 3', 5'-trliodothyronine (rT,) and tetraiodothyroacetic acid (tetrac) induced metamorphosis of Haliotis discus discus, Haliolis gigantea and Haliotis iris at concentrations as low as 0.1-1 |jiM, while 3, 5-diiodothyronine (3. 5-T2) and diiodotyrosine (DIT) were inactive (Fukazawa et al. in press, and unpubl. data). Thyroid hormones are involved in the metamorphosis, development or me- tabolism of vertebrates and some invertebrates (e.g., Galton 1992, Bales 1997, Johnson 1997). The abalone data support the case for an ancient evolution of the regulatory functions o( thyroid hor- mones (Johnson 1997). Algal extracts are often toxic to abalone larvae and can inhibit normal larval attachment, metamorphosis or shell growth (Morse & Morse 1984a, Roberts & Nicholson 1997). In the case of the brown alga Dilophus okamurai and Haliotis discus lumnai the toxic compounds were found to be two spatane-type diterpene alcohols (Taniguchi et al. 1989, ShiraishI et al. 1990). Various environmental contaminants and temperature can also interfere with normal abalone larval settlement (Morse et al. 1979b. Stein- beck 1980. Conroy et al. 1996, Raimondi et al. 1997). Certain antibiotics have also caused positive or negative interference in abalone settlement assays (see above). DISCUSSION Do Alliuliiiuiil ami Melainoiphosis Occur Spontaneously? The normal behavior of competent abalone larvae Includes pe- riods of crawling and substrate testing (Seki & Kan-no 1981a). Thus, it is not surprising that even clean, young negative controls with antibiotics often have a small proportion of larvae attached (e.g., Roberts & Nicholson 1997). This background level of at- tachment probably occurs in the absence of any cue but is quite distinct from spontaneous metamorphosis. It is common to record a very low percentage of metamorphosis in negative controls, even in the presence of antibiotics (Table 3). Varying levels of apparently spontaneous metamorphosis do occur in aging abalone (Roberts & Lapworth 2001, Takami et al. unpubl. data. Searcy-Bernal pers. comm.). Given that microbes can Induce metamorphosis, it is difficult to be confident that any metamor- phosis occurred spontaneously unless the larval cultures are axenic. Abalone larvae become Increasingly sensitive to metamorphic cues during the week or so after first attaining competence (Morse 1984b. Barlow 1990, Moss & Tong 1992a) probably because of continued synthesis of chemical receptors by the aging larva (Deg- nan & Morse 1995). This will serve to heighten the larva's sensi- tivity to sub-optimal cues and may Increase the chance of meta- morphosis occurring without any obvious trigger. A more critical period for abalone comes later, when larvae are nearing the end of their life expectancy and are laced with the choice between spon- taneous metamorphosis or death. There is no Information on changes in settlement-related biochemical pathways In abalone larvae during this period. In some animals metamorphosis is regu- lated by internal inhibitory compounds as well as external cues (Youson 1997). If such a mechanism existed in abalone, the deg- radation of the Inhibiting compound in aging larvae would offer a possible explanation for spontaneous metamorphosis. Timing and Endpoint of the Settlement Response — Are There Multiple Cues? The timing and end point of the abalone settlement response varies greatly among cues. Attachment can be uncoupled from metamorphosis in at least several abalone species (Leighton 1972, Kawamura & KikuchI 1992, Roberts & Nicholson 1997. Bryan & QIan 1998. Moss 1999). A wide array of chemicals and organisms Induce attachment of abalone larvae within a day. but few give rapid metamorphosis. It Is possible that larval attachment and metamorphosis can be triggered by separate cues (Roberts & Nicholson 1997, Seki et al. 1997, Bryan & Qian 1998) acting via independent or interacting biochemical pathways. The slow increase in percent metamorphosis commonly seen among larvae Incubated with bacteria (Morse et al. 1979b) and many diatom sti'ains (e.g., Kawamura & KIkuchI 1992) could be explained by Infrequent encounters between larvae and sufficient quantities of inducers. The cues may be continually produced and degraded by microbes (Morse et al. 1 979b, Kaspar et al. 1 99 1 , Searcy-Bemal et al. 1992b, Kaspar & Mountfort 1995) but not present In sufficient quan- tity to induce a synchronized settlement response. An interesting aspect of induction by K* Is that metamorphosis Induction only occurred if the treatment was applied for at least 20 hours (Baloun & Morse 1984). In the case of coralline algae, or other fast-acting cues, metamoiphosis begins within several hours (e.g.. Barlow 1990, Seki 1997). Such differences (if not caused by confounding effects such as larval age) may provide clues about the biochemical pathways of settlement induction. The GABA-mlmetic receptor is strict In Its stereochemical specificity (Morse et al. 198()b). Many of the molecules that induce abalone attachment or metamorphosis probably trigger the mor- phogenetic pathway downstream of the GABA-mimetic receptor (Morse 1992). The activity of other molecules, such as dibro- momethane and thyroid hormones, cannot be readily explained by the existing model (A. Morse 1991. D. Morse 1992) and probably act at larval receptors not yet characterized. The diverse array of natural substances capable of inducing some degree of abalone larval attachment or metamorphosis suggests that either ( 1 ) there are multiple cues/receptors that can trigger settlement, or (2) the cues are widely available in marine biota. Synergistic Effects Between Cues The best described example of .synergy between abalone larval settlement cues is the interaction between the regulatory and mor- phogenetlc pathways proposed for Haliolis ntfesceus (Trapido- Rosenthal & Morse 1985. Trapido-Rosenthal & Morse 1986a, Trapldo-Rosenthal & Morse 1986b, Morse 1991). Diamino acids 582 Roberts can amplify larval sensitivity to GABA-mimetic molecules by tip to 100 fold. Other examples of synergy involve GABA plus cor- alline algal extract. GABA plus diatom film (Roberts & Nicholson 1997) and dibromomethane plus mucus (Seki et al. 1997. Fig. 4). The biochemical basis of these latter synergies is not known. Both cases involving GABA have a crude organic matrix (diatom film or crude extract) as the second cue. This matrix could yield di- amino acids that act through the regulatory pathway. Alternatively, the synergy may involve inducers/receptors not yet characterized. Settlement Cues for use in Hatcheries Unpredictable settlement rates complicate post-larval culture by giving variable post-larval densities. Low or intermediate rates of settlement are very unpredictable, so hatcheries should aim for consistently high percentage metamorphosis. Ideally metamorpho- sis would also occur very rapidly so that full water flow and full aeration can be resumed as soon as possible, without risking loss of larvae. Very few settlement cues are able to deliver on these requirements. Crustose coralline algae meet these settlement requirements but are generally regarded as unsuitable for use in hatcheries because they may harbor predators, are slow growing, must remain wet and lack established culture methods. Their potential use has not been fully evaluated, and selected species may prove valuable with further research. Very thin crusts grow relatively quickly (Mc- Shane 1996) and do not host the predatory worms that live within thicker crusts (Morse et al. 1979b. Naylor & McShane 1997). The food value of corallines for post-larval abalone is generally low (Leighton 1989. Kitting & Morse 1997), so they would need to be supplemented with diatom food. Some corallines will inhibit the growth of diatoms (A. Krsinich. pers. comm.) but corallines need only cover a small proportion of plates to trigger settlement. Con- trolled culture of certain corallines is feasible (S. Daume. pers. comm.) and growth rates are not critical if only a small proportion of the surface needs to be covered at settlement. GABA is a very convenient and simple cue. and is used in some commercial hatcheries, particularly in Mexico and the USA (Morse 1992, Searcy-Bemal 1994. Searcy-Bemal & Anguiano- Beltran 1998). Its performance appears to vary among abalone species, and between occasions (e.g.. Table 3). GABA is suscep- tible to degradation by microbes but this can be counter-acted to some extent by doubling the GABA concentration (Searcy-Bernal & Anguiano-Beltran 1998) or by settling larvae in clean tanks and adding diatom food a few days later (Searcy-Bemal et al. 1992a; Roberts et al. 2001). It may also become a non-issue if lar\al sensitivity could be affordably enhanced 10-fold by addition of diamino acids (Trapido-Rosenthal & Morse 1985), or if short ex- posures to high concentrations of GABA (e.g.. Shilling et al. 1996, Buchal et al. 1998) prove practical in hatcheries. GABA causes arrest of swimming activity, so larvae attach and metamorphose largely on non-vertical surfaces. This is impractical in many sys- tems (e.g.. Leighton 1989). including those employing stacks of vertical plates (although see Salas-Garza et al. 1994). Angled plates may catch a large proportion of larvae, but a more elegant alternative would be to utilize a combination of diatoms and GABA. A diatom film is likely to induce attachment over filmed plates or tank surfaces within a day or so. Once attachment had occurted. addition of GABA would trigger metamorphosis of at- tached larvae, provided the GABA had time to act before it was degraded by microbes. Ungrazed diatom films alone provide little certainty for larval settlement in hatcheries. Few strains induce consistently high metamorphosis, and the need for a dense film at settlement can rapidly lead to excessive biofilm and adverse conditions for post- larvae. By contrast, pre-grazed films generally induce high rates of attachment and metamorphosis, and the reduction of biomass by grazing will help avoid excessive biofilm. Films grazed for several weeks are expected to be more reliable because of the selection for certain algae that are favorable for settlement. However, even young films grazed for one to three days appear to be efficient settlement inducers. GABA could be used in conjunction with diatom films, as discussed above. Differences and Similarities Among Abalone Species Comparison between species is often complicated by differ- ences in methodology (see above). A few studies have compared abalone species using equivalent assay methods, and provide com- pelling evidence of real differences in settlement responses be- tween abalone species. Two examples (reviewed above) are the response of Haliotis iris and Haliolis virginea to GABA (Table 3, Fig. 5) and the response of Haliotis rubra and Haliolis laevii;ata to several diatom strains (Daume et al. 1999b). These examples warn of differences in chemical signaling that result in fundamentally different responses between abalone species. However, in many respects the larval settlement response appears to be conserved across abalone species. Examples reviewed above inckide the widely observed positive response to crustose coralline algae and pre-grazed diatom films. Ecological Relevance of Chemical Inducers Chemical extraction procedures liberate hundreds of com- pounds from the source material without preserving the spatial or stereochemical context of the intact surface. This creates the po- tential to identify inducers that are merely artifacts of the extrac- tion and purification process (e.g.. Morse 1990). It is accordingly difficult to prove that a compound isolated from a settlement sub- stratum is ecologically relevant. The extensive work by Daniel Morse and colleagues, working with Haliolis riifesceiis, presents a compelling case for a GABA-mimetic peptide from Lithotham- niiim californicum. However, this molecule could be just one of many inducers available in the marine environment. Dibro- momethane has also recently been suggested as an ecologically relevant settlement inducer for abalone (Seki et al. 1997). but further evidence is required to demonstrate this. Most other pure chemicals that induce abalone settlement are not proposed to be ecologically important. Is Settlement Exclusive to Coralline Algae in the Wild? This review has noted the close association between juvenile abalone and crustose coralline algae, and the consistently excellent inducing activity of corallines in laboratory settlement experi- ments. However, experiments show that abalone larvae can also settle in response to many substrata that are present in abalone habitat including other macroalgae, abalone trail mucus, and bio- films (reviewed above). These experiments are artificial in that they usually confine larvae with a substrate. In the wild, larvae may be more likely to settle on first contact if they land on cor- alline algae rather than some less effective cue. This contention is supported by the rapidity with which abalone metamorphose on coralline algae compared to many other cues in laboratory assays. Settlement choice experiments in laboratory conditions would be Review of Abalone Settlement Cues 583 instructive, and have already demonstrated that larvae exhibit pref- erences for certain coralline algal species (Daume et al. 1999a). Artificial collectors conditioned with algal films or pre-grazed films then placed in abalone habitat do contain abalone post-larvae after 1-8 weeks in the sea (Keesing et al. 199.'i. Nash et al. 1993). These data suggest that abalone larvae do naturally settle on sur- faces other than corallines, but it is possible that corallines had colonized the plates during conditioning or deployment and trig- gered larval settlement. Settlement deterrents from algae (e.g.. Taniguchi et al. 1989) and sessile invertebrates may play a major role in discouraging settlement on many surfaces, although Morse and Morse (1984a) rejected the idea that abalone's preference for corallines was due to the toxicity of other algae tested. The distribution of post-larvae or juveniles on natural surfaces can be influenced by post-settlement survival or migration, so it is not necessarily a reliable indicator of settlement sites. Corallines do not provide a perfect surface for abalone post-larvae. Corallines appear to give relatively slow growth rates of post-larval abalone when compared to benthic diatoms (Kitting & Morse 1997, Takami et al. 1997, Kawamura et al. 1998), they can harbor preda- tors (Morse et al. 1979b, Shepherd & Daume 1996. Naylor & McShane 1997) and they generate moderately steep gradients in boundary layer water chemistry (Kaspar 1992). However, they may still represent a better option than other surfaces. Research Needs The ability of this re\ie\v to generalize about the effectiveness of abalone settlement cues has been hampered by differences in methodology between experiments, and often by the lack of data on metamorphosis as distinct from larval attachment. It is hoped that this review will encourage the use of appropriate terminology and methodology. Most abalone larval settlement experiments will aim to either: (1) separate treatment effects from all possible al- ternatives in tightly controlled laboratory conditions; (2) determine the usefulness of selected treatments to commercial abalone farm- ing; or (3) attempt to describe settlement patterns or processes in the sea. Small-scale experiments attempting to mimic hatchery or field conditions risk falling between two aims. Laboratory experi- ments will continue to be necessary and infonnative in research on abalone settlement cues, but there is currently a disproportionate amount of laboratory data and a corresponding need for laboratory results to be assessed in commercial abalone farming systems or natural habitats. Despite many years of commercial abalone farming around the world there are very few published data describing settlement suc- cess in full-scale commercial conditions. The rigor of laboratory studies needs to be coupled with the reality of commercial scale hatchery conditions, and quantitative, microscopic observations made over time. This will create challenges in terms of controlling experimental conditions and obtaining adequate experimental de- signs and sub-samples, but when achieved will yield very valuable information. Further research on the hatchery application of cor- alline algae, trail mucus and commercially available chemicals such as GABA would be valuable. A chemical inducer that could be bound to plastic surfaces would be ideal for the many hatcheries that use vertical plates. Hatchery experiments should aim to assess cost-effectiveness as well as biological success. Issues such as practicality and predictability will influence the cost-benefit equa- tion. The cheinistry of abalone settlement induction remains very poorly known, with the exception of the GABA-mimetic system studied in Haliotis nifescens. It would be interesting to screen for proposed "natural" cues (i.e., GABA-mimetics and dibro- momethane) on various inductive substrates, to determine whether known cues explain the activity of diverse substrates, or whether there are multiple alternative pathways for settlement induction in abalone. The chemical basis of settlement induction by biofilms should be a priority, as biofilms induce settlement in a very wide range of invertebrates, and have economic importance in abalone farming and biofouling control. The mechanisms underlying the synergy between cues are academically interesting and have po- tential to aid abalone farming. The role of bacteria in settlement induction by diatom films and trail mucus is unknown. Bacteria will certainly be degraders of organic inducer molecules, but may also play a critical role as producers or modifiers of inducers. Settlement assays conducted under axenic conditions would be useful in separating microbial effects from other cues, and in determining whether abalone can metamorphose spontaneously. Field experiments offering a range of natural substrata should help clarify the apparent paradox between the wide range of po- tential inducers for abalone settlement and the tight association between juvenile abalone and crustose corallines in the wild (Mc- Shane 1995). The extent of settlement deterrence by algae and sessile invertebrates (McShane 199.3). and further examination of preferences for certain coralline species (McShane 1996). could be incoiporated in these studies. Observations from larval settlement through the post-larval period would provide direct evidence about the relative contribution of settlement and post-settlement pro- cesses in determining spatial variation in juvenile recruitment be- tween microhabitats. These studies would be u.seful in site selec- tion for larval reseeding programs. ACKNOWLEDGMENTS Special thanks to Hideki Takami and Tomohiko Kawamura for supplying literature and unpublished data, and for help with inter- preting Japanese papers. Thanks to Henry Kaspar for many useful discussions, and comments on the manuscript. Thanks to Shirley Plant and two anonymous referees for critical review of the manu- script. Sabine Daume, Anton Krsinich, Daniel Jackson, Regina Counihan. Bemie Degnan, Ricardo Searcy-Bernal. Scoresby Shep- herd, and Tetsuo Seki all contributed by supplying data. Thanks to Tomohiko Kawamura for supplying Figures 2 and 3. This re\ lew was supported by funding from contracts CAW 801 and CAW X0004 with the New Zealand Foundation for Research Science and Technology. LITERATURE CITED Akashige, S., T. Seki. H. Kan-No & T. Nomura. 1981. Effects of -y-ami- nobutyric acid and certain neurotransmitters on the settlement and iht- metaiiiorphosis of the larvae of Haliotis discus hamuli I no (Gas- tropoda). Bull. Tohoku Reg. Fish. Res. Lab. 43:37-t5. Baloun. A. J. & D. E. Morse. 1984. Ionic control of settlement and meta- morphosis in larval Haliotis nifescens (Gastropoda). Biol. Bull. 167: 124-1.38. Barlow. L. A. 1990. Electrophysiological and behavioural responses of 584 Roberts larvae of the red alialonc {Hciliaiis riifescens) to settlement-Inducing substances. Bidl. Mar. Sci. 46:537-554. Ba.xter, G. & D. E. Morse. 1987. G protein and diacylglycerol regulate metamorphosis of planktonic molluscan larvae. Proc. Null. Acciil. Sci. 84:1867-1870. Berkett, N. M.. R. D. Roberts. D. O. Mountfort & H. F. Kaspar. 1994. Detection of gamma-aminobutyric acid and related substances using an enzyme linked immunosorbant assay and abalone settleinent bioassay. Cawthron Report No. 551. Nelson. New Zealand: Cawthron Institute. 22 pp. Bryan. P. J. & P. Y. Qian. 1998. Induction of larval attachment and meta- morphosis in the abalone Haliotis cliversicolor (Reeve). J. E.\p. Mar. Bial. Eiol. 223:39-51. Buchal, M., J. Levin & C. Langdon. 1998. Dulse Palinaria mollis as a settlement substrate and food for the red abalone Halioiis riifescens. Aquaciiltitre 165:243-260. Cariolou, M. A. & D. E. Morse. 1988. Purification and characterisation of conchiolin shell peptides from the marine mollusc. Haliotis riifescens. as a function of development. ./. Camp. Biacheni. Physiol. B 157:717- 729. Carpizo-ltuarte. E. & J. D. Rosa-Vele/.. 1993. L-thyro.xine induces meta- morphosis in two species of marine gastropods. Anicr. Zoal. 33(51:41 A (Abstract only). Castro-Galvez. R. & R. Searcy-Bemal. 1997. Settlement of Haliotis ful- geiis with GABA in a Mexican hatchery. Monterey: Abstracts of the 3"' International Abalone Symposium. 26-31 October. 1997. p.l3. Chen, H.-C. 1989. Farming the small abalone. Haliotis iliversicalor siiper- le.xta. in Taiwan.. In: K. O. Hahn. editor. Handbook of culture of abalone and other inarine gastropods. Boca Raton. FL: CRC Press, pp. 265-283. Conroy. P. T.. J. W. Hunt & B. S. Anderson. 1996. Validation of a short- term toxicity test endpoint by comparison with longer-term effects on larval red abalone Haliotis riifescens. Environ. To.xicol. Clieni. 15: 1245-1250. Daume, S., S. Brand-Gardner & W. J. Woelkerling. 1999a. Settlement of abalone larvae (Haliotis laevigata Donovan) in response to non- geniculate coralline red algae (Corallinales. Rhodophyta). / &/). Mar. Biol. Ecol. 234: 125-143. Daume, S., S. Brand-Gardner & W. J. Woelkerling. I999h. Preferential settlement of abalone larvae: diatom films \ s non-geniculate coralline red algae. Aciiiaciiltiire 174: 243-254. Daume, S., A. Krsinich, S. Farrell & M. Gervis. 20(10. Settlement and early growth of Haliotis rubra in response to different algal species. ./. Appl. Phycol. 12:479^88. Day, L. & G. M. Branch. 2000. Relationships between recruits of the abalone Haliotis miclae. encrusting corallines and the sea urchin Parechiniis angiilosus. S. Afr. J. Mar. Sci. 22:145-156. Degnan. B. M. & D. E. Morse. 1995. Developmental and morphogenetic gene regulation in Haliotis riifescens larvae at metamorphosis. Ainer. Zool. 35:391-398. Degnan. B. M.. J. C. Groppe & D. E. Morse. 1995. Chymotrypsin mRNA expression in digestive gland amoebocytes: cell specification occurs before metamorphosis and gut morphogenesis in the gastropod, Hali- otis riifescens. Roiix's Arch. Dev. Biol. 205:97-101. Degnan, B. M., S. M. Degnan & D. E. Morse. 1997. Muscle-specific regulation of tropomyosin gene expression and myofibrillogenesis dif- fers among muscle systems examined at metamorphosis of the gastro- pod Haliotis riifescens. Dev. Genes Evol. 206:464-471. Eales. J. G. 1997. Iodine metabolism and thyroid-related functions in or- ganisms lacking thyroid follicles: are thyroid hormones also vitamins? Proc. Sac. Exp. Biol. Med. 214:302-317. Ebert, E. E. & J. L. Houk. 1984. Elements and innovations in the cultiva- tion of red abalone Haliotis riifescens. Aciiiaciilture 39:375-392. Fenteany, G. & D. E. Mor.se. 1993. Specific inhibitors of protein synthesis do not block RNA synthesis or settlement in larvae of a marine gas- tropod mollusk (Haliotis riifescens). Biol. Bull. 184:6-14. Fukazawa, H.. H. Hirai. H. Hori. R. D. Roberts. H. Nukaya. H. Ishida & K. Tsuji. Induction of abalone larval metamorphosis by thyroid hormones. Fisheries Sci. (in press). Galton, V. A. 1992. The role of thyroid hormone in amphibian metamor- phosis. Trends Endcrinolot. Metab 3:96-100. Genade. A. B., A. L. Hirst & C. J. Smit. 1988. Observations on the spawn- ing, development and rearing of the South African abalone, Haliotis midae Linn. 5. Afr. J. Mar. Sci. 6:3-12. Goodwin. K. D.. J. K. Schaefer & R. S. Oremland. 1998. Bacterial oxida- tion of dibromomethane and methyl bromide in natural waters and enrichment cultures. App. Environ. Microbiol. 64:4629—1636. Grant, J. F. 1981. Abalone culture in Japan: development and current commercial practice. Tas. Fish. Res. 23:2-17. Groppe, J. C. & D. E. Morse. 1993. Molluscan chymotrypsin-like protease: structure, localisation, and substrate specificity. .Arch. Biochem. Bio- phys. 305:159-169. Groppe, J. C. & D. E. Morse. 1994. Sequence-independent detection of gene family homologs: identification of a transcript encoding a mol- luscan serine protease homologous to the pancreatic enzymes of ver- tebrates. Comp. Biochem. Physiol 110B:75-82. Hahn, K. O. 1989a. Culture of Haliotis tuberculata at the Argenton Ex- perimental Station, France. In: K. O. Hahn, editor. Handbook of Cul- ture of Abalone and Other Marine Gastropods. Boca Raton, FL: CRC Press, pp. 285-294. Hahn, K. O. 1989b. Larval development of abalone. In: K.O. Hahn, editor. Handbook of Culture of Abalone and Other Marine Gastropods. Boca Raton. FL: CRC Press, pp. 71-99. Hahn. K. O. 1989c. Induction of settlement in competent abalone larvae. In: K. O. Hahn, editor. Handbook of Culture of Abalone and Other Marine Gastropods. Boca Raton, FL: CRC Press, pp. 101-1 12. Hone, P. W., S. M. Madigan & A. E. Fleming. 1997. Abalone Hatchery Manual for .Australia. Adelaide South Australian Research and Devel- opment Institute. 34 p. Ishida. T.. T. Akutsu & K. Torisawa. 1995. Effects of monocultured benthic diatoms on the metamorphosis of veliger larvae and on the growth of juveniles of abalone. Haliotis gigantca. Bull. Shi:.iioka Pref Fish. E.\p. Stn. 30: 17-21. Jensen, R. A., D. E. Morse, R. L. Petty & N. Hooker. 1990. Artificial induction of larval metamorphosis by free fatty acids. Mar. Ecol. Prog. Ser 67:55-71. Johansen, H. W. 1981. Coralline Algae: a First Synthesis. Boca Raton, FL: CRC Press. 239 p. Johnson. C. R. & D. C. Sutton. 1994. Bacteria on the surface ofcrusto.se coralline algae induce metamorphosis of the crown-of-thorns starfish Acainhaster planci. Mar. Biol. 120:305-310. Johnson. L. C. 1997. Thyroxine's evolutionary roots. Perspecl. Biol. Med. 4:529-535. Kaspar, H. F. 1992. Oxygen conditions on surfaces of coralline red algae. Mar. Ecol. Prog. Ser 81:97-100. Kaspar, H. F. & D. O. Mountfort. 1995. Microbial production and degra- dation of 7-aminobutyric acid (GABA) in the abalone larval settlement habitat. FEMS Microbiol. Ecol. 17:205-212. Kaspar, H. F., D. O. Mountfort & V. Pybus. 1991. Degradation of gamma aminobutyric acid (GABA) by marine microorganisms. FEMS Micro- biol. Ecol. 85:313-318. Kawamura, T. 1994. Taxonomy and ecology of marine benthic diatoms (in Japanese). Marine Fouling 10:7-25. Kawamura, T. 1995. Mechanisms of variations in the benthic diatom com- munities (in Japanese). Kaiyoh 27:591-596. Kawamura, T. 1996. The role of benthic diatoms in the early life stages of the Japanese abalone (Haliotis discus hannai). In: Y. Watanabe, Y. Yamashita & Y. Oozeki, editors. Survival strategies in eariy life stages of marine resources. Rotterdam: A. A. Balkema. pp. 355-367. Kawamura, T. & S. Kikuchi. 1992. Effects of benthic diatoms on settle- ment and metamorphosis of abalone larvae. Suisan;oshoku 40:403- 409. Kawamura. T. & H. Takami. 1995. Analysis of feeding and growth rate of Review of Abalone Settlement Cues 585 newly metamorphosed abalone Halioris discus hiiniini led on four spe- cies of benlhic diatom. Fisheries Sci 6l:357-3.'iiS. Kawamura, T., R. D. Roberts & H. Takami. 1998. A review of the feeding and growth of post-larval abalone. J. Shellfish Res 17:615-625. Keesing, J. K.. R. Grove-Jones & P. Tagg. 1995. Measuring settlement intensity of abalone: results of a pilot study. Mm: Frcshwiiler Res. 46:339-543. Kitting, C. L. & D. E. Morse. 1997. Feeding effects of post-larval red abalone Huliolis riifescens (Mollusca: Gastropoda) on encrusting cor- alline algae. Moll. Res 18:183-196. Leighton, D. L. 1972. Laboratory observations on the early growth of the abalone, Halimis sorenseni. and the effect of temperature on larval development and settling success. Fi.sh Bull. 70:1 137-1 145. Leighton, D. L. 1989, Abalone (Genus Haliotis) Manculluie on the North American Pacific Coast. Fish. Bull ^l-.b'i'i-lOl. Liu, L., L. Chu, K. Chang & F. Chia. 1986. Negative effect of g- aminobutyric acid (GABA) on the settlement of larvae of the small abalone Haliolis cliversicolor superlexta Lischke. Hull. Iiisi. Ziml. Aea- demiii Sinica 25:1-5. Matthews, I. & P. A. Cook. 1995. Diatom diet of abalone post-larvae {Hiiliinis midae) and the effect of pre-grazing the diatom overstorey. Mar. Freshwaler Res. 46:545-548. McKinney, M. & E. Richelson. 1986. Blockade of N-lE-1 15 murine neu- roblastoma muscarinic receptor function by agents that affect the me- tabolism of arachidonic-acid. Biocheni. Pharm. 35:2389-2398 (Ab- stract). McLachlan, J. 1964. Some considerations of the growth of marine algae in artificial media. Can. J. Microbiol l(.):769-782. McShane, P. E. 1992. Early life history of abalone: a review. In: S. A. Shepherd. M, J. Tegner & S. A. Guzman del Pnio, editors. Abalone of the worid: biology, fisheries and culture. Oxford: Blackwell. pp. 12f)-138. McShane, P. E. 1995. Recruitment variation in abalone: its importance to fisheries management. Mar. Freshwater Res. 46:555-570. McShane. P. E. 1996, Recruitment processes in abalone (Haliotis spp,). In: Y. Watanabe. Y. Yamashita & Y. Oozeki, editors. Survival strategies m early life stages of marine resources. Rotterdam: A. A. Balkema. pp. 315-324. McShane, P. E. & M. G. Smith. 1988. Measuring abundance of juvenile abalone Haliolis ruber Leach (Gastropoda: Haliotidae) — comparison of a novel method with two other methods. Ausr. ,/. Mar. Freshwater Res. 39:331-336. Morse. A. N. C. 1991. How do planktonic larvae know where to settle? Amer Sci. 79:154-167. Morse, A. N. C. & D. E, Morse, 1984a, Recruitment and metamorphosis of Haliotis larvae induced by molecules uniquely available at the surfaces of crustose red algae, J. Exp. Mar. Biol. Ecol. 75:191-215, Morse. A, N, C. & D. E, Morse. 1984b. GABA-mimetic molecules from Porphyra (Rhodophyta) induce metamorphosis of Haliotis (Gas- tropoda) larvae. Hydrobiologia 116:155-158. Morse, A. N, C. C, A. Froyd & D. E. Morse. 1984. Molecules from cyanobacteria and red algae that induce larval settlement and metamor- phosis in the mollusc Haliotis rufescens. Mar Biol. 81:293-298. Morse, D. E. 1984a, Biochemical and genetic engineering for improved production of abalones and other valuable molluscs. Aquaculture 39: 263-282. Morse. D, E. 1984b. Biochemical Control of Larval Recruitment and Ma- rine Fouling. In: J. D. Costlow & R. C. Tipper, editors. Marine Bio- deterioration. Annapolis. MD: Naval Institute Press, pp. 13-1-140. Morse, D. E. 1985. Neurolransmitter-mimetic inducers of larval settlement and metamorphosis. Bull. Mar. Sci. 37:697-706. Morse. D. E. 1990. Recent progress in larval settlement and metamorpho- sis: closing the gaps between molecular biology and ecology. Bidl. Mar. Sci. 46:465-483, Morse. D, E. 1992. Molecular mechanisms controlling metamorphosis and recruitment in abalone larvae. In: S.A, Shepherd. M,J. Tegner & S.A. Guzman del Proo (eds.). Abalone of the world: biology, fisheries and culture, Oxford: Blackwell, pp. 107-1 19. Morse, D. E., N, Hooker. H. Duncan & L. Jensen. 1979a. g-aminobutyric acid, a neurotransmitter, induces planktonic abalone larvae to settle and begin metamorphosis. Science 204:407^10. Morse. D, E., N, Hooker. L, Jensen & H, Duncan. 1979b. Induction of larval abalone settling and metamorphosis by g-aminobutyric acid and its congeners from crustose red algae: II: Applications to cultivation seed-production and bioassays. principle causes of mortality and inter- ference. Proceedings of the World Mariculture Sociery 10:81-91. Morse, D. E., H. Duncan, N, Hooker. A, Baloun & G, Young, 1980a, GABA induces behavioral and developmental metamorphosis in plank- tonic molluscan larvae, Couipar. Pharm. Aquatic Species 39:3237-3241. Morse, D. E., N. Hooker & H, Duncan. 1980b. GABA induces metamorphosis in Haliotis. V: Stereochemical specificity. Brain Res. 5:381-387. Morse, D. E., M. Tegner, H, Duncan. N. Hooker, G. Trevelyan & A. Cameron, 1980c, Induction of settling and metamorphosis of plank- tonic molluscan (Haliotis) larvae: Part III): Signaling by metabolites of intact algae is dependent on contact. In: D, Muller-Schwarz & R. M. Silverstein, editors. Chemical Signaling In Vertebrate And Aquatic Animals. New York: Plenum Press, pp. 67-86. Moss, G, A. 1999. Factors affecting settlement and early post-settlement survival of the New Zealand abalone Haliotis australis. NZ J. Mar. Freshwater Res. 33:271-278. Moss, G. A. & L. J. Tong. 1992a. Effect of stage of larval development on the settlement of the abalone Haliotis iris. NZ J. Mar. Fresh. Res. 26:69-73. Moss, G, A, & L, J. Tong, 1992b, Techniques for enhancing larval settle- ment of the abalone Haliotis iris, on artificial surfaces, NZ.J. Mar. Freswater Res. 26:75-79. Nash, W, J„ J, C. Sanderson, J. Bndlcy, S. Dickson & B. Hislop. 1995. Post-larval recruitment of hiacklip abalone (Haliolis rubra) on artificial collectors in southern Tasmania. Mar. Freshwater Res. 46:531-538. Naylor. J. R. & P. E. McShane. 1997. Predation by polychaete worms on larval and post-settlement abalone Haliotis iris (Mollusca: Gas- tropoda). / E.xp. Mar. Biol. Ecol. 214: 283-290. Ohgai, M., M, Wakano & S. Nagai, 1991, Effect of attached microalgae on the settlement of larvae and growth of juveniles in abalone, Haliolis discus hamuli Ino. Suisanzoshoku (H3):263-266. OMeley. C, & M. Daintith. 1993. Algal cultures for marine hatcheries. Aquaculture sourcebook. Launceston, Australia: University of Tasmania, Pearce, C. M. & R. E. Scheibling. 1994. Induction of metamorphosis of larval echinoids (Sirongylocentroius droebachiensis and Echinarach- nius parma) by potassium chloride (KCI). Invert. Reprod. Develop. 26:213-220. Raimondi, P. T., A, M, Barnett & P. R, Krause, 1997, The effects of drilling muds on marine invertebrate larvae and adults. Environ. Toxi- col. Chein. 16:1218-1228. Roberts, R, D. & C, M. Nicholson. 1997. Variable response from abalone larvae (Haliolis iris. H. virginea) to a range of settlement cues. Moll. Res 18:131-141. Roberts, R, D. & C, Lapworth. 2001. Effect of delayed metamoiphosis on larval competence, and post-larval survival and growth, in the abalone Haliotis iris Gmelin. ,/. E.xp. Mar Biol. Ecol. 258:1-13. Roberts. R.. H. Kaspar. D. Mountfort & N. Berkelt. 1994. Induction of settlement and metamorphosis in paua (Haliotis spp.) by algae and algal extracts. Abstracts of the New Zealand M;irine Sciences Society Annual Conference. 31 Aug. - 3 Sep. 1994. Hamilton, New Zealand, pp. 61. Roberts, R. D„ T, Kawamura & C. M, Nicholson, 1997a, Settlement of abalone larvae, Haliotis iris, in response to benthic diatom films. Monterey: Abstracts of the 3"' International Abalone Symposium. 26- 31 Oct. 1997. pp. 55. Roberts, R, D„ C. M, Nicholson & T. Kawamura. 1997b. Abalone larval settlement on coralline algae and marine biofilms, Monterey: Abstracts of the 3rd International Abalone Symposium, 26-31 Oct, 1997. pp, 54, Roberts, R. D., R, Searcy-Bernal & P. A. Cook. 1998. A workshop on the culture of larval and post-larval abalone. Third International Abalone Symposium, Oct. 26-31. Monterey. California, USA, Cawthron Report No, 477, Nelson. New Zealand: Cawthron Institute. 5 pp. 586 Roberts Roberts. R. D.. C. J. Lapworth & R. Barker. 2001. Effecl of starvation on the growth and survival of post-larval ahalone iHiilioiis iris). Aqiiacul- nire. 200:323-338. Rodriguez. S. R.. F. P. Ojeda & N. C. Inestrosa. 1993. Settlement of benthic marine invertebrates. Mar. Ecol. Prog. Ser, 97:193-207. Salas-Garza. A.. E. Vazquez-Moreno. E. Carpizo-Ituarte & L. S. McAnally-Salas. 1994. Htiliofis fidgens larvae induced to metamor- phose with GABA and other inducers on vinyl plates. Hobart. Austra- lia: Abstracts of the 2nd International ."Xbalone Symposium. 7-1 1 Feb- ruary 1994. Searcy-Bemal, R. 1994. Current use of gamma-aminobutyric acid (GABAl in massive-.scale abalone seed production. Abstracts of the 2nd Inter- national Abalone Symposium. 7-1 1 February 1994. Hobart. Australia. Searcy-Bernal, R. 1999. Settlement and post-larval ecology of the red abalone Hcilioiis nifescens in culture systems. PhD Thesis. Davis. CA: University of California. Davis, and San Diego State University. Searcy-Bernal. R. & C. Anguiano-Beltran. 1998. Optimising the concen- tration of gamma-aminobutyric acid (GABA) for inducing metamor- phosis in the red abalone Haliotis nifcsccns (Mollusca: Gastropoda). J. World Aquacult. Sac. 29:463^70. Searcy-Bernal. R.. A. E. Salas-Garza & R. A. Flores-Aguilar. 1992a. Re- search in Mexico on the critical stage of abalone {Halioiis spp.) seed production. In: S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the World: Biology, Fisheries and Culture. Oxford: Blackwell. pp. 547-560. Searcy-Bernal. R.. A. E. Salas-Garza. R. A. Flores-Aguilar & P. R. Hino- josa-Rivera. 1992b. Simultaneous comparison of methods for settle- ment and metamorphosis induction in the red abalone {Halioiis nife- scens). Acjuaculture 105:241-250. Seki. T. 1980. An advanced biological engineering system for abalone seed production. In: International Symposium on Coastal Pacific Marine Life. Bellingham: Western Washington University, pp. 45-54. Seki, T. 1997. Biological studies on the seed production of the northern Japanese abalone. Halioiis discus luiiwai Ino. (in Japanese with English abstract). Bull. Tohoku Nail. Fish. Res. In.st. 59:1-71. Seki, T. & H. Kan-no. 1977. Synchronized control of eariy life in the abalone. Halioiis discus lunmai Ino. Haliotidae. Gastropoda (in Japanese with English abstract). Bull. Tohoku Reg. Fish. Res. Lab 38:143-153. Seki. T. & H. Kan-no. 1981a. Observations on the settlement and meta- morphosis of the veliger of the Japanese abalone. Halioiis discus han- nai Ino, Haliotidae. Gastropoda. Bull. Tohoku Reg. Fish. Res. Lah 42:31-39. Seki. T. & H. Kan-No. 1981b. Induced settlement of the Japanese abalone Halioiis discus hannai veliger by the mucous trails of the juvenile and adult abalones. Bull. Tohoku Reg. Fish. Res. Lah. 43:29-36. Seki, T. & K. Taniguchi. 1996. Factors critical to the survival of herbivo- rous animals during settlement and metamorphosis. In: Y. Watanabe, Y. Yamashita & Y. Oozeki. editors. Survival strategies in early life stages of marine resources. Rotterdam: A. A. Balkema. pp. 341-354. Seki. T.. K. Taniguchi & K. Kurata. 1997. The metamorphosis inducing role of dibromomethane on the Japanese abalone, Halioiis discus han- nai Ino. Monterey: Abstacts of the Third International Abalone Sym- posium. 26-31 Oct. 1997. 60 pp. Shallow Seafarming Research Institute in Kosong. 1990. Tiaining manual on artificial breeding of abalone {.Halioiis discus hannai) in Korea DPR. Kosong: Shallow Seafarming Research Institute. 107 pp. Shepherd, S. A. & S. Daume. 1996. Ecology and suivival of juvenile abalone in a crustose coralline habitat in South Australia. In: Y. Wa- tanabe, Y. Yama.shita & Y. Oozeki (eds.). Survival strategies in early life stages of marine resources. Rotterdam: A.A. Balkema. pp. 297-313. Shepherd. S. A. & J. A. Turner. 1985. Studies on southern Australian abalone (genus Halioiis). VI. Habitat preference and abundance and predators of juveniles. J. E.xp. Mar. Biol. Ecol. 93:285-298. Shilling. F. M.. O. Hoegh-Guldberg & D. T. Manahan. 1996. Sources of energy for increased metabolic demand during metamorphosis of the abalone Halioiis rufescens (Mollusca). Biol. Bull. 191:402—112. Shiraishi. K.. K. Tamgucht. K. Kurata & M. Su/uki. 1990. Effects of methanol extracts from the brown alga Dilophus okamurui on avoid- ance behavior of young abalone {Halioiis discus hannai). Bull. Tohoku Nat. Fish. Res. Insl. 52:13-15. Slattery, M. 1992. Larval settlement and juvenile survival in the red aba- lone (Halioiis rufescens): an examination of inductive cues and sub- strate selection. Aquaculture 102:143-153. Steinbeck. J. R. 1980. Temperature effects on gamma amino butyric-acid induced settlement in Haliotis rufescens (Swainson). Anwr. Zool. 20: 955 (Abstract only). Stevenson. J. & A. Melville. 1999. Settlement and recruitment of the abalone Haliotis cyclohaics Peron. 1816. Mar. Freshwater Res. 50: 229-234. Suzuki. H., T. loriya. T. Seki & Y. Aruga. 1987. Changes of algal com- munity on the plastic plates used for rearing the abalone Haliotis discus hannai. Nippon Suisan Gakkaishi 53:2163-2167. Takahashi, K. & A. Koganezawa. 1988. Mass cultui'e of Ulvclla lens as a feed fo|- abalone. Haliotis discus hannai. NOAA Tech. Rep. NMFS 70:29-36. Takami. H.. T. Kawamura & Y. Yamashita. 1997. Contribution of diatoms as food sources for post-larval abalone Haliotis discus hannai on a crustose coralline alga. Moll. Res. 18:143-151. Takami. H.. T. Kawamura & Y. Yamashita. 2000. Starvation tolerance of newly nietamoi"phosed abalone Halioiis discus hannai. Fisheries Sci. 66:1180-1182. Taniguchi. K., K. Shiraishi, K. Kurata & M. Suzuki. 1989. Inhibitory effects of the settlement and metamorphosis of the abalone Halioiis discus hannai veligers by the methanol extracts from the brown alga Dilophus okamurui. Nippon Suisan Gakkaishi 55:1 133-1 137. Taniguchi. K.. K. Kurata. T. Maruzoi & M. Suzuki. 1994. Dibro- momethane, a chemical inducer of larval settlement and metamorphosis of the sea urchin Strongylocenlrotus nudus. Fish. Sci. 60:795-796. Tong. L. J. & G. A. Moss. 1992. The New Zealand culture system for abalone. In: S.A. Shepherd, M.J. Tegner & S.A. Guzman del Proo, editors. Abalone of the world: Biology. Fisheries and culture. Oxford: Blackwell. pp. 58.V59I. Tong, L. J., G. A. Moss, P. Redfearn & J. Illingworth. 1992. A manual of techniques for culturing paua. Haliotis iris, through to the early juve- nile stage. New Zealand Fisheries Tech. Report No. 31. 21 pp. Trapido-Rosenthal. H. G. & D. E. Morse. 1985. L-a.u-Diamino acids facilitate G.ABA induction of larval metamorphosis in a gastropod mol- lusc {Haliotis rufescens). J. Comp. Physiol. 155:403^14. Trapido-Rosenthal. H. G. & D. E. Morse. 1986a. Availability of chemosen- sory receptors is down-regulated by habituation of larvae to a morpho- genic signal. Proc. Natl. Acad. Sci. 83:7658-7662. Trapido-Rosenthal. H. G. & D. E. Morse. 1986b. Regulation of receptor- mediated settlement and metamorphosis in larvae of a gastropod mol- lusc {Halioiis rufescens). Bull. Mar. Sci. 39:383-392. Wassermann. N. H., E. Bartels & B. F. Erianger. 1979. Conformational properties of the acetyl choline receptor as revealed by studies with constrained de-polarizing ligands. Proc. Natl. Acad. Sci. 76:256-259. Wodicka, L. M. & D. E. Morse. 1991. cDNA sequences reveal mRNAs for two Ga signal transducing proteins from larval cilia. Biol. Bull 180: 3 1 8-327. Woelkeriing. Wm. J. L. M. Irvine &. A. S. Harvey. 1993. Growth-forms in non-geniculate coralline red algae (Corallinales, Rhodophyta). Ausi. Sy.tt. Bot. 6: 277-293. Yang. Y. & B. Wu. 1995. Induction of larval settlement and metamorpho- sis of Haliotis discus hannai Ino (Gastropoda. Mollusca). Chinese J. Ocean. Limn. 13:71-77. Yool. A. J., S. M. Grau. M. G. Hadfield. R. A. Jensen. D. A. Markell & D. E. Morse. 1986. Excess potassium induces larval metamorphosis in four marine invertebrate species. Biol. Bull. 170:255-266. Youson, J. H. 1997. Is lamprey metamorphosis regulated by thyroid hor- mones?/Imfr. Zool. 37:441—160. Joiinuil (>t Shellfish Rcwarch. Vol. 20, No. 2, 5S7-S91, 2(101. EFFECT OF BIOFILM DENSITY ON GRAZING AND GROWTH RATES OF HALIOTIS FULGENS POSTLARVAE RICARDOSEARCY-BERNAL,* LUIS A. VELEZ-ESPINO, AND CASANDRA ANGUIANO-BELTRAN Instituto lie Investigaciones Occaiioloiiicas. Apurtadd Postal 453, Euscuada 22H60, Baja Culiforuki. Mexico ABSTRACT Grazing rates oi Haliolis fuli>cns postlarvae of different ages (7. 15, ,10, 45 and AO days), feeding on tlie cultured diatom Naviciila iiiccrla. were estimated in 10 ml sterile plastic dishes previously inoculated with different densities of the diatom ( 100—1,000 cells/mnr). Postlarvae (3-7 per dish) were allowed to graze for two to three hours and video recordings were taken to estimate postlarval size and grazing rates by digital image analysis, Seawater was changed every other day and postlarvae were measured again after six to eight days to estimate growth. Grazing and growth rates of postlarvae older than 15 days increa.sed linearly with biofilm density. The highest grazing rales for seven and 60 day-old abalone were 79 and 10,999 cells/postlarva/hour, respectively. The most important increase in grazing activity occurred between ages 45 and 60 days when postlarvae reached 1.5-2.0 mm and started the formation of the first respiratory pore. Implications for the management of production systems are discussed. KEY WORDS: abalone postlarvae, Haltnus fiilgen.s. grazing rates, biofilms INTRODUCTION 111 culture systems, abalone [Huliotis \pp.) postlarvae feed on biofilms dominated by benlhic diatoms atid bacteria (Hahn 1989. Kawamura 1996). A key problem in this production stage is how to keep an adequate balance between postlarval feeding and dia- tom supply. If grazing is excessive, starving conditions may se- verely impact postlarval survival and growth. In the opposite situ- ation, overdeveloped biofilms may also create deleterious condi- tions leading to massive losses of postlarvae (Ebert & Houk 1984, Hahn 1989. Searcy-Bernal et al. 1992. Tong & Moss 1992. Searcy- Bernal 1996). Most studies on abalone postlarval feeding, have focused on the effect of different diatoms on postlarval growth, on the factors affecting their food value (Kawamura et al. 1998a. Roberts et al. 1999). and on changes in diatom selectivity and digestion as post- larvae grow (Norman-Boudreau et al, 1986. Matthews & Cook 1995. Seki & Taniguchi 1996. Kawamura et al. 1998b. Takami et al. 1998. Roberts et al. 1999). The quantitative assessment of postlarval grazing is crucial to understand these complex culture systems and to suggest better management options. However, despite its itnporlance. this issue has received little attention. Marti'nez-Ponce and Searcy-Bernal ( 1998) provided the first estimates of abalone postlarval grazing, by measuring cleared areas by H. nifescens postlarvae fed known densities of the benthic diatom Navicula incerta. using the digital analysis of video-recorded images. Roberts et al. ( 1999) estimated consumption rates of W. iris postlarvae fed different diatom strains, by determining fecal production. Both studies concluded that graz- ing increases rapidly as postlarvae grow, after the first one to two weeks after settlement. Body size is one of the main factors de- termining grazing rates of juvenile abalone (Marsden & Wil- liams 1996), Preliminary data suggest that, besides postlarval size and dia- tom species, the density of biofdm may also affect postlarval graz- ing and growth rates. This article explores this relationship for Haliotis fulgens (Philippi 18.54) fed the diatom N. incerta. Accord- ing to general concepts on the response of consumers to food *Corresponding author. E-mail: rsearcy@faro.ens.uabc.mx. abundance (Valiela 199.5). grazing rates were expected to increase with diatom density until saturation values were reached (type II functional response), MATERIALS AND METHODS Veliger larvae were produced at the hatchery of Cooperativa Emancipacion. S.C.L. (Puerto Nuevo. B,C,S., Mexico) and trans- ported to the Instituto de Investigaciones Oceanologicas (I.I.O.) (Ensenada. B.C.). where they were reared in l-|j.m filtered, UV treated seawater with chloramphenicol (lOtng/l) added, changed every day. Competent larvae were induced to metamorphose in 2()L plastic containers with 1.5 jxM gamma-aminobutyric acid (Searcy-Bernal and Anguiano-Beltran, 1998). The cultured diatom Navicula incerta (-200 cells/mm") was added at day 2 after settle- ment. This benthic diatom (-13 ixm length) was isolated from local coastal waters and is currently used in research and commercial operations in Mexico, Po.stlarvae were cultured in these containers with seawater (treated as described) changed every two days for the first two weeks, and with flowing 5-p,m filtered seawater (-300 nil/min) thereafter. Constant fluorescent illumination (-50ji,E/m"/s) and gentle aeration were provided. Temperature fluctuated be- tween 17 and 19°C. Postlarvae were collected from the culttire systetn at days 7, 1 5. 30, 45 and 60 after settlement and introduced to containers ( 10-ml wells of sterile culture dishes. 3-7 postlarvae per well), previously inoculated with N. incerta at target densities of 100. 500. 750. 1000. 2000. and 4000 cells/mm" (with the exception of trials with seven-day-old postlarvae where the higher density was substituted by 1.500 cells/mtii"^). Before the preparation of experimental den- sities, batch cultures of this diatom were iinmersed in an ultra- sound bath for 3-5 min and filtered through a 20 p.m sieve to disperse cells, A randomized block design with three replicates per diatom density was originally considered for each postlarval age (with the exception of age 45 days where only two replicates were used). Two to three hours after the introduction of postlarvae to the experimental units, video-images were recorded (Sony SSC-C374 high resolution camera) on an inverted microscope (Meiji Techno) and digitally analyzed (NIH Image 1.59, PowerPC Macintosh computer) to measure postlarval shell length, and to determine 587 588 Searcy-Bernal et al. cleared areas and actual diatom densities. Grazing rates (cells/ postlarva/h) were estimated by multiplying cleared areas by dia- tom densities. Previous experience had shown that postlarvae start feeding activity almost immediately after their introduction in ves- sels (i.e. in less than 2 min) and that N. inceila growth and re- colonization of cleared areas are negligible during this 2-3h period (Marti'nez-Ponce & Searcy-Bernal 1998). Experimental containers were maintained for six days (19 ± 1°C, 4 (j.E/m-/s) with seawater (treated as described) changed every 48 hours (with the exception of 7-day-old postlarvae). Post- larval shell length (SL) was measured at the end of this period to determine growth rates (fjim/d). Because of the high variability in diatom density among repli- cates (despite the efforts to minimize it), data (average growth and grazing rates) were analyzed considering wells individually. To estimate the maximum grazing and growth rates for each postlar- val age, the three highest values were averaged. RESULTS Figure I shows the shell length of H. fiilgens postlarvae col- lected from the culture system at the different ages tested. In the first 15 days growth was slower (14.9 )xni/d) than in the period from days 15 to 60 (33.4 |j.ni/d). Mortality of postlarvae of all ages in wells was negligible at all diatom densities during the experi- mental period (<5'7f ). Grazing rates of seven-day-old postlarvae averaged 54.5 cells/ postlarva/li and were independent from diatom density (r = 0.3 19. P > 0.05) (Fig. 2). However, at age 15 days grazing rate increased dramatically, averaging 362 cells/pl/h and showing a positive and significant linear association with diatom density (r = 0.773. P < 0.01) (Fig. 3a) which was not significant for postlarval growth rates (r = 0.345, P > 0.05) (Fig. 3b). At postlarval ages of 30, 45 and 60 days, the dependence of grazing and growth rates on diatom density was stronger (Figs. 4. 5, 6 respectively). Within the range of densities tested, most of the.se relationships were adequately described by linear regres- sions, although for 30-day-old postlarvae an asymptotic trend for grazing at higher diatom densities was observed (Fig. 4a). A loga- rithmic tit to these data resulted in an r = 0.829 which is hiaher ^ 120 J ^ ♦ Q. 100 -- 0 80 -- Q) 60-- ♦ ro ra 40- 1 20- 0 ♦« 1— • 1 •— I — • 1 1 0 500 1000 1500 2000 2500 diatom density (cells/mm^) Fijjure 2. Grazing rates of seven-day-old postlarvae of H. fiilgens at different densities of the diatom Saricula iiuerta. Data are means for 5-7 postlarvae in experimental units. than the \alue for the linear model (r = 0.655). In addition, at this age. linear correlation between growth rate and diatom density was not significant because of an outlier (slow growth at the highest density). 700 -r jr 600 - - L). U) 500 -- (D O^ 400 -- 0) ro 300 -- O) 200 -- N P 100 -- O) (a) y = 0.0719x + 241.41 r = 0.7728, P< 0.01 ■+- ■+■ ^ 1000 2000 3000 4000 5000 diatom density (cells/mm^) 2.5", 0 10 20 30 40 50 60 age (d) Figure I. Shell length (mml of H fiilgens posllavae collected from the culture system at different ages (days after settlement) Data are aver- ages of 36-90 postIar>ae. 30 X 5 25- E 3 20- - t_ ♦ rate CJ1 ♦ - ♦ ♦ growth en o - y = 0.0008x-t- 19.716 r = 0.3446, P> 0.05 (b) + + -+- 0 1000 2000 3000 4000 5000 diatom density (cells/mm^) Figure 3. Grazing (a) and growth (b) rates of 15-day-old postlarvae of H. fiilgens at different densities of the diatom Saviciila incerta. Data are means for 5-7 postlarvae in experimental units. Abalone Postlarval Grazing and Growth Rates 589 ro i— a> c N (D i_ CD 6000 J 5000 - 4000 3000 -- 2000 -- 1000 -I-/ 0 y = 962.47Ln(x)- 3892.1 r = 0.8290, P< 0 01 (a) y = 0.9132x+ 1415.3 r = 0.6548, P< 0.01 1000 2000 3000 diatom density (cells/mm ) 4000 0) ro c N ro 8000 J 7000 -- 6000 -- 5000 -- 4000 -- 3000 -- 2000 -- 1000 -- 0 y= 1.658X- 75.982 r = 0.8974, P < 0.01 1000 2000 3000 4000 diatom density (cells/mm^) 60 -r 50-- -n b 40 -- u 30 -- m 20-^ o O) 10-^ ♦ ♦^ (b) y = 0.0063x+ 19.766 r= 0.3739, P> 0.05 1000 2000 3000 4000 diatom density (cells/mm ) Figure 4. (grazing (a) and growth (bl rates of .V)-dav-old postlarvae of H. fulgen,s at ditTcrent densities of the diatom Xaviciila incerta. Data are means of 3-5 postlarvae in experimental units. The slopes of linear regressions of grazing rates on diatom density increased with postlarval age, but this trend was not ob- served for postlarval growth rates (Figs. 2-6). Estimates of niaxinuim grazing and growth rates are presented in Table 1. Ma.ximum grazing rates ranged between 79 and 10,999 cells/pl/h for postlarval ages of seven days and 60 days, respec- tively, and increased non-linearly with postlarval shell length, fol- lowing a trend consistent with a power function (Fig. 7). Maxi- mum growth rates did not follow a similar pattern, reaching values of 25.6. 47.6. 37.4 and 39.6 p.m/d for ages at 15, 30, 45 and 60 days, respectively. DISCUSSION Effects of Biofilm Density on Postlarval Grazing Rates This study shows that grazing rates of H. fulgens postlarvae increase abruptly and become strongly dependent on diatom den- sity the first one to two weeks after settlement. Grazing by earlier postlarvae (seven days) was minimal and independent of food abundance, in agreement with observations on postlarvae of this and other abalone species (Matthews & Cook 1995, Kitting & Morse 1997, Kawamura et al. 1998b. Marti'nez-Ponce & Searcy- Bemal 1998, Roberts et al. 1999). Only for 30-day-old postlarvae, results are consistent with a (b) 0.0071X-H2.645 0.7435, P< 0.01 H 1000 2000 3000 4000 diatom density (cells/mm ) Figure 5. Grazing (al and growtli (Ix rales of 45-dav-old postlarvae of H. fulgens at different densities of the diatom Xaviciila incerta. Data are means of 3-5 postlarvae in experimental units. type II functional response of grazing rates to diatom density (Fig. 4a). Grazing rates level off at about 3.500 cells/pl/h, corresponding to an approximate diatom density of 2,000 cells/mm". This type of functional response is common in marine consumers (Valiela 1995) and has been recently documented for juveniles (3-4 cm SL) of the abalone H. asinina feeding on macroalgae (Tahil & Juinio- Menez 1999) and for the periwinkle Liltorina lillorea feeding on biofilms (Sommer 1999). In a previous study with one-month-old H. nifescens postlarvae an asymptotic trend in grazing rates was also detected, leveling off at around 1 .000 cells/pl/h when N. incerta reached densities of ca. 1.000 cells/mm" (Searcy-Bernal & Cook, in prep.). Differences with H. fulgens might be explained by the smaller size of H. nifescens postlarvae (-750 |jiM SL). the lower temperature (~I6°C) or might reflect actual inter-specific differences. In high diatom densities (~>2.000 cells/mm") many of the H. nifescens postlarvae crawled to the walls of wells almost immediately after their introduction into the bottom of containers. This behavior was never observed in H. fulgens regardless of the postlarval age or diatom density tested, which suggests that this species is able to cope with denser biofilms. At the other ages tested — except at age seven days — grazing rates by H. fulgens postlarvae increased linearly with biofilm den- 590 Searcy-Bernal et al. Q. o^ a> to O) c N i- O) 14000 T 12000 10000 -- 8000 -- 6000 -- 4000 -- 2000 -- y = 2.3467X + 284.02 r = 0.8930, P< 0.01 (a) ^ 50 45 T3 40 b 35 0) 30 m ?5 u. x: 20 "5 15 Q O) 10 5 0 1000 2000 3000 4000 5000 diatom density (ceils/mm^) y = 0.0043x+ 15.044 r = 0.5923, P< 0.05 (b) 1000 2000 3000 4000 diatom density (cells/mm^) 5000 Figure 6. Grazing lal and growth (b) rates of 60-day-old postlarvae of H. fiilgeiis at different densities of tlie diatom Navicula incerla. Data are means for 3-5 postlarvae in experimental units. sity and no evidence of leveliiig-off was observed. For older post- larvae (43- and 60-day-old). this pattern may indicate that densities tested were not high enough to produce an asymptotic trend in grazing activity. However, grazing rates of 15-day-old postlarvae would have been expected to level off at a diatom density lower than the con'espondent for 30-day-old abalones. This result may be partially influenced by the low grazing activity by 15-day-old post- TABLE 1. Maximum grazing and growth rales i<( H. fiilgens postlarvae of different ages. Data are means of the three highest values and standard errors are shown in parenthesis. Length Max. Grazing Rate Max. Growth Rate Age(d) (Mm) (cells/pl/h) (pm/d) 7 363.4 78.9 (2.1) (15.4) 15 473.-S 553.5 25.6 (S.O) (35.6) (1.5) 30 9fi3. 1 3,903.1 47.6 (16.1) (752.9) (3.1) 45 1,563.7 4.984.5 37.4 (37.6) (2,646.7) (3.9) 60 2,000.7 10,998.7 39.6 (22.91 (2,110.9) (3.3) i^uuu - * r 12000 - 10000 - - 8 8000 - - 6000 - _ c M 4000 - _ 5 2000- 0 - - .25491 y = 0.00005 X" r= 0.9547, P< 0.01 500 1000 1500 postlarval length (microns) 2000 Figure 7. Relationship hetween maximum grazing rate and length in postlarvae of H. fiilgeiis. larvae, which increased dramatically (-10-fold) at age 30 days old (Fig. 3a and Fig. 4a). Effects of Biofilm Density on Postlarval Growth Rates In general, postlarval growth rates also increased linearly with diatom density, following similar trends than grazing rates (Figs. 3-6). This may reflect the obvious relationship between food con- sumption and growth, despite the fact that grazing rates correspond to the first two to three hours of the experiment and growth rates to the total six-day period. Since diatom density at the end of that period was usually lower than the initial (although probably pro- portional), the quantitative relationships between growth and bio- fihn density can only be considered as a first approximation. It should be noted, however, that moderate differences of dia- tom densities may produce dramatic differences in growth. In trials with 30. 45, and 60 days postlarvae. growth rates after 6 days at 2,000 cells/mm" were more than 509f higher than those at 1.0(30 cells/mm" (Figs. 4b-6b). This short-term effect, in conjunction with patchiness of postlarvae and biofilm densities (and composi- tion) in culture systems, may create local areas of differential growth, influencing the size heterogeneity commonly observed since early culture stages (Hahn 1989). This initial size variability would be further maintained and even increased by intra-specific competition. Therefore, mechanisms underlying growth heteroge- neity in abalone postlarvae, may not be only genetic or related to metamorphosis induction as commonly suggested (Hahn 1989. Morse 1992). but may also relate to early feeding history. The effect of diatom density on abalone postlarval growth, should also be considered for research purposes. For instance, a study may conclude that postlarvae grow better in diatom A than in diatom B. only because diatom A was provided at a more adequate density. Grazing and (irtiHtli Rales of Postlarvae as a Function of Shell Length (SL) According to the power function fitted, (Fig. 7) maximum graz- ing rates were proportional to SL" ^"^. which indicates a closer relationship to postlarval volume or biomass. Since, as discussed above, the maximum grazing rates calculated here probably un- derestimate asymptotic values for postlarval ages of 45 and 60 Abalone Postlarval Grazing and Growth Rates 591 days the actual coefficient of that equation was probably higher. Roberts et al. ( 1999) found that fecal production (as an indicator of ingestion rate) of H. iris postlarvae was proportional to SL"^'^, which is consistent with our results. Estimates of maximum growth rates were not proportional to postlarval age or length. The lowest value (25.6 ixm/d), corre- sponded to 15-day postlarvae and the highest (47.6 jxm/d) to 30- day postlarvae, reflecting an increased effect of diatom grazing. Again, potential growth rates for 45-day and 60-day postlarvae were probably underestimated as a consequence of the range of diatom densities tested. The low variability in growth rate of post- larvae during development, is consistent with the nearly linear increase in SL after the first one to two weeks of abalone postlarval culture, reported here (Fig. I ) and in other studies (Searcy-Bernal et al. 1988. Flores-Aguilar 1989, Kawamura et al. 1998b, Mar- tinez-Ponce & Searcy-Bernal 1998). ACKNOWLEDGMENTS The authors are grateful to Cooperativa Emancipacion S.C.L. (Puerto Nuevo, B.C.S,. Mexico) for the donation of H. fulgens larvae and to Enrique Valenzuela, head of the I.I.O. Microalgae Laboratory, for the supply of the diatom Naviaila incerta. This study was partially funded by the Mexican government (CONA- CyT grants 4023P-B; SNl grant 5532) and the University of Baja California (grant 4040). LITERATURE CITED Ebert, E. E. & J. L. Houk. 1984. Elements and innovations in the cultiva- tion of the red abalone (Haliolis nifescens). Aijiunitlnire 39:375-392. Flores-Aguilar, R. A. 1989. Estudio sobre la inducclon del asentamiento y nielaniorfosib de larvas de dos especles de abuk'm i/iuliolis rufcsccns y Haliolis ciirnii>citu). Licenciatura thesis. Facultad de Ciencias Marinas, Ensenada B.C.: Universidad Autiinoma de Baja California, pp. -'i7 Hahn, K. O. 1989. Handbook of culture of abalone and other marine gastropods. Boca Raton, FL: CRC Press, pp. 349. Kawamura, T. 1996. The role of benthic diatoms in early life stages in the Japanese abalone Haliolis discus bannai. In: Y. Watanabe, Y. Ya- masliita. & Y. Oozeki, editors. Survival strategies in early life stages of marine resources. Rotterdam: A. A. Balkema. pp. 355-367. Kawamura, T., R. D.Roberts & C. M. Nicholson. 1998a. Factors atTeciing the food value of diatom strains for postlarval ah;ilone Halimis iris. Aquaculture 160:81-88. Kawamura. T., R. D. Roberts & H. Takami. 1998h. A review of the feeding and growth of post-larval abalone. / Shellfish Res. 17:615-625. Kitting, C. & D. E. Morse. 1997. Feeding effects of postlarval red abalone, Haliolis nifescens (Mollusca:Gastropoda) on encrusting coralline al- gae. Moll. Res. 18:183-196. Marsden, I. D. & P. M. J. Williams. 1996. Factors affecting the grazing rale of the New Zealand abalone Haliotis iris Martyn. J. Shellfish Res. lS:4ni-+()6. Martinez-Ponce. D. R. & R. Searcy-Bernal. 1998. Grazing rates of red abalone { Haliolis nifescens) postlarvae feeding on the benthic diatom NavicuUi incerta. J. Shellfish Res. 17:627-630. Matthews, I. & P. A. Cook. 1995. Diatom diet of abalone post-larvae (Haliolis midae ) and the effect of pre-grazing the diatom overstorey. Mar. Freshwater Res. 46:545-548. Morse, D. E. 1992. Molecular mechanisms controlling metamorphosis and recruitment in abalone larvae. In: S. A. Shepherd, M. J. Tegner. & S. A. Guzman del Proo, editors. Abalone of the World. Oxford: Fishing News Books, pp. 107-119. Norman-Boudreau, K., D. Bums, C. A. Cooke & A. Austin. 198(i. A simple technique for detection of feeding in newly metamorphosed larvae. Aquaculture 51:313-317. Roberts, R. D., T. Kawamura & C. M. Nicholson. 1999. Growth and survival of po.stlarval abalone (Haliolis iris) in relation to development and diatom diet. / Shellfish Res. 18:243-250, Searcy-Bernal, R. 1996. Boundary layers and abalone postlar\al culture: Preliminary studies. /A(//(((CH/f»)f 140: 129-137. Searcy-Bernal, R. & C. Anguiano-Beltran. 1998. Optimizing the concen- tration of gamma-aminobutyric acid (GABA) for inducing larval meta- morphosis in the red abalone Haliotis rufescens (Mollusca: Gas- tropoda). J. World Aquaculture Soc. 29:463^70. Searcy-Bernal, R., A. E. Salas-Garza & R. A. Flores-Aguilar. 1988. Cre- cimiento de postlarvas y juveniles de abulon azul (Haliotis fulgens) en un laboratorio mexicano. Ciencias Marinas 14:57-72. Searcy-Bernal, R., A. E. Salas-Garza, R. A. Flores-Aguilar & P. R. Hino- josa-Rivera. 1992. Simultaneous comparison of settlement and meta- morphosis induction ui the icil abalone iHiilmiis rufescens). Aquacul- ture 105:241-250. Searcy-Bernal, R. & P. A. Cook, in prep. Grazing rates of Haliolis rufe- scens postlarvae in different diatom densities. Seki. T. & K. Taniguchi. 1996. Factors critical to the survival of herbivo- rous animals during settlement and metamorphosis. In: Y. Watanabe. Y. Yamashita & Y. Oozeki, editors. Survival Strategies in Early Life Stages of Marine Resources. Rotterdam: A. A. Balkema. pp. 341-354. Sommer, U. 1999. The susceptibility of benthic microalgae to periwinkle (Littorina litlorea. Gastropoda) grazing in laboratory experiments. Aquat Bol 63:11-21. Tahil, A. S. & M. A. Juinio-Menez. 1999. Natural diet, feeding periodicity and functional response to food density of the abalone. Haliolis as- inina., (Gastropoda). Aquacult Res 30:95-107. Takami, H., T. Kawamura & Y. Yamashita. 1998. Development of polysaccharide degradation activity in postlarval abalone Haliolis dis- cus Imnnai. J. Shellfish Res. 17:723-727. Tong. L. J. & G. A. Moss. 1992. The New Zealand culture system for abalone. In: S.A. Shepherd, M.J. Tegner, & S.A. Guzman del Proo, editors. Abalone of the World. Oxford: Fishing News Books, pp. 583- 591. Valiela, I. 1995. Marine Ecological Processes. Berlin: Springer- Verlag. 686 pp. Jdumal oj Shellfish Rcsauxh. Vol. 20. No. 2. 5y.V6()l. 2(101. LARVAL DEVELOPMENT IN HALIOTIS ASININA LINNAEUS SAOWAPA SAWATPEERA.' * E. SUCHART UPATHAM," ' MALEEYA KRUATRACHUE," YAOWALAK P. CHITRAMVONG," PICHAI SONCHAENG,' TANATE PUMTHONG/ AND JINTANA NUGRANAD^ ' Insiinitc (if Marine Science. Bnraplui Universiry. Chonbuii 20131. Thailand: ~ Department of Biology, Faculty of Science. Maliidol University. Baniikok 10400. Thailand: 'Faculty of Science. Bitrapha University. Choithuri 20131 , Thailand: Coastal AqHaculture Development Center. Prachiiap Kliiri Khan 77000. Thailand ABSTRACT Larval development of HalioUs usiniiui Linnaeus was observed under water temperatures of 25°C. 28°C. 31°C, and 34'C from fertilization to the formation of the fourth tubules on the cephalic tentacles. The larvae had 42 stages of development. The time-period of larval development depended on the water temperature: it was 65, 49, 41 and 4 1 h at water temperatures of 25°C. 28''C, 3 1 °C, and 34^C, respectively. After settlement, the development of post-larvae through the formation of the first respiratory pore was ob.served under room temperatures (28°C-35°C). The velum was shed and the mantle began to secrete a new shell. Mouth, radula and digestive organs were developed by the third day of settlement. The heart was seen on the fourth day. The prominent structure on the roof of the mantle cavity was the ctenidium. The second pair of epipodial tentacles began to form and the eyestalks were completely developed on the eighth day after settlement. From day ^ to day 24, the post-larvae increased in shell size and number of epipodia and tubules on the cephalic tentacles. The ctenidium became more developed. The first respiratory pore began to form on day 24 to 30. KEY WORDS: ahalone. larsal development. Hulidtis usiiiiiui INTRODUCTION The life cycle of abakine is composed of larva, post-larva, juvenile and adult stages. The larval stage begins with fertilization and ends with the formation of the fourth tubule on the cephalic tentacles (Hahn 1989a). Then, the larvae begin to settle and un- dergo metamorphosis. The deposition of the peristomal shell marks the transition from the larval to the post-larval stage (Hahn 1989a), The post-larval period continues until the formation of the first respiratory pore (notch stage). From the notch stage until sexual maturity, the abalone is called a juvenile (Leighton 1974). Larval development is a gradual process that does not occur in discrete stepwise stages. However, various stages can be recog- nized during larval and post-larval development (Hahn 1989a), In Haliotis discus hannai (Ino 1952), there are 41 distinct larval stages with recognizable external features, from fertilization until initiation of metamorphosis (Seki & Kan-no 1977). These mor- phological features and the order of their appearance in larval developmental stages provide guidelines for studies of larval de- velopment in other abalone species. After the larvae settle, I I recognizable external features appear following the appearance of the cilia in the mantle cavity. These are confirmed to be the char- acteristics occurring before metamorphosis (Seki 1997). The larval development rate is measured by the time required for larvae to exhibit features distinctive to each stage (Hahn 1989a). Water temperature is the major factor that affects the larval development rate (Leighton 1974, Ebert & Houk 1984, Owen et al. 1984, Hahn 1989c). Abalone seed production in Thailand has re- mained in a primitive and small-scale stage for a long period. This is partly because of the difficulty in planning spawning, and the high mortality in the early developmental stage (Seki 1997). Stud- ies of developmental morphology, behavior and the biological re- quirements of abalone larvae during settlement and metamorphosis can be applied to improve the technology for the abalone seed production systems. An understanding and knowledge of larval development can thus lead to improved techniques and manage- *Correspondmg author. E-mail: saowapaCs bims.buu.ac.th ment in abalone hatchery systems. The objective of the present study was to study the larval development of Haliotis iisiiiimi Linnaeus. MATERIALS AND METHODS Ctiiiititioniiig of Hroodstocks The broodstocks of abalone H. asiniiui (81.4 cm in shell length, 132,6 g in weight) were naturally conditioned in 10 tons concrete race-way tanks at the Coastal Aquaculture Development Center, Prachuap Khiri Khan Province, Thailand. The water temperature was 30-33°C. A flow-through system with a flow rate of 2-3 1/min was employed. The broodstocks were fed macroalgae such as GracUaria fisheri. G. tenuistipitata and Acantliophora spp. Every month, the abalone were examined for gonad maturation, which was determined by size and color according to Ebert and Houk ( 1984). When gonad maturation reached stage 2-3 (Ebert & Houk 1984), the abalone were transfeiTed to the spawning room for induced spawning. Induced Spawning Spawning of H. asinina was induced during April to August, 1998, by using UV-filtered seawater. Males and females were separated in 500 L rectangular flow-through system tanks of fil- tered seawater with a flow rate of 2-2,5 L/min. The spawning tanks were illuminated by a 40-watt cool white fluorescent lamp (1000 lux). Each tank contained 50-60 abalone. The spawning room was set up with the following photoperiod: light (6:00 pm to 6:00 AM) and dark (6:00 am to 6:00 pm). The abalone were reared under this reversed photoperiod for .seven to ten days before spawning occurred. At 1 1 .00 am, the tanks were cleaned and the UV-filtered seawater was put in under a static condition. Released eggs and sperms were collected separately in a 5 L container. The fertilization period was 30 minutes after spawning. The water tem- perature during the fertilization period was 30-33°C. Observation of Larval Developmental Stage The observation of larval development of H. asiiiiini. from fertilization to initiation of settlement, followed 41 stages of larval 593 594 Sawatpeera et al. development described by Seki and Kan-no (1977) and Seki ( 1997). All experiments were performed with pooled larvae in 5-1 containers until the larvae reached the last stage of larval devel- opment. Immediately after fertilization, the eggs were placed in the experimental container. Each container contained 5 eggs/ml. These were then placed under four temperature conditions; 25°C and 28°C (in controlled room temperature). 31"C and 34°C (in water baths) and at ambient temperatures of 2S-35"C which served as controls. Each experiment was performed in triplicate. The observation of larval development, carried out under a transmission compound microscope, started immediately after fer- tilization and was continued every five minutes until the embryos reached stage 10 or gastrula stage. After the gastrula stage. lar\al development was observed every 30 minutes until the embryos reached stage 4 1 , which w as the last stage of larval development. The observation of larval development continued once a day after settlement until the abalone produced the first respiratory pore. This was done in a 250 L settlement and hatchery tank containing diatom plates (Nitzschia sp.). The developmental stages after settlement through to the beginning of the juvenile stage followed those of Seki and Kan-no (1977) and Seki (1997). Effects of Temperature on Ijinal Development Biological Zero Point The biological zero point is the critical water temperature at which the abalone larvae can develop completely until settletnent. It was calculated from the development rate of some developmen- tal stages (Seki & Kan-no 1977), such as hatch-out, formation of the larval retractor muscle. 90-degree torsion, formation of the epipodial tentacle and formation of the fourth tubule on the ce- phalic tentacles. Effective Accumulative Temperature The effective temperature is the difference between the water temperature and biological zero point and the quantity above the biological zero point that has an additive effect toward lar\al de- velopment. The summation of effective temperature value during larval development is called the effective accumulative tempera- ture (EAT) which was calculated according to Seki and Kan-no (1977). RESULTS Spawning According to the experiment, the reversal of photoperiod af- fected the spawning time of H. asiiuna. That is, they spawned during the day (12:00 am-1:30 pm) instead of at night (12:00 PM-1:30 am). The males spawned between 12:15 am and 1:30 pm and the females spawned between 12:30 am and 1:30 pm. iMrval Size The eggs of H. asinina were dark green in color. The pigments were concentrated at the vegetal pole, while the animal pole was transparent yellowish. The average size of unfertilized eggs was 88.0 ± 8.0 [jLm in width, 102.4 ± 3.6 p.m in length and the nuclear size of eggs was 70.4 ± 6.7 |xm. After fertilization, the egg size increased: their width, length and nuclear size were 126.4 ± 17.1. 148.2 ± 10.4 and 99.5 ± 16.3 (xm, respectively (Table 1 ). Egg size did not change until hatch-out stage. The width and length of H. TABLE I. The size of H. asinina larvae (fim), from fertilization to Tormation of the first respiratory pore. Size (fim ± S.E.) Stage Width Length Unfertilized eggs 88.0 ±S. 102.4 ±3,6 Fertilized eggs to hatch out 126.4+ 17. 148.2 ± 104 Trochophore larvae I0.s.3±4. 1.^0.7 ±4.8 Completion of velum 108 ±3. 135.3 + 2.9 Larval retractor mu.scle 110.7 ±4. 139.5 ±4.2 Completion of larval shell 174.5 ±6. 242.0 ± 64 Formation of operculum 175.5 + 6. 239.8 ± 5.9 Fourth tubule on cephalic 178.3 ±5. 245.7 ± 4.6 tentacle First dav of settlement 178.8 ±5. 247.7 ±2.5 Beginning of first respiratory 11424 ±94. 15-^0.2 ± 140.7 pore formation 30 asinina trochophore larvae were 105.3 ± 4.1 and 130.7 ± 4.8 (xm. respectively (Table I ). After the larvae developed to completion of the larval shell, the size of the larval shell did not change until secretion of the juvenile shell. The first respiratory pore started building after they reached the size of 1142.4 ± 94.6 (width) x 1530.2 ± 140.7 |j.m (length) (Table 1). Larval Development The larvae of H. asinina had 42 stages of development (Table 2, Fig.l, Fig. 2). Shortly after fertilization (stage 1 ), the first polar body was discharged (stage 2) followed quickly by discharge of the second polar body (stage 3). Cleavage began after discharge of the polar bodies and development progressed to the gastrula (stages 4-1 1 ). Cilia grew along the top of the embryo forming the prototrochal girdle and apical tuft, and began beating (stage 12). The cilia caused the embryo to rotate intermittently inside the egg membrane. The stomodeum formed (stage 13) and cilia along the prototrochal girdle were completely formed (stage 14). At this stage, the embryo was classified as a trochophore larva. The larva began moving more frequently inside the egg membrane, the egg membrane became thinner, and finally burst (stage 15). The apical cilia aided the larva in bursting the egg membrane during hatch- out. The hatched lar\a immediately swam to the water surface. Soon after hatch-out. the larval shell began to be secreted at the back of the larva (stage 16). The trochophore larva continued to develop until it becaine a veliger larva. The larva was classified as veliger when the apical region became flat and the velum was completely developed with long cilia present (stage 17). The larval retractor muscle formed (stage 18), followed by formation of an integumental attachment to the larval shell (stage 19). The foot mass protruded from the top of the shell (stage 20) at completion of the larval shell (stage 21). During torsion, the cephalo-pedal mass rotated a 90-degree angle followed by the top of the mantle membrane tearing off from the top of the larval shell (stage 22). The velum and cephalo-pedal mass rotated relative to the region of the body covered by the larval shell. The region destined to becotne the mouth and the foot continued to rotate, until the cephalo-pedal mass was rotated at a 180-degree angle from its original position (stage 23). Larval Development of Abalone 595 TABLE 2. The rate of larval development (hours) of H. asinina at water temperatures of 25. 28, 31, 34"C and room temperatures (28-35 C). Larval development stage Time period of larval dei telopment (h) Sequence 25°C 28°C ire 34°C 28-35°C 1 Fertilization 0.00 0.00 0.00 0.00 0.00 2 First polar body 0.25 0.25 0.25 0.25 0.25-0.33 3 Second polar body 0.33 0.33 0.42 0.33 0.25-0.42 4 First cleavage (2 cells) 0.50 0.50 0.42 0.50 0.33-0.58 5 Second cleavage (4 cells) 1.(1(1 0.92 0.83 0.92 0.75-1.00 6 Third cleavage (8 cells) 1.33 1.25 1.00 1.17 1.00-1.17 7 Fourth cleavage (12 cells) 1.42 1.33 1.25 1.25 1.17-1.33 8 Fifth cleavage (16 cells) 1.50 1.42 1.33 1.33 1.33-1.50 9 Morula (32 cellsl 2.83 2.42 1.83 2.00 2.00-2.50 in Blastula 3.83 3.17 2.83 3.00 2.50-3.00 II Gastrula 5.50 4.00 3.50 3.50 3.00_;.00 12 Prototrochal cilia 6.00 5.00 4.17 4.00 4.50-5.00 13 Siomodeuni 7.00 5.50 4.50 4.50 5.00-5.50 14 Prototrochal girdle 8. 00 3.50 5.00 5.00 5.00-5.50 15* Hatch oul 9.50 7.50 6.00 6.00 5.50-7.00 16 Beginning of larval shell 10.50 8.00 6.50 7.00 6.00-7.00 17 Completion of velum I2.(.)0 9.50 8.00 8.00 7.50-8.00 18* Larval retractor muscle 13.50 11.00 9.00 9.00 10.00-11.00 19 Integumental attachment 14.50 11.50 10.50 10.50 10.5(5-12.00 20 Protrusion of foot mass 16.00 12.50 10.50 10.50 11.00-12.50 21 Completion of larval shell 16.50 13.50 11.00 11.00 11.00-13.00 22* Torsion (90° twisting) 18.00 14.00 11.50 11.50 12.00-14.00 23 1 80° rotation of foot mass 21.00 18.00 13.50 13.50 13.00-15.00 24 Operculum 25.00 22.50 16.00 16.00 16.00-16.50 25 Long spines on end of nietapodiuni 29.00 23.00 18.50 17.00 18.00-19.00 26 Fine cilia on foot 30.50 25.00 20.50 20.50 19.00-20.00 27 Vertical groove in \elum 35.00 27.00 21.00 21.00 21.00-22.50 28 Eye spot 40.00 28.00 21.00 21.00 21.00-23.00 29 Propodium 41.00 29.50 24.00 24.00 24.00-24.50 30 Cephalic tentacle 43.50 30.50 25.00 24.00 24.00-25.50 31 Cilia on propodium 45.00 32.00 26.50 26.00 25.00-27.00 32* First epipodial tentacle 45.50 36.00 28.00 28.00 27.00-29.00 33 Cilia in mantle cavity 48.00 36.00 30.00 29.00 31.00-32.00 34 Apophysis on propodium 5 1 .00 39.00 30.00 30.00 32.00-33.00 35 Otolith 55.00 41.00 34.00 33.00 32.00-33.00 36 Short spines on cephalic tentacle 57.00 42.00 34.00 34.00 33.00-34.00 37 Two tubules on cephalic tentacle 59.00 42.00 35.00 35.00 34.00-35.00 38 Third tubule on cephalic tentacle 60.00 43.00 37.00 36.00 36.00-37.00 39 Snout protrusion 62.00 45.00 39.00 38.50 38.00-39.00 40 Ciliary process in mantle cavity 64.00 46.00 40.00 38.50 39.00-41.00 41 Retractor muscle draws in mantle cavity 64.00 47.00 40.00 40.00 40.00-i2.00 42* Fourth tubule on cephalic tentacle 65.00 49.00 41.00 41.00 42.()0-+4.00 * the stages used for calculating the biological zero point. There were three pairs of long spines at the posterior end of the metapodium after torsion (stage 24). The operculum formed (stage 25) and at this time, the cephalo-pedal mass could be retracted into the shell. In succession, fine cilia developed on the foot sole and began beating (stage 26), a vertical groove formed in the velum (stage 27). the eye spot appeared (stage 28), the propoditim tomied (stage 29), a cephalic tentacle formed on the velum (stage 30), and cilia began growing on the propodium (stage 31). Cilia formed in the mantle cavity up to the anterior edge of the velum and the cilia began beating (stage 32). The propodium twisted to the side and an apophysis appeared on the propodium (stage 33). A pair of epi- podial tentacles formed on both sides of the foot under the oper- culum (stage 34) and crawling on a suiface with its fool was now possible at this stage. The otolith formed and was clearly visible (stage 35), short spines appeared on the cephalic tentacles (stage 36), production of the snout began from underneath the velum (stage 37), two tubules appeared on the cephalic tentacles (stage 38), a ciliary process formed on the roof of the mantle cavity (stage 39), and a third tubule formed on the cephalic tentacles (stage 40). The larval retractor muscle attached to the larval shell drew the enlarged mantle cavity toward the back of the shell (stage 41). Larval de- velopment was completed with the formation of the fourth tubule on the cephalic tentacles (stage 42). The total time of larval development depended on the water temperature in the larval tank (Table 2). This was 65 h at 25°C. 49 h at 28°C, and 41 h at 31°C and 34°C. In the control group (28-35°C), the total lime of development was 40—41 h. 596 Sawatpeera et al. Sperm t^ Egg membrane Figure 1. The development of//, asinina larvae from egg to the development of the foot mass. (I, unfertilized egg; 1, fertilized egg: 2, discharge of first polar body; 3. discharge of second polar body; 4, first cleavage (2 cellsl; 5, second cleavage (4 cellsl; 6, third cleavage (8 cells); 7, fourth cleavage (12 cells, top view); 8, fifth cleavage (16 cellsl; 9. morula; 10, blastula; ll,gastrula; 12, appearance of prototrochal girdle and cilia along the prototrochal girdle*: 13, appearance of stomodeum (rotation): 14, prototrochal girdle and cilia along the prototrochal girdle are completely formed.*; 15, hatch - out; 16, beginning of larval shell formation; 17, the apical region becomes fiat and the velum is completely developed «ith long cilia present; 18, appearance of larval retractor muscle; It, appearance of integumental attachment; 20, development of foot mass. * There are 24 cilia surrounding the velum along the prototrochal girdle. Larval Development of Abalone 597 Figure 2. The development of H. asinina larvae from the complelion of larval shell stage to the appearance of the fourth tubule on cephalic tentacles. 21. completion of larval shell; 22. cephalo-pedal mass 90 degree torsion; 23, foot muscle 180 degree torsion; 24. appearance of spine on end of metapodium; 25, operculum; 26. appearance of fine cilia on foot sole; 27, vertical groove formation in velum; 28, appear- ance of eye spot; 29, appearance of propodium; 30, appearance of cephalic tentacles; 31, appearance of cilia (m the propodium; 32, ap- pearance of cilia in mantle cavity up to the anterior edge of the velum: 33, formation of apophysis on propodium; 34, formation of first epi- podial tentacle; 35, appearance of otolith: 36, appearance of spines on cephalic tentacles; 37, protrusion of snout underneath the velum; .^8, appearance of two tubules on cephalic tentacle; 39, ciliary process forms on the roof of the mantle cavity; 40, third tubule appearance on cephalic tentacle; 41, the retractor muscle attached to the larval shell draws the enlarged mantle cavity toward the back of the shell: 42, fourth tubule appearance on cephalic tentacles. After 44 hours, or when the larvae had already formed the fourth tubule on the cephalic tentacles, they were placed in the settlement tank. On the first day of settlement, they were still in the exploration and inspection stage. They were repeatedly swimming with the velum and temporarily crawling over the surface of the diatom plate. On the second day. most larvae were shedding the velum and the mantle membrane moved to the shell edge. Some larvae had already secreted the new shell. The settled larvae had already formed the fifth and sixth tubules on the cephalic tentacles and the ciliated process was already formed near the right cephalic tentacle. At this stage, all were crawling on the plate. In the mantle cavity, the cilia on the roof of the mantle were beating to circulate the water and functioned as a ctenidium. However, some larvae still had the velum and were still in the exploration and inspection stage. On the third day of settlement, the juvenile shell increased in length. The post-larvae began to feed on diatoms that were smaller than 10 (j.m. On the fourth day, the heart began to form and was beating. On the roof of the mantle was a prominent structure that developed into the ctenidium. The juvenile shell increased in length. On day eight, the second pair of epipodia began to form and the number of tubules on the cephalic tentacles increased. The eye stalks were completely developed. From day nine to day 24 of the post-larvae stage, the number of ctenidia and epipodia in- creased. The juvenile shell increased in size. The first respiratory pore began to form on day 24 to day 30. Effects of Water Temperature on iMrral Developmeiil Biological Zero Point The biological zero point was calculated from the developmen- tal rates of hatch-out, formation of larval retractor muscle, 90 degree torsion, formation of epipodial tentacles and the fourth tubule on the cephalic tentacle stage of H. asinina at temperatures of 25°C. 28' C. 3r='C. and 34 C. In this experiment, the biological zero point of the larval development rate at 34"C was not calcu- lated because it was close to that of the development at 3I"C. The average biological zero point of H. asinina was 15.0 ± 0.3' C (Table 3). Effective Accumulative Temperature The important stages of larvae in the aspect of abalone aqua- culture management are hatch-out, completion of larval shell, for- mation of the first epipodial tentacle and the fourth tubule on the cephalic tentacles (Hahn 1989a). The effective accumulative tem- peratures at water temperatures of 25°C, 28°C. 3 1 °C. and 34°C were 95. 98, 96 and ll4"C-hour, respectively, in the hatch-out stage; 165, 176, 176 and 209°C-hours, respectively, in the comple- TABLE 3. Equation of relationship between water temperature and time inversion of beginning formation of the fifth larval developmental stage (1/time) and biological zero point of the important larval stages of H, asinina. Larval developmental Equation of Biological stage relationship Zero Point HatL-h (Hit 1/t' = 0.0102T''-O.I5Lens Philippi. beige in H. sorenseni Bartsch. olive in H. coniiiiuta Wood (Hahn 1989a). violet in H. coccinea canan'iisis Nordsieck (Harrison & Grant 1971) and green or blue-green in H. cyclobates Peron & Lesueur. H. ku'vigata Donovan. H. scahiris Leach and H. roei Gray (Shepherd & Laws 1974). In some abalone species, the color of the eggs varies in each stage of development, such as in H. rubra Leach (Shepherd & Laws 1974). The pigments in the eggs of H. asinina are concentrated at the vegetal pole, while the animal pole is transparent yellowish. This feature is in contrast with those of other abalone species, where pigmentation of the eggs is dark at the animal pole and light at the vegetal pole (Hahn 1989a). In H. coccinea canariensis. the animal pole is violet and the vegetal pole is slightly yellowish (Harrison & Grant 1971). The pigments in H. asinina also appear in the visceral mass, so that the visceral mass TABLE 4. The effective accumulative temperature (EAT) of larval development of H. asinina al water temperatures of 25, 28, 31, 34°C and room temperatures (28-35 C). l.ar\al development stage HE.\T (X - h) Sequence 25 C 28 C 31 C 34°C Mean I Fertilization (1(1 0.0 0.0 ().() 0.0 2 First polar body 2.5 3.3 4.0 4.8 3.6 3 Second polar body 3.3 4.3 6.7 6.3 5.1 4 First cleavage (2 cells) 5.0 6.5 6.7 9.5 6.9 5 Second cleavage (4 cells) 10.0 12.0 13.3 17.5 13.2 6 Third cleavage (8 cells) 13.3 16.3 16.0 T> -> 16.9 7 Fourth cleavage (12 cells) 14.2 17.3 20.0 23.8 18.8 8 Fourth-fifth cleavage ( 12-16 cells) 15.0 18.5 21.3 25.3 20.0 9 Morula (32 cells) 28.3 31.5 29.3 38.0 31.8 10 Blastula 38.3 41.2 45.3 57.0 45.4 11 Gastrula 55.0 52.0 56.0 66.5 574 12 Prototrochal cilia 60.0 65.0 66.7 76.0 66.9 13 Stomodeuni 70.0 71.5 72.0 85.5 74.8 14 Prototrochal girdle 80.0 45.5 80.0 95.0 75.1 15 Hatch out 95.0 97.5 96.0 114.0 100.6 16 Beginning of larval shell 105.0 104.0 104.0 133.0 111.5 17 Completion of velum 120.0 123.5 128.0 152.0 130.9 18 Larval retractor muscle 135.0 143.0 144.0 171.0 148.3 19 Integumental attachment 145.0 149.5 168.0 199.5 165.5 20 Protrusion of foot mass 160.0 1 62.5 168.0 199,5 172.5 21 Completion of larval shell 165.0 175.5 176.0 209.0 1814 22 Torsion (90' twisting) 180.0 182.0 184.0 218.5 191.1 23 180° rotation of foot mass 210.0 234.0 216.0 256.5 229.1 24 Operculum 250.0 292.5 256.0 304.0 275.6 25 Long spines on end of metapodiuni 290.0 299.0 296.0 323.0 302.0 26 Fine cilia on foot 305.0 325.0 328.0 389.5 336.9 27 Vertical groove in velum 350.0 351.0 336.0 399.0 359.0 28 Eye spot 400.0 364.0 336.0 399.0 374.8 29 Propodium 410.0 383.5 384.0 456.0 408.4 30 Cephalic tentacle 435.0 396.5 400.0 456.0 421.9 31 Cilia on propodium 450.0 416.0 424.0 494.0 446.0 32 First epipodial tentacle 455.0 468.0 448.0 532.0 475.8 33 Cilia in mantle cavity 480.0 468.0 480.0 551.0 494.8 34 Apophysis on propodium 510.0 507.0 480.0 470.0 516.8 35 Otolith 550.0 533.0 544.0 627.0 563.5 36 Short spine on cephalic tentacle 570.0 546.0 544.0 646.0 576.5 37 Two tubules on cephalic tentacle 590.0 546.0 560.0 665.0 590.3 38 Third tubule on cephalic tentacle 600.0 559.0 592.0 684.0 608.8 39 Snout protrusion 620.0 585.0 624.0 73 1 .5 640.1 40 Ciliary process in mantle cavity 640.0 598.0 640.0 731.5 652,4 41 Retractor muscle drawn in mantle cavity 640.0 611.0 640.0 760.0 662.8 42 Fourth tubule on cephalic tentacle 650.0 637.0 656.0 770.0 680.5 Larval Development of Abalone 599 of larvae is dark green, while the foot, velum and cephalic parts are transparent yellowish. H. sorenscni larvae are beige in color, while the velar margin is yellowish (Hahn 1989a). In some abalone species, the pigments seem to be concentrated at the velum: for example. H. fulaens larvae are generally brown with green velar margins, H. cornigata larvae are light yellow-green with a velar fringe of a darker shade of green {Leighlon 1974). H. coccinca canarieiisis is orange in visceral mass and violet in foot, velum and cephalic mass (Harrison and Grant 1971). The size of fertilized abalone eggs also varies from species to species. The diameter of the fertilized egg of H. asinimi was 125.4 X 148.2 |j.m. It is 103 |jLm in H. coccinea canariensis (Pena 1986), 150-225 jxm in H. cyclnbates (Shepherd & Laws 1974). 200-250 |j.m in H. kievigula (Shepherd & Laws 1974), 200 (xm in H. soreiueni (Leighton 1972) and H. rubra (Harrison & Grant 1971), 210 pm in H. iiiherculala Linnaeus (Koike 1978). 230 p.m in H. ilisais Reeve (Ino 1952) and H. iris Gmelin (Har- rison & Grant 1971), 270 |j.m in H. gigantea Gmelin (Ino 1952) and 280 |j.m in H. scibohlii Reeve (Ino 1952) (Table 5). The size of the trochophore and veliger also varies depending on species (Table 5). There are 41 stages of larval development in H. discus luinmii (Seki & Kan-no 1977). H. diversicolor sitpertexm Lischke (Oba 1964), H. sii'holdii (Ino 1952) and H. discus (Ino 1952). Only 32 stages were described in H. nijcscens Swainson (Hahn 19S9a). There are 42 stages of larval development in H. asininii. The previous studies on larval development of several species of aba- lone showed that there were no differences in the development from fertilization to trochophore larva in H. discus haninii (Seki & Kan-no 1977). H. siel'nidii ilno 1952). H. discus (\no 1952) and H. diversicolor superte.xta (Oba 1964). However, there are certain differences between larvae of these abalone species and that of H. asiniiui. In other abalone species, the fourth cleavage produced 16 cells, while 12 cells were observed in the fourth cleavage (stage 7) and 16 cells were observed in the fifth cleavage (stage 8) of H. asiuuui. H. asininii larvae had both long and short cilia on the protolrochal cell (stage 12). while larvae of other abalone species had only long cilia. In addition. H. asinina trochophores appeared to lack the apical tuft cilia that were reported in the trochophores of other abalone species. The larval development of H. asinina. from trochophore stage to torsion, is similar to that of H discus liaiuiai, H. sicboldii. H. discus and H. diversicolor supertexui. The larval retractor muscle (stage 18) was formed prior to formation of the integumental at- tachment to the larval shell (stage 19). In addition, protrusion of the foot mass (stage 20) and development of the operculum (stage 24) occurred after these stages (Seki & Kan-no 1977. Ino 1952. Oba 1964). Similarly. H. asinina. H. discus hanuai. H. sicboldii. H. discus and H. diversicolor superte.xta formed cilia during early larval development and the cilia were retained (Seki & Kan-no 1977). There are many differences in lar\al development from torsion to metamorphosis. As in H. discus hannai. the larval shell of H. asinina was completed (stage 21) before torsion (stage 22) oc- curred (Seki & Kan-no 1977). In H. sieboldii. Ino (1952) found that torsion occurred first and then the larval shell was completed about 13 hours later. In H. asinina. the spine at the end of propo- dium (stage 29) was very fine and very difficult to observe. It could be seen after the formation of the operculum. In H. discus luutiuu. the spine was found at the end of the propodium before the formation of the operculum. In addition, like H. sieboldii.. the spines on the metapodium (stage 25) of H. asinina were formed after the operculum (Ino 1952). In H. discus hannai. these spines were formed before the operculum (Seki & Kan-no 1977). In H. asinina. the cilia in the mantle cavity (stage 33) up to the anterior edge of the velum were seen after the formation of the first epipodial tentacle (stage 32). In H. discus hannai. the formation of the first epipodial tentacle occurred before the appearance of apophysis and cilia in the mantle cavity. In addition, in H. asinina. the ciliary process on the roof of the mantle cavity (stage 40) was formed after the formation of the third tubule on the cephalic tentacles (stage 38). In H. discus hannai. the third tubule on the cephalic tentacles was seen before the formation of the ciliary process on the roof of the mantle cavity. H. asinina and H. di.'icus hannai developed all the larval stages from stage 34 to 42 {H. asinina). or stage 32 to 41 {H. discus hannai) before metamorphosis. As in H. discus hannai. the otolith (stage 35) was observed after the formation of the cephalic tentacle (Seki & Kan-no 1977). The duration of the pelagic stage in gastropods is variable, and the larvae may spend from two to 14 days in the water column (Leighton 1974). Also the abalone larvae are pelagic and the du- ration of development varies from species to species. In general, tropical abalone species such as H. asinina and H. diversicolor TABLE 5. The size of egg, trochophore and larva of some Haliotis spp. Completion Egg Trochophore of larval shell Species (fini) (pm) ((im) References H. asinina 12b X 14S 105 X L^l \15 X 242 The present study H. coccinea canariensis 103 201 X 160 264 X 206 Pena (1986) H. cycliiljates 150-225 Shepherd and Laws (1974) H. laevigata 200-250 Shepherd and Laws (1974) H. discus 230 178 X 200 290 Ino (1952) H. rufescens — 160 X 195 210x270 Ebert and Houk (1984) H. sieboldii 280 Ino (1952) H. gif>aniea 270 Ino (1952) H. ins 230 Harrison and Grant ( 1971 ) H. rubra 200 Harrison and Grant (1971) H. sorenseni 200 Leighton (1972) H. tuherciilala 210 Koike (1978) 600 Sawatpeera et al. siipevtexta usually have a short larval development (Oba 1964), In H. asinina. the time durations for hatch-out. torsion and formation of the first epipodial tentacle were 6.0. 11.5 and 28.0 h. respec- tively (at 31 -C) and 7.5, 14.0 and 36.0 h. respectively (at 28°C) (Table 6). These durations are quite similar to those in H. diver- sicolor supertexta, i.e., 6.0, 13.0 and 38.0 h for hatch-out. torsion and formation of the first epipodial tentacle, respectively (at 26.2°C) (Owen et al. 1984) (Table 6). The first epipodial tentacle of H. cisinina appeared before the fonnation of the apophysis on the propodium, and the larvae started to crawl on the bottom after the formation of the apophysis. In H. discus hannai larvae, the apophysis on the propodium was formed before the appearance of the first epipodial tentacle and the larvae started to crawl on the bottom when the first epipodial tentacle was formed (Seki & Kan- no 1977). The induction of settlement could be done in H. asinina after the formation of the apophysis on the propodium. However, the larvae were not able to metamorphose at this stage until they reached the last stage of development. The abaione larvae are competent to metamorphose when they have four branches on the cephalic tentacles, the fully developed foot is able to pull the larvae upright and they can move by ciliary action (Seki & Kan-no 1977. Ebert & Houk 1984). The duration of complete larval development of H. asinina was 41.0 to 65.0 h or only 1.7 to 2.7 days (Table 6). while the duration of larval development of temperate abaione species was from 99.0 to 360.0 h or 4.1 to 15 days depending on the species and water temperature (Table 6). The duration of larval development in H. asinina at water temperatures of 3I°C and 34°C was identical. However, from observation, water temperature at 34°C is rather high and may cause abnormal development and high nioilalily of lar\ae. The optimum water temperature of larval de- velopment in H. asinina was 28''C to 3I°C. Temperature is an important factor in many stages of develop- ment such as gonad maturation, spawning (Hahn 1989b. Hahn 1989c) and larval development (Seki & Kan-no 1977. Hahn 1989a). Normally, the animals grow faster in a higher temperature. Larval development in Hiiliolis spp. also occurs at a faster rate at higher water temperature (Hahn 1989a). The development of H. asinina larvae at water temperature of 3I°C (41 h) was faster than that at 25°C (65 h). In Haliotis nifescens Swainson. the hatching time in water temperature of 1 7°C to 1 8" C was faster than that at 15°C (Table 6). A successful culture and high yield are assured by changing the rearing conditions at three important stages during larval develop- ment: hatch-out. development of larval shell, and settlement (Hahn 1989a). So the information on the biological zero point and the effective accumulative temperature (EAT) is necessary for hatch- ery management. The biological zero point varies among abaione species, depending on water temperature in each location (Hahn 1989a). For example, the biological zero point of H. nife.tcens is 0.9''C which is below the ambient water temperature (Hahn 1989a). In H. asinina. the biological zero point (15.0°C) is also below the ambient water temperature (20°C). Biological zero point is one of the values used to calculate EAT. The average EAT of larval development of H. asinina is lower than those of temperate abaione species such as H. discus luinnai and H. nifescens (Table 7). The important EAT values in hatchery are during hatch-out, completion of larval shell formation and competency of settlement stage. After hatch-out, the healthy trochophore larvae will swim toward the surface immediately and can be collected near the water surface and transferred to the other tank (Shepherd 1975). The trochophore larvae of W. asinina should be decanted as soon as possible after they reach the EAT I01°C- hours. in order to reduce bacterial contamination from decompo- sition of egg membranes and unfertilized eggs. Moreover, the en- zymes secreted by the larvae to facilitate hatching may contribute to the lowering of water quality (Hahn 1989a). From the tro- chophore stage to completion of the larval shell stage, the larvae TABLE 6. The time of hatch out. 90 torsion, first epipodial tentacle formation and fourth tubule on cephalic tentacle formation of several abaione species. Larval First Fourth tubule Temperature Hatch out retractor Torsion epipodial on cephalic Species CC) (h) muscle Ihl (90 twisting) (h) tentacle (hi tentacle (h) References H. discus lumiiai 20 12.6 23.4 32.1 64.9 99 Seki and Kan-no (1977) H. discus 16-17 20 45^6 120 Hahn (1989c) H. sic'boldii 16-17 18 29 35 96 Ino (1952) H. giganlea 16-18 21-22 40-43 Ino (1952) H. rufescens 15 20 144 Ebert and Houk (1984) 15 20 175 Hahn (1989c) 17 24 44 Leighton (1974) 17 25 36 48 Hahn (1989c) 18 16-24 Owen et al. (1984) H. fulgens 17 14 48 360 Leighton (1974) H. conugata 16 17 48 192 Leighton (1974) H. sorenseni 15 24 240 Leighton (1972) H. tuherculata 20 13 20.5 84 120 Koike (1978) H. coccineu canariefisis 15 21 52 Pena{1986) H. diversicolor superlexia 26.2 6 13 38 Oba (1964) H. asinina 25 9.5 13.5 IS.O 45.5 65.0 The present study 28 7.5 11.0 14.0 36.0 49.0 The present study 31 6.0 9.0 11.5 28.0 41.0 The present study M 6.0 9.0 11.5 28.0 41.0 The present study Larval D[:velopmknt of Abalone 601 TABLE 7. The biological zero point (C) of some Haliotis spp. Biological zero point Species ( Cl References H. ilisctt.s hanihii 7.6 Seki and Kan-no (1977) H. discus 8.5 Hahn (1989c) H. gigwitea 9.0 Seki and Kan-no (1977) H. rufescens 8.5 Seki and Kan-no (1977) H. rufescens 0.9 Hahn (1989c) H. fulgens 9.0 Leighton ( 1974) H. {isiiiiiui 1.5.0 The preseni study are very fragile, because they do not have a protective organ. The water in the tank should be changed after formation of the larval shell, or when the EAT of larval development reaches 181.4°C- hours. The EAT of settlement stage of H. asinina larvae was 680.5°C-hours when the larvae reached the formation of the third tubule on the cephalic tentacles. Settlement occurs rapidly on wavy-plate substrata if the larvae are healthy and competent to settle and metamorphose (Hahn 1989a). ACKNOWLEDGMENT This study received financial support from the Thailand Re- search Fund BRG/ 12/2542 and PG 2 1 (J 13/2539. LITERATURE CITED Ebert. E. E. & J. L. Houk. 1984. Elements and ninovation in the cultivation of red abalone. Haliotis rufescens. Aquaculture. 39:375-392. Hahn. K. O. 1989a. Larval development of abalone. In: K. O.Hahn. editor. Handbook of Culture of Abalone and other Marine Gastropods. Boca Raton: CRC Press, pp. 71-98. Hahn. K. O. 1989b. Survey of the commercially important abalone species in the world. In: K. O. Hahn, editor. Handbook of Culture of Abalone and other Marine Gastropods. Boca Raton: CRC Press, pp. 4-1 1. Hahn. K. O. 1989c. Nutrition and growth of abalone. In: K. O. Hahn. editor. Handbook of Culture of Abalone and other Marine Gastropods. Boca Raton: CRC Press, pp. 135-156. Harrison. A. J. & J. F. Grant. 1971. Progress in abalone research. Tasma- nian Fish. Res. 5:1-10. Ino, I. 1952. Biological studies on the propagation of the Japanese abalone (genus Haliotis). Bull. Tokai Reg. Fish. Res. Lab 5:10-102. Koike. Y. 1978. Biological and ecological studies on the propagation of the ormer. Haliotis tuherculata Linnaeus. I. Larval development and growth of juveniles. La Men 16:124-136. Leighton. D. L. 1972. Laboratory observation on the early growth of the abalone, Haliotis sorenseni, and the effect of teinperalure on larval development and settling success. Fish. Bull. 70:373-381. Leighton. D. L. 1974. The influence of temperature on larval and ju\enile growth in three species of southern California abalones. Fish. Bull. 72:1137-1145. Oba. T. 1964. Studies on the propagation of the abalone Haliotis diversi- color supertexta Lischke. II. On the development. Bidl. Jpn. Soc. Sci. F/.S/; 30:809-819. Owen. B.. L. H. Disalvo. E. E. Ebert & E. Fonck. 1984. Culture of the California red abalone Haliotis rufescens Swainson (1922) in Chile. Veliger. 42:49-51. Pena. J. B. 1986. Preliminary study on the induction of artificial spawning in Haliotis coccinea caiiariensis Nordsieck (1975). Aquaculture. 52: 35^1. Seki. T. 1997. Biological studies on the seed production of the northern Japanese abalone Haliotis discus hannai Ino. Bull. Tohoku Natl. Fish. Res. Inst. 59:1-71. Seki. T. & H. Kan-no. 1977. Synchronized control of early life in the abalone. Haliotis discus hannai Ino, Haliotidae. Gastropoda. Bull. To- hoku Reg. Fish. Res. Lah. 38:143-153. Shepherd. S. A. & H. M. Laws. 1974. Studies on southern Australian abalone (genus Haliotis): II. Reproduction of five species. .Aust. J. Mar. Fresim: Res. 25:49-62. Shepherd. S. A. 1975. Distribution, habitat and feeding habits of abalone. ,\ust. Fish. .34:1-1. Journal «t Shtlljisli Rcscanli. Vol. 20, No. 2. 603-61(1. 2001. ISOLATION AND GROWTH OF EIGHT STRAINS OF BENTHIC DIATOMS, CULTURED UNDER TWO LIGHT CONDITIONS JUAN GABRIEL CORREA-REYES,' * MARIA DEL PILAR SANCHEZ-SAAVEDRA,' DAVID ALFARO SIQUEIROS-BELTRONES,- AND NORBERTO FLORES-ACEVEDO' Aquacullurc Department, Centro cle Invcstigacidn Cieiitifica v cle Educuvioii Superior de Ensenada (C/CESE). Km. 107 Carretera Transpeuinsular Tijiiana-Ensemida, Ensenada, Baja California. Mexico, C.P. 22^30: ' Departamento de Pldncton y Ecologi'a Marina, CICIMAR-IPN Univer.'iidad Aiitonoina de Baja California Siir, Apartado Postal I9-B. La Paz. Baja California Snr. Mexico. C.P. 23081 ABSTRACT We isolated seven strains of benthic diatoms from lliree abalone commercial fishery areas in the state of Baja California and in the proximity of several hatcheries and nursery. In all the cases, we isolated several strains of the genus Nilz-nhia and one strain o( Ampliiproni puludosa var. hyolina. Monospecific and non-axenic cultures of each strain were maintained on batch cultures under white and blue light at high (150 p,E m~" s"') photon lluence rates. For all the diatoms we observed that the growth rate (ja.) in the second or third day of culture had high values of duplicated strains per day. We compared the growth rate of these strains with a control specie Niiviciilii incerla. a strain used as food in several abalone farms in Mexico, and found similar growth rates in some of the cases. For each strain, a covariance analysis was used to compare growth using white versus blue light wavelength. No significant differences in growth were observed between the two light conditions. We obtained high growth rate for Ainpliiprora paludosa var. hyalina with almost double values of those obtained with of the control species N. incerta and the others strains: N. thermalis var. minor. N. fonticola var. pelagica and N. frustulum var. pennimiia. The sub-strains of N. laevis exhibit equal growth rate and nearly 50% less than that of N. incerta. Our results suggest that some strains of benthic diatoms offer a good potential for use in feeding abalone post-larvae, because they can be cultured under high light irradiance apparently without photoinhibition and may exhibit high growth rates and high cell concentrations. KEY WORDS: benthic diatoms, growth rates, light quality, abalone, post-larvae. INTRODUCTION Benthic diatoms play an important role in abalone cultures, acting as inductors for settlement of the larvae and serving as the main source of nourishment during the early juvenile stage (Ebert & Houk 19S4. Kawamura et al. 1995). To be used as nourishment in abalone cultures: the diatoms must be preferably of a size less than 10 |jim. exhibit adequate nutritional level, and adhere strongly to the substrate so that they may be readily ingested by the abalone post-larvae and juveniles (Hahn 1989. Fleming et al. 1996, Kawa- mura et al. 1998). Rearing of post-larva abalone under culture conditions relies on two different strategies to obtain the first microalgae food supply. The first method involves allowing the proliferation of natural populations of phytoplankton. This culture tnethod has the disad- vantage of being unpredictable, and results in uncontrolled species introduction and densities. The growth of the postlarvae may be insufficient or inadequate when natural populations of diatoms or other phytoplankton species are used, due to the seasonal variabil- ity that they present (Ebert & Houk 1984. Kawainura & Hirano 1992. Voltolina 1994). The second method uses controlled condi- tion cultures of one or several benthic diatom species. This strategy has the advantage that the food densities can be controlled pro- ducing a more predictable growth rate, and a higher survival of abalone post-larvae. However, in some cases the species of benthic diatoms used in abalone culturing farms are non-endemic to the region, which may result in the introduction of species and the displacement or succession of local species. Therefore, it is im- portant for farms to use benthic diatoms isolated from the area, to guarantee the quantity and quality of nourishment while avoiding the ecological unbalance of the organisms of the region. *Corresponding author. PC Box 434844. San Diego, CA 9214.^: E-mail: cjuan@cicese.mx Algae cultures may be influenced though the manipulation of environmental variables. Light is the driving force of photosyn- thesis. As such, its quantity and spectral composition and its fluc- tuation within the culture control biomass production rates (Du- binsky et al. 1995). The potential of this light-induced manipula- tion of the cultures has not been sufficiently explored in the context of aquaculture. Surprisingly, only few studies have used light manipulation to obtain different growth rates and nutritional values of plankton microalgae in culture (Flaak & Epifanio 1978. Correa-Reyes 1993. Sanchez-Saavedra & Voltolina 1994. Sanchez-Saavedra & Voltolina 1995). The effect of light irradiance on the growth rate of benthic diatoms has been poorly studied (McBride 1990, Flores-Vergara 1998). It is generally considered that the light intensity and quality suitable for rapid growth, is related to the type of habitat utilized in the natural environmental. Usually for benthic diatoms culture for abalone post-larvae only a low light intensity is supplied, and in abalone farms, light is provided with fluorescent lamps at 80 (xE m~" s~' on the water surface for less than a 12- hour period (Hahn, 1989). In several studies, the light regiinen used for the culture of benthic diatoms was between 20 and 200 (xE m"" s"' (Thompson et al. 1992. Cota-Sanchez 1998, Roberts et al. 2000). In the natural environment, the spectral characteristics of light changes as function of depth, red light is the first wavelength that is absorbed in the water column. In the benthic environment, blue light is the light quality that is usually available and hence used by photosynthetic organisms. These are the conditions found in the habitat shared by benthic diatoms and abalone. Surprisingly, there is no previous work, relating the effect of light quality on growth of benthic diatoms; all studies have been conducted on planktonic diatoms (Gostan et al. 1986. Senger 1987). The aim of this study was to compare the growth rate of several strains of benthic diatoms species isolated from different zones 603 604 Correa-Reyes et al close to ubalone farms in Mexico, and to culture these under two different light conditions: white and blue light. MATERIALS AND METHODS We collected biofilm samples of benthic diatoms from three different regions of fishery production and within the proximity of abalone farms: two in Baja California (Ejido Erendira and Isla de Cedros) and one in Baja California Sur (Puerto Nuevo) in Mexico (Fig. I ). The i-solation techniques used to obtain the different spe- cies or strains of benthic diatoms, involved micropipetting and serial dilution as described by Hoshaw and Rosowski (1973) and by Voltolina (1994). The different benthic diatom strains were preselected based on morphology and attachment capacity. During the isolation process of the different stocks and with the purpose of eliminating bacteria and fungi, baths were applied with a mixture of 50 ppm penicillin and 100 ppm streptonncin during an eight- hour period. For the taxonomic identification of the diatoms, we proceeded to eliminate the organic matter present in the cultures of each diatom strain. While preserving the ornamental structures by wash- ing with concentrate H^SOj. K^HPOj IM. 3% glutaraldehyde and different percentages of ethanol in a sequence of 10. 25. 50. 75. 90 and 100 (Nalewajko 1989). An aluminum film permanent slide was mounted for observation under scanning electron microscope (SEM) for each diatom strain. Phase contrast (lOOx) and planapo- chromatic (63x) microscopy was used in the taxonomic analysis. Taxonomic criteria were based on morphology of the frustule. using the classic literature and recent works (Navarro 1981a, Na- 120° 34- 32° 28" 26' Erendira Cedros Island Puerto Nuevo Pacific Ocean van-o 1981b. Navarro 1982, Navarro 1983. Round et al. 1990, Simonsen 1987, Siqueiros-Beltrones 1994, Siqueiros-Beltrones 1999, Siqueiros-Beltrones 2000. Siqueiros-Beltrones & Sanchez- Castrejon 1999, Siqueiros-Beltrones & Voltolina 2000). Curatorial records on cards for each taxon were made, including SEM pho- tographs. The permanent slides were included in the "Diatom Col- lection of the Museo de Historia Natural" of the Universidad Au- tonoma de Baja California Sur. Mexico. We also used as control species a strain of Naviciila incerta isolated from coastal waters of Bahi'a de Todos Santos, Mexico by the Microalgae Laboratory staff of the "Institute de Investiga- ciones Oceanologicas" Universidad Autonoma de Baja California, Mexico. That strain was obtained from the abalone farm "Abu- lones Cultivados S. de R.L. de C.V." in Ejido Erendira. Baja California, Mexico. Non-axenic monospecific batch cultures of each strain were placed in ten replicate 250-ml Erlenmeyer flasks with 150 ml of Guillard and Rhyther's (1962) "f medium. The densities of the inoculum used to grow each strain varied as a result of differences in cell size (Table 1 ). However for comparison purposes, each one of the strains of diatom species on the experimental condition of light received the same quantity of inoculum. The culture condi- tions were salinity 34±1 %, and temperature 22 ± I'C provided by air conditioning. These values are similar to those used in labora- tory abalone production. The pH was not controlled and values ranged between 8.2 and 9.7. Every other day we checked the cell concentrations. Ultrasound was applied to the culture flasks for fifteen to sixty seconds, de- pending on the age of culture, to dislodge the cells that were firmly attached to the walls (Voltolina 1985). Two flasks from each light condition and strain were used to determine cell concentrations; these were discarded to a\oid the possible effect of the ultrasonic method on the growth of the diatoms. The evaluation of cell con- centration was measured through direct counts using a haemacy- tometer. The growth rate and generation time was determined by TABLE \. Benthic diatom strains isolated from tliree different zones close to abalone farms in Baja California (B.C.) and Baja California Sur (B.C.S.), .Me.xico. Procedence, isolation techniques and average size in microns (length and Hidth). Figure \. Mew of different localities sampled to obtain benthic dia- toms strains in Ejido Erendira. Cedros Island on Baja California, and Puerto Nuevo, in Baja California Sur. Mexico. Isolation Size Strain Procedence Techniques (microns) Nitzschia ihennalis var. Ejido Erendira. Micropipette 21.0x7..S minor (.4) B.C. Nitzsdua laevis IB) Ejido Erendira. B.C. Micropipette 7.5 X 6.0 Nirzscliici liieris (C) Puerto Nuevo. B.C.S. Serial dilution 8.0 X 5.5 Nitzscliia laevis (D) Puerto Nuevo. B.C.S. Serial dilution 7.0 X 4.0 Naviciila incerta (E) Ejido Erendira. Serial dilution. 14.5x5.5 BC. Donated Ninscliia cf. fimlicola Isla de Cedros. Serial dilution 16.4x4..^ var. pelagica (F) B.C. Amphiprora paludosa Isla de Cedros. Micropipette 13.2x4.6 var. hyalina (Donk.) B.C. Reimer IG) Nitzschia fnistiihim Isla de Cedros. Serial dilution 10.3x3.6 var. penniniita IH) B.C. Benthic Diatoms Growth Rath 605 calculations the log, cell concentration at different times for each strain as described by Sorokin (1973). We exposed each strain to two continuous light conditions. The first one was obtained using normal laboratory Cool White 40 W fluorescent lamps (Sylvania F40CW) that provided a spectral emission with two main peaks of emission principally, one in the blue-green wavelengths (433-300 nm) and the other within the red (550-620 nm). In the second, ambient Blue Light was provided with 40 W General Electric F40B, and the peak of emission was primly within the blue-green light (430^70 nm). For the two light qualities, a photon fluence rate of 130 p-E m"~ s"', because this irradiance is commonly used in several commercial abalone hatch- eries in Mexico. Photon fluence rates values were measured at the center of the culture surface with a QSL-100 (Biospherical Instru- ments) 4 TT quanta meter. The spectral characteristics of the light- ing sources were determined as described by Sanchez-Saavedra and Voltolina (1996a). A covariance analysis and two ways ANOVA were used to compare cell concentration among treatments and culture days. A p-level of 0.05 was used to evaluate significance with the "Statis- tica 5.0 for Windows" program (StatSoft Inc., 1996). RESULTS As result of the sample collections, six stocks of benthic dia- toms were isolated (Stocks: A through G, Fig. 2, Fig. 3, Fig. 4, Fig. 5), and one strain previously isolated was obtained by donation. In the majority of the localities, stocks belonging to the Nirz.scliia genus were isolated. The other one was identified as Ampiiiproni pahidosa var. hvdlina. Size ranged between 7 and 20 p, (Table 1 ). In all cases, the mean cell concentration for a given strain grown under white or blue light conditions (Table 2), did not exhibit significant differences A: p = 0.971, F = 0.001; B: p = 0,231, F = 1.534; C: p = 0,070, F = 3.695; D: p = 0.120, F = 2,644; E: p = 0.491, F = 0.494; F; p = 0.231, F = 1.534; G: p = 0.070, F = 3.693). Due to differences in length and width of the cells, we obtained different cell concentrations between the eight strains. However the highest cell densities (for white and blue light respectively) were ior Amphipvora pahulosa var. hyalina and the size of this strain was one of the highest (48.4 and 44.38 lO'' cell ml"'), two substocks of Nitzschia laevis (4.85 and 41.43 io" cell mr' and 37.98 and 53.94 10' cell ml"'), and Navicula inccna (30.64 and 28.98 lO'^ cell mP'). The lowest cell concentrations were found for Nitzschia ihermalis var. minor (2.50 and 3.02 10' cell ml"') (Table 2), and the strain with the largest cell size (21.0 X 7.5 p.) (Table 1). The mean cell concentration values for each of the eight benthic diatoms cultures under white or blue light for ten days, yielded significant differences in the covariance analysis (for white light p = 0.000; F = 25.662 and for blue light p = 0.000; F = 25.075). The growth rate for each strain was similar (Table 3) and with equal values under the two light treatments examined. The highest growth rates occurred during the first two days of culture. We obtained high growth rate values for Amphiprora paliidttsu var. Inalina (2.70 and 2.69), which was twice as high as that found for the control species Navicula incerta (1.62 and 1.58). Similar values were obtained for N. thermalis var. minor, N. fonticola var. pelagica and N. fru.stuhim var. perminiila (1.18 to 1.62^. For the sub-strains of N. laevis. we measured the same growth rate for each one; the values were nearly 50% less than that for Navicula incerta. Spitl 2gm Figure 2. Images a{ Nitzschia Iheniuilis vcir. minor (.\( and A';V-.vf/i/V; laevis (B), isolated close to the post-larvae abalone farm .^bulones Cultivados S. de R.L. de C.V. in Kjido Erendira, Baja California, Mexico. Under the two light conditions, the generation time obtained for the eight strains as expected were lower than or close to one for the first two days and increased as corresponded to cell concentrations values. The lower value of generation time under white or blue light for the first days (0.37 for day two and 2.04 for day four), confirmed the higher cell concentration values for Amphiprora paludosa var. hyalina (Table 4). For days six to ten for both light conditions and for the eight .strains, the inconsistency with the values of generation time, were due to minimum values of differ- ences on cell concentrations on the last days of culture. DISCUSSION For the selection of benthic diatom species with a potential to be used in nutrition of abalone post-larvae, several suitable char- acteristics have to be considered. Among the suggested ones are: an adequate size, many authors recommend sizes around 10 p., the degree of adhesion, the frustulum width or degree of silicification 606 Correa-Reyes et al 2pm Figure 3. Images of Nilzschia laevis (C) and Nitzschia laevis (D), iso- lated close to the laborator> for ahalune seed production SCPP Eman- clpacion in Puerto Nuevo, Baja California Sur, Mexico. and motility capacity (Kawamura et ai. 1998). However for other cell sizes, apparently the abalone radula can cause the cell to become deformed and create physical rupture of the diatom cell, as described by Daume et al. (1997). who observed that abalone post-larvae .settling over coralline algae were capable of breaking the theca of Cocconeis spp., and were capable of better assimilat- ing the cellular content. This ability of the radula of abalone post- larvae is corroborated by the work of Roberts el al. ( 1999). In our results, the lowest values of cell concentrations corresponded to largest cell size. e.g. Nitzchia rhennalis van minor. Smaller cell sizes corresponded to medium cell concentrations, e.g. the three different strains of Nitzschia laevis. The highest cell concentrations and largest cell sizes corresponded to .Aiiipliipmra paludosa var. hyalina and Nitzschia fnistiduni var. perininuta. We can isolated six strains of Nilz.cliia spp., because several Nitzschia species have fragile cellular walls and in some cases strong adhesion to substrates, making them susceptible to grazing by abalone post-larvae (Siqueiros-Beltrones. 1999). In addition Siqueiros-Beltrones. 1999 suggested that the diatoms may break 4pm Figure 4. Images of Navicula incerta (E) isolated from coastal waters of Bahia de Todos Santos by the staff of niicroalgae laboratory of the Institute) de In\estigaciones Oceanologicas of the I'niversidad .\u- tononia de Baja California. Mexico. For this work this strain was obtained from the abalone farm Abulones Cultivados S. de R.L. de C'.V. in Ejido Erendira, Baja California, Mexico. when grazed, and lead to a better assimilation of the cellular ma- terial. However, for the eight diatom strains isolated in this study there is no current information available regarding degree of si- licification, adhesion and digestibility by abalone post-larvae. It has been proposed that the species of the Nitzscliia genus may provide good nourishment for abalone post-larvae due to their high growth rates and dominance in successions in culturing systems (Siqueiros-Beltrones. 1999). In addition, the substratum from which benthic diatoms were isolated should be considered (Siqueiros-Beltrones 200(.), Siqueiros-Beltrones & Voltolina 2000). The habitat were they are isolated determines the hardness of cellular walls. All the strains isolated for this study, were ob- tained of zones with lowest water movement or from tanks on abalone farms. In this work, we observed high growth rates for all species during the first two days of culturing. Other studies (Cota-Sanchez 1998, Cuevas-Rocha 1998, Flores-Vergara 1998), recorded maxi- mum growth rates between days four and five for benthic diatoms of the genera Nitzschia. Ampliora and Navicula. However, for other benthic diatoms as such Navicula cf. cincta and Nitzschia sp. cultured individually or mixed with Ampliora cf. catenula, the maximum growth rales were observed between days two and three. These studies proposed that the release of extracellular products could inhibit or promote growth (Cota-Sanchez 1998. Cuevas- Rocha 1998). Since these studies and ours used different species and culture conditions (light, nutrients, temperature, pH, salinity, and water movement), the growth rates obtained can be influenced directly. Thus, we propose that culture conditions are an important fact than must be considered, for explaining different growth rates and consequently differences on chemical composition and nutri- tional value. Several abalone nurseries discarded the use of benthic diatoms when the mean generation time was higher than one day (Voltolina 1985, Voltolina 1994, Cuevas-Rocha 1998). All the diatom strains Benthic Diatoms Growth Rate 607 2Mm ■Iptn Figure 5. Images of Xilzschia cf. fonlicola vcir. pelagica (F). Am- phiprora paludosa rar. hyalina ((i) and Mlzscliia fnistiiliim rar. per- ininuta (H», isolated near tlie laboratory for abalone seed pniduction SCPP Pescadores Nacionales de Abiilcin In Isia de Cedros, Baja Cali- fornia, Mexico. used on this work had mean generation time closest to one. by this we can recommend using the eight isolated strains and experiment on this work for abalone postlarvae nutrition. High variations on generation time are due to the minimum difference obtained on cell concentration values. However. The sample strategy used may produce lower differences on cell concentration values due to minimum differences between Erienmeyer flask culture replica- tions. Since light provides [lie energy that drives photosynthesis other parameters such as photoperiod. light energy or light intensity, and the light quality must be considered. Some algae adapt to very low light intensities and can live at depths as much as 200 m (Fallu 1991). No previous works about the effect of light t|ualily on benlhic diatoms or abalone ethology are available. However, it should be considered that research on this subject is needed to understand the ecology and physiological response of alga com- munities (macro and micro) and of abalone. Abalone have negative phototaxis and is usually found in refuges or cavities in their natural environment or in culture tanks (Hahn 1989). Controlling the level of light available to the culture tanks has been the most important and most effective means of influencing the alga community on which abalone post-larvae feed. Commonly high light le\els produce filamentous algae, this originates dis- placement of benthic diatom species and occasional loss of post- larva with the water circulation system exchange. Usually this practice is accomplished by utilizing greenhouse shading material, or black polyethylene plastic (McMulIen & Thompson 1989). De- pending on cell density, pigmentation and cell geometry, the op- tical density and the wavelength of light can be changed within the culture, depending on the light source of vessel configuration and pond depth (Dubinsky et al. 1995). Benthic diatoms may be prolific at great depths < I 'f irradiance (Cahoon et al. 1993). where the main component of the light spectrum is blue wavelength. Abalone also dwells under low light conditions were benthic diatoms serve as their main food source //; situ, especially for their postlarvae and juveniles. However, other works reported good biomass production at high irradiance (around .^01) |jlE nV' s"') (Kromkamp et al. 1998). It was determined recently that they are in fact not reaching their optimal biomass production under low light conditions. The satu- ration light intensity may vary among diatom species. The results presented here suggest that light intensity was not limiting for growth, because we obtained good growth rates when compared to other reports for other benthic diatoms (McBride 1990. Simental- Trinidad 1999). With an adequate light quality we believe growth rate can be increased and metabolic changes induced by effect of deviation of carbon metabolism and cell number or biomass quan- tity increased as described for plankton diatoms (Flaak & Epifanio 1978. Sanchez-Saavedra & Voltolina 1996b). However, for the eight diatoms, strains cultured under blue light for this work, high growth rates or increase biomass production was not obtained compared with white light. These important differences in physi- ologic response could be related to the vertical natural distribution of benthic diatoms, and other factors may be the efficiency of light availability, thai can be assumed different for each species accord- ing to its preference. The results obtained in this study differ from those reported by Sanchez-Saavedra and Voltolina (1994. 1996a. 1 996b), who kept cultures of Chaetoceros sp. under blue and white light and observed significant differences in growth, finding greater cellular concentrations when cells were cultured under white light. 608 Correa-Reyes et al TABLE 2. Mean values (n = 3) of cell concentration (1 x Uf cell ml') and standard deviation (in parenthesis) cultured throughout 10 days with white (a) and blue light (b) tor Xilzschia Ihemialis var. minor (A); \ilzscliia laeris (B): Xilzscliia laevis (C): .Vitzxchia laevis (l)»: \ayiciila iiicerta (E); Nilzschia if. fonticola var. pelagica IF); .Amphiprora paludosa var. hyalina (Ci) and .\'ilzscliia friistalum var. peniiinuta (H). a Cell concentration of benthic diatoms (1 x 10'^ cell ml"') (days) A B C D E F G H 0.30 3.00 3.00 3.00 0.90 2.00 0.45 3.00 0 (O.OO)a (O.OO)bd (O.OO)bd (O.OO)bd (O.OO)b lO.OO)b (O.OO)c (O.OO)cd 1.93 9.10 9.51 9.83 8.51 14.70 18.96 18.45 2 (0.43 )a (0.51 )bd (0.57)bd (0.04)bd (0.98)b (0.64)b lO.OO)c (2.08)cd 1.89 19.61 10.17 16.99 12.37 20.75 37.46 28.49 4 (0.60)a (4.56)bd (0.36)bd (0.18)bd (O.lO)b (0.01 )b (0.73)c (l.OO)cd 3.31 24.69 22.54 28.70 13.99 21.35 28.35 34.37 6 (0.85 )a (0.81)bd (2.49)bd (5.59)bd (1.56)b (0.42 )b (O.OO)c (0.15)cd 2.44 33.09 31.20 36.21 21.65 22.15 46.38 30.74 8 (0.24)a (8.14)bd (6.15)bd (2.67)bd (8.89)b (0.77)b (4.67)c (2.l8)cd 2.50 3 i .68 34.S5 37.98 30.64 26.45 48.40 37. 1 8 10 (().(l.S);i (3.22)bd (3.S0)bd (5.18)bd (1.47 lb (:,h9)h (8,32)c (2.69 )cd b Time Cell concentration of benthic diatoms (1 x 10^ cell ml"') (days) A B C D E F G H 0.30 3.00 3.00 3.(.)0 0.90 2.00 0.45 3.00 0 (O.OO)a (O.OO)b (O.OO)bc (O.OO)bcd (O.OO)b (O.OO)b (O.OO)c (O.OO)bcd 2.04 10.28 12.69 15.37 8.06 12.18 18.65 17.69 2 ((1.27)a (0.69)b (0.17)bc (0.89)bcd (0.01 )b (2.44)b (O.lO)c (0.69)bcd 2.29 16.91 13.57 20.44 14.90 18.62 35.63 28.44 4 (0.55 )a ( 1 .06 )b (1.54lbc (3.29)bcd (2.21)b (0.57)b (2.69)c (O.SDbcd 2.31 19.98 32.64 26.19 13.53 27.28 35.44 32.29 6 (0.57)a (4.7 l)b (12.18)bc (2.12)bcd (1.40)b (1.25)b (7.00)0 (l.91)bcd 2.45 24.90 30.55 32.73 17.12 26.53 41.51 32.82 8 (0.92)a (5.17)b (3.84)bc (0.08)bcd (1.12)b (3.08)b (l.Ollc (0.26)bcd 3.02 35.15 41.43 53.94 28.98 25.23 44.58 41.88 10 (0.4 Da (2.79)b (4.50)bc (10.35)bcd (O.OO)b (1.45)b (0.65 )c (1.73)bcd Some letters indicate lack of significant differences as determined with a two-way ANOVA and Duncan a poncnoir test ^- = 0.05. a < b < c. TABLE 3. Group rates (// = 3) throughout ten culture days under white (a) and blue light (b) for Nilzschia ihermalis var. minor (A); Nitzschia laevis (B); Nilzschia laevis (C); Nilzschia laevis (D); Naviciila incerta (E): Nilzschia cf. fonticola var. pelagica (F); Amphiprora paludosa var. hyalina {C) and Nilzschia frusliilum var. perminuta (H). Growth rate of benthic diatoms (divisions by day) (days) A B C I) E F t; H 2 1.33 0.80 0.83 0.86 1.62 1.44 2.70 1.31 4 -0.02 0.54 0.05 0.40 0.27 0.25 0.49 0.32 6 0.41 0.18 0.57 0.37 0.09 0.02 -0.21 0.14 8 -0.21 0.20 0.23 0.17 0.29 0.03 0.35 -0.08 10 b 0.02 -0.02 0.08 0.03 0.28 0.13 0.03 0.14 Grow th rate of benthic diatoms (divisions by day) days A B C D E F G H 2 1,38 0.89 1.04 1.18 1.58 1.30 2.69 1.28 4 o.os 0.36 0.05 0.20 0.44 0.31 0.47 0.34 6 0.00 0.11 0.61 0. 1 8 -0.07 0.28 -0.01 0.09 8 0.03 0.1(1 -0.02 0.16 0.14 -0.02 0. 1 2 0.01 10 0.17 0.2(1 0.22 0.35 -0.03 0.05 0.17 Benthic Diatoms Growth Rate 609 TABLE 4. Generation time throughout ten culture days under white (a) and hiue Hghl (b) for \'itzscliia thennalis var. minor (\): S'ilzschia Uteris (B): Nilzscliia liuvis (C'l; \itzschia laetis (Dl; Xaviciilii inccrtu (Kl: Sitzschia cf. fonticnla var. piliigivu (Fl; Ampliipritra paludtisa var. hyaliiia {(.'•) and Sitzschia frustiilum var. perminutu (H). Time (days) Generat on time of benth c diatoms (days) A B C D E F G H 2 0.76 1.23 1.20 1.17 0.62 0.70 0.37 0.77 4 2.m 1.98 21.95 2.53 3.82 4.05 2.04 3.22 6 2.44 14.24 1.76 2.88 18.43 65.40 -4.74 7.53 8 -5.61 10.23 8.24 15,50 5.08 18,44 2.89 -15.79 10 b -5.02 0,71 -4 S.s 15 (ill 6.56 S53 -1.45 7.29 Time (days) Generation time of benth c diatoms (days) A B C D E F G H 2 0.73 1.13 0.96 0.85 0.63 0.78 0.37 0.78 4 28.36 2.88 -22.04 5.35 2.35 3.72 2.16 2.96 6 -377.06 -9.88 1.73 5.77 5.63 3.63 2.56 14.45 8 -0.28 -6.19 0.37 6.58 7.88 27.92 82.45 -23.41 10 12.43 4.20 4.55 3.06 2.41 -15.54 23.18 6.06 One other important consideration is. all the isolated strains are conditioned to local en\ironmental niictiiations. Usually, many of the strains kept in laboratories are rarely suited to cooler environ- ments and many ot the algae are planklonic, therefore the disad- vantage of poor culture conditions or lack on benthic habitat of abalone post-larvae. It is often considered desirable to have pure strains of microalgae. This may allow farmers complete control over what they feed their spat and controlled experimentation or commercial production. It is not generally desirable to feed a single species of microalgae. several pure strains can be mixed to improve the nutrition of abalone post-larvae (Fallu. 1991). Con- sideration of the chemical composition as protein, carbohydrates and lipid content of the strains, and abalone post-larvae require- ments, could lead to the production of a more nutritional food value from diatoms promoting higher survival and growth rates of abalone. The eight strains of diatoms used in this study offer a good alternative to be used in the nutrition of abalone post-larvae. These strains are acclimated to local environmental fluctuations (tem- perature and light principally). However, it is necessary to evaluate the nutritional value of these local diatom strains for abalone post- larvae, and relate their biochemical composition, adhesion to the substrate and digestibility with respect to the survival and growth of abalone post-larvae. ACKNOWLEDGMENTS We are grateful to Consejo Nacional de Ciencia y Tecnologia (CONACYT) who granted a Ph.D. scholarship to the first author. We thank Israel Gradilla-Martinez and Nestor Vallez-Villareal for technical assistance on the preparation of the scan electron pic- tures. We also thank Juan Pablo Lazo-Corvera for manuscript review and two anonymous reviewers. This project was supported by funding from Centro de Investigacion Cientifica y de Educa- cion Superior de Ensenada (CICESE) projects 623101 and 623108. REFERENCES Cahoon, L. B.. G. R. Beretich. C. J. Thomas & A. M. McDonald. 1993. Benthic microalgal production at Stellwagen Bank. Massachusetts Bay. Mar. Ecol Progr. Ser. 102:179-185. Correa-Reyes. J. G. 1993. Alimentacion de .\ricmia frunciscana con nii- croalgas cultivadas bajo diferentes tipos de luz. Bach. Thesis. Facultad de Ciencias Marinas, Baja California. Mexico: Universidad AiiUinoma de BaJa California. 56 pp. Cota-Sanchez. F. 1998. Culti\o mono y pluriespeciTico de tres diatomeas bentonicas en condiciones de luz y temperatura constantes. Bach. The- sis. Facultad de Ciencias del Mar. Sinaloa. Mexico: Universidad .''lU- tononia de Sinaloa. 90 pp. Cuevas-Rocha. F. 1998. Competicion e inhibicion del creciniiento de tres diatomeas bentonicas de la laguna de Venecia. Bach. Thesis, Facultad de Ciencias Marinas. Si'naloa. Mexico: Universidad Autonoma de Si- naloa. 92 pp. Daume, S.. S. Brand & W. J. Woelkerling. 1997. Effects of post-larval abalone Halii>ti.i rubra grazing on the epiphytic diatom assemblage of coralline red algae. .Mp//. Res. 18(2): 1 19-130. Dubinsky. Z.. K. Matsukawa & I. Karube. 1995. Photobiological aspects of algal mass culture. / Mar. Biolecli. 2(2):61-65. Hbert, E. E. & J. L. Houk. 1984. Elements and innovations in the cultiva- tion of red abalone Halions nifescens. Aqitacitltiire 39:375-392. Fallu. R. 1991. Abalone Farming. Oxford: Fishing News Books. 196 pp. Flaak, A. R. & C. E. Epifanio. 1978. Dietary protein levels and growth of the Oyster Cra.'ssostrea virginica. Mar. Biol. 15:157-163. Fleming. A. E., R. J. Van Barmeveld & P. W. Hone. 1996. The develop- ment of artificial diets for abalone. .Acpiaculiure 140:1—4. Flores-Vergara. C. 1998. Crecimiento y composicion bioqui'mica de mi- croalgas bentonicas cultivadas bajo diferentes condiciones de temper- atura e intensidad de lu/. Thesis. Baja California, Mexico: Centro de Investigacion Cientffica y de Educacion Superior de Ensenada (CICESE). 101 pp. Gostan. J.. C. Lechuga-Deveze & L. Lazzara. 1986. Does blue light effect the growth oi Chaetoceros prouiberans (Bacillariophyceael. ./. Phycol 22:63-71. Guillard. R. L. L. i; J. H. Rhyter. 1962. Studies on marine planktonic 610 Correa-Reyes et al diatoms 1. Cyclotelki nana Hustedt and Dttonida confenacea (Cleve) Gran. Can. J. Microbiol. 8:229-239. Hahn, K. O. 1989. Handbook of Culture of Abalone and Other Marine Gastropods. Boca Raton. FL: CRC Press. 348 pp. Hoshaw. R. W. & J. R. Rosowski. 1973. Methods for microscopic algae. 53-68. In: J.R. Stein, editor. Handbook of Phycological Methods. Cul- ture Methods and Growth Measurements. Cambridge, UK: Cambridge University Press. 448 pp. Kawamura. T. & R. Hirano. 1992. Seasonal changes in benthic diatom communities colonizing glass slides in Aburatsubu Bay, Japan. Diarom Research 7(2):227-239. Kawamura, T., T. Saido, H. Takami & Y. Yamashita. 1995. Dietary value of benthic diatoms for the growth of post-larval abalone Haliotis discus hannai. J. Exp. Mar. Biol. Ecol. 194:189-199. KawaiTiura, T.. R. D. Roberts & C. M. Nicholson. 1998. Factors affecting the food value of diatoms strains for post-larval abalone Haliotis iris. Aqiiacnilure 160: 81-88. Kromkamp. J., C. Barranguet & J. Peene. 1998. Determination of micro- phytobenthos PSII quantum efficiency and photosynthetic activity by means of variable chlorophyll fluorescence. Mar Ecol. Prog. Ser. 162: 45-55. McMullen. J. & T. Thompson. 1989. Abalone culture in an urban envi- ronmental. 227-238. In: K.O. Hahn, editor. Handbook of Culture of Abalone and Other Marine Gastropods. Boca Raton, FL: CRC Press. 348 pp. McBride, S. C. 1990. The relative effects of diet and irradiance on the growth and survival of post-larval red abalone, Haliotis nifescens. The- sis. San Jose, CA: Faculty of The Department of Biology, San Jo.se State University. 45 pp. Nalewajko, C. 1989. Notes of Training Course of Physiological and Bio- chemical Techniques in Microalgal Research. Scarborough Campus of The University of Toronto. 54 pp. Navarro, N. 1981a. A survey of the marine diatoms of Puerto Rico. I. Suborders Coscinodicinae and Rhizosoleniineae. Botanica Marina 24: 615-630. Navarro, N. 1981b. A survey of the marine diatom of Puerto Rico. II. Suborder Biddulphiniineae: families Biddulphiaceae and Eupodis- caceae. Botanica Marina 24:615-630. Navarro, N. 1982. A survey of the marine diatom of Puerto Rico. III. Suborder Biddulphiniineae: families Chaetoceraceae. Botanica Marina 24:305-319. Navarro. N. 1983. A survey of the marine diatoms of Puerto Rico. Botanica Marina 26:393^08. Roberts. R., T. Kawamura & H. Takami. 1999. Morphological changes in the radula of abalone [Haliotis iris), during post-larval development, / Shellfish Res. 8(2):637-644. Roberts. R., T. Kawamura & H. Takami. 2000. Diatoms for Abalone Culture: a workshop for abalone farmers. Prepared for 4th International Abalone Symposium. Cawthron Report No. 547. New Zealand: Caw- thron Institute. 28 pp. Round, F. E.. R. M. Crawford & D. G. Mann. 1990. The Diatoms. Cam- bridge, UK: Cambridge University Press. 747 pp. Sanchez-Saavedra. M. P. & D. Voltolina. 1994. Cultures of Pavlova liilbcn (Droop) Green (Prymnesiophyta) in diluted wastewater. J. .\ppi Phy- col. 6(3):285-288. Sanchez-Saavedra. M. P. & D. Voltolina. 1995. The effect of different light quality on the food value of the diatom Chaetoceros sp. for Arteinia franciscana Kellogg. Riv. Italian. Aqnacolt. 30:135-138. Sanchez-Saavedra. M. P. & D. Voltolina. 1996a. Effect of blue-green light on growth rate and chemical composition of three diatoms. J. Appl. Phycol. 8:131-137. Sanchez-Saavedra, M.P. & D. Voltolina. 1996b. The effect of different photon fluence rates of blue-green light on the biomass quality of a coastal diatom in pilot scale, semicontinuous cultures. Monographic Book "Underwater Light and Algal Photobiology". 5c/. Mar. 60(Supl. 1 ):267-272. Senger. H. 1987. Blue Light Response: Phenomena and Occurrence in Plants and Microorganism. Boca Raton. FL: CRC Press. Vol. I. 160 pp.. Vol. II. 169 pp. Simental-Trinidad. J. A. 1999. Produccion masiva de diatomeas bentonicas empleando fertilizantes agricolas y su efecto en la calidad de biomasa y en la composicion proximal. Thesis. Baja California. Mexico: Centro de Investigacion CientiTica y de Educacion Superior de Ensenada. 1 15 pp. Simonsen, R, 1987. Atlas and Catalogue on the Diatom Types of Hustedt I-III. Kramer, Berlin. 238 pp. Siqueiros-Beltrones, D. A. 1994. Estudios sobre diatomeas bentonicas en litorales de la Peninsula de B.C. In: D. Siqueiros-Beltrones. editor. Mem. IX Simp. Int. Biol. Mar. La Paz. Mexico: Universidad Autcinoma de Baja California Sur. pp. 65-79. Siqueiros-Beltrones, D. A. 1999. Analisis Taxonomico y sucesional en peliculas de diatomeas bentonicas utilizadas como alimento de poslar- vas de abulon {Haliotis spp.) en cultivo. Thesis, Universidad Autonoma de Baja California. Baja California, Mexico: Facultad de Ciencias Ma- rinas. Posgrado en Oceanogratla Costera. 93 pp. Siqueiros-Beltrones. D. A. & E. Sanchez-Castrejon. 1999. Association structure of benthic diatom assemblages from a mangrove environment in a Mexican subtropical lagoon. Biotropica 31(1 ):48-70. Siqueiros-Beltrones, D. A. 2000. Benthic diatoms associated with abalone (Haliotis spp.) on a rocky substratum from Isia Magdalena, Baja Cali- fornia Mexico. Ocechiides l5(l):35-46. Siqueiros-Beltrones, D. A. & D. Voltolina. 2000. Grazing selectively of red abalone poslarvae on benthic diatoms films under culture conditions. J. World Aquacolt. Soc. 31(2):239-246. Sorokin, C. 1973. Dry weight, packed cell volume and optical density. In: J.R. Stein, editor. Handbook of Phycological Methods. Culture Meth- ods and Growth Measurement. New York: Cambridge University Press, pp. 321-343. StatSoft. Inc. 1996. Statistica for Windows. Computer Program Manual. Tulsa. OK: StatSoft. Inc.. Thompson, P. A., P. J. Harrison & J. N. C. Whyte. 1992. Influence of irradiance on the fatty acid composition of phytoplankton. J. Phycol. 28:481-488. Voltolina. D. 1985. Biomass evaluation in cultures of benthic diatoms: an experimental review of methodology. Coastal Marine Science Labora- tory Internal Manuscript Series 85:4. Victoria. B.C.: Royal Roads Mili- tary College, 86 pp. Voltolina. D. 1994. Cultivos de microalgas bentonicas para alimentacion dejuveniles de abulon (Haliotis spp.). In: D. A. Siqueiros-Beltrones, D. Voltolina & A. M. Jaramillo-Legorreta, editors. Topicos Selectos Sobre Microalgas. Serie Cientifica. Universidad Autonoma de Baja California Sur, 145 pp. Jdiinwl of Shclltlsh Rcecinli. Vol. 20. No. 2, 611-617. 2001. BIOCHEMICAL COMPOSITION OF BENTHIC MARINE DIATOMS USING AS CULTURE MEDIUM A COMMON AGRICULTURAL FERTILIZER J. A. SIMENTAL-TRINIDAD, M. P. SANCHEZ-SAAVEDRA, AND J. G. CORREA-REYES L(ih(>ra!(iriii de Biologfa y Citliivo de Micnuiliias del Departaincnio de Acuicidtura. Centra de liuestlgcicidn Cientifica y de Ediicack'm Superior de Ensenadu. Apdo. Postal 2732. Ensenada. Baja California, Mexico. C.P. 22800 ABSTRACT Three strains of benthic marine diatoms (Nitzschia thermalis var. minor. Nitzschia laevix and Navicula imerta) were grown Indisidually in batch systems with 10 liters of non-conventional culture media formulated with three common agricultural fertilizers. The quantity and quality of the hiomass produced using non-conventional culture media were compared to those obtained with a traditional culture media. ' 172". The objective of this study was to consider a low cost and alternative media to culture the benthic diatoms commonly used in cominercial abalone hatcheries. No significant differences were found among the hiomass produced with non-conventional or control media. The trend in chemical composition of each diatom culture had a significant difference during 10 days of experimentation for all species. For the cultures of diatoms Nitzschia thermalis var. minor. Nitzschia laevis and Navicula incerta. the chemical composition did not differ between media. Our results suggested non-conventional media did not affect biomass or biochemical composition and it cost eight times less than "f/2". A£>' WORDS: Benthic diatoms, agricultural fenilizer. chemical composition, mass culture INTRODUCTION When undertaking laboratory research on molluscs, and/or when commercially producing these, large quantities of microalgal biomass are needed as the main source of food recjuired for their development and normal growth (Fabregas et al. 1987). Benthic diatoms play an important role in abalone culture, acting as induc- tors for larval settlement, and as food for the early juvenile stages before these can ingest macroalgae (Searcy-Bemal et al. 1992. Kawamura et al. 19951. The biochemical composition of niicroaigae is dependent on numerous factors, including the nutrient concentration and com- position of the growth medium (Fabregas et al. 1986, Wilkfors 1986), temperature (Thompson et al. 1992), light intensity and wavelength (Figueroa et al. 1996, Sanchez-Saavedra & Voltolina 1996), photoperiod (Sicko-Goad & Andersen 1991) and growth stage at time of harvest (Dunstan et al. 1991 ). It can. therefore, be substantially altered by manipulating their culture conditions (Du- binsky et al. 1978. Ben-Amot/ et al. 1987, Geldenhuys et al. 1988). Generally, where benthic diatoms are used in the produc- tion of abalone postlarvae, the nutrient enrichment of culture me- dium is not used to avoid increase of production costs (Ebert & Houk 1989). However, algal biomass production obtained is sub- stantially less than that produced within an enriched medium. To obtain high quantities of diatom biomass of suitable nutritional quality, the development of niicroaigae cultures using alternate or non-conventional simplified and economical culture media be- comes necessary. One possible source of economical nutrients for microalgae cultures is agricultural fertilizers. They have been used successfully to culture Isochrysis aff. galbana and Chaetoceros mulleri (Valenzuela-Espinoza et al. 1999). One important consideration in the formulation of culture me- dia is that they must have an adequate proportion of nutrients in relation to the nutritional requirements of the microalgae (Valen- zuela-Espinoza 1997). It is also important that the production of benthic diatoms be of high density and of good nutritional quality for the abalone postlarvae. All this should be achieved at low cost (De la Cruz & .Alfonso 1975. Gonzalez-Rodriguez & Maestrini 1984), The culture media prepared with laboratory grade reagents are expensive (Simental-Trinidad 1999). It is also not practical to use these for aquacultural purposes since the conditions of com- mercial culture differ from laboratory conditions. One possible solution to this problem is to use the more economical agricultural fertilizers as culture media (Gonzalez-Rodriguez & Maestrini 1984, Simental-Trinidad 1999). Commercial production of mollusc larvae implies the produc- tion of massive cultures of high quality microalgae, since these are indispensable for the adequate nourishing of larvae, postlarvae or adults. This is where a "bottle neck" is found in commercial farms. Thus, several alternatives are being sought to decrease the produc- tion costs of culture media for microalgae. The aim of this work was to evaluate the chemical composition of three benthic diatoms Nitzschia thermalis var. minor. Nitzschia laevis and Navicula in- certa cultured in an alternative media made with commercial ag- ricultural fertilizer as a supplement of nitrogen, phosphate and silica compounds required to grow diatom cultures. MATERIALS AND METHODS For this work, we used the benthic diatoms Nitzschia thermalis var minor (29.57 p.), N. laevis (7.56 p.) and Navicida incerta ( 12.82 (x) (Fig. 1 ). Both Nitzschia species are potentially useful in abalone aquaculture and were isolated from production farms in the locality of Ensenada, Baja California, Mexico (Correa-Reyes et al. in preparation b). Navicula incerta (12.82 p.) was isolated in coastal waters off Bahi'a Todos Santos, Mexico by the staff of the Microalgae Laboratory of the "Instituto de Investigaciones Oceanologicas, Universidad Autonoma de Baja California, Mexico". The strain of Navicula incerta was obtained from the abalone farm "Abulones Cultivados", Ejido Erendira, Baja Cali- fornia, Mexico. The commercial medium "f/2" (Guillard 1975) was used as the control culture medium in all the experiments. Nutrients available fiom non-conventional cultured medium lacked trace metals, and vitamins were added in an equal concentration to that indicated in the formulation of the control medium "f/2"" (Table I). We used liquid agricultural fertilizers for the preparation of the non- conventional medium. The fertilizers were obtained from '"Insu- mos y Servicios Agrologicos S.A. de C.V."", Ejido Camalii, Baja California. Mexico. Nitrogen density was 1.37 grP', total nitrogen 611 612 Simental-Trinidad et al. Figure 1. Benthic diatoms Nitzscliia laevis (A), Nilzscliia thcrmalis var. minor (B) and Narkulu incerla (C). was 32% and the cost US$0.39 1"'. Phosphorous density was 1 .69 grl"', total phosphorous was 72% of phosphoric acid and the cost US$0.78 1~'. SiHca was as sodium niethasilicate with 0.107 moles r' and was obtained from a distributor "Quimica Intemacional de Tecate S.A. de C.V.". Mexico, at USS0.85 P'. Cultures were prepared from the production of inoculum in progressive volumes of 10 ml, 250 ml and 900 ml. Each of the three species was acclimated to the cultured medium before ex- perimentation. These cultures were maintained under controlled light conditions (150 jJiEm"" s~'), environmental temperature (21 ± I T) and pH (between 8.7 and 9.0). Batch cultures were kept in circular 18 1 plastic containers of 38 cm by 30 cm. White plastic containers were used for the assay. The containers were provided with a lexon acrylic lid with two orifices, one for the incoming air and the other for the outcoming air. Air was not injected to the culture medium and it was only used on the surface to create a circulation cell to maintain a constant tempera- ture. The seawater for the culture medium was disinfected chemi- cally following the methodology described by Hemerick ( 1973). A sampling strategy, without sampling replacement, was made for biomass quality and quantity, due to the benthic diatoms char- acteristics and the presence of a biofilm on the bottom of the container. Sampling for each species and culture medium consisted of taking two containers evei7 48 hours and harvesting the whole culture to determine the proximal composition and biomass in triplicate. To determine microalgal biomass, microalgae cells were counted in a 0.1 -mm depth haematocytometer with the aid of a compound microscope. For the proximal analyses, the microalgae were harvested from each container, concentrated and placed in ultrasound water to eliminate possible lumps. Subsamples were taken from the con- centrate and were reconcentrated in Poretics GF-C glass fiber fil- ters (2.5 cm diameter) previously rinsed with distilled water, and incinerated at 470 °C for four hours. After filtering, the samples were stored at -20 °C as recommended by Cordero-Esquivel et al. (1993). The analytical technique used for total and ash free dry weights was as those described by Sorokin (1973). Samples were dried at 60 "C to constant weight in a conventional oven, and ashed in a muffle furnace at 470 "C for eight hours. Analysis of proteins were performed by varying the extraction time and NaOH concentrations for optimum extraction, as pro- posed by Correa-Reyes et al. (in preparation a). Protein determi- nation was undertaken using the protocols proposed by Lowry et al. (1951). Carbohydrates were extracted following the Whyte method ( 1987) and their determination was made according to the methodology of Dubois et al. (1956). Lipids were extracted fol- lowing the methodology of Bligh and Dyer ( 1959) and their quan- tification was estimated by the method of Pande et al. (1963). Calibrations curves were obtained for each one of the analyses following the indications described by the colorimetric techniques specified for each case, using a Shimadzu spectrophotometer model UV-1201. Bovine albumin (98%) was employed as a pro- tein standard, glucose (99%) as a carbohydrate standard and tri- palmitin (99%) as a lipid standard. We used a covariance statistical analysis to determine differ- ences between the productions obtained within different culture media and the three strains. RESULTS Nitzschia laevis The percentage of protein content (or Nitzschia laevis (Table 2) cultured in "■f/2"" medium, had a maximum (31.257) on the eighth day of culture, while that cultured In the experimental media showed the highest protein percentage (25.470) on the fourth day. There were significant differences in the protein percentage be- tween both treatments (F = 18.821, P = 0.000). There were no significant differences in carbohydrate (F = 2.726, P = 0. 115) and lipid percentage (F = 0.878, P = 0.360) for this strain cul- tured on the experimental medium. Organic dry weight and ash content increased significantly with culture age and were similar to those obtained when the diatom were grown on the experimental medium (Table 2). Nitzschia thermalis var. minor Nitzscliia thermalis var. minor (Table 3), showed a maximum protein percentage (33.505) on the fourth day of culture in the control medium. The microalgae cultured in the experimental me- dium reached the maximum protein percentage (35.579) on the second day of culture. There were no significant differences be- tween both treatments (F = 3.233, P = 0.084). I Benthic Diatom Culture with Fertilizer 613 TABI.E 1. Chemical composilion and cost for 1000-1 of prepared culture media used to grow benthic diatoms. Agricultural fertilizers were used as experimental media and "f/2" media as Guillard and Ryther (19751 for control medium. Agricultura Fertilizer "f/2" Medium Source Chemical form Ionic form Cost (US $) Chemical form Ionic form Cost (US $) (fimol) Nilrale NO3 Nitrooen Animonium NH4 3.09 X 10 " Sodium nitrate NaNO 1.01 X 10-- Urea CO(NHj; 882.0 Phosphorus Phosphoric acid H,POj 2.26 X 10"" Sodium phosphate NaHjPOj ■ H,0 5.90 X lO--* 36.3 Silica Sodium methasilicate Na,SiO, ■ 9H,0 4.2.S X \i.)-' Sodium methasilicate Na,SiC , ■ 9H,0 6.80 X 10-^ 107.0 Common chemicals for the tw( media Source Chemical form Ionic form Cost (US $) (pniol) Ferric c hloride FeEDTA 1.05 X 10-' 23.3000 Copper sulphate CuSO_ • 5H,0 l.lOx 10-" 0.0395 Trace metals Zinc su Iphate ZnSOj •7H,0 2.30 X 10-'' 0.0765 Cobaltous chloride CoCl, •6H,0 1.90 X 10-" 0.0425 Manganese chloride MnCl, ■4H,0 2.14 X 10-^ 0.9000 Sodium molihdate Na,Mo04 ■ 2H,0 1.23 X 10-=^ 0.0260 Biotin C|||H| ,N,0,S 4.20 X 10-' 1.336 X 10-" Vitamins Cianocobalamin Q„H«sCoN„P 3.90 X 10-' 0.393 X 10-" Thiamin C„H, CINjOS 4.90 X 10-' 0.18726 The carbohydrate content in niicroalgae cultured in the control medium had a ma.ximum peicentage (23.923) on the second day of culture, while the experimental group did not show significant differences along the whole culture period. There were differences between both treatments only on the second day of culture in which the cultured niicroalgae in the control medium obtained the highest percentages (Table 3) (F = 7.694. P = 0.012). The lipid percentage in the experimental culture showed the highest values (21.06) on the tenth day of culture, while the cul- tured microalgae in the experimental medium showed the highest values from the sixth (34.247) to the tenth day (44.549). There were significant differences between both treatments onlv on the tenth culturing day. on which the cultured microalgae in the con- trol medium had a higher percentage (F = 6.791, P = 0.015). The fraction of organic dry weight and ash content showed a similar trend to those of Nitzschia laevis. Their final concentrations were also similar to those obtained for that species (Table 3). Navicula incerta Navicida incerta showed a maximum protein percentage during days two (35.153) and four (35.405) when cultured in the control medium, whereas when cultured in the non-conventional medium, maximum percentage was obtained during the second day (33.530) TABLE 2. Average proximal composition (expressed in percentage) based on the dry weight organic of Nitzschia laevis cultured in 18-1 plastic containers with ■■f/2" culture media (Guillard & Ryther 1975) and with agricultural fertilizers. Culture medium Dav Cell m|- ^f Proteins % Carbohydrates Vf Lipids % Ash 2 144687a 22.533a (30937) (0.71) 4 374100b 25.625a "f/2" medium (96638) (1.68) 6 73945()b 27.221a (136000) (0.11) 8 1133300c 31.257b (51901 1 (0.46) 2 155600a 18.493a (19632) (0.53) 4 47165()b 25.470b Agricultural fertilizers (95147) (0.93) medium 6 736850c 22.731b (149437) (0.81) l()65750d 27.632b 8 (18879) (0.51) 5.310a (0.49) 7.074a (0.48) 8.676a (0.85) 13.563b (2.21) 15.907a 60.560a (1.48) (7.25) 18.339a 54.121a (1.62) (7..55) 13.183a 46.016a (1.29) (4.23) 14.356a 43.686a (0.98) (0.69) 13.380a 58.385a (1.39) (3.55) 21.735a 51.722a (1.54) (8.04) 12.060a 53.588a (2.16) (4.91) 21.913a 39.414a (3.05) (9.71) 7.813a (1.37) 8.997a (1.90) 8.739a (1.59) 14.824b (1.78) The standard deviation is mcluded in parenthesis. The different letters on the side of the quantities indicate significant differences (two-way Anova and Tuckey's 1/ poslcridii test, a = 0.051: a < b < c. 614 Simental-Trinidad et al. TABLE 3. Average proximal composition (expressed in percentage) based on the dry weiglit organic of Nitzschia thennalis rar. minor cultured in 18-1 plastic containers with "f/2" culture media iGuillard & Ryther, 1975) and with agricultural fertilizers. Culture medium Day Cell density % Proteins % Carbohydrates % Lipids % Ash ■'f/2" Medium 10 65468a * (11782) 83125a 33.505b (12311) (0.90) 88750a 25.473a (16250) (0.98) 73750a 25.497a ( 1 3994) (0.99) 69674a 29.697a (12521) (3.68) 34843a 35.579b (15732) (4.96) 77812a 28.930a (7739) (0.55) 63112a 29.700a (15909) (1.43) 94687b 26.557a (34248) (4.08) 88906b 26.404a (1767) (1,1)0) 23.923b (2.75) I2.11()a (1.35) 8.679a (0.74) 7.255a (0.61) 8.321a (0.38) * 52.537a (3.03) 21.374a 48.205a (1.12) (0.28) 39.808b 33.731a (1.66) (4.09) 41.201b 34.788a (1.82) (2.87) 58.601c 38.192a (841) (6.11) 24.370a 50.150c (2.44) (1.15) 18.973a 44.427a (1.72) (2.23) 34.247b 34.875b (2.22) (5.98) 41.340b 31.041b (0.67) (3.12) 44.549b 31.993b (3.12) (5.08) Agricultural fertilizers medium 10 1 1 .453a (1.86) 12.831a (1.87) 10.695a (1.54) 9.037a (0.95) 8.875a (0.57) *Not measured The standard deviation is included in parenthesis. The different letters on the side of the quantities indicate significant differences (two-way Anova and Tuckey's a posteriori test, a = 0.05): a < b < c. TABLE 4. Average proximal composition (expressed in percentage) based on the dry weight organic of Naviciiia iiiccrla cultured in 18-1 plastic containers with "f/2" culture media (Guillard & Ryther 1975) and with agricultural fertilizers. Culture medium Day Cell density Vf Proteins fc Carbohydrates 'ii Lipids % Ash "f/2'" Medium 10 9715()a 35,153a (43838) (0,98) 45965b 35.405a (37825) (4.31) 464262c 30.518a (6884) (5.16) 5302S7d 30.973a (68634) (1.27) 5.35625d 26.336a (88838) (7.24) 62150a 33.530a (22899) (6.03) 263100b 30.159a (34791) (2.14) 433725c (106521) * 525937c 25.618a (80247) (2.15) ,345387b 23..344a (17677) (1..36) 7.182a (1.09) 11.129a (1.45) I 1 .02 1 a (0.81) 12.082a (2.03) 13.693a (1.84) 9.242a 39.997a (1.16) (3.93) 13.288a 31.089a (0.62) (10.85) 14.289a 3 1 .835a (1.30) (5.96) 15.161a 31.447a (1.02) (5.58) 1 3.783a 28.007a (0.72) (4.17) 9.820a 39.813a (1.13) (7.06) 12.205a 40.727a (1.42) (11.62) 13.034a 41.107a (0.95) (0.63) 10.692a 35.6-36a (0.58) (3.97) Agricultural fertilizers medium 10 7.589a (1.18) 13.193b (0.77) 10.838b (0.31) 12.886b (0.74) * Not measured The standard deviation is mcluded in parenthesis. The different letters on the side of the quantities indicate significant differences (two-way Anova and Tuckey's a posteriori test, a 0.05): aacioncs Ma- rinas 17:1-25. Dortch, Q., P. A. Thompson & P. J. Harrison. 1991. Short-term mteraction between nitrate and ammonium uptake in Tlialassiosira pseiiclonana: effect of preconditioning nitrogen source and growth rate. Mar. Biol. 110:183-193. Dubinsky. Z.. T. Berner & S. Aronson. 1978. Potential of large-scale algal culture for biomass and lipid production in arid lands. Bimechnol. Bioeng. Symp. 8:51-58. Dubois, M.. K. A. Gilles. J. K. Hamilton, P. A. Rebers & F. Smith. 1956. Colorimetric method for determination of sugars and related sub- stances. Analyt. Chen,. 28:350-356. Dunstan, G. A., J. K. Volkman. S. M. Barret & CD. Gadand. 1993. Changes in lipid composition and maximization of the polyunsaturated fatty acid content of three microalgae grown in mass culture. J. Applied Phyeol. 5:71-83. Ebert, E. E. & J. L. Houk. 1989. Abalone Cultivation Methods Used at California Department of Fish and Game's Marine Resources Labora- tory. 239-254 pp. In: K. O. Hahn, editor. Handbook of Culture of Abalone and Other Marine Gastropods. Boca Raton: CRC Press. 348 pp. Fabregas, J., C. Herrero, B. Cabezas & J. Abalde. 1986. Biomass produc- tion and biochemical composition in mass culture of the marine mi- croalga Isochrysis galbana Parke at varying nutrient concentration. Aquaculture 53:101-1 13. Fabregas, J., L. Toribio, J. Abalde, A. Cabezas & C. Herrero. 1987. Ap- proach to biomass production of the marine microalgae Tetraselmis suecica (Kylin) Butch using common garden fertilizer and soil extract as cheap nutrient supply in batch cultures. Aquacul. Engin. 6:141-150. Figueroa, F. L., C. Jimenez. J. L. Perez-Llorens & F. X. Niell. 1996. Un- derwater Light and Algal Photobiology. Scienlia Marina 60 (Supl. 1): 1-343. Floras- Vergara, C. 1998. Crecimiento y composicion bioquimica de mi- croalgas bentonicas bajo diferentes condiciones de temperatura y luz. M.Sc. thesis. Baja California, Mexico: Centro de Investigacidn Cien- tifica y de Educacion Superior de Ensenada. 101 pp. Geldenhuys, D. J.. R. D. Walmsley & D. F. Toerien. 1988. Quality of algal material produced on a fertilizer-tap water medium in outdoor plastic- enclosed systems. Aquaculture 68:157-164. Guillard. R. L, L. & J. H. Ryther. 1962. Studies on marine planktonic diatoms I. Cyclotella nana Hustedt and Detomila confervacea (Cleve) Gran. Can. J. Microbiol. 8:229-239. LITERATURE CITED Guillard, R. L. L. 1975. Culture of phytoplankton for feedmg marme m- vertebrates. 29-60. In: M. L. Smith & M. H. Chanley. editors. Culture of Marine Invertebrates Animals. New York: Plenum Press, 338 pp. Gonzalez-Rodriguez, E. & S. Y. Maestrini. 1984. The use of some agri- cultural fertilizers for the mass production of marine algae. Aquacuhure 36:245-256. Gracida-Valdepefia, M. L. 1999. Produccion de tres especies de diatomeas uulizando una mezcla de fertilizantes agri'colas. M.Sc. thesis. Baja California, Mexico: Centro de Investigacidn Cientifica y de Educacion Superior de Ensenada. 108 pp. Hemerick, G. 1973. Mass culture. 255-273. In: J, R. Stem, editor. Hand- book of Phycological Methods. Culture Methods and Growth Measure- ment, New York: Cambridge University Press. 448 pp. Kawamura, T. 1996. The role of henthic diatoms in the eariy life stages of the Japanese abalone (Haliotis discus hannai). In: Y. Watanabe, Y. Yamashita & Y. Oozeki, editors. Survival Strategies in Eariy Life Stages of Marine Resources. Yokohama, Japan: Proceedings of an International Workshop, 12 pp. Kawamura, T. & R. Hirano. 1992. Seasonal changes in henthic diatom communines colonizing glass slides in Aburatsubu Bay. Japan. Diatom Res. 7(21:227-229. Kawamura. T., T. Saido, H. Takami & H. Yamashita. ^Ws. Dietary value of benthic diatoms for the growth of post-larval abalone Haliotis di.scus hannai. J. Exp. Mar. Biol. Ecol. 194:189-199. Ldpez-Eli'as, J. A. & D. Voltolina. 1993, Cultivos semicontinuos de cuatro especies de microalgas con un medio no convencional. Ciencias Ma- nnas 1 9(2): 169- 180. Lowry, O. H.. N. J. Rosehrought, A, L, FaiT & R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chein. 193:265- 275. McAnally-Salas, L. S., F. Ocampo-Aranda & L, E. Garcia-Pamanes. 1992. Efecto de la microalga Pavlova lutlieri (Droop) cultivada con fertili- zantes agri'colas en el crecimiento y supervivencia de larvas y postlar- vas del mejillon Mylilus edidis (L). Ciencias Mariims 18(41:57-74. Pande, S. V„ R. P. Khan & T. A. Venkitasubramanian. 1963. Microdeter- minanon of lipids and serum total fatty acid. Analyt. Biochem. 6:415- 423. Parsons, T R., Y. Malta & C. M. Lalli. 1984. A Manual of Chemical and Biological Methods for Seawater Analysis. Oxford: Pergamon Press. 173 pp. Perera-Carbonell, C, 1994, Uso de fertilizantes para cultivos hidroponicos como medio de crecimiento de dos especies de microalgas marinas. Rev. Invest. Mar. 15(2):135-137. Roberts, R., T, Kawamura & H, Takami. 1999. Morphological changes in the radula of abalone (Haliotis iris), during post-larval development. J. Shellfish Res. 8(21:637-344. Round, F. E.. R. M. Crawford & D. G. Mann. 1990. The Diatoms: Biology and Morphology of the Genera. Cambridge: Cambridge University Press. 747 pp. Sanchez-Saavedra, M. P. & D. Voltolina. 1996. Effect of blue-green light on growth rate and chemical composition of three diatoms, J. Applied Phvcol. 8:131-137. Searcy-Bernal, R., A. Salas-Garza, R. A. Flores-Aguilar & P. R. Hinojosa- Rivera. 1992. Simultaneous comparison of methods for seUlement and metamorphosis induction in the red abalone {Haliotis rufescens). Aqua- culture 105:241-250. BhNTHic Diatom Culture with Fertilizer 617 Sicko-Goad. L. & N. A. Andersen. 199L Effect of growth and light/dark cycles on diatom lipid content and composition. J. Phycol. 27:710-718. Simental-Trinidad. J. A. 1999. Producciiin masiva de diatomeas bentonicas enipleando fertilizantes agn'colas y su efecto en la cantidad de biomasa y en la composicion proximal. MSc. thesis. Baja California, Mexico: Centro de Investigacion Cientifica y de Educacion Superior de Ensenada. 1 15 pp. Sorokin. C. 1973. Dry weight, packed cell volume and optical density. In: i. R. Stein, editor. Handbook of Phycological Methods. Cuhure Meth- ods and Growth Measurement. New York: Cainbridge University Press. 44S pp. StatSoft, Inc. 1996. Statistica for Wmdows. Computer Program Manual. Tulsa: StatSoft, inc. Syrett. T. 1981. Nitrogen Metabolism of Microalgae. In: T. Piatt, editor. Physiological Bases of Phytoplankton Ecology. Can. Bull. Fi.sh. Aqmit. Sci. 236 pp. Thompson, P. A.. G. Ming-Xing & P. H. Harrison. 1992. Effects of varia- tion in temperature on the biochemical composition of eight species of marine phytoplankton. J. PIncol. 228:481-188. Valenzuela-Espinoza. E. 1997. Uso de un medio alterno al "172'" para el cultivo de hochrysis tiff', aalhana (clone t-ISO). M.Sc. thesis. Facultad de Ciencias Marinas. Baja California, Mexico: Universidad Autcinoma de Baja California. 51 pp. Valenzuela-Espinoza. E., V. Gendrop-Funes, R. Perez-Castafieda & J. Wil- burn-Gonzalez. 1999. Larval survival of Litopcnnuens vannumei (BOONE) fed Cluwloceros muelleri produced with agricultural lertil- izers. Ciencias Marinas 25(3):423^37. Werner, D. 1977, The Biology of the Diatoms. Botanical Monographs. Vol. 13. Berkeley: University of California Press. 498 pp. Wheeler, P. .\. 1983. Phytoplankton Nitrogen Metabolism. In: E.J. Car- penter & D. G. Capone. editors. Nitrogen in the Marine Environment. New York: Academic Press. Whyte, J. N. C. 1987. Biochemical composition and energy content of six species of phytoplankton used in mariculture of bivalves. Aquaculuire 60:231-241. Wilklors. G. H. 1986. Altering growth and gross chemical composition of two microalgal molluscan food species by varying nitrate and phos- phate. .Aqiiaciiltiire 59:1-14. Jimiihil (ij Shillfish Research. Vol, 20, No, 2, (i|y-626. 2(M)I. ECOLOGY AND POST-SETTLEMENT SURVIVAL OF THE EZO ABALONE, HALIOTIS DISCUS HANNAI, ON MIVAGI COASTS, JAPAN RYO SASAKI' * AND SCORESBY A. SHEFHERD" ^Miyagi Prefecliinil Sea Fanniiii; Center. Yagawa Beach. Oshika Town. MIxagi Prefecture, 986-2402 Japan: -Soiitli Australian Research and Deveh>pnient Institute. FO Box 120, Henley Beach. South Australia 5U22 ABSTRACT Settlement and survival of the ezo abalone. Haliofis di-icus Imitnai were examined in near-shore habitats at exposed to sheltered sites on Miyagi coasts. Japan from l^yb-lQ^g, Larvae settled mainly in the boundary zone at 2-7 m depth according to site between shallower Eiseiiiu forests and deeper crustose coralline flats at densities of ~200-800/m-. Growth rates were 30—40 |j.m/day. Instantaneous natural mortality rates. M. in the first 55 days after settlement were 3,7-5.5/nionth. Sampling of the habitat during the week after settlement showed a 29-37% incidence of atrophied post-larvae of -500 pim shell length indicative of starvation. Laboratory experiments suggested that post-larvae exhaust the natural food supply on crustose coralline algae in the first week after settlement and begin to starve at a size of -450-500 (i.m in the absence of a suitable diatom diet. One-year-old juveniles were mainly found in the boundary zone but with increasing age juveniles moved into the shallower Eisenia forest where food algae is abundant. KEY WORDS: settlement, recruitment. sur\ival. density-dependent mortality, crustose coralline algae, isoyake. Hulioiis discus hannai INTRODUCTION An understanding of the ecology and dynamics of recrtiitment in abalone populations is crucial to management of their fisheries (McShane 1992. McShane 1996). especially when those fisheries are in decline through overfishing. Numerous studies have shown the importance of crustose coralline algae (CCA) as a preferred substratum for larval settlement of abalone (Saito 1981. Shepherd & Turner 1985) but few studies have examined the survival of settlers (Shepherd & Daume 1996) or the causes of mortality. Two hypotheses, predation and food availability, have been proposed to explain the early high mortality (M) of post-larval abalone, Morse et al. (1979) and Naylor and McShane (1997) suggested that infaunal polychaete predators controlled the abun- dance of abalone post-larvae and McShane (1991) thought that bulldozing by urchins may also inflict high M. A corollary of this hypothesis is that the characteristics of the CCA substratum may be critical, either by providing a protective micro-habitat, e.g. in the lumpy growth-form of certain CCA (Shepherd & Daume 1996). or conversely by favoring loss from wave-shear on smooth forms of CCA (McShane 1996). Alternatively, food availability has been considered to play a key role in the grow th and survival of post-larvae ( Kawamura et al. 1998). While growth and survival on a CCA substratum is superior to those on any other substratum in the first few days to a week after settlement (Morse 1984. Hooker & Morse 1985. Daume et al, 1999a). thereafter the quality, type and abundance of diatoms be- come critical for growth and survival. Both factors, predation and food availability, together were considered by Shepherd et al. (2000) to be the main determinants of the carrying capacity of the CCA habitat. Our eariier study on the spawning, larval dispersal and settle- ment of Haliolis discus hwmai Ino on Miyagi coasts (Sasaki & Shepherd 1995) showed that spawning was epidemic and induced by typhoons. This enabled the prediction of spawning events from the occurrence of typhoons and the estimation of the time of larval settlement. We monitored larval settlement of H. d. Iiannai fol- *Corresponding author. lowing annual typhoons from 1995 to 1999 in Samenoura Bay. Specifically we asked what the characteristics were of the habitat in which larvae of this abalone settled, in what density did they settle and were post-settlement processes limiting recruitment to wild populations. In this article we describe the preferred zone of settlement within the near-shore rocky reef ecosystem, post-larval survival and subsequent movement and habitat of juveniles to age five years. We also carried out ancillary experiments to elucidate the cause of post-settlement mortality. MATERIALS AND METHODS Description uj Study Sites The coast of Samenoura Bay in northeast Honshu Island where the study was done (Fig, 1) ranges from sheltered to exposed. Following the description of Taniguchi ( 1991 ) we recognize three distinct sublittoral zones according to algal dominants, named A-C in this article, on rocky bottom. Zone A is an Eisenia bicyclis forest going from low water to a depth of 2-4 m. The zone is shallowest in sheltered conditions and its depth increases with increasing exposure to wave action. Sar- gassuin yezoense is a sub-dominant at the shallowest depths. Zone B is a boundary transition zone between the Eisenia forest and the deeper CCA /one C. and is characterized by Acrosorium polyiieunim. and. to a less extent, by Clioiidnis occllatus and Dilo- phus oliamurai. Zone C. called the CCA flats (in Japanese isoyake). comprises boulders with high CCA cover but bare of erect algae and grazed intensively by the sea-urchin Slrongyloceiilrotus nudus. We have no data on the species of CCA present, but in southern Hokkaido the main species are (with mean percentage cover in brackets): Lithophyllum yessoense (69%). L. okamurai (59f ). Litholliaiiiniiini japonicum (8%) and Neogoniolithcm sp. (6%) (Noro et al. 1983), The distribution of common algal dominants (Taniguchi 1991) in the three zones is shown in Figure 2 and illustrated in Figure 3. Posl-lMnal Sampling We chose three sampling sites, numbered 1-3, to provide a range of exposures to wave action (Fig. I ). The three zones and the same algal flora were present at each site, but the zones differed in 619 620 Sasaki and Shepherd Pacific Ocean • Enoshima Is Samenoura Bay Pacific Ocean IM 20m depth Figure I. Tht coast of Miyagi Prefecture llefti and the study sites in Samenoura Bay (right! in northeast Honshu island, Japan (inset). horizontal extent. Site 1 is the most exposed, facing the Pacific Ocean, with a gentle slope of 4/100. Site 2 at the entrance of Samenoura Bay is moderately exposed with a bottom slope of 6/100 and Site 3 within the Bay has a slope of 6/100 (see Figs. .^. 4, 5 for the horizontal extent of the respective zones). At each site we estimated the abundance of newly settled aba- lone spat by taking, with the aid of SCUBA. 4-6 boulders, each 20-30 cm diameter, per sample at regular sampling intervals along a transect normal to the coast and intersecting each of the three zones. The boulders were taken to the laboratory and immersed in 0.4% formaldehyde solution; then the spat was washed off the substratum with a water jet, examined under a binocular micro- scope, and individual shell lengths measured. We also recorded those spat showing atrophy (i.e. shrinkage) of the soft tissue within the shell. The length, width and depth of each boulder were mea- sured and the total area of the sides and top of a rectangular box of those dimensions was taken as an estimate of the area of the boulders. Densities were converted to numbers/m". e ■3 4 - a. a Sargassum yezocnse ^^^^^^^^^^^^^^^^^ EJscnis ^Vl-w/i^^ Picfocladia .-spsllacea r.^,,a,am aman«, ChoDiinis ocellatus ~~' — -,,,.__^ Pachvmeaiopsis c Hi plica ^^"^^^ Calhanhfon ^'c^wt'^5c Aaosonuw polvueurum 60 80 100 120 Distance froin shoreline (m) Figure 2. Schematic horizontal distrihution with depth of major benthic algae at Site 1 (after Taniguchi 19911, The sloping line repre- sents rocky bottom, the thickness of the distribution lines indicates % cover and the dashed line at 6 m depth is the boundary betneen Zones B and C, Sampling of boulders for newly settled spat was done annually for 2 y at Site I, for 3 y at Site 2, and for 2 y at Site 3. In 1997, boulder sampling at Site Iwas repeated four times over 55 days and the natural mortality rate of spat estimated by regressing the natural logarithm of density versus time. At Site 1 we also deployed strings of 80 scallop shells anchored just above the bottom as artificial collectors at regularly spaced intervals to independently monitor spat settlement. The surface area of the shells was estimated and the results presented as num- bers/nr. Post-Larval Survival Experiments Two experiments, one using dried boulders in the field, and the other with cobbles covered with living CCA in aquaria, were car- ried out to measure post-larval survival in different conditions. In the former experiment we placed sun-dried boulders in Zone B at Site 1 and compared the number and size of post-larvae on dead substrata and nearby natural control substrata. We placed the boul- ders in the field on 22 September 1997 on the same day that settlement began, and retrieved them four days later. In the second experiment (called the aquarium experiment) 2,500 H. d. luiimai larvae were released into each of five 10-L aquaria with living CCA cobbles of -10 cm diameter set on the bottom and their survival monitored over 14 days. The cobbles had been previously brushed and washed in filtered sea-water to re- move predators and microalgae including diatoms. The aquarium water temperature was 1 8-20°C. The contents of the aquaria were preserved in formalin successively at two- to three-day intervals and the empty shells and those living to that time measured and counted. The aquarium experiment was repeated twice with the same protocols. Juvenile Sampling The abundance of juvenile (one to five y) abalone was estimated at each site in January 1998 by searching within each zone within a 5 x 5 m quadrat with three replicates. All such surveys were done by one of us (RS) to standardize searching Ecology and PosT-SETTLEMnNT Survival of the Ezo Abalone 621 Figuif 3. I'hdtographs of the bottom at Site 1 at the ciilraiice to Samenoura Bay. a: Zone A showinf; canop> of Kiscnia hicyclis; b: boundary Zone B showinj; low cover of the macroalgae, Acriisorium polyneiirum and Dilophiis okamurai. and c: Zone C showing sea-urchin Slron^ylocentrotus midiis in isoyake area. efficiency. Small abalone were aged by the annual growth check which is conspicuous in this species (Sakai 1960). RESULTS Distribution and Density of Post-Larvae The distribution of density of post-larvae along transect lines in relation to depth and habitat is shown in Figures 4-6. At Site 1 (Fig. 4) settlement was monitored on scallop-shell collectors and showed a maximum density at a distance of 160 m from shore at ."i m depth in the boundary zone (B). Mean post- larval shell length was 41. "i |xm (SE 60 \x.m) five days after settle- ment. In 1996 post-lar\al density ranged from 0-30/m" in Zone A to 22()/m~ in Zone B with intermediate values in Zone C. Mean shell length was 425 |xm (SE 75 \\.m). In 1997 with increased sampling intensity at 15 m intervals there is a clear modal density of 340 spat/ni" at a distance of 1 80 m from shore at the outer edge of Zone B. Mean shell length was 510 ijliii (SE 40 |jim). At Site 2 (Fig. 5) there was a peak in spat density in Zone B in 1997 (mean shell length 340 (xm, SE 20 ixm), nearly uniform spat density in 1998 in all zones to a depth of 6 m (mean shell length 910 (Jim, SE 830 \x.m). and in 1999 a modal density at 80 m from shore in Zone C (shell length 380 (xm. SE 45 fim). At Site 3 (Fig. 6) there was a weak settlement in 1998 with a peak density in Zone C (mean shell length 695 p.m. SE 575 |jLm). and a stronger settlement in 1999 with a peak density at the inner edge of Zone C (mean shell length 370 |xm, SE 50 jaiti). In summary the results show trends of declining spat density with increasing shelter and peak densities at the boundary between Zones B and C with lower densities further in- and off-shore at each site. Post-Liirral Growth and Mortality Length-frequency data at Site 1 after the 1997 settlement (Fig. 7). shows modes at 390 pm and 410 p.m on 25 September in- creasing to 490 (xm and 560 p.m respectively 6 days later, and finally a single apparent mode at 1-2 mm 13 days after that. In contrast, the "dead"" boulders showed a single peak at 290 ixm early after settlement. The length-frequency distribution of spat on "dead"" boulders differed significantly from that on natural boul- ders obtained in the same period (x"^ = 42.2; P < 0.001 ). but the mean size of spat did not differ significantly between the two sets of boulders (t = \.6: P = 0.1) although mean size was 10 |xm smaller on the dead boulders. The density of post-larvae at Site I sampled up to day 55 after putative settlement declined exponentially (Fig. 8). The estimated instantaneous monthly mortality rale, M, for the first 27 days was 5.3 (SE 0.7) and for the whole 55 days was 3.7 (SE 0.6). Length- frequencies of post-larval samples obtained at Site 1 after the 1996 and 1997 settlements (Fig. 9) showed that the mean length of spat with atrophied foot was 504 (xm (SE 78 (xm) and that of normal spat was 387 p,m (SE 56 p.m) in 1996. and 508 jxm (SE 44 jxm) and 512 |xm (SE 41 jxm) respectively in 1997. The spat with atrophied foot comprised 37% of the 1996 sample (N = 92) and 29'7r of the 1997 sample (N = 304). In the first aquarium experiment 98% of the released larvae settled on the cobbles and survived to day 5, but declined sharply and more or less linearly after day 7 to 53% on day 14 (Fig. 10). Mean shell length increased at a daily growth rate of -40 |xm to a mean length of 478 |xm (SE 13 [x.m) and then practically ceased, reaching a mean length of 499 |xm (SE 22 (xm) at day 14 when the experiment ended. The mean shell length of spat that died was 435 p.m at five days and increased only slightly after that; the mean shell length of "dead"" shells was 468 (xm (SE 19 p.m) for the whole 14 days. In the second experiment, survival was 85% after 14 days and in the third experiment 27% after 16 days; in both the last two experi- ments survival was 99% for the first week and declined steeply after that. 622 Sasaki and Shepherd 300 n 200 J 100- I ■ , 22-27 Sep 1995 Shell-collector 28 Sep, 1996 CCA boulders jIiIiIu 1 Oct, 199 7 CCA boulders Distance (m) Figure 4. \'ertical distrihutiiin of densitj of Halintis posl-larvae in numhers/m- along a transect normal to the coast at Site 1 in 1995, 1996 and 1997, with depth profile of the bottom and horizontal extent of Zones A, B and C. Distribution of Jii veiiiles Length-frequency distributions of juvenile ahalone in Zones A. B, and C for all sites combined (Fig. 11) show that the largest mean size and the highest densities of juveniles were in the Eisenia forest (A) and the lowest densities were in the CCA flats (C). The mean density of juveniles was 1.06/m" in Zone A. 0.83/m" in Zone B and 0.27/m- in Zone C. Mean densities in Zones A and B did not differ significantly (t = 1.3; NS) but did differ significantly be- tween Zones B and C (t = 3.1; P. 0.01). Size distributions in Zones A and C differed significantly from those in Zone B (for AB and BC comparisons x" = 1 1^ and 35 respectively; P < 0.001 ). Mean shell lengths in the three zones were 79 mm. 46 mm and 62 mm respectively. The age distribution of juveniles in the three zones combined (Fig. 12) shows that the density of the O-t- and 1 + 28 Sep. 1999 CCA boulders 29 Sep 1998 CCA boulders 25 Sep. 1997 CCA boulders Distance (m) Figure 5. Vertical distribution of density of Haliotis post-larvae in numbers/m- along a transect normal to the coast at Site 2 in 1997, 1998 and 1999 with depth profile and zones as in Figure 4. years classes was 6-7 times higher in Zone B than in the other zones. The 0+ and 1+ year classes combined comprised b'Vc of all abalone on Zone A. 49% in Zone B and 25% in Zone C. The age distributions in Zones A and C also differed significantly from that in Zone B (for AB and BC comparisons x" = 486 and 123 re- spectively; P < 0.001). A change in the ratio of the density of each year class in Zone A to the density of the same year class in Zone B (Fig. 13) indi- cates the likely direction of migration between the two zones as- suming that there is no differential M between zones. The low ratios for the 0+ and 1+ year classes shows high settlement and retention of those year classes within the zone, and the sigmoid curve for older year-classes suggests strong initial inigration of the 2+ year-class from Zone B to A lexelling off in the older year- classes. DISCUSSION Post-Lai^al Distribution This paper provides the first published account of the dynamics of settlement of H. d. Iiciiiiiai. survival of spat, and subsequent iklUui. 28 Sep 1999 CCA boulders 29 Sep 1998 CCA boulders Figure 6. a transeci Distance (m) Vertical distribution of density of Haliotis post-larvae along normal to the coast at Site 3 in 1998 and 1999 with depth profile and zones as in Figure 4. Ecology and Post-Settlement Survival of the Ezo Abalone 623 22~26 Sep 1997 Artificial boulders 250 275 300 325 350 375 400 425 450 4 75 500 ShelMenglh (um) 25 Sep 1997 I boulders 250 2 75 300 325 350 3 75 400 425 450 475 500 Shell-lenglh (um) 1 Oct. 1997 Natural boutders 425 450 475 500 525 550 575 600 625 650 Shell-length (urn) 40 o 10 1 13 Oct. 1997 Natural boulders m^- ^ 0 |S«5fift 2 3 Shell length (mm) Figure 7. Length-frequency distributions of Haliotis post-larvae on dried and natural boulders at successive time periods after settlement on about 22 September 1997 at Site 1. movement of juveniles in relation to the near-shore habitats in north-east Honshu. Our earlier study (Sasalae (circles) and size of dead shells (crosses) over time. Bars are standard errors. also varies according to these factors and other local contingencies such as presence and abundance of predators. The dominant poly- chaete in CCA at our study sites is the detritus-feeding cirratulid Dodecaceria conchanim (Hayashi et al. 1982) and we have no Distance -~^/^>Tv, ' ' 1 - ^"B^^^->_^__^ i C^^ —-^ ^_^ A 60 -| ^ ^!55^ Eisenia algal forest S L : 7 9 mm W. BW ; 7 4 g !^- z 0- c 1 — p^^(^ yy// zw, '/-y- 1 ~\^ 20 40 60 80 100 120 size Class (mm) 60-1 B Boundary zone S L : 4 6 mm g40- BW : 1 4 g d20- b u 0 20 40 60 80 100 120 Size Class (mm) Crustose coralline alg c S L - 6 2 mm 20- |io- BW ; 3 5 g — 1 — 1 1 — 1 1 1. , 0 20 40 60 80 100 120 Si eC las s (m m) Figure 11. Lenglh-frequencv distributions of abalone for all sites com- bined for each of the three zones. A, B and C sampled in .January 1998. SL = mean shell length and BW = mean in-shell weight. Density 0.2 0 4 0.6 0.8 i^^ i i Eisenia algal forest Boundary zone Crustose coralline algal area Figure 12. Mean density of five age classes of//, d. hannai in the three zones in 1998. evidence that they consume abalone spat as in McShane's experi- ments. Surxiviil Our M \ alues of 3.7-5.3/nionth according to the period exam- ined are higher than our earlier values in Kesennuma Bay (1.4- 2.7/month). hut initial densities were ~900/m" compared with only ~200/m" in the former study. Shepherd ( 1998) and Shepherd et al. (2000) reviewed earlier studies on post-larval M and showed strong density-dependent M in the range 0.6-2.7/month; our present study is consistent with this hypothesis. But what is the mechanism for density-dependent M? Our data suggest that starvation is a major contributing factor to the high M at our study site, and is how we summarize the evi- dence. First, the very high growth rates and low M on CCA cobbles show that CCA is the optimal substratum for settling post- larvae for the first five to seven days after settlement. Experiments by Dauine et al. (1999a. 1999b) also favor this conclusion. The cessation of growth at -450-500 |jLm thereafter suggest that food became limited and induced the increasing mortality that occurred from day eight to day 14. This is a critical phase in the life of a post-larval abalone, and diatoms are essential for further growth and survival (Kawamura et al. 1998. Daume et al. 2000). The highly skewed distribution of length-frequency data on dried boul- ders in the field experiment (Fig. 7) suggest that relatively few larvae grew beyond the size at settlement or. alternatively, that M of larger post-larvae was higher in comparison with natural boulders. m \ Figure 13. Plot of the ratio of density of each age class in the Eisenia forest (.\) to the density of the same age class in the boundary zone (B). Ecology and Post-Settlement Survival of the Ezo Abalone 625 Lastly, the apperance of atrophied post-hirvae at modal sizes of 475-530 10.111 (Fig. 9) suggests that the high densities of post-larvae exhausted the natural diatom supply soon after the critical size when diatoms dominate the diet of post-larvae (Kawaniura et al. 1998). In independent experiments with plankton trap-nets set at Site 1 in October and in other years, we have also recorded atro- phied post-larvae of mean length 740 |a,ni (SE 240 p.ni) as well as other small gastropods presumably washed off the substratum by wave turbulence. This suggests that starving post-larvae are sus- ceptible to mortality from wave-shear as proposed by McShane (1996). Shepherd et al. (2000). in reviewing the evidence for den- sity-dependent M of post-larvae, suggested that in southern Aus- tralian habitats, the carrying capacity of CCA boulders was -100 post-larvae/m". From the results of this study and our earlier one we hypothesize a carrying capacity of <200/m" in terms of food availability on Miyago coasts. It is also possible that Tci;iilci spp. may compete as post-larvae with post-larval abalone and so further reduce the carrying capacity of the habitat for abalone. Distribution of Juveniles The pattern of distribution of juveniles is initially similar to that of post-larvae, although in practice we found that highest juvenile densities occurred within the boundary zone (B) but just shoreward of the area of highest density of post-larvae. Because the depth of the Eiseiiia forest decreases with increasing shelter, the band of highest juvenile density is shallower in more sheltered waters. What is the adaptive significance of shoreward migration of this abalone? Ino (1966) and Nie (1992) stated that this abalone migrated into shallow water before spawning, implying some adaptive advantage such as increased larval dispersal. If so, this behavior is analogous to the movement of other abalone species into the swell or into increased current flow (Shepherd 1986. Cla- vier & Richard 1984). While this may be true at our study site, the obvious explanation is that Eisenia and associated algae are a preferred food (Taniguchi 1991. Shepherd & Steinberg 1992). Hence movement to this habitat must be advantaaeous for srowth and reproduction. A similar size-related migration by the related species H. I<.amtschalkana from deeper water where larvae settle into shallower kelp habitat is also recorded in Alaska and Canada (Sloan and Breen 1988, S. A. Shepherd unpublished data). Such migration is quite different from the aggregation and movement of spawning adults onto the highest points of reefs in shallow water for spawning (Breen & Adkins 1980. StekoU & Shirley 1993. unpublished observations). CONCLUSIONS Our study has important implications for the management of the declining H. d. Iniiiiiai fishery in north-east Honshu. Abalone larvae advect shoreward according to local hydrodynamic condi- tions and settle at densities that depend on coastal topography and the extent of suitable habitat. Post-larval densities of both this abalone and Tegula spp. are highest in the boundary zone between the near-shore Eisenia forest and the deeper CCA habitat. How- ever, post-larval survival appears to be food-limited because sig- nificant numbers of post-larvae five to seven days old show an atrophied condition. This raises the possibility that the abundance of Tegiila post-larvae in the same habitat may affect the sur\'ival of post-larval abalone. With increasing age juvenile abalone migrate into the Eisenia forest which provides an abundant food supply for adult abalone. To achieve an understanding of the causes underlying the decline of the fishery, studies should now focus (a) on the relation between recruitment strength of abalone into the fishery and the extent of the boundary zone, and (b) the competitive relations between aba- lone and Teiiiila post-larvae. ACKNOWLEDGMENTS We thank Dr. K. Taniguchi (University of Tohoku) for his valuable advice on the algal flora at our study sites and for per- mission to use the data in Figure 2. Dr. T. Seki (National Research Institute of Aquaculture) gave helpful advice and encouragement. We thank Dr. Sabine Daunie for improvements to the manuscript. LITERATURE CITED Breen. P. A. & B. E. Adkins. I9l SlicUlisli Rc-sicinh. Vol. 20, No. 2. f)27-fi36, 2001. SUITABILITY OF AUSTRALIAN FORMULATED DIETS FOR AQUACULTURE OF THE TROPICAL ABALONE HALIOTIS ASININA LINNAEUS D. JACKSON,' - K. C. WILLIAMS," AND B. M. DEGNAN' * Departmciir of Zoology and EiUoniology. University of Queenskmd. St. Lucia 4072. Australia; -CSIRO, Division of Marine Research. P.O. Bo.\ 120. Cleveland 1 163. Australia ABSTRACT The performance of Hali(:ii.\ usiiiina Liniuicus led tour commercially available diels formulaled for temperate Austra- lian species was assessed relative to a natural diet of Gracilana ccliitis. Parameters measured included growth in terms of total wet body weight (TWBW) and shell length (SL). survival, food conversion ratio (FCR). condition index (CI), and se.xual maturation. The results of these measurements indicated that two of the formulated diets could be used for commercial culture of H. a.sininu. Nutritional and physical parameters of these diets were used in a multivariate regression analysis to identify variables that explained the most variation in growth between diets. Parameters that were selected by this analysis were the ratio of crude protein (CP) to gross energy (GE) and the rate that protein leached from the formulaled diets. The resulting regression equation was highly significant in its explanation of the variation in growth {P = 2,2 x lO""), To determine the quality of the dietary CP in each diet, the essential amino acid profile of each diet was compared to that of whole soft body tissue in H. asininu. Methionine seemed to be deficient in all of the diets. Animals fed G. cihilis did not seem to suffer from a lack of n-3 PUFAs. KEY WORDS: Halioiis usinina. formulated diet, gnnuh. amino acid, fatty acid, sexual maturation, protein leaching INTRODUCTION Many abalone fisheries around the world are in decline (Hahn 1989a). and. as a result, reseai-ch programs aimed at the efficient artificial production of a range of haliotid species are underway. One important aspect of such research is the identification of the nutritional requirements of cotnmercially inipoilant abalone spe- cies (Fleming et al. 1996). This allows the formulation of cost- effective diets that provide a consistent and precise method of supplying the nutrients necessary for maximal growth. Diet for- mulations can also be easily modified as the nutritional require- ments of an anitiial changes during development (e.g,. postsettle- ment. growout, and broodstock diets). Traditionally, the tnajority of abalone diet research around the world has focused on temperate species, because these form the majority of abalone exports worldwide (Hahn 1989a). However, recent research has indicated a growing demand for tropical aba- lone (Nateewalhana ct Hylleberg 1986. Singhagraiwan & Doi 1993. Jarayabhand & Paphavisit 1996). Jarayabhand and Pa- phavisit (1996) report that in Taiwan these smaller tropical species are preferred to the temperate species because of their size and delicate flavor. The distribution, ecology, reproductive biology, and natural diet of the tropical abalone Haliotis asinina Linnaeus differ from Australian teinperate species. H. asinina is found on intertidal coral reef flats throughout the Indo-Pacific (Gosliner et al. 1996); whereas, Haliotis rubra Leach and Haliotis laevigata Donovan are subtidal animals (Prince & Shepherd 1992) found on rocky tetii- perate reefs. Populations of//, asinina on Heron Island (23°27'S: 15r55'E) spawn fortnightly in cue with lunar and tidal cycles from mid-October to mid-April (Counihan et al, 2001 ); whereas. temperate species in culture are artificially induced to spawn be- cause of their relatively unpredictable and infrequent natural spawning patterns (Hahn 1989b). Temperate Australian abalone occur on reefs where macroalgae are relatively abundant (Dayton 1994); whereas, the abundance of frondose macroalgae on coral reefs is limited (Littler & Littler 1994). Sawatpeera et al. (1998) demonstrated that the gut contents of adult H. asinina collected from coral reefs of Samed Island. Thailand contained over 80% diatoms. These significant differences between temperate and tropical abalone suggest that diets formulated for temperate Aus- tralian abalone, H. rubra and H. lacviiiota. may not be suitable for H. asinina. Research on tropical abalone nutrition is limited. Capinpin and Corre (1996) have investigated the growth rates produced by H. asinina fed a Japanese-forinulated diet over 90 days, but per- formed minimal nutritional analyses on the diets used. Bautista- Tereul and Millamena (1999) also conducted a 90-day growth experiment and reported on the effects of dietary energy and pro- tein on the growth of //. asinina. but not on the fatty acid (FA) profiles of the diets. These studies did not include a nutritional acclimatization period, and were run for only 90 days. Day and Fleming (1992) highlight the effects that previous diets may have on growth and recommend an initial period of nutritional acclima- tization before any growth experiments, even for artificially pro- duced animals that have a comtnon diet history. Furthermore, they suggest that, in the case of H. rubra, a growth trial of less than 100 days will not be informative. Here we report on the suitability of fonnulated diets that were produced for temperate Australian species (//. rubra and //. lae- vigata) for the culture of H. asinina. We determined: (I ) the water .stability of the diets; (2) rates of growth in terms of total wet body weight (TWBW) and shell length (SL) increase; (3) sexual matu- ration; (4) food conversion ratio; (5) dry feed intake; and (6) the condition index (CI) of the abalone after being fed these diets for 6 mo after a nutritional acclimatization period of 49 days. Growth rates promoted by the various diets are related to a range of nu- tritional parameters, including the crude protein (CP), total lipid (TL), gross energy (GE), fatty acid (FA), and essential amino acid contents of the diets fed. MATERIALS AND METHODS Diets *Corresponding author. E-mail: bdegnanfeizen. uq.edu.au Samples of four formulated diets from two Australian abalone feed companies were obtained for comparison with a natural diet 627 628 Jackson et al. of the red seaweed, GraciUiriu cduits. Animals were fed nightly to satiation, and uneaten food was removed the following morning. This procedure was done only once on weekends. G. echilis was chosen as the reference diet, becatise Gniciliiria species have been reported to promote high growth in H. cisiniiici {Singhagraiwan & Sasaki 1991, Singhagraiwan & Doi 1993, Capinpin & Corre 1996. Bautista-Teruel & Millamena 1999, Capinpin et al. 1999). Animals and Management Juvenile H. asiiuna used in the following experiments were derived from spontaneous spawnings by wild broodstock collected from Heron Island, Great Barrier Reef. Australia (23°27'S: I51°55'E). Abalone at the start of the experiment ranged in SL from I 1 .8 mm to 23.4 mm (average 18.3 ± 2.76 mm) and TWBW from 0.40 g to 2.75 g (average 1.32 ± 0.577 g). Individuals were blot dried, weighed, measured, tagged, and assigned to an experi- mental unit. Abalone were distributed among replicates so that there were no significant differences in SL or TWBW. Each diet treatment was replicated six times, and consisted of a 1 .7-L lidded plastic container (21x14x7 cm) housing four abalone. A "spare" replicate of each diet treatment was al.so maintained. Animals in these spare replicates were used to replace any that died during the experiment. This maintained a constant density of animals per experimental unit for the duration of the experiment. Filtered (10 Jim) seawater was heated to 28°C and delivered to individual containers at a rate such that the entire volume was exchanged once every 3 niin. Seawater was not recirculated. Abalone were maintained in a I2L:I2D light cycle for the duration of the ex- periment. Abalone were acclimatized to their respective diet treat- ments before the first TWBW and SL measurement for 49 days to reduce the residual effects of accumulated body stores from pre- \ ious natural diets. Growth Data Collection SL and TWBW measurements of every animal were made at four weekly intervals. The trial was run for 24 wk. Growth pro- moted by each diet treatment was expressed as the increase in SL and TWBW over 24 weeks. All animals were examined for the presence of gonad tissue from the third month of the experiment onward. If gonad tissue was present, its color was recorded, al- lowing the sex of individual animals to be determined. As animals were individually tagged, a statistical comparison between male and female growth rates was also made. Condition Index A simple CI was used to e\'aluate the health of each animal at the end of the experiment using the equation: 'TWBW ^ "'0 where TWBW is the total wet body weight (g), and SFBW is the shell free body weight (g). SFBW was estimated by subtracting the shell weight (SW) from the TWBW. Because the animals used in the feeding trial were to be used in other experiments. SW was estimated indirectly according to the equation: SW = 3 X M)-* X SL-'-^' where SL equals shell length. This relationship had an R- value of 0.99 and was derived from over 300 shell measurements. Food Conversion Ratio The food conversion ratio (FCR) of animals reared on each diet was estimated during wk 18 to 20. Each night the wet weight of diet added to each experimental unit was weighed to within 0. 1 mg. Twelve h later, uneaten feed was collected by passing the entire contents of each experimental unit carefully through a I -mm- filter to remove feces, and then through preweighed What- rnann filter paper. The wet feed was then vacuum filtered and dried at I05°C overnight. For each experimental unit, the nightly dry feed intake (DFI) was calculated as follows: DFl = f, X°^/,„„ -('=-"•%„) where f, is the wet weight (g) of feed added each night, DM is the percentage of dry matter for each diet, f-, is the weight (g) of uneaten dried feed, and WS is the water stability (%) of each diet. Water stability was calculated as the dry weight retained after immersion in seawater for 1 2 h under conditions identical to those at which the growth experiment was conducted. The nightly DFI over the 2-wk period was summed to give the total food intake (TFI) for each replicate. TWBW measurements were taken at the beginning and end of the 2-wk period, allowing the FCR to be calculated as follows: FCR = "^^p, where WG is the TWBW gain (g) over the 2-wk period. Gonad Index and Histology During the final TWBW and SL measurements the gonad index (GI) of each animal was ranked following the method described by Singhagraiwan and Doi (1992). At the end of the growth experi- ment, the gonads from the two most mature animals of each sex on each diet were dissected, fixed in Bouin's fixative for 6 days, and washed three times in 70'^ ethanol over 3 days. Gonads were then embedded in paraffin, sectioned at 6 txm on a Spencer 800 micro- tome and stained with hematoxylin and eosin. Representative digi- tal photos were taken with an Olympus DPIO digital camera at- tached to a BX60 microscope. Nutritional Analyses Representative samples of each diet were analyzed, usually in triplicate, for nutrient composition. The dry matter of each diet was determined by drying at 105°C for 12 h. Ash content was deter- mined by incinerating dried samples at 200°C for 2 h, and then at 550'-'C for 12 h. Nitrogen content was analyzed by the Kjeldahl procedure, and CP was estimated by multiplying the nitrogen con- tent by 6.25. TL was extracted with chloroform-methanol using the method of Folch (Folch et al. 1957). Saponification and esterifi- cation of the TL extract was performed following a modified method of van Wijngaarden (1967). Two replicate extracts of each diet were analyzed on a Hewlett Packard GC for FA content. GE was measured with an IKA Labortechnik Calorimeter. For amino acid analysis, six wild adult abalone. ranging in TWBW from 63.6 to 149.3 g and in SL from 63 to 100 mm, were shucked, and the foot and viscera were blended together into a homogenate before being freeze dried. Two replicates of each diet were freeze dried and finely ground using a mortar and pestle. Samples were hydrolyzed in 6 M HCI at I05"C for 24 h. Free amino acids were derivitized using phenylisothiocyanate and sepa- rated using high-power liquid chromatography (HPLC) on a C-18 column with detection at 254 nm. Instrumentation was a Bio-Rad Formulated Diet for Culture oe Tropical Abalone 629 2800 solvent delivery system with a Bio-Dimension UV/Vis moni- tor, controlled by a Series 8()0HPLC and v2.30.la software. To compare diets directly, each amino acid was expressed as a pro- portion of the corresponding amino acid in H. asinina whole soft body tissue. Protein Leaching The rate of protein leaching from each diet was measured by the Bradford micromethod of protein estimation (Bradlbrd 1976) using the Bio-Rad protein assay kit. 100 niL of 10 |xm filtered seawater was placed into a clean conical tlask and a background reading (before addition of any diet) was taken in triplicate. Ap- proximately 1 g of each diet was weighed to within 0.1 mg and added to a flask. Triplicate optical density readings were taken at exactly 30 and 60 min after the addition of a diet treatment and thereafter hourly for a further 6 h using a Hitachi U-2000 spec- trophotometer. Flasks were gently vortexed (approximately 120 cycle.s/min) in a Ratek orbital mixer at room temperature to simu- late the effects of water flow and physical disturbance. A control consisting of 100 niL of seawater was run in parallel to the dietary treatments. All treatments were replicated three times, and optical density readings were related to a standard curve constructed from 10 concentrations of bovine serum albumin dissolved in seawater. The total protein leached (i.e.. the protein concentration after 7 h) was expressed as a percentage of the total protein added to the flask (calculated from the protein content of each diet and the weight of feed added to each tlask). It was found that G. edidis could not be included in this experiment, because it interfered with the dye-binding reaction of the Bradford method, resulting in pro- tein concentrations less than the control. Therefore, to be included in the multiple regression analysis (see below) it was assumed that G. ediiUs leached no protein over a 7-h period. Statistical Analyses All statistical analyses were performed on STATISTICA for Macintosh (Statsoft 1994). All datasets were tested for violation of the assumptions of homogeneity of variance, normality, and inde- pendence of errors before further analysis. Datasets that failed any one of these assumptions were transformed and re-analyzed. The effect of tank was tested for all analyses where appropriate and was never found to be significant. Where significant differences between diet treatments were detected by one-way analysis of variance (ANOVA). a Student's-Newman-Keuls test (SNK) was performed for ((/Jo.vff /■/(«/ comparison of treatment means. Growth data were analyzed by comparing the total increase in SL and TWBW between diets for the entire duration of the trial by one- way ANOVA. TWBW data were square root transformed before analysis to satisfy the assumptions of normality, independence of errors, and homogeneity of variance. SL data did not require trans- formation. The effect of diet and sex on Gl was tested by a Kruskal- Wallis ANOVA by ranks. A forward stepwise multiple regression analysis was used to determine the nutritional param- eters that explained the most variability in TWBW growth. Vari- ables included in this analysis were: CP. TL, GE. the ratio of CP to GE. and the rate that protein leached from the diets (this rate was assumed to be zero for G. ediilis. see above). Those variables that produced an F ratio in excess of 1 .0 were included in the regression equation, and the tolerance level was set at 0.01. The significance level for all statistical analyses was set at a = 0.05. RESULTS Water Stability anil .\iitritiiinal Composition of Diets The water stability and nutritional composition of four formu- lated diets and G. edidis were analyzed (Table I ). Diets were arbitrarily labeled I to 4. Diet 4 was significantly (P < 0.05) more water stable than all other formulated diets. When compared on a dry matter basis to G. edidis. formulated diets had significantly (P <0.05) higher CP (21.8-40.0 vs 16.4%), TL (4.1-5.7 vs 0.8%) and GE ( 18.7-19.4 vs 9.0 MJ/kg) contents, respectively. The ash con- tent and CP:GE ratio of diet 4 was significantly lower (P <0.05) than all other diets (4.8-5.4 vs 3.3% for ash and 17.0-20.8 vs 1 1 .7 mg/kj for CP:GE). The rate of loss of protein from each of the formulated diets was essentially linear over 7 h. but seemed to begin to plateau by the final measurement, most noticeably in diet 4 (Fig. 1). The total loss of protein (expressed as mg dissolved protein per niL of leachate per gram of feed) over 7 h ranged from 9.1 to 18.0 (Fig. 1 ). with diet 4 losing the least protein and diet 3 the most. However, when the total amount of protein leached was expressed as a percentage of the total available protein (a more meaningful measure because of the different protein contents of each diet), diet 1 leached the lowest proportion of its protein con- tent (9.6% ). and diet 4 leached the highest proportion of its protein content (13.4%). These differences were significant between all diets, except diets 3 and 4 (P <0.05; see Table 1). Amino Acids Relative to the formulated diets, the essential amino acid (EAA) composition of G. editlis protein contained higher amounts TABLE I. Water stability and gross nutritional composition of four formulated diets and Gracilaria edidis led to Haliotis asinina. Water Stability Protein Leached .Moisture Ash CP TL GE CP:GE (%) (nig/kg CP) (g/kg) (g/kg) (g/kg) (g/kg) (MJ/kg) (g/MJ) Diet 1 76.6 ( 0.67 r' 96(2.0)' 111 (O.IS)' 5-i (0.7)-' 400(3.7) 45(3.9)" 19.2 (0.1 !)■■■" 20.8 Diet 2 75.0 (O.IS)"" 113(1.6)" 124(0.25)" 53 (0.3)-' 347 (0.4) 57(1.3)" 19.4 (0.09)' 17.9 Diet .^ 77.8 (0.09)" I26(l.7r 113(0.16)" 48 (2.3)" 322 (6.4) 41 (O.S)" 19.0(0.10)'" 17.0 Diet 4 82.9 (1.32)'-' 134(4.3)" 92(0.12)' 33 (0.2)' 218(5.8) 55(1.2)" 18.7(0.07)" 11.7 G. ediilis 95.4(1.70)'' ND 890(1.02)'' 59 (2.3)'' 164(1.6) 8(1.4)'- 9.0 (0.24)' 18.2 Water stability is expressed here as the percentage of dry weight remaining after overnight immersion in seawater. Protein leached is expressed as the weight of dissolved protein per kg of CP after 7 h immersion in seawater. Moisture is expressed as the weight (g) lost after 12 h at 105°C. Ash. crude protein (CP). and total lipid (TL) values are expressed as g/kg of dry diet. Gross energy (GE) is expressed as MJ per kg of dry sample. Numbers in parentheses are standard deviations of the mean. Values in the same column that share the same superscript letter do not differ significantly (P > 0.05). 630 Jackson et al. 16 , Diet 1 14 □ Diet 2 4 Diet 3 U , o Diet 4 10 , _x- Control TABLE 3. Fatty acid content of four formulated diets and Gracilaiia ediilis. 0 50 100 150 200 250 300 350 400 time (minutes) Figure 1. Rate of protein leaching measured in four formulated diets over 7 h. \ alues are expressed as milligrams of dissolved protein x mL of leachate"' x gram of feed '. Error bars are omitted for clarity of presentation. The control consisted of three flasks of seawater treated in exactiv the same wav as the diet treatments. of isoleucine. leucine, methionine, threonine, and valine and a lower amount of histidine (Table 2). Of the formulated diets, diet 4 contained the lowest amounts of all essential amino acids except for lysine (Table 2). Fatty Acids Large differences were found in the FA profile of G. ediilis as compared with the four formulated diets (Table 3). The largest differences were between 20:4n-6 (33-79 times less abundant in the formulated diets) and 18:2n-6 (40-53 times more abundant in the formulated diets). Other large differences were seen for 18: 3n-3. 20:ln-9 and 22:6n-3, which were less abundant and for TABLE 2. Concentrations of amino acid residues in Haliotis asiiiiiia tissue and four formulated diets and Gracilaria ediilis. H. G. Nonessential asiiiiiia" ediilis'' Diet 1" Diet 2" Diet i"" Diet 4" Alanine Aspartic acid Glutamic acid Glycine Proline Serine Tyrosine Essential Arginine Histidine Isoleucine Leucine Lysine Methionine Phenylalanine Threonine Valine 44.21 25. 1 y 6.59 96.6.^ 47.44 37.34 11. no 32.72 5.99 17.11 30.55 21.80 12.00 9.22 31.79 27.19 7.55 6.62 9.36 7.65 4.98 6.33 0.69 3.05 0.83 3.10 5.29 2.39 1.11 1.59 4.05 5.30 4.16 6.18 15.59 5.18 7.62 5.49 0.49 2.91 1.67 2.45 4.56 2.30 0.81 1.61 3.81 4.12 4,91 7.06 14.92 6.02 6.93 5.53 0.68 2.52 1.77 2.38 4.73 2..54 1.01 1.46 3.35 4.38 5.05 7.83 18.91 5.71 7.23 5.79 2.14 3.16 1.65 2.42 4.80 2.69 0.81 1.44 3.85 4.28 4.46 5.39 18.20 5.54 7.54 4.83 022 2.51 1.27 1.91 4.27 2.49 0.53 1 .35 2.94 3.55 Fatty Acid Diet 1 Diet 2 Diet 3 Diet 4 G. ediilis 14:0 3.8 2.4 4,0 3.6 1.2 16:0 20.8 19.1 21.1 18.8 43.2 16:ln-7 3.0 1 -) 3.5 3.3 1.4 18:0 3.8 3.5 3.8 3.2 1.8 18:ln-9 12.1 14.5 12,0 15,7 6.4 18:ln-7 2.1 2.0 -) T T -^ 3.8 lS:2n-6 31.5 32,2 29,8 24,0 0.6 l8:3n-3 3.8 3.4 3.7 2.5 0.4 18:4n-3 0.9 0.9 1.1 1.4 0.0 20:ln-9 3.4 3.8 3.8 5.1 0.2 20:3n-6 0.0 0.0 0.0 0.0 3.2 20:4n-6 0.6 0.4 0.6 0.5 31.7 2():4n-3 0.6 0.6 (1.7 1.0 0.0 2():5n-3 3.8 34 4.2 5.0 1.9 22:ln-9 1.9 2.5 1.7 2.7 0.0 22:5n-3 0.8 o,q 0.9 1,2 0.5 22:6n-3 5.1 5.7 5.0 7.1 04 Summarv Total H-3 PUFAs 3.9 3,5 3.8 2,7 0.4 Total n-6 PUFAs 31.7 32,6 30. 1 24.4 4.2 Total «-3 HUFAs 11.2 11,5 11.9 15.7 2.8 Total «-5 HUFAs 0.7 0,4 0.7 0.6 32.8 Total H-3 15.1 15.0 15.7 18.4 3.2 Total n-b 324 33.0 30.8 25.0 37.0 SFA 29.6 26.2 Ml\ 26.7 47.6 MUFA 22.8 25.7 23.6 30.0 12.3 PUFA 35.5 36.1 33.9 27.0 4.6 HUFA 12.0 12.0 12.5 16.2 35,5 Values are the means of six abalone satnples and two diet samples. ■■ Values expressed as g/kg. '" Values expressed as g/100 protein. Values are presented as the percentage of total lipid and ate the means of two samples. 20:3n-6. which was more abundant in G. edidis as compared to the formulated diets. These differences culminated in total n-3 FAs being approximately five times higher in the formulated diets. Total n-6 FAs occurred at similar concentrations in all diets; whereas, saturated FAs (SFAs) were higher in G. ediilis relative to the formulated diets. Differences in the FA profiles between for- mulated diets were less pronounced than those between C. edidis and formulated diets. The most obvious difference was with diet 4, which was higher in total n-3 and lower in total n-6 FAs compared to diets 1 to 3 (Table 3). Feeding Study Animals reared on C. ediilis. diet 1 and diet 2 produced sig- nificantly {P <0.05) higher TWBW growth rates than animals reared on diets 3 and 4 (Table 4). G. ediilis produced significantly (P <0.05) higher SL growth than all other diets, followed by diet 1, which was significantly higher than all the other fortnulated diets (P <0.05). Diet 2 produced significantly (P <0.05) higher SL growth than diets 3 and 4, which did not differ from each other (Table 4). No difference was detected between male and female growth in terms of TWBW and SL growth (data not shown). Although the cuniulati\e monthly measures of TWBW and SL show a fairly linear growth rate over 24 weeks across all diets (Fig. 2), the monthly growth rates (Fig. 3) revealed a considerable amount of intermonthlv \ariation. Most noticeable was the eleva- Formulated Diet for Culture oi- Tropical Abalone 631 TABLE 4. Increase in total "it Ixidj weight lyl and shell length (mm), mean condition index, mortality rates, food conversion ratio, and ni!>litl> dry food intake of abalone reared on four formulated diets and Gracilaria ediilis for six months. TVVBW Growth SL (;ro\vth CI Mortality FCR DFI Diet 1 }^)2 lL4l8r' 7.2 (L40)'' 87.1 (Lie)" 0.25 1.4 (0.87)'' 0.67 (0.090)" Diet : 2.30 (0.fi64)- .'^.3 (0.98)'^ 85.3 (1.40)" 0.38 1.8 (0.80)- 0.77(0.073)"" Diet 3 LSO (0.661)'' 4.2 (1.00)" 85.0(1.15)" 1.25 1 .5 (0.2 1 )- 0.66(0.163)-' Diet 4 0.96(0.493)" 3.2 (1.11)" 81.9 (1.90)'' 1.25 1 .6 (0.47)" 0.75 (0.099)"" G. ediilis 3.98 (1.068)" 9.6 (1.24)"' 85.0(1.16)" 1.00 1.6(0.34)° 0.85(0.100)" Total wet body weight (TWBW) growth data was square root transformed before any statistical analysis was performed, but values presented here are untfansfonned. Condition inde.\ (CI) measures are the weight of soft tissue (viscera and muscle) expressed as a proportion of the total wet weight (soft tissue plus shell weight). Mortality is expressed as the number of deaths recorded for each diet as a proportion of that recorded for G. edulis. Food conversion ratio (FCR) was calculated as the ratio of dry feed eaten to wet weight gained. Dry food intake (DFI) is the weight of dry food eaten each night, expressed as a percentage of wet body weight. Values in the same column that share the same superscript letter do not differ significantly. Numbers in parentheses are standard errors of the mean, except for FCR and DFI where they are standard de\'iations. tion in growth over mti 2 to 4 relative to mo 0 to 2. followed by the subsequent decline over months 4 to 6. Animals fed diet 1 had significantly higher (f <0.0.5) condition indices than those reared on all other diets; whereas, animals fed diet 4 had significantly lower condition indices than those reared on all other diets (Table 4). FCRs did not differ significantly between diet treatments (Table 4). However, the power of this ANOVA to detect a differ- ence was very low (approximately \6%). FCRs for the formulated diets ranged from 1.4 to 1.8; whereas, that for G. edulis was 1.6. During the period of FCR measurement, abalone fed G. edulis ate A. , Diet 1 • D Diet 2 • ^ Diet 3 • oD'ef , G edulis • X D D ' ♦ 5 a A • A ♦ o z i O o A o A ! B. ' X ^ Diet I ^ ^, □ Diet 2 • -"■. 4 Diet 3 o Diet 4 X o a Q X G edulis • o A A ■n ♦ A s A O fl A o o * 0 I') 17 Figure 2. .\verage (h = 6) cumulative growth of ahalone fed four formulated diets and (/. edulis over 24 « k after a nutritional acclima- tization period of 7 wk (shown as week 0 to week 7). Measures were taken every 4 wk. (irowth was measured in terms of (A) total wet body weight and (B) shell length. Error bars are omitted for clarity of presentation. significantly more (P <0.05) than animals fed diets 1 and 3; whereas, those fed diets 2 and 4 did not differ from G. edulis in the weight of food eaten. Animals fed diets 1 to 4 did not differ significantly in the weight of food eaten (Table 4). The results of the multivariate regression analysis identified two nutritional parameters that explained a significant proportion of the variation in growth observed: CP:GE ratio and the rate of protein leaching. Overall, the regression equation was highly sig- nificant (P <0.01 ) with an r' value of 0.63 (Table 5). The CP:GE ratio carried the most weight in explaining growth with a beta coefficient of 0.55; whereas, the inverse relationship between pro- tein leaching and growth had a beta value of -0.39. The normal probability plot of residuals revealed a very strong linear associa- ■ Diet 1 Q Diet 2 a Diet 3 ■ Diet 4 ■ G edulis monlhs'.l-l months 1-2 months 2-1 months 3^ monlhs 4-5 months 5-h Figure 3. Average in - 6) monthly growth rates of abalone fed four formulated diets and G. edulis over 24 wk after a nutritional acclima- tization period of 7 wk. Monthly growth rates are presented in terms of (.\) total wet body weight and (B) shell length. Error bars are the standard error of the mean. 632 Jackson et al. TABLE 5. Results of forward stepwise multivariate regression analysis. Beta B Variable Coefficient SE Beta coefficient SEB Tolerance P-Value CP:GE 0.55 0.1.^ o.oy o.o: 0.83 0.0003 Protein leaching -0.39 0.13 -0.04 0.0 1 0.83 0.006 Regression equation G -owlh = 0.37 + O.Oy CP:GE - 0.04 Protein leachnig 2.2 X 10-" tion, indicating that all \ariables were ntirmally distributed. A scatter plot of observed versus predicted residuals revealed that a linear regression was a good description of the relationship be- tween growth and the three selected variables. No outliers were detected. Gniiad Index and Histology Males were significantly more mature than females by the end of the trial (P = 0.029). When separated by sex. the GI of females fed different diets did not differ significantly {P = 0.23): whereas, for males the effect of diet on GI was significant (P = 0.015); diet 1 produced the most mature males, followed in order by diets 2, 3, 4, and C. cdidis (Table 6). The smallest animal with visible gonad tissue had an SL of 16.4 mm. weighed 0.92 g, was male, and had been fed diet 4 for 84 days. When examined microscopically, all animals reared on G. edit- lis and diets I and 2 seemed to have healthy gonads when com- pared to gonadal sections of H. asinina taken by Apisawetakan et al. (1997) and Jebreen et al. (2000). Sections of gonads from animals reared on diets 3 and 4 showed evidence of substantial amoeboid cell activity and large areas of degenerated oocytes; animals reared on diet 4 seemed to be more affected than animals fed diet 3 (Fig. 4). DISCUSSION Growth Experiment This study .sought to compare the growth of H. asiniiui reared on commercially a\ailable Australian temperate abalone feeds relative to a Gracikiria diet, a genus of red algae known to sustain high growth in this haliotid (Upatham & Sawatpeera 1998, Bau- tista-Teriiel & Millamena 1999, Capinpin et al. 1999). Comparing the growth rates achieved by H. asinina fed the commercial feeds to that of indi\ iduals fed G. ediilis demonstrates that two of these formulated diets may be suitable for the aquaculture of H. asinina (Table 4 and Fig. 2). Although the highest SL growth was obtained on G. edidis. TWBW growth was not significantly different be- TABLE 6. Gonad index observations made at the end of the growth experiment. Diet 1 Diet 2 Diet 3 Diet 4 G. ediilis P-Value Male GI Female GI 1.5 I 1.5 1 0.015 0.225 Values are medians and range from 0 lo 3. with 3 being fully mature. P-values were obtained from a Kruskal-Wallis analysis of variance (ANOVA) by ranks. iween G. edidis. and diets I and diet 2. From an aquaculture perspective, weight gain is considered lo be more iinportant than SL increase (see Fleming et al. 1996). Furthermore, the CI of animals reared on diet 1 (87.1%) was significantly higher than that of animals reared on G. ediilis (85.0%), indicating that diet 1 produced healthy animals of a high product quality (Table 4). The diets and experimental conditions employed within this study produced growth rates lower than similar studies on H. as- inina: Capinpin et al. (1999) recorded an SL growth rate of 60 mm/year (equivalent to 164 |xm/day) with animals reared in sea cages and fed Giaciliaria luiilinae. Bautista-Tereul and Millamena (1999) achieved growth rates ranging from 222 to 247 |j.m/day over 90 days when 20 juvenile H. asinina (of a similar size to those used in the present study) were housed in large 60-L containers and fed either of three formulated diets. In the present study, H. asinina were reared in small (1.7 L) enclosures, which may have contributed significantly to the low growth rates ( 19-56 ixm/day) when compared to these studies. Although the results are not di- rectly comparable because of differing experimental conditions. Capinpin et al. ( 1999) found a significant density dependent effect on growth across four stocking densities: 43, 88. 130. and 175 abalone/m"^. The density used within this experiment equated to approximately 72 abalone/m". and implies that growth may have been affected by density. Furthermore. Mgaya and Mercer ( 1995) demonstrated a size grading effect on growth in H. tiihercidala such that small abalone reared in the presence of large abalone did not grow as quickly as their size-graded counterparts. In our study, strict size grading did not occur. Growth may have been less variable and possibly higher had such a regimen been einployed. No significant differences were detected in the DFI by animals fed the formulated diets (Table 4), suggesting that differences in growth between these diets were attributable to nutritional factors rather than diet preferences (see Vandepeer et al. 1999). However, animals fed G. edidis ate significantly more than animals fed diets 1 and 3. Therefore, the possibility that animals fed G. ediilis grew faster in terms of TWBW than animals fed diet 3 because of a higher food consumption cannot be ruled out. Despite the short period over which FCR measurements were taken, the recorded values were low ( 1.4—1.8), and similar to those reported by Bautista-Tereul and Millamena (1999). The formu- lated diets used by these authors produced very high growth in juvenile H. asinina of a similar size to those used in the current study. Furthermore, Fleming et al. (1996) reviewed the food con- version efficiency (FCE) of seven diets formulated for temperate abalone by various feed manufacturers around the world and re- ported an FCR range of 0.77 to 3.33. The FCR values obtained in the present study compare favorably (Table 4). All four formulated diets leached a considerable portion of their protein content when gently agitated at room temperature for 7 h (Fig. 1 and Table I ). The only other similar assay for abalone diets Formulated Diet for Culture of Tropical Abalone 633 l?^^'*:.;.t^^i- (Viana et aL 1996) reported a leaching rate of 90.1 to 98.7mg/g over 12 h at 16"C, indicating that those diets were more protein- stable than those employed within the present study. The results of the protein leaching assay contradict the water stability results, because diet 4 (the most water-stable diet) leached the highest proportion of its protein; whereas, diet 1 (no different from the least water stable diet) leached the lowest proportion of its protein. This discrepancy between nutrient leaching and dry matter water stability has been noted by others (Viana et al. 1996, Edwards & Cook 1999). The multiple regression analysis identified CP:GE and protein leaching rate as variables that explained a significant amount of the variation in growth between the five diets. The correct balance between energy and protein is es.sential for a diet to promote maxi- mum growth (Fleming et al. 1996). Although CP;GE ratios are not Figure 4. Representative ovary structures of H. asinina reared on a natural diet and four formulated diets. (.\-Bl Sections of ovary taken froni an animal reared on G. ediiliK for 6 mo. .Animals reared on G. edulis were not as mature as those reared on the formulated diets. This is reflected by the relatively thin layer of ovary that overlies the di- gestive gland. DG, digestive gland; t, trabeculae; (I, Gonad. (C-D) Sections of ovary taken from an animal reared on diet 1 for 6 mo. Note the similarity in the over-all structure of the gonad in C, and in the size and shape of an individual oocyte and germinal vesicle in D. (E-F) Sections of ovary taken from an animal fed diet 2 for 6 mo. Note the similarities between diets 1 and G. edulis. (G-H) .Sections of ovary taken from an animal fed diet 3 for 6 mo. Note the lack of over-all integrity of the gonad in (J, with large populations of macrophages (ml. In H. an oocyte that has had its vitelline envelope (\ E) ruptured is undergoing phagocytosis by macrophages. Volk granules (VGl are also evident. (I-Jl Sections of ovary from an animal fed diet 4 for 6 mo. Note extensive voids left by degenerated gonad tissue (arrowhead) in I, and extensive macrophage activity in J. Scale bars in .\, C, E, G, and I are IIMI ^m, and in B, D, ¥, H, and J are III fim. as useful as digestible protein:digestible energy ratios (DP:DE). they were able to explain a significant amount of the variation in growth recorded in this study. The beta coefficients in Table 5 indicate the relative contribution of each variable to the regression equation. The standardized beta value of the CP:GE variable was higher than the protein leaching variable, indicating that CP:GE contributed the most toward the explanation of growth. The high- est TWBW growth was achieved on diets G. edulis and diets I and 2, which had CP:GE ratios of 18.2. 20.8, and 17.9 MJ/kg, respec- tively. We expect that further research concerning the DP:DE ratio will allow the production of cheaper and more effective feeds (see Britz 1996, Britz & Hecht 1997, Fleming et al. 1996, Vandepeer et al. 1999). The different methods and ingredients used to bind a diet will presumably affect the rate that protein is lost from a diet. Because protein is essential for soft tissue growth in abalone and is an expensive component of a formulated feed, the efficiency with which a diet binds its protein could conceivably determine its over-all quality. The results of the multivariate regression analysis suggest that this is an essential characteristic of a formulated diet. The two major differences in the FA content of G. edulis and the formulated diets lay in the amounts of 20:4n-6 and 18:2n-6 (Table 3). 20:4n-6 was 53 to 79 times more abundant in G. edulis than the formulated diets: whereas, 18:2n-6 was 40 to 54 times less abundant in G. edulis. Floreto et al. (1996) suggested that abalone may not be able to synthesize 20:4n-6 from the n-6 series of lower FAs and that it may be a nutrient that affects growth. This is supported by other workers (Dunstan et al. 1996, Mai et al. 1996. Uki et al. 1986). Although the percentage of 20:4n-6 in the fonnulated diets was much lower than in G. edulis. the higher measures of TL in diets I to 4 may have meant that this FA was adequately supplied. l8:2n-6 has been shown to enhance growth in other haliotids significantly (//. discus hanuai. Floreto et al. 1996. Mai et al. 1996, and H. tuherculaia. Mai et al. 1995). The obvious deficiencies of 18;2n-6 and total n-3 PUFAs in G. edulis may provide an expla- nation for the reduced growth rates noted in the present study in comparison to others studies on H. asinina that employed G. haili- nae (see Bautista-Teruel & Millamena 1999. Capinpin et al. 1999). It could also partially explain the reduced growth rates noted dur- ing months 5 to 6 (Fig. 3A and B). It would be informative to analyze the FA profile of G. hailinae for comparison with G. 634 Jackson et al. Fijjure 5. Essential amino acid profiles of 5 formulated diets and G. ediilis (diet A was taken from Bautista-Teruel & Millaniena 1999). The amount of each amino acid is expressed as a unitless proportion of the corresponding amino acid found in H. asiiiina soft body tissue. edulis. It is also interesting to note that among the formulated diets, the amounts of 18:2n-6 in diets 1 and 2 (which promoted signil'i- cantly higher TWBW growth) were higher than in diets 3 and 4. Bameveld et al. ( 1998) concluded that Australian abalone feed formulators are oversupplying the lipid content of their feeds. Not only does lipid contribute to the energy content of a diet formu- lation (Fleming et al. 1996) affecting the DP:DE ratio, it may also inhibit the utilization of other nutrients (Bameveld et al. 1998). TL content was not detected by the multivariate regression analysis as being a nutritional parameter that affected growth. However, its putative inhibitory effect on nutrients (such as vitainins, minerals, and digestible energy) was not considered in this analysis. Diets 2 and 4 contained the highest amounts of TL. Furthermore, diet 4 contained the lowest CP content of the formulated diets. The low growth rate produced by diet 4 may be a result of the inhibitory effect of a high lipid inclusion on digestible protein. G. cdiilis contained a lipid concentration of 0.8% and produced high grow th rates. This suggests that a diet with a blend of appropriate FAs may contain a minimal lipid content. The Japanese feed produced by the company Nihon Nosan Kogyo K.K. (NNKKK). which is re- ported to be a highly nutritious and well-balanced feed for abalone. contains a lipid content of 1.5% (Fleming et al. 1996), supporting evidence that abalone do not require a diet high in lipids. Allen and Kilgore ( 1975) determined the essential amino acids required by H. nifescens. These 10 amino acids have since been assumed to be essential for other haliotid species (Mai et al. 1994. Knauer et al. 1995, King et al. 1996), and are assumed to be essential for H. asinina in the current study. The ideal ratio of amino acids in dietary protein has been shown to reflect the ratio of amino acids in the body tissue of fish (Wilson & Poe 1985). This approach to abalone diet formulation has been used in several studies (Knauer et al. 1995, reviewed in Fleming et al. 1996. King et al. 1996). although its accuracy when applied to abalone nutri- tion has not been rigorously tested. However, a comparison of the amino acid profile of the formulated diets and G. edulis to that of H. asinina soft body tissue seemed to explain some of the observed variation in growth between diets. The essential amino acid pro- files of the five diets were compared by expressing each amino acid (g/kg) as a proportion of that found in the soft tissue of H. asinina (Fig. 5). Included in the comparison is the amino acid profile of a formulated diet (diet A) that was reported by Bautista- Tereul and Millaniena ( 1999) to produce a very high growth (247 |jLm/day) in similar sized H. asinina. This diet was included to test the validity of this comparison. According to this method of com- parison, a diet with consistent ratios of essential amino acids will provide adequate amounts of each residue: a diet with a flat profile, will be more balanced than a diet with any large fluctuations between residues. Such factors as the amount of protein in each diet, the feed intake, digestibility, and water stability will affect the over-all amount of each amino acid a\ ailable to the abalone. and hence the height of each column. From Fig. 5. none of the diets within the current study had essential amino acid profiles that seemed to be very balanced. The essential amino acids methionine and arginine seemed to be limiting in diets 1 to 4 and G. edulis. A deficiency in arginine and methionine was reported by Mai et al. ( 1994) in six species of macroalgae, and also seems to apply to G. edulis. Lysine also seemed to be deficient in G. edulis. The rela- tively high concentrations of methionine, phenylalanine, and lysine in diet A as compared to diets 1 to 4 and G. edulis is obvious and may explain the high growth rates achieved by Bautista-Tereul TABLE 7. Essential amino acid profiles expressed relative to lysine for seven haliotid species. Amino Acid H. asinina H. discus" H. iris'' H. midue^ H. tuherculata' H. rubra'' H. rufescens" Arginine 1 .-^0 1,3.^ 1 (I'l 1.27 1.22 1.70 1.21 Threonine 1.4b 0.b5 U.73 0.80 0.68 0.80 0.77 Leucine 1.40 1.14 1.24 1.11 1.15 1.16 1.31 Lysine 1.00 1.00 1.00 1.00 1 .00 1.00 1 .00 Valine 1.25 0.76 0.69 0.74 0.78 0.80 0.79 Isoleucine 0.78 0.70 0.63 0.66 0.74 0.70 0.63 Methionine 0.55 0.31 0.26 0.34 0.32 0.38 0.43 Tyrosine 0.51 0.51 — ■ — 0.59 — — Phenylalanine 0.42 0.59 0.71 0.60 0.58 0.60 0.68 Histidine 0.27 0.29 0.36 0.29 0.29 0.34 0.33 Cystine — 0.17 — — 0.24 — — Met -H cys — — 0.50 0.72 — 0.65 0.55 Phe + tyr 0.9.^ — 1.25 — — 1.13 1.32 Tryptophan — — 0.19 0.13 — 0.18 0.06 •' Mai et al. ( l'-W4). 'Tromak Technology (NZ) Ltd. as cited in Fleming ct al. ( 1946). "Knauer et al. ( 1995). ''King et al. (1996). "Allen and Kilgore ( 1975). Formulated Diet for Culture of Tropical Abalone 635 and Millamena (1999). However, the different experimental con- ditions between these two studies inust be considered when com- paring these growth rates. When the essential amino acids of H. asiniua are expressed relative to lysine, the profiles of all haliotids seem to be similar, except for higher amounts of threonine, valine, and methionine and lower amounts of phenylalanine H. usinina (Table 7). This sup- ports the conclusions of previous comparisons between other spe- cies (reviewed by Fleming et al. 1996) and suggests that if diets formulated for temperate abalone can be supplemented w ith threo- nine, valine, and methionine, they will be suitable for H. asiiiiiui. Gonad Index and Histology The effect of diet on sexual maturation was not initially in- tended to be a measure of the nutritional quality of each of the diets investigated in this study. However, after some animals displayed gonad tissue early on in the trial, it was decided to make further observations. Although the effect of diet on gonad maturation, as assessed by a noninvasive method, was not statistically significant for females (Table 6). the size at which females on the formulated diets started to mature was smaller than females reared on G. edulis. The size at which animals started to mature was well below that reported by other workers (Singhagraiwan & Doi 1992. Cap- inpin et al. 1998); the smallest animal recorded with the first signs of gonad tissue was a male with an .SL of 16.40 nun and weighing 0.917 g. Subsequent monitoring of individual abalone revealed monthly fluctuations in the presence and absence of gonad tissue (data not shown), indicating the ability of H. asiniim to resorb germinal tissue. This was reflected in the results of the histological analysis. Female animals reared on diets 3 and 4 showed marked differences in the over-all gonad integrity when compared to those on diets 1. 2. and G. edulis. The lack of structure of gonad tissue among female animals on diets _^ and 4 (Figs. 4D and E) is re- flected by the large populations of amoeboid cells evident in these gonads (Figs. 41 and J). In normal ovaries, amoeboid cells only proliferate after a spawning event and seem to be responsible for the degradation of unspawned mature oocytes (Jebreen et al. in press). Because of the random distribution of amoeboid cells and immature eggs within the ovaries of animals fed diets 3 and 4. it is assumed that this resorption of eggs is not a postspawning event, but is rather a response to either stress or malnutrition. Female animals reared on diets I. 2. and G. edulis did not possess exten- sive populations of amoeboid cells in the gonads. There were no apparent histological differences in the gonads of male animals reared on different diets. The precocious maturation by H- usinina maintained on for- mulated diets is an obvious concern for potential H. asininu farm- ers and feed formulators. Bautista-Teruel and Millamena (1999). achieved very high growth rates with H. usinina maintained on three formulated diet and noted no sexual maturation during the course of the 90-day experiment. Conceivably, a longer experi- ment would have seen reduced growth rates and yielded mature animals. This possibility is supported by Capinpin and Corre (1996). who found that all juveniles reared on a formulated feed and G. heterodada were sexually mature at the end of a 120-day experiment. They attributed reduced growth rates toward the end of the experiment to a channeling of energy into gonad maturation. Fleming et al. (1996) suggested that the nutritional lipid require- ment of abalone may increase during gonad development. Given the relatively high lipid content of the formulated diets tested within this study as compared to G. edulis. it is not surprising that female animals reared on the formulated diets were more mature than animals fed G. edulis. Ideally, feed formulations for "grow- out"" diets should maximize somatic growth and minimize germi- nal growth. Further research in this area with more appropriate experimental designs (i.e.. increased sample size and reduced han- dling stress) should yield a better understanding of the relationship between growth and reproduction in haliotids. Conclusion The results of this study indicate that H. asinina can be effi- ciently farmed on Australian diets formulated for temperate aba- lone. Efficient binding of nutrients into the diet to increase water stability, although retaining their biological availability for diges- tion still remains a challenge for feed formulators and seems to be an area of research that has not received the attention it deserves. The rates of protein loss from the formulated diets recorded in this study highlights the need for appropriate on-farm feeding strate- gies. Further research and development of these diets in terms of the DP:DE ratio, the amino acid requirements, and FA require- ments will increase growth, making the culture of this species more profitable. On the basis of these results. H. asinina seems to be an ideal species for aquaculture within Australia; its highly frequent, synchronous and predictable natural spawning pattern (Counihan et al. 2001). very high meat to shell weight ratio (>859'f ), and high growth rate potential in culture (Bautista-Teruel & Millamena 1999. Capinpin et al. 1999) meet the major require- ments for the economic production of any haliotid (Hahn 1989a). ACKNOWLEDGMENTS We thank Lina Daddow for expert histological advice and Maggie Barclay and Chris Wood for FA and amino acid analysis. This work was funded by a SPIRT grant to B. M. Degnan, K. C. Williams, and N, Preston. LITERATURE CITED Allen. W. V. & J. Kilgore. 1975. The essential amino acid requirements of the red abalone. Haliotis rufescens. Comp. Bio. Pliysiol. 50A:11\-11^. Apisawetakan, S.. A. Thongkukiatkul & C. Wanichanon. 1997. The ga- metogenic processes in a tropical abalone. Haliotis asinina Linnaeus. J. Sci. Sol: Thai. 23:225-240. Barneveld. R. J. V.. A. E. Reming. M. E. Vandepeer. J. A. Kruk & P. W. Hone. 1998. Influence of dietary oil type and oil inclusion level in manufactured feeds on the digestibility of nutrients by juvenile greenlip abalone {Haliotis Uwvigala). J. Slu-llfish Res. 17:649-655. Bautista-Teruel. M. N. & O. M. Millamena. 1999. Diet development and evaluation for juvenile abalone, Haliotis asinina: protein/energy levels. Aquaculture. 178:117-126. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of proiein-dve bindmg. Anal. Biocliem. 72:248-254. Britz. P. J. 1996. Effect of dietary protein on growth performance of South .African abalone. Haliotis mitlae. fed fishmeal-based semipurified diets. .Aquaculture. 140:55-61.5. Britz. P. J. & T. Hecht. 1997. Effect of dietary protein and energy level on growth and body composition. Aquaculture. 156:195-210. Capinpin. E. C. & K. G. Corre. 1996. Growth rate of the Philippine aba- lone, Haliotis asinina fed an artificial diet and macroalgae. Aquacul- ture. 144:81-89. 636 Jackson et al. Capinpin. E. C. J., V. C. Encena 11 & N. C. Bayona. 1W8. Studies on the reproductive biology of the donkey's ear abalone. Halioris asiiiiiui Linne. Aciuaculnire. 166:141-150. Capinpin, E. C. J.. J. D. Toledo. V. C. 1. Encena & M. Doi. 1999. Density dependent growth of the tropical abalone Hulintis usiniiia in cage cul- ture. Aquaculnire. 171:227-235. Counihan. R. T.. D. C. McNaniara, D. C. Souter. E. J. Jebrean. N. P. Pre- ston, C. R. Johnson & B. M. Degnan. 2001. Pattern, synchrony, and predictability of spawning of the tropical abalone HuUotis asiniiui from Heron Reef, Australia. Mar. Ecol. Plug. Ser. 213:193-202. Day. R. W. & A. E. Fleming. 1992. The determinants and measurement of abalone growth. In: S. A. Shepherd. M. J. Tegner & S. A. Gijzman del Proo, editors. Abalone of the world. Biology, fisheries, and culture. Oxford, UK: Fishing News Books. Oxford. UK: Fishing News Books, pp. 141-181. Dayton, P. K. 1994. Kelp forests. In: C. S. Lobban & P. J. Harrison, editors. Seaweed ecology and physiology. Cambridge. UK: Cambridge Uni- versity Press, pp. 75-77. Dunstan. G. A.. H. J. Baillie. S. M. Barrett & J. K. Volknian. 1996. Effect of diet on the lipid composition of wild and cultured abalone. Aqiia- ailnire. 140:115-127. Edwards. S. & K. Cook. 1999. Rapid leaching of minerals from abalone dietary binders is not determined by ionic mobility and competition between cations. Dry matter loss is not a good indicator. Proceedings of the 6th annual abalone aquaculture workshop. Sydney. Fleming. A. E.. R. J. Van Barneveld & P. W. Hone. 1996. The develop- ment of artificial diets for abalone: A review and future directions. Aciuuciihure. 140:5-53. Floreto. E. A. T.. S. Teshima & S. Koshio. 1996. The effects of seaweed diets on the lipid and fatty acids of the Japanese disc abalone Haliotis discus hamiai. Fish. Sci. 62:582-588. Folch. J., M. Lees & G. H. Sloane-Stanley. 1957. A simple method for the isolation and purification of total lipids from animal tissue. J. Biol. Chan. 226:497-509. Gosliner. T. M., D. W. Behrens & G. C. Williams. 1996. Coral reef animals of the Indo-Pacific. Monterey, CA: Sea Challengers. Hahn, K. O. 1989a. CRC handbook of culture of abalone and other marine gastropods, Boca Raton, FL: CRC Press. Hahn, K. O. l9S9b, Survey of the commercially important abalone species in the worid. In: K. O. Hahn. editor. Handbook of culture of abalone and other marine gastropods. Boca Raton. FL: CRC Press, pp. 3-39. Jarayabhand. P. & N. Paphavisit. 1996. A review of the culture of tropical abalone with special reference to Thailand. Aquaaillure. 140:159-168. Jebreen. E. J.. R. T. Counihan. D. R. Fielder & B. M. Degnan. 2000. Synchronous oogenesis during the semilunar spawning cycle of the tropical abalone Halioris asinina. J. Shellfish Res. 19:845-851. King. R. A., C. J, Rayner, M, Kerr, H, K, Gorfine & P, E. McShane. 1996, The composition and amino acid balance of abalone {Halioris rubra) tissue. Acfiaciilturc. 140: 109-1 1 3. Knauer, J., D. Brady, J. R. Duncan & T. Hecht. 1995. Amino acid, fatty acid, and mineral element profile of juvenile South African abalone, Halioris midae L. Aqua. Res. 26:283-288. Littler, M. M. & D. S. Littler. 1994. Tropical reefs as complex habitats for diverse macroalgae. In: C. S, Lobban & P, J, Harrison, editors. Sea- weed ecology and physiology. Cambridge. UK: Cambridge LIniversity Press, Cambridge, pp. 72-75. Mai, K,. J. P. Mercer & J. Donlon. 1994. Comparative studies on the nutrition of two species of abalone, Halioris niherculaia L. and Halioris discus haiinai Ino IL Amino acid composition of abalone and six spe- cies of macroalgae with an assessment of their nutritional value. Aqua- culnire. 128:11.5-130. Mai, K., J. P. Mercer & J. Donlon. 1995. Comparative studies on the nutrition of two species of abalone. Halioris ruherculara L. and Haliotis discus hannai Ino. 3. Response of abalone to various levels of dietary lipid. Aquaculrure. 1 34:65-80. Mai. K.. J. P. Mercer & J. Donlon. 1996. Comparative studies on the nutrition of two species of abalone. Halioris ruherculara L. and Halioris discus hannai Ino. V. The role of polyunsaturated fatty acids of mac- roalgae in abalone nutrition. Aquaculrure. 139:77-89. Mgaya. Y. & J. P. Mercer. 1995. The effects of size grading and stocking density on growth performance of juvenile abalone. Halioris ruhercu- lara Lmnneui. Aqiuiculrure. 136:297-312. Nateewathana. A. & J. Hylleberg. 1986. Thai species of abalone (Halioris spp.). Is aquaculture feasible? Thai Fisheries Ga:. 39:10-17. Prince. J. D. & S. A. Shepherd. 1992. Australian abalone fisheries and their management. In: S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: Biology, fisheries, and culture. Oxford, UK: Fishing News Books, pp. 407-426. Sawatpeera, S„ E. S. Upatham, M. Kruatrachue. V. Ingrsisawang. T. Sing- hagraiwan. Y. P. Chitramvong & K. Parkpoomkamol. 1998. Determi- nation of gut contents of Thai abalone Halioris asinina Linnaeus. / Shellfish Res. 17:765-769. Singhagraiwan. T. & M. Doi. 1992. Spawning pattern and fecundity of the Donkey's ear abalone, Halioris asinina Linne. observed in captivity. Thai Mar. Fish. Res. Bull. 3:61-69. Singhagraiwan, T. & M. Doi. 1993. Seed production and culture of a tropical abalone, Halioris asinina Linne. In: The research project of fishery resource development in the kingdom of Thailand. Government of Thailand, pp. 32, Singhagraiwan, T. & M. Sasaki. 1991. Growth rate of Donkey's ear aba- lone, Halioris asanna Linne, cultured in tank. Thai Mar. Fish. Res. Bull. 2:95-100. StatSoft, 1994. STATISTICA for the Macintosh, 4.1, StatSoft, Oklahoma Uki, N.. M. Sugiura & T. Watanabe. 1986. Requirement of essential fatty acids in the abalone Haliotis discus hannai. Bull. Japan. Soc. Sci. Fish. 52:101.3-1023. Upatham, E. S., S. Sawatpeera, M. Kruatrachue, Y. P. Chitramvong, T. Singhagraiwan, T, Pumthong & P. Jarayabhand. 1998. Food utilization by Halioris asiniiui Linnaeus. J. Shellfish Res. 17:771-776. Vandepeer, M. E., P. W. Hone, R. J. V. Barneveld & J. N. Havenhand. 1999. The utility of apparent coefficients for predicting comparative growth performance in juvenile greenlip abalone Halioris laevigara. J. Shellfish Res. 18:235-241. Viana, M. T., L. M. Lopez, Z. Garcia-Esquivel & E, Mendez. 1996. The use of silage made from fish and abalone viscera as an ingredient in abalone feed. Aquaculrure. 140:87-98. van Wijngaarden, D. 1967. .Anal. Chan. 39:848-849. Wilson. R. P. & W. E. Poe. 1985. Relationship of whole body and egg essential amino acid patterns to amino acid requirement patterns in channel catfish. Co/u/j. Bio. PInsiol. 80B:38.5-388. Joiiival oj Shellfish Rescuirh. Vol. 20. No. 2. 637-64.'S, 2001. THE PARTIAL AND TOTAL REPI ACEMENT OF FISHMEAL WITH SELECTED PLANT PROTEIN SOURCES IN DIETS FOR THE SOUTH AFRICAN ABALONE, HALIOTIS MIDAE L. T. A. SHIPTON AND P. J. BRITZ Department of Iclitliyology and Fisheries Seience. Riwdes University. Grahanistinvn 6140. Sontli Africa ABSTRACT As protein is the most expensive nutrient to supply in abaione leeds. it is necessary to evaluate the relative performance of (Jifferenl protein sources in formulated diets. Fourteen diets were formulated to contain 34% crude protein and 69r lipid. These were fed to juvenile abaione, HaUolis miilae (initial shell length; 10.6 ± 0.1 mm). Dietary fishmeal was stibstituted at 30. 50. 75, or 100% with plant protein concentrates, and the growth and nutritional parameters were recorded over a 180-day growth period. No significant differences were found in the growth rates between the control diet ( 100% fishmeal) and diets in which 30% of the fishmeal component had been replaced by either soy or sunflower meals, or torula yeast {P > 0.05). Substitution of 50% fishmeal with either soya meal or spirulina did not affect growth rates iP > 0.05). Replacement of either 75 or 100% of the fishmeal with plant protein sources had a significant affect on growth ^P < 0.05). Pearson product moment correlations between dietary lysine levels and either growth rates or protein efficiency ratios revealed positive correlations (r = 0.75. /=< 0.002; r = 0.88. />< 0.0001. respectively), suggesting that lysine may have been the first limiting amino acid in these diets. Carcass analysis revealed that dietary protein source had no significant affect on body composition (P > 0.05). KEY WORDS: abaione nutrition, fishmeal replacement, plant proteins. Halinus miilae INTRODUCTION Increasing concern about the future supply and demand for fishmeal (Barlow 1989. Rumsey 1993) has increased efforts to reduce its use as the major protein source in commercial aquacul- ture feed formulations. The reduction of the fishmeal component in feeds has met with considerable success in a number of species. and suggests that its partial substitution with alternative protein sources need not compromise growth. The most notable success to date is illustrated by the channel catfish industry, where the sub- stitution of fishmeal with soy. meat, and blood meals has reduced current fish meal inclusion rates to as low as 0^% as compared with >30% in pre-1975 formulations (Stickney et al. 1996). Al- though the level of substitution has not been as dramatic in other sectors of the industry, it has been demonstrated that significant reductions in dietary fishmeal are possible. Among others, these sectors include the salmonids (Pongmaneerat & Watanabe 1992. Moyano et al. 1992). tilapias (Jackson et al. 1982. Lim 1989a). and shrimp (Lim 1989b). With respect to abaione nutrition, a number of studies have compared fishmeal-based diets with other single protein formula- tions (Uki et al. 1983a. Uki & Wantanabe 1986. Uki & Watanabe 1992. Viana et al. 1993. Britz 1996). and although these studies give an indication of which proteins may have potential to become fishmeal substitutes. Fleming et al. (1996) reviewed the nutritional composition of twelve currently available commercial and experi- mental dietary formulations, and found that of the eight diets that contained fishmeal, all contained supplementary proteins. How- ever, no indication of the effects that these alternative protein sources had on growth and nutritional indices is available. Never- theless, considering the success of previous growth trials using single plant protein sources, and the fact that Haliotids are natu- rally herbivorous animals with a digestive and enzymatic physiol- ogy equipped for processing plant materials (Knauer 1994), the prospects for replacing dietary fishmeal with plant protein sources are promising. The aim of this study was therefore to determine the effects that the partial and total replacement of dietary fishmeal with locally available plant protein sources has on growth, nutri- tional indices, and body composition of juvenile H. midae. The protein sources chosen for investigation were: soy and sunflower meals, spirulina. torula yeast, and corn gluten meal. MATERIALS AND METHODS Experimental Animals and System The experiment was conducted at the Port Alfred Laboratory using water drawn from the mouth of the Kowie River estuary on the Indian Ocean (33°45'S; 26°0()'E). A 10.000 L semirecirculat- ing system was employed. Water quality was maintained using biological filtration and the replacement of 25% of the system volume per day. A 1 2 light; 1 2 dark photoperiod was used through- out the experimental period. Temperature and salinity were main- tained at 20 ± IC and 35 ± 1 ppt. respectively. Animals were housed in 24 fiberglass experimental containers (30 x 50 x 26 cm. 39 L volume, four water exchanges per hour). Within experimental containers, three replicate groups of animals were housed in per- forated 2-L plastic jars. Water flow through the jars was main- tained by the use of airstones. Juvenile hatchery-reared abaione (10.6 ± 0.1 mm shell length, 0.26 ± 0.01 g weight) were used in the experiment. Before experi- mentation, the animals were fed exclusively on a fishmeal-based artificial diet (34'^/f crude protein, 4.5% lipid) and acclimated to the experimental system for a I mo period. Three replicate groups of 30 abaione were assigned to each dietary treatment, and placed in the experimental system according to a randomized block design. The duration of the experiment was 180 days. Whole body wet weight and shell length measurements were taken at monthly in- tervals throughout the experimental period. Weight was recorded to the nearest 0.01 g using an electronic balance and shell length to the nearest 0.01 mm using vernier calipers. Before these mea- surements were taken, the animals were starved for a 24-h period to ensure that their last meal had been processed and did not affect the animal" s weight. A 10% (w/w) solution of magnesium sulfate was used to anesthetize the animals during the measurement pro- cess. Animals were fed daily at 1700 h. and the feed consumed over the experimental period was recorded and conected for leaching. Correction factors for total solids leached were determined for each diet by placing 4 g of feed in a control container containing 637 638 Shipton and Britz no abalone and placing it in the aquaria for a 12-li period. Total solids leached were calculated as dry weight loss over this period. Feed conversion ratios (grams dry feed consumed /grams wet weight gain) and protein efficiency ratios (grams wet weight gain/ grams protein consumed) were calculated for all treatments. Feed consumption and condition factors for all replicate treatments were calculated according to Britz (1996). Dietary Fonmilatioiis Thirteen experimental diets were formulated to contain 34% crude protein. 4.5% lipid; a fourteenth diet containing 34% LT- fishmeal and 4.5% lipid was used as a control (Table 1 ). Formu- lations between experimental diets varied such that the fishmeal component of the control diet was substituted at either 30, 50, 75, or 100% with the selected plant protein sources. The diets were bound using agar according to the method described by Knauer ( 1994). In suinmary, 20 g agar was boiled in 1,200 mL water for 30 sec and allowed to cool to 45°C. The agar was then mixed with 80 g of dietary mixture, spread into a flat sheet 0.5-cm thick and allowed to set at room temperature. Once set. the diet was cut into squares (0.5 x 0.5 cm), oven dried at 60°C for 24 h and stored frozen at -20°C until required. Carcass and Dietary Analysis Proximate analyses of the dietary treatments and the abalone at the beginning and end of the experimental period were undertaken. ."Mthough triplicate 5-g samples of feed were used to determine the proximate coinposition of the dietary treatments; thirty animals at the start of the experiment and five animals from each replicate at the end of the experimental period were used to determine the proximate composition of the abalone. The abalone were initially frozen at -20°C for subsequent analysis. On defrosting, the ani- mals were weighed and shucked, and the soft body material was homogenized using an Ika Turrax T25 homogenizer. With the exception of lipid, the proximate composition of both abalone and the diets were measured according to the standard methods of the AOAC (1984). Thus, crude protein was determined as total Kjeldahl nitrogen (N x 6.25), ash as the residue following the combustion of samples at 550°C for 12 h, and moisture by oven drying to constant weight at 104 C. Lipid was determined by the chloroform/ methanol extraction method (Folch el al. 1957), and the nitrogen-free extract was estimated from the difference. Statistical Analysis Growth rates were determuied as the slopes of the natural log- transformed weight data. The slopes of the growth models were compared using analysis of covariance (ANCOVA) followed by a one-tailed F-test (Zar 1984). One-way analysis of variance and Student's-Newman-Keul's multiple comparison procedure were used to determine significant differences in feed consumption, condition factors, food conversion ratios (FCR), and protein effi- ciency ratios (PER) between dietary treatments, and, in addition, the effect of dietary treatment on the proximate composition of the abalone at the end of the experimental period. Comparisons be- tween the essential amino acid profiles of the dietary treatments and the whole soft tissue of H. midac were undertaken using Pearson product moment correlations. In addition, Pearson product moment correlations were used to determine the strength of asso- ciation between growth rates, protein efficiency ratios, and the levels of individual essential amino acids across all dietary treat- ments. RESULTS The partial and total replacement of dietary fishmeal with the alternative protein sources significantly affected abalone growth over the experimental period (Tables 2 and 3). Although five of the TABLE L Ingredient composition (% inclusion) and nutrient analysis of experimental diets; analysis expressed as a dry weight basis. Control Fishmeal/ Fishmeal/ Fishmeal/ Fishmeal/ Fishmeal/ Fishmeal/ fishmeal soy meal soy meal sunflower meal sunflower meal torula yeast torula yeast Diet 100 (%) 70:30 ( % ) 50:50 ( % ) 70:,^) ( '^c ) 50:50 ( % ) 70:30 (%) 50:50 ( % ) Danish LT fishmeal' 47.7 33.0 23.5 33.0 23.5 33.0 23.5 Soy inear — 22.5 37.5 — — — — Spiriilina''^ — — — — — — — Suntlowcr meal'' — — — 28.6 47,7 — — Torula yeasC' — — — — — 23.8 39.7 Starch 3L3 21.5 15.4 15.8 5.9 n ") 15.0 Lipid" — 2.0 2.6 1.6 1.9 0.9 0.8 Vit/min mix^ 1.0 1.0 1.0 1,0 1.0 1.0 I-O Agar' 20.0 20.0 20.0 20.0 20.0 20.0 20.0 Analysis calculated Protein 34.2 34.2 34.2 34.2 34.2 34.2 34.2 Lipid 4.5 4.5 4.5 4.5 4.5 4.5 4.5 Analysis (measured) Protein 35.1 35.6 33.0 31.5 33.3 31.4 32.2 Lipid 4.5 4.4 4.4 4.5 4.3 4.3 4.2 Ash 7.8 7.5 7.0 7.8 8.1 9.0 10.4 Moisture 5.2 6.4 6.5 6.3 5.3 6.8 6.2 Gross energy (MJ/Kg) 19.62 19.51 19.17 18.52 19.42 18.81 18.94 Leaching factor (soHds) 0.46 O.Sl O.Sl 0.')] 0.9.^ 0.91 0.92 continued on next page FisHMEAL Replacement in Abalone Diets 639 formulations produced growth rates that were not significantly different to the control, the remaining eight formulations produced significantly lower growth rates. Growth rates and increases in abalone weight over the experimental period ranged between 1 .09 to 1.41% body weight/day"' and 600 to 1.2.^0% respectively. FCRs and PERs ranged from 0.80 to 1 .25 and 2.\5 to .^.75 respec- tivel). With respect to the control diet (lOO'/r fishmeal). only torula yeast significantly improved feed conversion at both the .^0 and 507f fishmeal substitution levels (P < 0.03. Table 2). However, although a 30% substitution of fishmeal with torula yeast did not significantly affect growth, at 50% substitution, growth was sig- nificantly lower than the control {P = 0.00027. Table 3). With respect to protein efficiency, a 30% substitution of fishmeal for torula yeast significantly improved efficiency (P < 0.05). but at a 50% substitution level, protein efficiency did not differ signifi- cantly from the control. In addition, consumption rates did not differ significantly between the control and the diets containing either 30 or 50% torula yeast. In comparison with the control, the inclusion of suntlower meal at 30% or soy meal at either 30 or 50% did not significantly affect growth rates iP > 0.016. Table 3). Nevertheless, the inclusion of com gluten meal at either 30 or 50% or sunflower meal at 50% produced significant reductions in growth rates iP < 0.016. Table 3). In comparison with the control diet, feed conversion was sig- nificantly improved when soy meal was used at the 30% substi- tution level (P < 0.05. Table 2) but did not differ significantly at a 50% level of substitution. In contrast, the opposite was found to be the case when either sunflower or corn gluten meals were used as substitutes. In both cases, a 30% fishmeal substitution produced an insianificant reduction in feed conversion ratio. At the 50% sub- stitution level, feed conversion was further reduced, and in both cases, found to be significantly lower than control (P < 0.05). With respect to protein efficiency, inclusion of either soy. sunflower, or com gluten meals produced a significant reduction in efficiency at both levels of inclusion. It is. however, interesting to note that at the 30% tlshmeal substitution level, there were no significant dif- ferences in the protein efficiency ratios between the three protein sources. However, at the 50% substitution le\el. the protein effi- ciency ratios of the suntlower and corn gluten formulations were significantly lower than the soy meal formulation. At the 30% fishmeal substitution level, the consumption rates of these diets were higher than those observed for the control; however, it was only when inclusion rates were increased to 50% that they were significantly higher than the control (P < 0.05). One exception was the diet containing corn gluten meal, where a comparison with the control diet produced no significant difference in consumption rates (P > 0.05). As respective consumption rates and final body weights were generally found to increase and decrease across all the diets in which fishmeal had been substituted at either 30 or 50% with soy. suntlower. or corn gluten meals, a Pearson product moment correlation was performed between the consumption rates and final body weights to determine whether consumption rates were related to animal size. No significant correlation was found (r = -0.42. P > 0.05, II =7); thus, it seems that differences in animal size cannot adequately account for the observed consumption rates. Differences in consumption rates may therefore be attributed to dietary factors. The results using spirulina as a partial fishmeal replacement, demonstrated that a 50% substitution of fishmeal w ith spirulina produced neither a significant reduction in growth rate (P > 0.0125, Table 3). nor a significant reduction in feed conversion ratio (P TABLE 1. continued Fishmeal/ Fishmeal/ Spirulina/ Spirulina/ Fishmeal/ Fishmeal/ soy meal/ soy meal/ soy meal/ soy meal/ Fishmeal/ corn gluten corn gluten sunflower meal sunflower meal sunflower meal sunflower meal spirulina Diet 70:30 ( % ) 50:50 ( % ) 50:25:25 (%) 25:37.5:37.5 (%) 50:25:25 ( % ) 25:37.5:37.5 (%) 50:50 (%) Danish LT fishmeal'' 33.0 23.5 23.5 II.S — — 23.5 Soy" — — 18.8 28.1 18.8 28.1 — Spirulina'^^ — — — — 30.0 15.0 30.0 Sunflower'' — — 23.9 35.8 23.9 35.8 — Com gluten'' 15.9 26.6 — — — — — Starch 28.3 26.7 10.5 0.5 3 0.8 23.9 Lipid^' 1.8 -) 1 2.3 2.8 3.3 3.3 1.6 Vit/min mix^ 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Agar' 20.0 20.0 20.0 20.0 20.0 16.0 20.0 Analysis (calculated) Protein 34.2 34.2 34.2 34.2 34.2 34.2 34.2 Lipid 4.5 4.5 4.5 4.5 4.5 4.5 4.5 Analysis (measured) Protein 33.4 35.6 33.4 33.4 30.6 30.7 30.1 Lipid 4.7 4.8 4.2 4.2 4.2 4.1 4.2 Ash 5.9 5.0 7.8 7.6 11.2 9.5 11.4 Moisture 6.1 4.9 5.8 6.1 6.7 5.3 5.4 Energy (MJ/Kg) 20.29 19.36 19.65 19.23 19.65 19.23 18.27 Leaching factor (solids) 0.93 0.89 0.92 0.88 0.81 0.87 0.83 Dietary components supplied by: ■'Danish 999 tlshmeal: S.A.; ''■50/50 mixture of sunflower oil and "Marinol R" ;. Esbjerg Fiskeindustri. Denmark; ''EPOL. S.A.; ^Western Cape Tanneries. S.A.; '■' fish oil: 'Continental Lab. Supplies. S.A.: ^^Composition from Uki et al. (1985b). Nova Suaar. 640 Shipton and Britz § 5 o 3 > 5 .2^ ^ = I' - i ^ " O — — r 1 O --r O r^i i^' ooooooooo ■ — — — ■ — ri n — O r-i r*-, r-, ri f^l c^l r*-. r^ (N f^j in — r- — ' r^ — O ovoovo-^or-p- — o — d d o o +1 +1 +1 +1 p -. — ■ p +1 +1 +1 +1 +1 O OC O Ov GC r^, r*-, r-1 n r-i f, f) CN +1 +1 +1 +1 +1 +1 +1 +1 +1 r\ r\ ri ri n ri ri ri r-\ — — ri r*-, -t r*-| r*-) -r)- -rf f<-, d d d d d d d +1 +1 +1 +1 +1 +1 +1 — CO ^- m r~- ■^ — ^ ^ -^ r 1 ri ^^^ r^ rJ oj rj (N ri ooooooooo +1 +1 +1 +1 +1 +1 +1 +1 +1 3^00\Dv-)^■^0^»noO ri ri +1 +1 +1 +1 s S- „_ _ i i o d o d iri o o o o s o o 4j "5 d d d iir. 5 2; o P S £ E :; ■S n 1' y U s s >^ >. Ji ^ - = c c — - >i >i c n P 5 ^ OJ l; u OJ c J= -n x: -c ^ y. ^ ^/: ^ E £ £ lO >% >-, r-^ g in g r*-) ^ ri ^ >r! ~ in c r-' rj r^i p d 3 »/~, ~ 'n ri — '.: — ' ; c tu tt. u. u. u. p 11 C3 1J < JT u ■i o 1) Ol) ri c O # :5 -3 * < ^ > 0.0 ^ -^^ Z J= OJ < ^ D- . 'oj => O „ .y c^ g S F A Control-fishnieal 100 (%) B Fishmeal/soy meal 70:30 (%) C Fishmeal/soy meal 50:50 (%) D Fishnieal/suntliiwer meal 70:30 ('i) E Fishmcul/suntlower meal 50:50 (% ) F Fishmeal/torula yeast 70:30 C/r) G Fishmeal/loriihi yeast 50:50 (%) H Fishmeal/eorn iikiten 70:30 {9c) 1 Fishmeal/eorn gluten 50:50 (%) J Fishmeal/soy meal/sunflower meal 50:25:25 (9M K Fishmeal/soy meal/sunflower meal 25:37.5:37.5 (%) L Spirulin;i/soy nieal/siinnovver meal 50:25:25 (%1 M Spirulina/.soy nieal/suntlower meal 25:37.5:37.5 (%) N Fishmeal/spiriilina 50:50 {'7c) In y = - .26 + 0.01 3Sx 628 0.89 In y = - .32 + 0.0131 X 621 0.86 4.72'^" 1 . 1 245 P = 0.03005 NS In y = - .24 + 0.0132X 622 0.90 5.59*"" 1.1246 P = 0.01823 NS 0.00"' 1 . 1 2.W P = 0.96 NS In y = - .38 + 0.0I37X 602 0.90 0.16^" 1.1226 P = 0.68 NS In y = - .32 + 0.0I22X 616 0.83 29.48*'^ 1.1240 P < 0.00001 * 27.63'^'^ 1 , 1 209 P < 0.00001 * In y = - .28 + 0.0141X 613 0.82 0.55^'" 1.1237 P = 0.46 NS In y = - .19 + ().0127x 614 0.85 13.31^° 1.1238 P = 0.00027 * 14.65"=° 1.122.^ P = O.(K)013 * Iny = - .28 + 0.012 Ix 621 0.85 34.49"" 1 . 1 245 P < 0.00001 * In y = - .34 + 0.0 KWx 616 0.82 103.36*' 1.124(1 /'< 0.00001 * 17.22"' 1.1233 P = 0.00003 * In y = - .23 + 0.01 lyx 621 0.82 39.04-^' 1.1245 P < 0.00001 * In y = - .34 + 0.01 I7x 637 0.81 47.()6'''" 1.1261 P < 0.00001 * 0.43"^ 1.1254 P = 0.51 NS In y = - .44 + O.0125X 604 0.86 20.19''' 1.1228 P < 0.00001 * In y — — .33 + O.0I22X 600 0.87 31.38-"' 1.1224 P < 0.00001 * In y = - .27 + O.OI28X 607 0.84 6.19''^ I.I2II P = 0.01298 NS 0.92''" 1.1200 P = 0.33 NS 2.76"''' 1.1187 P = 0.096 NS 6.74"^ 1.1283 P = 0.00956 * NS: Not significant. *: Significant. Growth models expressed as the natural log of the weight (y in grams) as a function of time (x in days) for the ahalone fed the 16 experimental diets. Superscripts next to the f-values indicate which models were compared (h = number of values in model. /' = coefficient of determination, df = degrees of freedom. P = significance). To avoid accumulated error, significance was denoted at 0.0166 for three model comparisons, and 0.0125 for four model comparisons. > 0.05. Table 2). Furthermore, allhoiiiili the animals fed this diet consumed significantly more feed than those animals fed the con- trol diet {P < 0.05. Table 2); protein efficiency was significantly reduced {P < 0.05). indicating less efficient protein utilization. It is interesting to note that in the four diets containing the soy meal/ sunflower ineal combinations and fishmeal or spirulina at either 25 or 50% inclusion; feed conversion and protein efficiency were greater in those diets containing the fishmeal component as op- posed to the spirulina. Nevertheless, a comparison between the diets containing spirulina or fishmeal at the 259^ inclusion level, revealed that there were no significant differences between the feed conversion or protein efficiency ratios {P > 0.05. Table 2). However, at the 50% fishmeal or spirulina inclusion level, feed conversion, and protein efficiency were significantly reduced in those animals fed the spirulina-based diets {P < 0.05, Table 2). A comparison between the consumption rates revealed that at the 50% inclusion rate, those animals fed the spirulina-based diet con- sumed significantly more than those fed the fishmeal-based diet; however, at the 259!- inclusion rate, consumption rates did not differ significantly. Coinparisons between the control and the diets containing the soy/sunflower combinations with fishmeal at either 25 or 50%. revealed that both formulations produced significant reductions in growth rates {P < 0.00001, Table 3). feed conversion, protein efficiency, and consumption {P < 0.05, Table 2). Nevertheless, it is interesting to note that between these formulations, no signifi- cant differences were found between either growth rates or the nutritional indices {P > 0.05, Tables 2 and 3). To determine whether the dietary essential amino acid balance had an effect on growth and protein utilization by the abalone. a preliminary analysis using Pearson product moment correlations between the essential amino acid profile of the soft body tissues of H. undue, and the calculated essential amino acid profiles of the dietary treatments was undertaken (Table 4). The coefficients (r) from these correlations indicated that, with respect to the amino acid balance of the commercial formulation, and with the excep- tion of those diets containing corn gluten, the amint) acid balances of all the dietary treatments produced equal or closer approxima- tions to the balance of amino acids found in the soft body tissues of the abalone. Secondaiy analyses, using Pearson product mo- ment correlations were performed between the dietary/abalone EAA correlation coefficients and either the respective growth rates or the protein efficiency ratios. Because no significant correlations were observed {P > 0.05, Table 5), it can be assumed that, in this 642 Shipton and Britz TABLE 4. Essential amino acid profile of the whole soft body tissues of H. midae, and the calculated essential amino acid profiles of the experimental diets. .Methionine Phenylalanine .4rginine Histidine Isolfucine Leucine Lysine + eysline Threonine + tyrosine Tryptophan Valine r //. midae 7.91 1.82 4.11 6.93 6.21 3.44 4.99 7.71 0.82 4.61 — Conlrol-fishnieal 100 5.62 2,92 3,97 7,25 7.52 3,66 4 13 6.89 1.15 4.78 0.90 Fishmeal/sov meal 70:30 5.<)l 2.74 4,1 1 7.21 7.00 3.38 3.97 7.07 1.22 4.66 0.93 Fishmeal/soy meal 50:50 6.20 2.65 4.26 7.28 6.75 3.24 3.92 7.29 1.29 4.64 0.95 Fishmeal/sunflower meal 70:30 6.74 2.82 4.20 7.49 6.44 3.52 4, in 7.19 1.19 4.99 0.96 Fishmeal/sunnower meal 50:50 7.58 2.79 442 7.76 5.83 3.48 4.14 7.49 1.23 5.20 0.97 Fishmeal/lorula yeast 70:30 5.49 2.83 4.48 7 16 7.48 3.37 4.46 7.72 1.11 5.11 0.91 Fishmeal/torula yeasi 50:50 5.49 2.81 4.89 7.21 7.56 3.23 4.75 8.38 1.11 54(1 0.90 Fishmeal/corn gluten 70:30 4.87 2.70 4.32 10.07 5.76 3.73 3.85 7.43 0.94 4.84 0.80 Fishmeal/corn gluten 50:50 4.44 2.59 4.61 12.09 4.69 3.83 3.73 7.89 0.82 4.95 0.70 Fishmeal/spirulina 50:50 6.04 2.41 4.80 7.89 6.10 3.49 4.27 7.23 1,22 5.61 0.93 Fishmeal/soy meal/sunllower meal 50:25:25 6.95 2.74 4.37 7.58 6.33 3.59 4.06 7.45 1.27 4,96 0,97 Fishmeal/soy meal/sunniiwcr meal 25:37.5:37.5 7.51 2.62 4.52 7.66 5,69 3.51 3.98 7.64 1.31 499 0.97 Spirulina/soy meal/suntlovvcr meal 50:25:25 7.31 2.21 5.16 8.17 4.88 3..39 4.17 7.73 1.33 5.75 0.93 Spirulina/soy meal/suntlower meal 25:37.5:37.5 7.73 2 36 4.94 798 4,97 3.43 4(15 781 1 35 5.41 0,95 Comparison between dietary essential amino acid profiles and the essential amino acid profile ofabalone are presented as the correlation coefficients between the dietary and abalone EAAs (r). Amino acid values are presented as a percentages of protein (dry weight). The amino acid profiles of H. muiac, spirulina, and torula yeast were taken from Britz ( 1995). The Ushmeal amino acid profile was supplied by 999 Esbjerg Fiskeindustri. Denmark, and those of soy meal, sunflower meal, com gluten, and maize taken from NRC 1993. Semolina was determined by the Department of Poultry Science. University of Natal. SA. Study, dietary e.ssential amino profiles could not be used to predict potential feed utilization by the abalone. In contrast. Pearson prod- uct moment correlations between the levels of individual dietary essential amino acids and either the growth rates or protein effi- ciency ratios (Table 5). revealed that dietary lysine levels were significantly and positively conelated to both growth and pro- tein efficiency (growth rate: r = O.T.'i. P < 0.002; PER: r = 0.88. TABLE 5. Pearson product moment correlations (rl between either the dietary essential amino acids or the correlation coefficients (r) between dietary/abalone EAA profiles, and growth rates and protein efficiency ratios. Factor Amino acid Groyyth rate Protein eff ciency ratio Arginine = -0.05 P > 0.05 r ^ -0.42 P > 0.05 HisticJine = 0.41 P > 0.05 r = 0.67 P < 0.05 Isoleticine = -0.38 P > 0.05 r = -0.48 P > 0.05 Leticine = -0.69 P<0.0\ r = -0.46 P > 0.05 Lysine = 0.75 P < 0.002 r = 0.88 P< 0.0001 Methionine +cvstine = -0.46 P > 0.05 r = -0.17 P > 0.05 Threonine = 0.47 P > 0.05 r = 0.46 P > 0.05 Phenylalanine + tyrosine = -0.43 P > 0.05 r = -0.22 P > 0.05 Tryptophan = 0.23 P > 0.05 r = -0.14 P > 0.05 Valine = -0.12 P > 0.05 r = -0.36 P > 0.05 Correlation coefficients (r) between diet try and abalone EAA profiles r = 0.41 P > 0.05 r = 0.09 P > 0.05 P < 0.0001). Furthermore, although leucine was found to be nega- tively correlated to the growth rate (r = -0.68. P = 0.003), and histidine positively correlated to protein efficiency (r = 0.67, P < 0.05). the other essential amino acids were not significantly cor- related with either growth or protein efficiency (P > 0.05). The body composition of the abalone at the end of the experi- mental period revealed that diet did not significantly affect the protein, lipid, carbohydrate (nitrogen-free extract), ash. or mois- ture levels in the soft body tissues of the animtils (/-" > 0.05, Ta- ble 6). Although condition factors were found to reduce in all treat- ments over the experimental period, at the end of the experiment, there were no significant differences in condition factors between the control diet and the experimental diets {P > 0.05, Table 2). The exceptions being those diets in which 50% of the fishmeal was substituted with either sunflower meal or corn gluten meal, and, in addition, those diets containing the mixture of soy and suntTower meals in combination w ith 50'/f fishmeal or 25% spirulina. In these diets, condition factors were found to be significantly reduced with respect to the control diet (P < 0.05). Mortality of the abalone was low, ranging from 0-4.4%, and was attributed to the small animal size and the consequent suscep- tibility to damage caused by handling stress during weighing and measuring. There was no evidence of disease in any of the repli- cates. DISCUSSION Without exception, the substitution of dietary fishmeal with plant protein sources increased coiisumptittn rates in the abalone. However, because feed consumption may be affected by a number of factors, it is difficult to determine causality. Among others, factors that have been shown to affect consumption in poikilo- FisHMEAL Replacement in Abalone Diets 643 TABLE 6. Proxiniatf composition ( 'i ) of tht soil tissue of tlie abalone at the end of the experimental period. Diet Protein ( % ) {II X 6.25) Lipid ( % ) Nitrogen-free extract ( % ) Ash ( % ) Moisture ( % ) Control-fishmeal 100 (%) Fishmeal/soy meal 70:30 C'r ) Fish meal/soy meal 50:50 ('{) Fishmeal/sunnower meal 70:30 (%) Fishmeal/sunflower meal 50:50 (%) Fishmeal/torula yeast 70:30 (%) Fishmeal/torula yeast 50:50 (%) Fishmeal/com gluten 70:30 (%) Fishmeal/corn gluten 50:50 (Vr) Fishmeal/soy meal/sunflower meal 50:25:25 (%) Fishmeal/soy meal/suntlower meal 25:37.5:37.5 (%) Spimlina/soy meal/suntlower meal 50:25:25 (%) Spirulina/soy meal/sunflower meal 25:37.5:37.5 (%) Fishmeal/spirulina 50:50 t''< I 55.4' 5.2-' 26.r 1 3.3- 74.6- 56.6^' 5.4" 23.4-' 14.6" 75.6" 54.9" 5.7" 26.5-' 12.9-' 76.6" 55.9" 5.1" 25.()-' 14.0-' 76.3" 52.8" 5.6" 28.I-' I3..5- 75.2" 55.3" 5.3" 25.1- 14.3" 75.6" 55.9" 5.9" 23.6-' 14.6-' 76.3" 54.6" 5.4" 25.8-' 14.2-' 77.9" 51.6" 5.6° 29.1" 13.7" 75.6" 54.6" 5.7" 25.8-' 13.9" 74.9" 56.9" 5.3" 23,5-' 14.3" 75.8" 55.3" 5.5" 24.6" 14.6" 76.1' 51.0" 5.3" 29.9-' 13.8" 77.2" 55.2" 5.4'' 26.4' 13.0" 76.4" With the exception of moisture, all values are expressed on a dry weight basis. Different alphabetic superscripts within a column indicate significant differences between treatments for the parameter concerned (it = 5. P < 0.05). Initial values for the abalone at the start of the experimentLiI period were: protein — 52.6%; lipid — 4.1%; ash — 15.2%; nitrogen-free extract — 28.1; moisture — 75.6% (n = 30). therms include dietary energy density (Lee & Putnian 1973. Weis- berg & Lotrich 1982). and feed attractiveness and palatability (Harada et al. 1996). If it is assumed that animals eat to satisfy their energy require- ments (Smith 1989). it is reasonable to suggest that the differences in digestible energy across the dietary treatments provide an ex- planation for the observed consumption rates. It is possible that high levels of indigestible carbohydrates present in the plant pro- tein sources may have reduced the dietary digestible energy den- sities. Thus, as the animals adjust their feed intake to satisfy their energetic requirements, they consume more of the fishmeal/plant protein diets that contain less available energy than the fishmeal- based diet. However, as it has also been demonstrated that dietary protein source may affect feed consumption in Haliotids (Uki & Watanabe 1986, Viana et al. 1994), the effect of differential at- tractiveness and palatability between protein sources cannot be discounted as factors that may have affected consumption. The efficacy of soy meal as a partial fishmeal replacement has been well documented in fish (Reinitz 1980. Jackson et al. 1982, Watanabe et al. 1992). and although it has been incorporated into a number of formulations for Haliotids (Fleming et al. 1996). this is the first detailed information suggesting that it is a viable fish- meal replacement for use in diets for H. midae. Although protein efficiency was reduced at both the 25 and .50%' substitution levels, growth rates were not significantly affected. Increasing dietary EAA deficiencies or lower bioavailability of nutrients may explain this reduction in protein efficiency. Although a number of studies have demonstrated that amino acid supplementation can improve nutritional indices in soy-based diets for salmonids (Dabrowska & Wojno 1977, Rumsey & Ketola 1975); methionine has been found to be deficient in soy-based formulations for Japanese eel (Lovell 1989) and carp (Murai et al. 1989). Therefore, it is probable that at the higher inclusion level, the soymeal-based diet may have been deficient in EAAs, and in particular methionine. However, it is not possible to verify whether this was the case until both the essential amino acid requirements of the species, and the availability of the amino acids in the soy and fishmeal have been established. The growth and nutritional indices using the sunflower meal as a fishmeal replacement demonstrate that it has good potential for future use in formulations. Although feed conversion increased and protein efficiency reduced with respect to the control, the higher consumption rates may have improved growth rates, which, at the 30% fishmeal substitution level was not significantly differ- ent from the control. However, the reduction in growth and protein efficiency that became apparent when the inclusion level of the sunflower was increased from 30 to 50% may have been attribut- able to either a reduction in digestible protein or energy, or im- balances in essential amino acids. Although there is no information concerning the effects of sunflower meal in diets for Haliotids. there have been a number of studies with fish. Using rainbow trout. Sanz el al. ( 1994) cited imbalances between leucine and isoleucine as the primary factor reducing protein efficiency both in their study and an additional study carried out previously by Tacon et al. (1984). It is well documented that a leucine/isoleucine imbalance has a negative impact on growth and nutritional indices in fish (Harper et al. 1970, Hughes & Rumsey 1983, Robinson et al. 1984). However, until more detailed information regarding the essential amino acid requirements of abalone is available, it is not possible to determine whether this was the case in this study. The combination of soy and sunflower meals as a partial fish- meal replacement significantly affected both feed conversion ratio and protein efficiency. These indices were respectively higher and lower than the control indicating poorer dietary utilization of the feed, which, with the exception of the 50% com gluten substitu- tion, produced the lowest growth rates in the study. Notwithstand- ing gross differences in digestible energy between the diets, these reductions in nutritional indices were probably caused by a com- bination of factors previously discussed when either the soy or sunflower meals were tested individually. It is. however, interest- ing to note that when the fishmeal component was completely removed and replaced with spirulina. there was an improvement in the growth rates. Not only does this suggest that spirulina is a viable fishmeal replacement — a result confirmed by the growth and nutritional indices obtained using the 50:50 fishmeal/spirulina 644 Shipton and Britz formulation — it also suggests that, as more inforination concerning the dietary requirements of the abalone and the availability of nutrients from plant protein sources becomes available, the com- plete removal of the fishmeal component may become a viable proposition. The substitution of fishmeal with torula yeast produced some of the best growth rates and nutritional indices in the study. This was most apparent at the 30"^ fishmeal substitution level where the diet out-performed all the alternative formulations. The high apparent protein digestibility coefficient of the torula yeast (83.4% pers. obs.), combined with the high protein and feed efficiency suggests not only high availability of the protein source to the animals, but that it contains a well-balanced essential amino acid profile that supports good growth in this context. Nevertheless, at the SO'/c substitution level, the growth and nutritional indices were reduced with respect to the 30% substitution level. As consumption rates between the two formulations were not significantly different, this reduction in efficiency was probably caused by an imbalance of EAAs at the higher torula yeast inclusion level. The results using the com gluten meal suggest that it has po- tential as a fishmeal replacement at levels below 30'/( . Comparable studies with fish have indicated that inclusion levels between 20- 30% dietary protein are viable, and do not have a negative impact on growth and feed efficiency (Gropp et al. 1979, Alexis et al. 1985, Robaina et al. 1997, Regost et al. 1999). However, in com- parison with other protein sources, reduced feed efficiency and growth at higher inclusion levels (>30%) have been related to its poor essential amino acid profile, and, in particular, low lysine levels (Robaina et al. 1997). This may account for the results in this study in which both feed conversion and protein efficiency were reduced at the 50% inclusion level. .Mthough Pearson product moment correlations between the dietary and abalone EAA profiles revealed a large variation in correlation coefficients (r = 0.70-0.97, Table 4), Pearson product moment correlations between the r-values and the respective growth rates or protein efficiency ratios revealed no significant correlations (Table 5). Therefore, it appears that the balance of dietary EAAs was not a good predictor of either growth or protein utilization in the abalone. In contrast, when similar correlations were performed between growth, protein efficiency, and the levels of individual dietary EAAs (Table 5); the levels of dietary lysine were significantly and positively correlated to both growth rates and protein efficiency. Thus, although the bioavailability of the EAAs was not taken into account in this study, it can be concluded that, as opposed to the dietary EAA profile, dietary lysine levels provide a better indication of which protein sources will promote growth and protein efficiency. Furthermore, the positive correla- tions between dietary lysine levels and growth and protein effi- ciency suggest that lysine may have been the first limiting amino acid in these diets. In conclusion, the results suggest that the potential to replace fishmeal with plant protein sources in commercial diets for H. midac is promising. However, although the commercial applica- tion of the torula yeast- and spirulina-based diets may be restricted by cuiTent high raw material costs from South African suppliers (respectively 340% and 250% higher than LT-fishmeal), the low raw material costs associated with the soy and sunflower meals make these protein sources particularly attractive to commercial feed formulators (respectively 60 and 75% lower than LT- fishmeal). Nevertheless, this study should be viewed as a prelimi- nary investigation into the partial and total replacement of fish- meal; and that, before detemiining the optimal protein mixtures required to optimize growth at the least cost, a considerable amount of further research is required. In particular, the digestibil- ity of essential amino acids and energy from the protein sources must be determined, the EAA requirements of the abalone. and the optimal dietary digestible protein: energy ratio. Only then can formulations be designed, which will optimize growth and feed utilization through the delivery of the required essential amino acids and appropriate digestible energy. ACKNOWLEDGMENTS We appreciate the financial support from Sea Plant Products Ltd., Rhodes University Joint Research Committee, THRIP and The National Research Foundation. LITERATURE CITED Alexis, M. N., E. Paparaskeva & V. Theochari. 1985. Formulation of practical diets for rainbow trout (Salmo gairdneri) made by partial or complete substitution of fishmeal by poultry by-products and certain plant by-products. Aquaculrure 50:61-73. AOAC (Association of Official Analytical Chemists). 1984. OtTicial meth- ods of the association of Official Analytical Chemists. 14th ed. Wash- ington: AOAC. pp. 152-160. Barlow. S. 1989. Fishmeal - World outlook to the year 2000. Fish Farmer. October:40-4.^. Britz. P. J. 1995. The nutritional requirements of Haliotis midue and de- velopment of a practical diet tor abalone aquaculture. PhD Thesis, Rhodes University, South .Africa. Britz, P. J. 1996. Suitability of selected protein sources for inclusion in formulated diets for the South African abalone. Haliotis inidae. Aqua- culture 140:63-7.^ Dabrowska. H. & T. Wojno. 1977. Studies on the utilization by rainbow trout (Salmo gairdneri) of feed mixture containing soybean meal and an addition of amino acids. Aquaculture 10:297-310. Fleming, A. E.. R. J. Van Baneveid & P. W. Hone. 1996. The development of artificial diets for ahalonc: a review and future directions. Aquacul- ture 140:5-53. Folch, J., M. Lees & S. Stanley. 1957. A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226:497-509. Gropp, J., H. Koops, K. Tiews & H. Beck. 1979. Replacement of fishmeal in trout feeds by other feedstulfs. In: Pillay, T. V. R. & W. A. Dill, editors. Advances in aquaculture. Papers presented at the FAO techni- cal conference on aquaculture, Kyoto, Japan. 21 May - 2 June. 1976. Farnham. UK: Fishing New Books. Harada. K.. T. Mikasaki, S. Kawashima & H. Shiota. 1996. Studies on the feeding attractants for fishes and shellfishes. XXVI. Probable feeding attractants in allspice PiuieiUu offiiialis for the black abalone Haliotis discus. .Aquaculture 140:99-108. Harper. A. E.. N. J. Benevenga & R. M. Wohlhueter. 1970. Effects of ingestion of disproportionate amounts of amino acids. Physiol. Rev. 50:428-558. Huges. S. G. & G. L. Rumsey. 1983. Dietary requirements of es.sential branch-chain amino acids hy lake trout. Trans. .Am. Fish. Soc. 112: 812-817. Jackson, A. J.. B. S. Capper & A. J. Matty. 1982. Evaluation of some plant proteins in complete diets for the tilapia Sarotherodon mossambicus. Aquacidture n:91-\m. FisHMEAL Replacement in Abalone Diets 645 Knauer. J. 1994. Development of an artificial weannig diet for the South African abalone. Huliotis /Hd/tic (Haliotidae: Gastropoda). MSc Thesis. Rhodes University. Grahamstown. South .Africa. Lee. D. J. & G. B. Putman. 1973. The response of rainbow trout to varying protein/energy ratios in a test diet. / Nutr. 103:916-922. Lim. C. 1989a. Practical Feeding - Tilapia. In; Lovell. T.. editor. Nutrition and feeding of fish. New York: Van Nostrand Reinhold. p. 180. Lim. C. 1989b. Practical Feeding - Penaeid shrimps. In: Lovell. T., editor. Nutrition and feeding of fish. New York: Van Nostrand Reinhold. p. 218. Lovell. T. 1989. Feed formulation and processing. In: Lo\ell, T.. editor. Nutrition and feeding of Fish. New York: Van Nostrand Remhold. pp. 107-127. Moyano. F. J.. G. Cardenete & M. De la Higuera. 1992. Nutritive value of diets containing a high percentage of vegetable proteins for trout. On- corhynchus mykiss. Aqiiat. Living Res. 5:23-29. Murai. T., W. Daozun & H. Ogata. 1989. Supplementation of methionine to soy flour for fingerling carp, iCypriinis carpio). Aquaculmre 11: 373-385. NRC (National Research Council). 1993. Nutrient requirements of fish. Washington. DC: National Academy Press, pp. 65-66. Pongmaneerat. J. & T. Watanabe. 1992. Utilization of soybean meal as protein source in diets for rainbow trout. Nippon Suisan Gnkkaishi 58:1983-1990. Regost, C, J. Arzel & S. J. Kaushik. 1999. Partial or total replacement of fishmeal by com gluten meal in diets for turbot {Psenu maxima). Aqua- culture 180:99-117. Reinilz. G. 1980. Soybean meal as a substitute for herring meal in practical diets for rainbow trout. Prog. Fisli Cult. 42:103-106. Robaina, L.. F. J. Moyano. M. S. Izquierdo. J. Socorro, J. M. Vergara & D. Montero. 1997. Com gluten and meat bone meal as protein sources in diets for gilthead seabream {Spanrs aurata): Nutritional and histologi- cal implications. Ac/KiJCH/rK/f 157:347-359. Robinson, E. H., W. E. Poe & R. P. Wilson. 1984. Effect of feeding diets containing an imbalance of branch-chain amino acids on fingerling channel cattish. Aquaculture 37:51-62. Rumsey. G. L. 1993. Fish meal and alternative sources of protein in fish feeds, update 1993. Fisheries 18:14-19. Rumsey. G. L. & H. G. Ketola. 1975. .'Xmino acid supplementation of casein in diets of Atlantic salmon (Salmo salar) fry and of soybean meal for rainbow trout (Salmo gairdneri) fingerlings. / Fish. Res. Board Can. 32:422^26. Sanz. A.. A. E. Morales. M. de la Higuera & G. Cardenete. 1994. Sun- flower meal compared with soybean meal as partial substitutes for fish meal in rainbow trout {Omorhynchus mykiss) diets: protein and energy utilization. Aquaculture I28:287-.3O0. Smith. R. R. 1989. Nutritional energetics. In: J.E. Halver, editor. Fish nutrition, 2nd ed. San Diego, CA: Academic Press, pp. 1-29. Stickney. R. R., R. W. Hardy. K. Koch & R. Harrold. 1996. The effects of substituting selected oilseed protein concentrates for fish meal in rain- bow trout Oncorhynchus mykiss diets. J. World Aqua. Soc. 27:57-63. Tacon, A. G. J., J. L. Webster & C. A. Maninez. 1984. Use of solvent extracted sunflower seed meal in complete diets for fingerling rainbow trout (Salmo gairdneri Richardson). Aquaculture 43:381-389. Uki. N.. A. Keniuyama & T. Watanabe. 1985a. Nutritional evaluation of several protein sources in diets far abalone Haliolis di.uus hannai. Bull. Jpn. Soc. Sci. Fish. 51:1835-1839. Uki. N.. A. Kemuyama & T. Watanabe. 1985b. Development of semipu- rified test diets for abalone. Bull. Japan. Soc. Sci. Fish. 51:1825-1833. Uki. N. & T. Watanabe. 1986. Effect of heat-treatment of dietary protein sources on their quality for abalone. Bull. Jpn. Soc. Sci. Fish. 52:1 199- 1204. Uki. N. & T. Watanabe. 1992. Review of nutritional requirements of aba- lone ( Haliotis sp. ) and development of more efficient artificial diets. In: S. A. Shepherd. M. J. Tenger. S. A. Gu/.man Del Proo. editors. Abalone of the world, fisheries, biology and culture. O.xford, UK: Fishing News Books, pp. 504-5 1 7. Viana, M. T.. L. M. Lopez & A. Salas. 1993. Diet development for juvenile abalone Haliotis fulgens. Evaluation of two artificial diets and mac- roalgae. Aquaculture 117:149-156. Viana. M. T., M. Cervantes-Trujano & R. Solana-Sansores. 1994. Attrac- tion and palatability activities in juvenile abalone, Haliotis fulgens. Aquaculture 127:19-28. Watanabe, T.. V. Viyakam, H. Kimura. K. Ogawa. N. Okamoto & N. Iso. 1992. Utilization of soya bean meal as a protein source in a newly developed soft-dry pellet for yellowtail. Nippon Suisan Gakkaishi 58: 1761-1773. Weisberg. S. B. & V. A. Lotrich. 1982. Ingestion, egestion, excretion, growth and conversion efficiency for the mummichog, Fundulus het- eroclitus. J. Exp. Mar Biol. Ecol 62:237-250. Zar, J. H. 1984. Biostatistical analysis. 2nd ed. Englewood Cliffs. NJ: Prentice-Hall, Inc. 7 1 8 pp. Jcmnuil oj Shellfish Ri-seanh. Viil. 20, No. 2. 647-631. 2001. THE USE OF STIMULANTS AS AN AID TO WEAN FISHERY CAUGHT BLACKFOOT ABALONE {HALIOTIS IRIS) TO ARTIFICIAL FOOD VICTORIA J. ALLEN,* ISLAY D. MARSDEN, AND NORMAN L. C. RAGG Departiiu'iit of Zoology. University of Caiiterbiiiy. Private Bag 4H0U. Christchurch. New ZeaUiiid ABSTRACT Abalone use a conihination of tactile and chemosensory feeding cues to detect suspended seaweed ni their natural environment. However, in a commercial situation, adult abalone (Haliotis iris) caught for broodstock or pearling, show reluctance to start feeding on stationary artificial food. If they cannot be induced to feed, they may lapse into a starvation phase that can last several weeks. Suhtidal adult H. iris (125 mm shell length) were collected during autumn, winter, spring, and summer from Banks Peninsula, Canterbury and held at Pendarves Abalone Farm Ltd. (South Island, New Zealand). Abalone were divided randomly into two groups. One group was offered a coinmercial pellet diet together with small quantities (0.03-0.05g dry weight L"') of suspended seaweed particles {Gracilaria spp.) to test its potential as a tactile phagostimulant. The other group (nonstimulant control) were supplied with the same commercial diet without the algal stimulant Each group consisted of three replicate vessels containing three individuals. Abalone behavioral responses and ingestion rates were monitored in both groups over a 3-wk period. Abalone in the stimulant experiment exhibited a typical feeding posture almost immediately, were more alert, and engaged in a greater range of feeding activities than the abalone that had been supplied with commercial food only. Those abalone without access to the algal stimulant remained either quiescent or alert in most seasons. Feeding activity of abalone in the stimulant treatment was significantly higher (P < 0.05) as compared with abalone from the nonstimulant control for all seasons except winter. Also, abalone consumed significantly more artificial food when tactile stimulants were present during autumn and winter. However, ingestion rates seemed highly sensitive to both season and acclimation time. Within each season, there was a consistent trend of increasing ingestion rate dunng the third week of the experiment, and this was most discernible in the stimulant treatment. KEY WORDS: Huliinis iris, stimulant, artificial food, behavior. New Zealand abalone INTRODUCTION The commercially \ aluable New Zealand abalone, Haliolis iris Matlyn. locally known as paua, is the main species of abalone harvested (Hooker & Creese 1995) and more recently, cultured in New Zealand (Clarke & Creese 1998). In addition to animals caught by the fishery for immediate processing, adult H. iris (S125 mm shell length) may be taken to land-based holding fa- cilities for later sale or, more critically, use in aquaculture as broodstock or for pearling. Newly fished abalone are inevitably stressed by handling and air transport, and a holding system must, therefore, permit initial recovery and subsequently promote the grow th and conditioning of gonads or allow abalone to become suitably physiologically robust to withstand the implanting of pearl nuclei. It is a major challenge to provide a suitable food source that will be accepted by the abalone throughout this period. Abalone are macroalgivores (Wee et al. 1992), using a combination of tactile and chemosensoi"y stimuli to detect suspended and attached seaweed in their natural environment (Fleming 1995). Adult H. iris are largely sedentary, favoring exposed rocky substrata and feeding on diift seaweed detected by the cephalic and epipodal tentacles and subsequently trapped by the anterior foot or shell margin (Poore 1972a). Few alga! harvesting permits have so far been granted for the commer- cial collection of seaweed in New Zealand (Schiel 1997), forcing fanners to rely upon artificial pellet feeds (certain artificial diets are also favored by pcari farmers fiir the resulting desirable nacre properties). Although nutritionally robtist and relatively stable in seawater (Fleming 1995). the dense artificial pellets are unable to simulate the dynamic motion of neutrally buoyant seaweed frag- ments (Fleming et al. 1996). In the absence of the tactile stimulus of moving algae, captive H. iris will usually lapse into a starvation phase, taking up to 6 wk to commence feeding on pellets (pers. *Corresponding author. E-mail: vjaqua_n2@yahoo.co.nz obs.); during this time, abalone inevitably suffer a loss of condition and may die. Qualitative observations have suggested that the addition of small quantities of seaweed particles to a tank with suitably dy- namic water movement will illicit a feeding response from adult H. iris, stimulating the abalone to graze upon static feed pellets that would otherwise be ignored on the tank bottom. This experiment is designed to characterize and quantify the nature of the abalone's response to the presence of a suspended tactile stimulant (dried Gracilaria spp. fragments). Seasonal variability in the feeding be- havior of abalone is well documented (Poore I972a.b. Shepherd 1973. Marsden & Williams 1996. Donovan & Carefoot 1998). hence the trials were repeated at 3-mo intervals over 1 yr and season treated as a fixed variable in subsequent analyses. MATERIALS AND METHODS Collection and Holding of Abalone Aduit Haliolis iris =:I25 mm shell length were collected using SCUBA from a depth of 6-10 m from a rocky reef within Whaka- moa Bay. Akaroa Heads. New Zealand (43'53'S. 172 53'E). Col- lections were made once each season beginning in autumn ( 1 1 May 1999) and including winter (30 August), spring (4 Novem- ber), and summer (22 December). The abalone were transported overland, in damp air at 5'C. to the Pendarves Abalone Fann Ltd.. Canterbury. New Zealand. Abalone were tagged and individual wet weight and shell length were recorded. Three animals were randomly assigned to each experimental vessel, an opaque 40-L tank receiving five ex- changes h~' of unfiltered seawater at ambient temperature (10- I6°C). pH 8.2 ± 0.2. salinity 32 ppt. Experimental Design Two dietary treatments were examined. The first simply offered the abalone static pellets of artificial feed, while the second in- 647 648 Allen et al. eluded Cracilaiia spp. fragments suspended in the water column as a tactile stimulant. There were three replicate vessels for the stimulant treatment and the control group. Behavioral responses of abalone were quantified in both groups for an initial 10-h period, and ingestion rates were measured at 3-day intervals for 3 wk. Tanks were covered with black polyethylene sheeting to pro- vide total darkness for the initial 96 h after stocking, providing the abalone with an undisturbed 4-day recovery period from handling and transportation. On day 5. the covers were removed and the tanks siphoned clean. To three tanks were added 3 g of pre- weighed. dry artificial food (Abfeed®, Sea Plant Products, South Africa), these are referred to as the nonstimulant treatment or control. A further three tanks were supplied with 3 g Abfeed. plus 1.5 g of dried, mulched Ciacilaria spp. particles (stimulant treat- ment) equivalent to 0.03-0.05 g L^' . The lank inflow jet was positioned to suspend the GracUaria spp. particles in the water column and leave the Abfeed pellets static on the tank floor. Behavioral Responses Abalone behavior was observed and recorded over a 10-h pe- riod beginning at approximately 1 p.m and finishing at II pm. A photoperiod of 12 h light/12 h dark was maintained with lights being turned off at 6 pm. Abalone were observed at night under red light. Abalone behavior was divided into the following categories: 1. quiescent (shell clamped, cephalic, and epipodal tentacles retracted); 2. alert (shell slightly raised, epipodal, and cephalic tentacles extended); 3. locomotion (crawling); 4a. feeding on artificial pellets; 4b. feeding on GntciUiiia spp. particles; and 5. active feeding posture (shell and foot raised, anterior lobes of foot and tentacles extended and waving). Abalone were examined every 15 min and their behavior re- corded. At the end of the 10 h period, nylon lids were placed on the tanks to minimize light-shadow disturbance. For the purpose of statistical analysis, behavioral responses were combined into feed- ing (categories 4a. 4b. and 5j or nonfeeding behavior (1-3). Ingestion Rates Tanks were cleaned and food and stimulant replaced every 3 days following the completion of the behavior assessment. Clean- ing consisted of the inflow being turned off; all solid material was then siphoned from the bottom of the tank and collected on GF/C filter paper. The solids were rinsed in fresh water, and all identi- fiable food particles were isolated from the GracUaria spp. and feces were then dried at 55°C for 24 h and weighed. Artificial feed was also stocked to a control tank containing no animals, the proportion of initial weight lost was used as a correction factor to compensate for passive leaching; in this way a correction factor was determined for each season using weight change over a 3-day period. Biomass-specific ingestion rates were calculated for each group of abalone during each 3-day feeding period using the fol- lowing equation: I [(W, XC..„„„,,,)-W„]^3 B 100 Where ingestion rate, I (percentage biomass ingested per day), is determined from W, and Wp, the initial and final dry weight of artificial food over a 3-day feeding period conected for C, the season-specific leaching correction, and standardized by biomass of abalone in the tank B. Abalone ingestion rates were calculated over 3 wk for each seasonal sample, but this was extended to 4 wk in winter because of low rates of food intake. Statistical tests were performed using Statistica"™' 5.1 (StatSoft Inc.. USA). RESULTS Behavioral Responses During all four seasons, the abalone tended to be largely inac- tive (categories I and 2 on Fig. I ), spending less than I09f of the time moving about the tank (category 3). The presence of the algal stimulant had little effect upon the amount of time engaged in each o 0) Autumn 100 80- 60 i 40 - 1 20 . 0 - I. ■ Non-Stimulant D Stimulant j5^ Winter 40 20 0 -- h.il ^ Spring Summer 60 i &} quiescent 3 crawling 4a 4b 5 feeding feeding on feeding on posture pellets stimulant Behaviour category Figure 1. Seasonal time budgels for Haliolis iris. beha\ ioral categories recorded over a 10-h period in the presence (Stimulant) or absence of algal stimulant (Nonstimulant). ;i = 9 individuals per treatment. CIraphs show mean ± .SK. Stimulants to Wean Abalone onto Artificial Food 649 nonfeeding activity cliirins; autumn and winter (Fig. 1). However, abalone receiving the stimulant treatment spent relatively little time clamped and unrecepti\e during spring and summer (20.1 ± 1.09c and 3.3 ± 1.39f. mean ± SE. respectively) as compared to their nonstimulant counterparts (62.6 ± 7.2% and 38.5 ± 6.6%). A two-way analysis of variance (ANOVA) was used to exam- ine the effects of stimulant presence and season upon the amount of time devoted to feeding activities (categories 4a. 4b and 5). Both the stimulant (F = 54.24. P < 0.001) and season (F = 21.99, P < ().()() 1 ) had a significant effect on the duration of the feeding response, with no interaction between the two factors (F = 1.13. P = 0.342). Tukey's "Honest Significant Difference"" pairwise comparison found the feeding activity of animals in the stimulant treatment was significantly higher (P < 0.05 ) than nonstimulant for all seasons except winter. Within each treatment, summer feeding activity was significantly higher than all other seasons. Nonstimu- lant abalone that exhibited \ irtually no feeding behavior between autumn and spring spent 22.0 ± 5.0% (mean ± SE) t)f their time in a feeding posture in summer. In contrast, abalone from the stimu- lant treatment spent 57.0 ± 6.0% of their time in this posture, as compared to 1 7.(1 ± 3.5% throughout the rest of the year. With the exception of isolated instances, where stimulant particles were successfully captured by the abalone. all feeding activity during the initial 10-h observation period was limited to a receptive pos- ture rather than actual intiestion. Autumn -Non-stimulant Stimulant E o 0.14 r 0-12 i °M 0 08 [ 0 06 • 0 04 ; 0 02 + 0 ^ 17-May 20-May 23-May 26-May 29-May Winter 2-Sep 5-Sep 8-Sep n-Sep 14-Sep 17-Sep 20-Sep 23-Sep 26-Sep 0.14 0.12 -- 0.1 0.08 0.06 0.04 0.02 - 0 Spring Ingestion Responses Abalone consumed significantly more artificial food when tac- tile algal stimulants were present during autumn and winter (ANCOVA, F = 4.763, P = 0.035. and F = 7.650. P = 0.008, respectively. It must be noted, however, that winter trials were extended for a fourth week because of very low ingestion rates. No significant differences were detected between stimulant and non- stimulant treatments in spring (F = 2.397, P = 0.130) and sum- mer (F = 2.159, Z' = 0.150). Ingestion rates seemed highly sen- sitive to both season and acclimation time (Fig. 2). Despite en-atic abalone consumption rates during autumn and very low initial rates in winter and spring, a consistent trend of increasing inges- tion rate occurred during the final week of all of the seasonal experiments. This was most discernible in the stimulant treatment. A similar potential trend in summer may ha\'e been obscured be- cause of a pump failure during the final 3 days of monitoring. Because the ingestion response attributable to the presence of stimulants seemed to become more pronounced over time, the final 3-day ingestion data from each season were considered to be the most representative predictor of long-term feeding responses. Sta- tistical analysis (two-way ANOVA) found that both treatment (F = 3.4, P < 0.001) and season (F = 14.6, P = 0.002) had a significant effect upon tuial ingestion rates; there was no signifi- cant interaction. A least-significant difference post hoc comparison of means (95% CI) found tmal ingestion rates were significantly higher in the presence of stimulants in autumn (0.13 ± 0.01% biomass d"' in stimulant, 0.07 ± 0.007% biomass d~' in nonstimu- lant). but no significant difference for any other season. An e,x- amination of seasonal effects alone, based upon final ingestion rate, ranked autumn significantly higher than summer, which was, in turn, significantly higher than winter and spring, which were similar. Summer 0 14 - 0.12 0.1 0.08 - 0.06 0.04 1 002 - 0 - Date Figure 2. Seasonal ingestion rate of Haliotis iris provided «ith an artitlcial diet in the pre.sence (Stimulant! or absence of algal stimulant (Nonstimulant). (;raph.s show mean + or -SK. DISCUSSION To understand the responses of wild abalone following capture and transfer to an artificial situation, such as a commercial facility, it is important to understand their feeding mechanisms and behav- ior. Initially thought to be browsing herbivores, haliotids tend to be largely sedentary except when food is scarce (Poore 1972a,b). Adult H. iris consume a wide range of seaweeds (Tunbridge 1967), feeding predominantly on drift algae in most localities (Schiel 1992). Haliotis ruber Leach and H. laevigata Donovan, when exposed to sufficient water fiow. raise the front of their shells to catch drift seaweed with their foot instead of grazing on benthic algae (Shepherd 1973). It was noted that this posture also occurred in the absence of drift algae. Hingham et al. (1998) also described the behavior whereby abalone were observed to form two ""hands'" with their foot and grasp food as it moved past, detected by contact with the epipodal tentacles. Poore (1972a) noted that H. iris re- sponded to drift seaweed in the same way. 630 Allen et al. Abalone use both chemosensory and tactile cues to detect food in the wild. Crofts (1929) observed that the cephalic, mantle, and epipodal tentacles appreciate even slight chemical changes in the water and that in all the tentacles, the sense of touch is particularly acute. The cephalic and epipodal tentacles play an important role in investigating the environment and guiding the abalone around their habitat (Crofts 1929, Upatham et al. 1998) and also respond to such tactile stimuli as drifting algal fronds and particles. The sensitivity to such stimuli is attributable to the cluster of sensory cells that reside beneath the epithelium on the tentacles, lips, an- terior edge of the foot, and mantle edge (Luchtel et al. 1997). According to de Vlieger (1968). the optimal stimulus for positive thigmotaxis (during creeping as well as when exploratory move- ments are carried out) is a nonmoving object occasionally met by the foot (e.g., food pellet). This results in the withdrawal of the tentacle or foot followed by a full tentative exploration of the initial stimulus (de Vlieger 1968). Abalone in both the control and algal stimulant treatments were observed to touch and feel around the edge and over the whole pellet with their foot after they had physically encountered it (they usually bumped into it, then stopped to explore). Some abalone crawled over the pellets after exploring them, but no attempt at ingestion was made. Hence, contact with the food pellets seemed to elicit a positive thigmot- actic response but did not stimulate a feeding response in freshly caught H. iris. Abalone were regularly observed in a typical feeding posture in the tanks where stimulants were provided: however, this same feeding posture was also observed in the nonstimulant tanks, es- pecially in summer. These abalone were found directly under the water in-tlow with their shells raised and feet extended, corrobo- rating the view taken by Shepherd (1973), who suggested that water movement is an important environmental factor affecting the feeding of those species that feed on algal drift. The increased tendency to adopt a feeding posture accounted almost entirely for the significantly elevated level of feeding behavior of H. iris in the stimulant treatment. Abalone consistently spent less time in the unreceptive quiescent/clamped state when stimulants were present (Fig. I ). Because clamping is usually regarded as an escape re- sponse in haliotids. this behavior may be considered a reaction to an unfavorable en\ironment. In the presence of stimulants, the amount of time spent quiescent seemed clo.sely related to water temperature. The same observation was found for H. kamtschat- kana Jonas (Donovan & Carefoot 1998) and may reflect metabolic limitation rather than an avoidance response. It is interesting to note that captive adult H. iris do not share the wild population's tendency to increase foraging (crawling) rate in response to hunger (Poore 1972a, Donovan & Carefoot 1998), nor to the presence of stimulants, locomotion tending to be limited to negative phototac- tic response only (pers. obs). The ingestion rates of H. iris feeding on a commercial pellet diet support the initial hypothesis that, with the addition of a tactile algal stimulant, abalone would be encouraged to feed on the arti- ficial food and increase the ingestion rate. In all the trials; how- ever, ingestion rates varied over time. Feeding rates were low or erratic for the first 2 wk of captivity, then increased steadily toward the end of each trial (Fig. 2,). This pattern occurred sooner and was better defined in the stimulant treatment than in the nonstimulant control. Ingestion rates increased up ti) 5i.n in autumn and winter and slightly less in spring and summer. The improvement in di- etary intake of abalone in the presence of the stimulant usually began after 9-12 days and was consistently later in nonstimulant abalone, typically taking 13-18 days (Fig. 2.). Extended trials using a larger sample of individuals to establish the steady-state ingestion rates of abalone when fully acclimated to captivity and artificial food would be valuable in determining the long-term role of seaweed particles as a phagostimulant. Differences in abalone ingestion rate between the four seasons can be attributed in part to elevated water temperature, which can increase metabolic rate and subsequently feeding rate (Marsden & Williams 1996). Hence, although it might be predicted that feeding rates should be higher during summer than in the autumn, food consumption may be affected by other metabolic demands, includ- ing the abalone reproductive cycle. Hutiotis iris from the collection site are known to spawn during the autumn (March to May) (Poore 1972b); hence, increased food intake may be neccessary to support rapid gonad development (gonad condition was not examined to minimize handling stress). Some apparent inconsistency in the ingestion data may also have been caused by the collection times within each season. Autumn, winter, and spring collections were made relatively late because of the uncompromising dive condi- tions, and an early collection was made in summer to exploit favorable sea conditions. Applied nutritional research on abalone has focused upon the use of attractants and chemical feeding stimulants. Harada et al. (1996) tested various novel ingredients, including herbs and spices for their attracting ability and Sakata and Ina ( 1992) used different algal methanol extracts to study the feeding behavior of young H. discus. Also, attempts have been made to design a neutrally buoy- ant feed that simulates algae in the culture tank (Fleming et al.l996). and more recently, the role of water movement in stimu- lating abalone feeding has become recognized (Hingham 1998). The results from the present study suggest that tactile algal par- ticles could provide an additional stimulant, complementing those already described. In conclusion, the addition of Gracilaria spp. particles to the water of a tank containing newly collected field specimens of Haliotis iris immediately stimulates a positive feeding response. This subsequently shortens the period to first acceptance of arti- ficial feed and seems to stimulate elevated ingestion rates. The nature of these responses; however, is highly seasonal and may be subject to a period of acclimation during which the feeding re- sponse can change. It is suggested that this aspect be examined further using a larger sample size and individuals of a range of shell lengths. Also, investigations into the effects of tactile pha- gostimulants on ingestion and subsequent meat, gonad, and nacre development may assist the future development of the abalone culture industry. ACKNOWLEDGMENTS We thank Pendarves Abalone Farm Ltd. for research space and facilities and the staff, including N. Williams. B. Williams, and C. Stanger for technical support and constructive advice. D. Brad- shaw, D. Tattle, and R. Bishop assisted with the collection of wild abalone. We also thank J. and J. Bradshaw for transport and access to Whakamoa Bay. and D. Bradshaw for discussion and comments on the manuscript. This research was undertaken as part of a Graduate Research in Industry Fellowship (awarded to V. Allen) funded by the Foundation for Research, Science, & Technology in New Zealand. Stimulants to Wean Abalone onto Artificial Food 651 LITERATURE CITED Clarke, C. B. & R. G. Creese. 19^8. On-grciwing cultured abalone (Haliolis Poore. G. C. B iria) in northern New Zealand. / Shellfish Res. 17:607-613. CrofLs. D. R. 1929. Ha/ioris. LMBC Memoirs on Typical British Marine Plants and Animals No. 29. 166 pp. de Vlieger. T. A. 1968. An experimental study of the tactile system of Lymnaea slagnalis. Neth. J. Zool. 18:105-154. Donovan. D. A. & T. H. Carefoot. 1998. Effect of activity on energy allocation in the northern abalone, Hciliaris kamtschalkaiKi (Jonas). J. Shellfish Res. 17:729-736. Fleming, A. E. 1995. Growth, intake, feed conversion efficiency, and chemosensory preference of the Australian abalone. Hiiliinis nihni. Aquaeidliire 132:297-311. Fleming. A. E.. R. J. Van Barneveld & P. W. Hone. 1996. The develop- ment of artificial diets for abalone: A review and future directions. Aqiutcultiire 140:5-53. Harada, K.. T. Miyasaki, S. Kawashima & H. Shiota. 1996. Studies on the feeding attractanis for fishes and shellfishes. XXVL Probable feeding attractants in allspice Pimeiua ojfieimilis for black abalone Haliaiis diseus. Aiiiuieiilliire 140:99-108. Hingham, J., P. Hone, S. Clarke, R. Baudinette & M. Geddes. 1 998. The effect of Oow on growth in juvenile Greenlip abalone, Haliolis laevi- gata (Donovan). Abalone Aquaculture Sub-program. Fisheries Re- search and Development Corporation. Proceedings of the 5 Annual Abalone Aquaculture Workshop, July 1998, Hobart, Tasmania, Aus- tralia, pp. 115-122. Hooker, S. H. & R. G. Creese. 1995. Reproduction of paua. Haliolis iiis Gmelin 1971 (MoUusca: Gastropoda), in northeastern New Zealand. Mar. Freslnvaler Res. 46:617-622. Luchtel, D. L., A. W. Martin, I. Deyrup-Olsen & H. H. Boer. 1997. Gas- tropoda: Pulmonata. In: F. W. Harrison & A. J. John, editors. Micro- scopic anatomy of invertebrates, vol. 6B: Mollusca II. New York: Wiley-Liss. pp. 459-718. Marsden, I. D. & P. M. J. Williams. 1996. Factors affecting the grazing rate of the New Zealand abalone Haliolis iris Martyn. J. SheUft.th Res. 15:401-406. 1972a. Ecology of the New Zealand abaloncs, Haliolis species (Mollusca: Gastropoda). I. Feeding. N.Z. ./. Mar. Freslnvaler. Res. 6:11-22. Poore. G. C. B. 1972b. Ecology of the New Zealand abalones, Haliolis species (Mollusca: Gastropoda). 2. Seasonal and diurnal niovement. N.Z. J. Mar. Freslmaler Res. 6:246-258. Sakata, K. & K. Ina. 1992. Algal feeding stimulants for abalone. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: Biology, fisheries, and culture. Oxford. UK: Blackwell Sci- entific, pp. 182-192. Schiel, D. R. T. 1992. The paua (abalone) fishery of New Zealand. In: S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: Biology, fisheries, and culture. Oxford. UK; Blackwell Scientific, pp. 427-437. Schiel. D. R. T. 1997. Review of abalone culture and research in New Zealand. Moll. Res. 18:289-298. Shepherd, S. A. 1973. Studies on Southern Australian abalone (Genus Haliolis). 1. Ecology of five sympatric species. Aiisi. J. Mar. Fresh- water Res. 24:217-257. Tunbridge. B. R. 1967. Feeding habits of paua {Haliolis iris Martyn). Fisheries Technical Report. N.Z. Marine Dept.. 20 pp. Upatham. E. S„ A. Thongkukiatkul. M. Kruatrachue. C. Wanichanon. Y. P. Chitramvong, S. Sahavacharin & P. Sobhon. 1998. Classification of neurosecretory cells, neurons, and neuroglia in the cerebral ganglia of Haliolis asinina Linnaeus by light microscopy. J. Shellfish Res. 17: 737-742. Wee, K. L., G. B. Maguire & S. M. Hindrum. 1992. Methodology for digestibility studies with abalone. I. Preliminary studies on the feeding and defecatory behavior of blacklip abalone, Haliolis rubra, fed natural and artificial diets. In: G. L. Allan & W. Dall. editors. Proc. Aquacul- ture Nutrition Workshop. Salamander Bay. 15-17. April 1991. NSW Fisheries. Brackish Water Fish Culture Research Station. Salamander Bay. Australia, pp. 192-196. Journal oj Shellfish Research. Vol. 20. No. 2. 6?3-fi57, 2001. COMPARATIVE PERFORMANCES OF JUVENILE HAUOTIS ROEI FED ON ENRICHED ULVA RIGIDA AND VARIOUS ARTIFICIAL DIETS SAM J. BOARDER' * AND MUKI SHPIGEL" ^Freinantle Maritime Centre, 1 Fleet St.. Fremaiitle, We.stern AiLstralia. 6160: ~I.srael Oceanoi;rapliic and Liniiiiiloiiicul Re.seiurli. National Center for Mariciilture. P.O. Bo.x 1212. Eiiat. Israel ABSTRACT GrovMh rates of juvenile Haliotis roei fed inorganically enriched Ulva rit^ida were compared with growth achieved on various artificial diets. Juvenile abalone (20—10 mm SL) were collected from reef platforms off the Perth (Western Australia) metropolitan area and assigned to one of seven different dietary treatments. All diets were fed atl libitum (3% bw day "' ) every second day, and growth rates were quantified over a 3-mo period. Specific growth rate (SGR) for both shell length and whole body weight indicated that growth of abalone fed enriched U. rii;itla was not significantly different {P > 0.05) to growth achieved from the best performing artificial diets. A 96-h salinity .stress test was also conducted on all treatments to as.sess the effects of diet on stress resistance. Diet significantly effected survival at 20%c. with the U. rigida fed animals exhibiting decreased tolerance to hypo-osmotic conditions (P < 0.05). This may have been caused by an interference with the cell volume regulatory mechanisms normally utilized by Haliotis spp. Enrichment of wild U. rigida increased the algal protein content from 1 1.4 ± 2% (dry weight) to 32.2 ± l.S^. perhaps partially explaining the difference in performance between this and other abalone feed trials utilizing Ulva spp. Results indicate that enriched U. rigida is a suitable feed for H. roei. providing similar growth to that achieved from several commercially available diets. KEY WORDS: abalone. Haliotis roei. Ulva rigida. nutrition, salinity tolerance INTRODUCTION The main constituents of aquaculture effluent are ammonia and phosphate, with ammonia known to be the main contributor to marine eutrophicalion when added to an oligotrophic marine en- vironment (Dosdat et al. 1993. Wu. 1993, Lemarie et al. 1999). Many shore-based aquaculture facilities utilize sedimentation ponds to remove particulate matter but ignore dissolved inorganics such as ammonia (NH^/NH/) and phosphate (POj'") because of the difficulty of their removal from high volumes of water. Use of macroalgal biofilters has been shown to remove a large propoilion of the dissolved nitrogen from aquaculture effluent when inte- grated into intensive fish culture systems (Neori 1996. Shpigel et al. 1996a). Species of sea lettuce {Ulvct spp.) have been shown to remove up to 907c of dissolved nitrogen from aquaculture effluent (Neori et al, 1998). Research in Israel has also shown that culture of Ulva spp. in nutrient-rich waters increases their protein content from 1 1-13% to over 33% of the dry weight (Shpigel et al. 1999). This enriched Ulva has been shown to provide good growth rates for Haliotis tuberculata (Neori et al. 1998, Shpigel et al. 1999), and Haliotis discus haimai (Corazani & lllanes 1998. Shpigel et al. 1999). Aquaculturists may be enticed to adopt effluent treatment procedures more readily if shown that enriched U. rigida can also be used as a feed for abalone species. Stress tests have previously been used as an assessment of general health or vigor in several aquatic species (Dhert et al. 1990a: Dhert et al. 1990b, Briggs 1992. Boarder & Maguire 1998. Samocha et al. 1998). Abalone are ostnoconformers and possess only limited osmoregulatory capabilities; thus, the effects of sa- linity fluctuations directly affect internal ionic composition (Brix 1983, Somero & Bowlus 1983). It has been shown previously that dietary history can directly effect resistance to low salinity in abalone (Boarder & Maguire 1998). This trial was conducted to assess the suitability of enriched U. rigida as a feed for Haliotis roei and to compare growth rates of this diet with crowth rates on various artificial diets. *Corresponding author. E-mail: sam_hoarder(^hotmail.com MATERIALS AND METHODS H. roei were collected from limestone reef platforms in the Perth metropolitan area in March 1999. Abalone were gently re- moved from the reef top and immediately placed into aerated seawater for transport. All animals were lagged (Hallprint Pty Ltd, FPN shellfish tags), and then randomly distributed into 28 trial tanks after being weighed to the nearest 0.01 g (whole wet body weight; WWBW). and measured with manual calipers (0.1 mm accuracy). The initial inean length and weight of the abalone were 31.9 ± 4.3 tiitii and 6.17 ± 2.28 g (mean ± SE) (ii = 360). Trial System Abalone were kept in rectangular 27-L plastic aquaria with artificial grass strips above the water line to prevent abalone from climbing out of the tanks. Twenty abalone were allocated to each trial tank (four replicates per diet), and all 28 tanks were arranged randomly in relation to experimental diet. The system was flowthrough and used a seawater bore as the water source. All tanks were individually aerated. The flow rate to each tank was maintained at 1 L min"' and checked weekly. Water quality pa- rameters (dissolved oxygen |DO]. pH. and temperature) were ran- dotiily measured during each feed with systetnatic analyses made in each tank weekly. Diets Diets used in this trial are listed in Table I. All artificial diets were commercially available (with the exception of the FRDC2 control diet), and enriched U. rigida was grown in high nutrient water (5 g N m"~ day"'; 0.6 g P m"" day"') on inorganic nutrients (NH4CI atid Na^HPOj) in accordance with Shpigel et al. ( 1999). The abalone were fed 3% bw/day. which is in excess of require- ments for greenlip abalone, H. laevigata (Tom Coote 1999 pers. comm.). Tanks were cleaned by siphon every second day. Feed rates for U. rigida were calculated on a dry weight basis. This was based on enriched U. rigida possessing a water content of 84%. The control diet, known as the FRDC2 formulation, was made available by the Fisheries Research and Development Council 653 654 Boarder and Shpigel TABLE 1. Artificial diets used. Protein Content Treatment Diet (<* Dry WeiHhIl 1 FRDC l—Auiluirs 35 2 AAFDG — Adam and Amos 27 3 Enriched Ulva rigida 32.2 4 Abfeed — Sea Plain Products 34.6 5 Haliogro — E.N. Hutchinson 30.0 6 Deakin Diet — Marine Feeds 7 Diet 7 — confidential Note: feed manufacturers are listed in italics. (FRDC). The dry ingredients included fislimeal, soyflour. semo- lina, sodium alginate, and CaCO,. The Deakin diet (Diet 6). and Adam and Amos (Diet 2) diet, were both cotiiniercially available within Australia. Haliogro (Diet 5) was from New Zealand, and AB-feed (Diet 4) was from South Africa. It should be noted that none of the diets used in this trial was specifically formulated for H. roei. Analysis of U. rigida Proximate analysis of U. rigiila was undertaken throughout the trial to quantify moisture, ash, protein (nitrogen x 6.25), fat, and carbohydrate levels. Moisture content was determined by gravi- metric weight loss on heating the samples to a constant weight at 105°C (Association of Official Analytical Chemists [AOAC] method number 950.46B). Ash was determined by gravimetric weight loss on ashing the samples in a muffle furnace at 550"C (AOAC method nutnber 920.153). Analysis of nitrogen content utilized the Kjeldahl method (AOAC method number 928.08), and fat/lipid levels were determined via acid hydrolysis and solvent extraction (AOAC method number 950.54). Consumption .\nalyses Consumption was assessed twice during the trial by manually siphoning uneaten feed froin tanks. Consumption was estimated by relating the dry weight of the uneaten food to the knovMi dry weight of the feed provided. Consumption data were corrected for dry matter weight loss attributable to leaching by leaching all diets over a 48-hr period and drying to a constant weight. Consumption estimates involved the collection of all uneaten food before clean- ing on the second day of the feed/clean cycle, through manual siphon onto lOO-jjim mesh screen. Collected feed was carefully washed onto a circular tllter mesh for drying and weighing. Feed was dried at 55-60°C for a period of 2 days before weighing. Salinity Stress Test A 96-h stress test for over-all health was conducted to assess the effect of diet on abalone health. Salinities for the stress test were derived from a prelitninary assessment of salinity tolerance for H. roei and the litnited literature regarding salinity tolerances of abalone (Nakanishi 1978. Mgaya & Mercer 1994, Jarayabhand & Paphavasit 1996, Boarder & Maguire 1998). Abalone were held in PVC cages (see Harris et al. 1997) and acclimatized in full strength seawater for 4 days (Boarder & Maguire 1998). Twelve abalone were randomly assigned to each cage. Upon acclimation, one cage from each diet was suspended vertically within an aerated l,()00-L tank of a particular salinity. Salinities used were 35, 25, and 207cc. and each salinity was replicated twice. Cages were checked every 4 h for 96 h by lifting the cage clear of the water and gently shaking. Abalone no longer attached to the cage and non- responsive to touch on the mantle were classified as mortalities. Animals responsive to touch and not attached to the cage were recorded as detachments and returned to the cages. RESULTS Growth Specific growth rate (SGR) was significantly affected by diet, as measured by either whole weight {P < 0.05) or by shell length (P < 0.01 ). Growth rates for H. roei fed inorganically enriched U. rigida (D3) were not significantly different (Tukey's test, P > 0.05) to growth achieved from the best performing artificial diets. The abalone grew best on the Adam and Amos ( D2 ) diet for both length and weight (SGR-L = 0.161 ± ().(X)4; SGR-W = 0.495 ± 0.2). However, growth was not significantly higher than on the FRDC2 (SGR-L = 0.139 ± 0.006: SGR-W = 0.371 ± 0.035), U. rigida (SGR-L = 0.142 ± 0.010; SGR-W = 0.379 ± 0.040), or AB-feed (SGR-L = 0.131 ± 0.006; SGR-W = 0.4113 ± 0.014) diets. This is shown in Figures I and 2, respectively. Consumption Analyses Consumption rate was significantly affected [P < 0.01) by diet for both consumption assessments. The two consumption analyses were also significantly different from each other (P < 0.01), as shown in Figure 3. Diets 2, 5, 6, and 7 were all observed to be less water stable than Diets 1, 3. and 4 after 2 days within the treatment tanks. Salinity Stress Test Abalone held in full strength seawater and 25%c salinity exhib- ited 100% survival at the conclusion of the 96-h trial, with the exception of abalone fed on Diet 3, which exhibited 929c survival over the same period in 257cc. Detachment rates for both of these salinities were negligible for all diets at the higher salinities. Diet significantly affected survival at 207i( (f < 0.05). No abalone fed Diet 3 (enriched U. rigida) survived in 207i< after 96-h (Fig. 4). Survival was significantly the same for all other diets (Tukey"s test P>0.05), with an average survival of 40.28 ±4.33%. 0.19 0.17 - c u -J 0.15 ^ 0.13 Qi O 0.11 - 0.09 - 0.07 ^ 0.05 ±S.E. ab abc be Dl D2 D3 D4 D5 D 6 D 7 DIET Figure 1. Effect of diet on specific growth rate for length (mm) for jmenile Haliotis roei (n = 4|. Note: Diets sharing the same letters are not signifkanllj different iTukey test, P > 0.05). Performances of H. roei Fed Artificial Diets 655 0.6 -| 0.5 - 04 0,3 0.2 0 1 0.0 -r ±SE ±S.E. abc ab ab be X Dl Si im D2 D3 D4 D5 D< DIET D7 Figure 2. Effect of diet on specific growth rate for weight Ig) for juvenile Halintis niei (n = 4). Xiite: Diets sharing the same letters are not significantlx different (Tuke> test, P > 0.05). Diet 3 was not included in the analysis of variance (ANOVA) because of zero variance caused by 09c survival. Mortality on Diet 3 was also observed earlier than all other diets (Table 2 1. Fifty percent of all animals fed enriched U. rigida died after 68 h. as compared with an average of 9.73 ± 2.06% for abaloiie fed other diets over the same time period. Most other diets reached 50% mortality after approximately 85 h. almost 20 h later than the U. rigida-ied abalone (Table 2). Abalone fed Diet 3 were also ob- served to detach earlier than those on other diets. DISCUSSION Growth rates of H. rod fed enriched U. rigida were not sig- nificantly different from growth rates achieved on the three best performing artificial diets (P > 0.05). Previous growth trials with Ulva spp. as a feed for abalone species have used "wild" Ulva spp. or Ulva spp. that had not been enriched (Mai et al. 1994. Mai et al. 1996. Simpson & Cook 19981. These growth trials have shown the effectiveness of U. rigida as a feed to be totally dependent on the S 0 035 so 0,03 ^ ^1 Sample I 1 I Sample 2 ±S,E o ■g 0.025 -I M 0.02 - be ^ij ^0 015 J •2 0 01 a. 1 0 005 2 0 m BC Dl D2 D3 D4 Diet D5 D6 D7 Figure 3. Kffect of diet on consumption rate (grams dry feed con- sumed/gram tank biomassi for juvenile Halintis roei at two separate sampling times {n = 4|. Xote: Diets sharing the same letters are not significantly different (Tukcy test. P > 0.05 1. 80 ^ 70 60 50 a 40 30 m 20 10 0 Dl D2 D3 D6 D7 Figure 4. Effect of diet cm survival after 96 h at Iti'^i salinity for juvenile H. roei (;/ = 2; 12 abalone per replicate). i\ole: Diets sharing the same letters are not significantly different (Tukey test, P > 0.05). abalone species to which it is being fed. Mai et al. ( 1996) found that although wild U. lactuca was a "inoderately good diet" for H. tubercidata. it was the worst of five diets for H. discus haiinai. Corazani and Illanes (1998) showed that H. discus hannai grew significantly faster on U. rigida than on other macroalgal diets, and the authors recommend juvenile abalone of this species be fed a diet of Ulva spp. in combination with an artificial diet. They felt that the artificial component of the diet was a supplement to the Ulva spp. and should be fed at levels no higher than 1% of total body weight day"'. Simpson and Cook ( 1998) investigated the growth of H. midae grown on various natural algal diets. They found that after 4 mo on trial diets, abalone fed on a diet of wild Ulva sp. were significantly shorter in shell length than abalone fed on most other macroalgal diets. t//ia-fed animals were also found to have the lowest wet weight to shell length ratios. However, when the animals were fed Ulva as a proportion of a mixed diet, they exhibited excellent growth rates, suggesting that Ulva provided essential nutrients not found in the other algae fed. The increase in protein content of U. rigida observed within this trial is supported by previous research in which Ulva spp. have been used as macroalgal biofilters (Tenore 1976, Neori 1996, Shpigel et al. 1996a. Shpigel et al. 1996b). In the current trial, wild U. rigida was found to have a protein content of 13% (% dry weight), compared with 32.2 ± 1.5% (% dry weight) for U. rigida. which had been enriched on inorganic nutrients. The high protein content throughout the trial could explain the differences observed between the growth achieved on wild Ulva spp. in other research and the growth achieved within this trial U. rigida. Fleming { 1995) found that the intake of digestible nitrogen directly influences the growth rates of H. rubra. This suggests that nitrogen may be a limiting factor for growth in Haliotis spp. Britz and Hecht (1997) support this view by stating that maximum growth can only be achieved when sufficient protein, in the correct proportions of amino acids, is supplied in the feed. Shpigel et al. (1996a: 1999) state that the good growth of H. tuberculata and H. fulgeiis when fed enriched U. lactuca was attributable to a consistent supply of high protein diet. The consumption results in this study do not adequately reflect the trends observed within the tanks. Some diets were observed to be more water stable than others, and thus were easier to fragment upon siphoning for collection. Diets 1 and 4 (FRDC2 and 656 Boarder and Shpigel TABLE 2. Effects of diet on time (liours ± SE) taken to reach 50% mortality for Juvenile H. rod held at im< saline (;/ = 2; 12 abalone per replicate). Diet 1 2 3 4 5 6 7 L,„ Time @ ICFi, 84 ± 4''' 86 ± 2- 68 ± 4" S"' + ~'^ S8±4'' 86 ± 2" >q6 Note: Diets sharing the same letters are not significantly different (P > 0.05). AB-feed) were both extremely water stable within the tank and. thus, held together well thioush the siphoning collection process. Diets 2. 5. and 6 (Adam & Amos. Haliogro and Deakin) were less water stable, but were still firm after 2 days. These diets frag- mented when they were siphoned, as did Diet 7. which was very soft after 2 days within the tanks. The data wei-e potentially biased against Diets 1.3. and 4 because of the fragmentation of other diets causing some loss of material during analysis. This would have reduced the amount of feed actually collected: thereby, increasing the apparent consumption. Differences in the various diets" water stability may be attributable to the diets having being formulated for colder water species. Stress determination for aquatic animals usually involves a quantitative evaluation of a variable that is directly affected by the stressor. The application of an environmental stress allows for an assessment of general health or vigor for animals from different tieatments. Salinity stress tests have been used previously to de- termine the quality of prawn post larvae (Briggs 1992. Samocha et al. 1998). fish larvae (Dhert et al. 1990a. Dhert et al. 1990b) and the effects of diet on abalone robustness (Boarder & Maguire 1998). Boarder and Maguire (1998) found that dietary vitamin levels directly deterinined the survival of H. laevigata at 239fc salinity. Animals fed on twice the normal dietary inclusion level of vitamin mix exhibited over 90% survival after 96 hours, as com- pared with approximately 50% survival for animals fed the normal dietary vitamin level. The poor survival at IQffoc for abalone fed U. rigida within this trial may be attributable to the osmoregulatory mechanisms uti- lized by Haliotis spp. rather than a general lack of health. Abalone species are known ionic conformers; therefore, blood osmolality and ionic concentration closely resemble those of the external environment (Somero & Bowlus 1983). The ability to regulate cell volume is also an important ability in soft-bodied osmoconfomiing marine animals (Tarr 1976. Burton 1983). because this prevents cells from rupturing upon exposure to low salinities. Cell \olume regulation within some osmoconfomiing mollusks is controlled by organic solutes, such as amino acids (Burton 1983. Mai et al. 1994). At low salinities, amino acids are excreted from the cell along with osmotically obligated water, thereby restoring cell vol- ume (Pierce & Amende 1981). Pierce and Amende state that there is a possibility that the physiological response of osmoconfomiing mollusks to low salinity may be directly related to their ability to maintain and control a large intracellular free amino acid pool. The most important free amino acids for cell volume regulation (dependant on abalone species) are taurine, glycine, alanine, and proline, with an emphasis on taurine (Burton 1983). Biosynthesis of the amino acid taurine from methionine is known to occur in gastropod mollusks (Mai et al. 1994). Mai et al. (1994) found that the levels of taurine in the viscera of H. discus hannai were sig- nificantly lower in abalone fed wild U. lactuca than abalone fed on all other diets. The authors also found that the viscera methionine levels within the abalone fed U. lactuca were extremely high, indicating some physiological inhibition in the biosynthesis of tau- rine from methionine. Toxic substances contained within U. lac- tuca. as detailed by Borowsky and Borowsky ( 1990), were thought to cause this inhibition. Johnson and Welsh (1985) found that an exudate excreted from U. lactuca killed lOOVc of estuarine crab larvae within 24 h. The authors also report that the bactericidal and fungicidal properties of illva spp. exudate have been known for some time. Because taurine is one of the most important amino acids for cell volume regulation, inhibition of taurine production from me- thionine could significantly affect the cell volume regulation abil- ity of abalone fed solely on Ulva spp. The comparatively good growth observed in H. roei fed U. rigida within this trial indicates that there was no feeding inhibition caused by antinutritional com- pounds or toxic substances contained w ithin the seaweed. Fleming (1995) found that antinutritional compounds within some algal species did not prevent preferences for algae containing high levels of digestible nitrogen. This suggests that, although the abalone fed on enriched U. rigida were capable of good gi'owth. they were not capable of adequate cell volume regulation when exposed to hypo- osmotic conditions because of a lack of taurine within their free amino acid pool. It is important to note that all other health pa- rameters (including growth) were excellent for abalone fed solely on enriched U. rigida. The results of this trial indicate the salinity tolerance of H. roei to be between 25 and 20%. This correlates closely with the limited literature on Halioiid salinity tolerances, which indicate short-term suivival is possible at salinities around 20% (Singharaiwan et al. 1992. Jarayabhand & Phapavisit 1996. T. McCormick 1997 pers. conini.. Boarder & Maguire 1998). The 96-h LS50, the salinity at which 50% of the test animals survive after 96 h. is relatively close to 20% for H. roei. Most diets displayed 50% survival at just under 96 h within this trial. This compares with only 10% sur\i\al after 96 h for W. diversicolor superte.xta ]\x\sn\\cs transferred froin 35 to 20% at similar temperature (Chen & Chen 2000). In conclusion, enriched U. rigida is a suitable feed for H. roei. producing coinparable growth rates to several commercially avail- able manufactured feeds. Diet directly affects survival of abalone under hypo-osmotic stress, and U. rigida may impair the ability to cope with this stress by affecting cell volume regulatory mecha- nisms. ACKNOWLEDGMENTS We thank the Natural Heritage Trust - Coasts and Clean Seas. South Metropolitan College of TAPE. Fisheries WA. and the Aquaculture Development Fund for funding this project. We also thank Mr. Arron Strawbridge for technical assistance and Mr. Gavin Partridge for reviewing the manuscript. Performances of- H. koei Fed Artificial Diets 657 ijteraturp: cited Boarder, S. J. cS: G. B. Maguire. 1998. Evaluation ol vitamin and mnieral requirements for greenlip abalone Halinns laevigata A progress re- port. In: P. W. Hone, editor. Proceedings of the 5th Abalone Aquacul- lure Workshop. July 3-6, Henley Beach, 1998. Hobart, Tasmania, Aus- tralia: FRDC. pp. 75-86. Borowsky. R. & B. Borowsky. 1990. Feeding inhibition of the salt marsh amphipod, Gammarus palu.srris Bousfield, 1969. by heat labile sub- stances in Ulva lacnica. Cnisuiceana 59(3):299-301. Briggs. M. 1992. A stress test for determining vigor of postlarval Penaeus monodon Fabiicnis. Aqiiaciili. Fish. Mgiut. 23:633—637. Brit/. P. J. & T. Hecht. 1997. Effect of dietary protein and energy level on growth and body composition of South African abalone, Halious iiii- dae. Aquaciihure 156:195-210. Brix, O. 1983. Blood respiratory properties in marine gastropods. In: O. Brix. editor. Environmental biochemistry and physiology, vol. 2: The mollusca. London: Academic Press, pp. 51-76. Burton. R. 1983. Ionic regulation and water balance. In: O. Brix. editor. En\ironmental biochemistry and physiology, vol. 11: The mollusca. London: Academic Press, pp. 273-315. Chen, J. C. & W. C. Chen. 2000. Salinity tolerance of Haliolis diversicoloi supertexia at different salinity and temperature levels. Aquaciilturc 181:191-203. Corazani, D. & J. E. Illanes. 1998. Growth of juvenile abalone, Haliolis discus haiinai Ino 1953 and Haliolis riifescens Swainson 1822. fed with different diets. / SMI. Res. 17(3):663-666. Dosdat, A., F. Gaumet & H. Chartois. 1995. Marine aquacuhure eftluent monitoring: Methodological approach to the evaluation of nitrogen and phosphorus excretion by t~ish. Ac/iiucidl. Enn. 14:59-84. Dhert. P.. M. Duray. P. Lavens & P. Sorgeloos. 1990a. Optimized feeding strategies in the larviculture of the asian seabass Lutes calcarifer. The Second Asian Fisheries Forum, 319-323. Dhert, P., P. Lavens, M. Duray & P. Sorgeloos. 1990b. Improved larval survival at metamorphosis of Asian seabass (Lates calcarifer) using w3 HUFA enriched live food. .Aquaciilture 90:63-74. Fleming. A. E. 1995. Digestive efficiency of the Australian abalone Halh Otis rubra in relation to growth and feed preference. Aquaculiure 134: 279-293. Harris. J. O.. G. B. Maguire. J. Edwards & S. M. Hindrum. 1997. Effect of nitrite on growth and oxygen consumption of juvenile greenlip abalone. Haliotis laevigata Donovan. / Shellfish Res. 16{2):395-401. Jarayabhand. P. & N. Paphavisit. 1996. A review of the culture of tropical abalone with special reference to Thailand. Aquacuhure 140:159-168. Johnson, D. A. & B. L. Welsh. 1985. Detrimental effects of Ulva lactuca (L.) exudates and low oxygen on estuarine crab larvae. / E.xp. Mar. Biol. Ecol. 86:73-83. Lemarie. G., J. L. M. Martin, G. Dutto & C. Garidou. 1999. Nitrogenous and phosphorous waste production in a flow-through land-based farm of European seabass [Dicentrarchus lahrax). Aquat. Living Resour. ll(4):247-254. Mai, K„ J. P. Mercer & J. Donlon. 1994. Comparative studies on the nutrition of two species of abalone, Haliotis tuberculata L. and Haliolis discus hatmai Ino. II. Amino acid composition of abalone and six species of macroalgae with an assessment of their nutritional value. Aquacuhure 128:115-130. Mai. K.. J. P. Mercer & J. Donlon. 1996. Ci)mparati\e studies on the nutrition of two species of abalone. Haliotis tuberculata L. and Haliotis discus haimai Ino. V. The role of polyunsaturated fatty acids of mac- roalgae in abalone nutrition. Aquacuhure 139:77-89. Mgaya. Y. D. & J. P. Mercer. 1994. A review of the biologs. ecology, fisheries, and mariculture of the European abalone Haliotis tuberculata Linnaeus 1758 (Gastropoda: Haliotidae). Proc. Renal Irish Acad. 94B(3):285-304. Nakanishi. T. 1978. Studies on the effects of the environment on the heart rate of shelltlsh. II. Effects of temperature, low salinity, and hypoxia on the heart rate of an abalone Haliotis (Nordotis) (//.vch.v hantuii Ino. Bull. Hokkaido Reg. Fi.sh. Res. Lib. 43:59-67. Neori, A. 1996. The type of N-supply (ammonia or nitrate) determines the performance of seaweed biofilters integrated with intensive fish cul- ture. Israeli J. Aquacuh. — Bamidgeh. 48( I ): 19-27. Neori, A., N. L. C. Ragg & M. Shpigel. 1998. The integrated culture of seaweed, abalone. fish, and clams in modular intensive land-based systems: II. Performance and nitrogen partitioning within an abalone (Haliotis tuberculata) and macroalgae culture system. .Aquacuh. Bug. 17(4):2l5-239. Pierce, S. & L. Amende. 1981. Control mechanisms of amino acid- mediated cell volume regulation in salinity-stressed mollusks. / E.xp. Zool. 215:247-257. Samocha, T. M., H. Guarjardo. A. L. Lawrence. F. L. Castille. M. Speed, D. A. McKee & K. I. Page. 1998. A simple stress test for Penaeus vannamei postlarvae. Aquacuhure 165:233-242. Shpigel. M.. Neori, A. & A. Marshall. 1996a. The suitability of several introduced species of abalone (Gastropoda: Haliotidae) for land-based culture with pond-grown seaweed in Israel. Israeli J. .Aquacuh. — Bamidgeh. 48(4): 192-200. Shpigel, M.. A. Mar.shall, I, Lupatsch, J, P. Mercer & A. Neori. 1996b. Acclimation and propagation of the abalone Haliotis tuberculata in a land-based culture system in Israel. / World Aquacuh. Soc. 27(4):435- 442. Shpigel. M.. N. C. Ragg, 1. Lupatsch & A. Neori. 1999. Protein content determines the nutritional value of the seaweed Ulva lactuca for the abalone Haliotis tuberculata and Haliotis discus hannai. J. Shellfish Res. 18(l):227-233. Simpson. B. J. A. & P. A. Cook. 1998. Rotation diets: .A method of improving growth of cultured abalone using natural algal diets. J. Shell- fish Res. 17(3): 635-640. Sinhagraiwan, T,, M. Doi & M. Sasaki. 1992. Salinity tolerance of juvenile donkey's ear abalone. Haliotis asiniiia Linne. Thai .Mar. Fish. Bull. 3:71-77. Somero. G. & D. Bowlus. 1983. Osmolytes and metabolic end products of mollusks: The design of compatible solute systems. In: O. Brix. editor. Environmental biochemistry and physiology, vol. 2: The mollusca. London: Acadeinic Press, 77-100. Tarr. K. 1976. An analysis of water content regulation in osmoconforming limpets (Mollusca: Patellacea). / Exp. Zool. 201:259-268. Tenore. K. 1976. Food chain dynamics of abalone in a polyculture system. Aquacuhure 8:23-27. Wu. R. S. S. 1995. The environmental impact of marine fish culture: toward a sustainable future. Mar Poll. Bull. 31(4-121:159-166. Jiminal ,if Slwllfisli Research, Vol. 20. No. 2. 659-665. 2UU1. SEASONAL ENERGETICS OF HAUOTIS FULGENS (PHILIPPI) AND HALIOTIS TUBERCULATA (L.) SUSAN C. MCBRIDE/ * EINAT ROTEM," DAVID BEN-EZRA," AND MUKI SHPIGEL" ' University of California. Sea Grant Extension Program. 2 Commercial Street, Enrelca, CA 95501: -Israel Oceanographic and Limnological Research. National Center for Mariciilture. Eilat, 88112, Is ■acl ABSTRACT Bioenergetics (food ingestion, absorption, growtli, respiration, ammonia, and mucus production) of the warm-water abaione Hatioiis fttlgens and tfie cold-water species H. ntberculata were studied under summer and winter conditions in the Gulf of Aqaba. Individuals of 38-+5 mm shell length and 7-11 g whole live weight were fed Ulva lactiica and Gracilaria conferta. Food ingestion was greater in summer but not sufficient to overcome the high maintenance, resulting in low (H. fulgens) or no (H. tiibenulata) growth in summer. Growth for H. fulgens and H. riiheirulaui durmg winter was 0.10 ± 0.05 nim/d and 0.09 ± 0.02 mm/d and 0.02 + 0.01 and 0.03 ± 0.01 g DW/d, respectively. Respiration rate and ammonia excretion were greater in summer for both species. In summer. H. tiiberculata had greater food absorption than H. fulgens, and the reverse was found in winter. Mucus secretion was only 1-9% of the energy budget. In winter, growth accounted for the majority of the energy budget of both species and in summer, respiration and ammonia were the mam components. Overall. H, fulgens seems lo be most suited to year-round conditions found in the Gulf of Aqaba. KEY WORDS: abaione. bioenergetics. Halimis fulgens. Huliolis luherculala INTRODUCTION Nonindigenous species must have certain characteristics if they are to be successfully cultured in a new environment. In particular, candidates for aquaculture must be acclimated and problems iden- tified quickly. It is essential to know the degree of compensation possible to environmental change and to understand how they function in their natural environment. An energy budget relates the growth of an animal to its physiological condition and is one way to understand the basic biology of an animal and observe its energy partitioning for maintenance and growth in the new environment. Previous research has shown that Haliotis fulgens and H. ui- berciilata are suitable for culture in warm-water environments (Shpigel et al. 1996a. Shpigel 1996b). However, comparison of the performance of these two species and description of the process by which ingested food is utilized for maintenance and growth is not known for warm seawater conditions. Growth and production characteristics can be measured using energy budget parameters. Food consumption, feces egestion. respiration, ammonia excretion, and mucus secretion are essential elements in an abaione energy budget (Peck et al. 1987. Barkai & Griffiths 1988. Donovan et al. 1998). Nutrient relationships are also useful in understanding an energy budget and indicate the role of digestive characteristics in energy utilization. Food absorption and food conversion efficien- cies have been used in other invertebrates to indicate interactions between metabolism and production (Lawrence 1975. Bayne et al. 1975). H. fulgens is a warm-water species found in seawater of 12- 23°C (Hahn 1989) but tolerates temperatures up to 28' C for short periods (Leighton et al. 1981 ). H. luberciilata is a cold-water spe- cies found in seawater of 5-18°C (Mgaya 1995) but shows accel- erated growth when cultured in seawater near 20°C (Koike et al. 1979. Shpigel et al. 1996a). H. fulgens and H. luberculata are benthic. motile to semisessile subtidal species relying on drift and attached macroalgal communities for food. The objective of this study was to determine the status of H. fulgens and H. luberculata by measuring energy budget parameters in summer and winter. METHODS *Corresponding author Two species of abaione. Haliotis fulgens and H. luberculata. were studied for two seasons, winter and summer. Individuals of 38-15 mm shell length and 7-1 1 g whole live weight were selected from cultured populations at the National Center for Mariculture in Eilat, Israel. Incoming seawater was filtered by sand and cartridge filtration to 1 p.. Seawater temperatures were 24-28°C and 20- 22°C in summer and winter, respectively. Dissolved oxygen was always near saturation, and salinity was 41 ± 1 ppt. Ambient photoperiod was 14 h of light in summer and 1 1 h of light in winter. Animals were held in glass aquaria of 25 L. For each species, five replicate aquaria held 10 individuals per season. Aba- lone were selected randomly from two aquaria for respiration, ammonia, feces, and mucus measurements. Individuals were used only once for metabolic measurements. Three aquaria per species were used to determine feeding and growth rates. These individu- als were not used for physiologic measurements. A preliminary energy budget including ingestion, assimilation, respiration, am- monia production, feces, mucus secretion, and growth was pre- pared showing the allocation of energy in joule/g DW/day in sum- mer and winter. Somatic Growth (G) H. fulgens and H. tuhcrcuhna were measured at the start and finish of each seasons' experiments to determine daily growth increments. Growth was measured as the change in shell length (mm) and whole dry weight (g) during the summer interval, 7 July to 5 October 1995, 9.3 days, and the winter interval 12 December 1995 to 5 March 1996, 83 days. Specific growth rate was calcu- lated for weight (g) and length (mm) as |(//; mean final abaione dry weight (length) - In mean initial abaione dry weight (length))/ days per experimental period] * 100. Growth as change in energy content (joule/g DW/day) was calculated for the energy budget using the dry tissue weight of the abaione. At the beginning of the study, samples of each H. fulgens and H. tuberculaia were weighed to the nearest 0.001 g. measured to 0.01 mm. and dried at 60°C for 7 days to determine whole wet weight/dry weight ratio of the abaione. The dry tissue/dry shell ratio was calculated after drying. The dried abaione tissue was 659 660 McBride et al. combusted in a Parr semimicro bomb calorimeter using benzoic acid as a standard to determine gross energy content (joule/g DW). Consumption (C) The abalone were fed Ulva lactiica and Gnuilaiia cimfeita ad lihiriim at a ratio of 3:1. Food introduced to the abalone and uneaten food removed was blotted on paper toweling to remo\e excess moisture. Dry food ingestion was measured over 1 -wk in- tervals four times each season. Algae held in aquaria without aba- lone had no significant weight change. Feed ingestion was calcu- lated directly as the difference between algae added and removed from the aquaria. The protein content, dry and ash weight of the algae were measured each season. Dry and ash weights were determined by drying to a constant weight at 60°C followed by incineration for 24 h at 55()°C. Crude protein was measured using the Kjedahl method and multiplying ii by 6.25. The gross energy content (cal/g DW) of U. lactiica and G. conferta was determined by combustion in a semimicro bomb calorimeter using benzoic acid as a standard. Gross food conversion efficiency was calculated using the food ingested in terms of energy as (mean final joule/g DW-mean ini- tial joule/g DW)/ (mean joule/g DW food consumed/abalone/ season) x 100. The dry matter content of the abalone tissue was used for food conversion efficiency. Net food conversion effi- ciency for joule/g DW/day was calculated using the food absorbed. Feces (Fl and Absorption Efficiency Feces were collected by placing individual abalone in a I L conical tank with a 1 mm mesh floor. The water containing the feces was filtered onto the dried, preweighed GFC filters, dried at 60°C for 24 h. Dry feces were incinerated for 3 h at 450°C in muffle furnace. The dry and ash weights were used to calculate the organic content of the feces. The absorption efficiency was calculated by the indirect method of Conover (1966): U' = F' - E7(1-E')(F'). where U' is the absorption efficiency, and F' and E' are the concentrations of organic matter in the food and feces, respectively. The energy expended in feces egestion was calculated from the food absorption values. The gross energy values of H. tiiherciilata feces, 2,817 cal/g dry weight, determined by Peck et al. (1987) were used for our energy budget calculations. Respiration (R) A 72()-mL round, Perspe.x respiratory chamber v\ Ith a magnetic stirrer submersed in a tlowthrough seawater bath was used to measure o,\ygen levels every minute until 50% oxygen saturation was reached. Test Point® software was used to measure the in- coming mg 0,/L with an accuracy of 0.01 mg O^/L. Each respi- ration determination lasted for approximately 3 h. Measurements were made during the day and night. Results are presented per 24 h. Five chambers were run concurrently. Respiration measure- ments were made over 2-wk intervals each season. Night respira- tion was completed between 2300 to 0200 h. Day respiration was completed between 1100 to 1400 h. For the energy budget, the daily oxygen consumption, mg Oi/g DW/day was calculated from the day and night rates. Oxygen consumption was converted to energy using 3.47 cal/mg O, (Bayne et al. 1975). Ammonia (U) In separate tests, individual abalone were placed in respiration chambers. Seawater collected from the chambers was analyzed with a Technicon II Autoanalyzer giving |j,M NH4*/h/abalone. Ammonia was analyzed following filtration through HCI washed Whatman GF/C filters according to the modified oxidation proce- dure of Neori et al. (1996). Three 1-mL aliquots were collected from each abalone. For the energy budget, a mean daily ammonia excretion rate jxM NHj*/" DW/day was calculated for day and night. Conversion factors used to calculate energy equivalents of ammonia excretion were |jl18 M ammonia/ jxg ammonia and 0.0249 joule/|j.g ammonia (Bayne et al. 1975). Samples were col- lected during the night and day as for respiration measurements. Each test lasted 2.8 to 3.5 h. Mucus (M) Individual abalone were moved to clean aquaria without feed and w ith low aeration for 48 h. The abalone were removed, and the walls of the aquaria were cleaned with a spatula. The mucus from the aquaria water and walls were collected on preweighed, dried. GF/C filters. The filters were rinsed with distilled, deionized water to remove salts, and dried at 60°C. Mucus samples were pooled for all aquaria per species. The caloric value of mucus used was 5,439 cal/g dry weight (Peck et al. 1987). Energy Budget An energy budget was calculated for H. fiilgcns and H. tiiher- ciilata during summer and winter based on joule/g DW abalone/ day using food ingestion (C), somatic growth (G). respiration rate (R), ammonia excretion (U), feces egestion (F), and inucus secre- tion (M) as parameters. The scope for growth, all maintenance energy (respiration, ammonia, mucus, and feces) subtracted from the total absorbed energy (A) (Widdows et al. 1989) was also calculated. Statistics Somatic growth, respiration, ammonia production, dry feed in- gestion, organic content of the feces, and total organic absorption were compared between species and between seasons by Student's f-test. All mucus samples were pooled to give one value per spe- cies per season. Specific growth rates for dry weight and shell length were calculated from means resulting in one value per spe- cies per season. RESULTS The whole wet weight (g)/whole dry weight ratio for H. fidgens and H. tiihercidata are 0.47 ± 0.3 and 0.43 ± 0.2. respectively in = 20 in each sample). The dry tissue/dry shell ratio is 0.36 ± 0.07 and 0.33 ± 0.3 for H. fulgens and H. tubercidata. The caloric content of dry tissue is 4,848 cal/g DW and 4,600 for H. fulgens and H. tiiherciilata. respectively. Somatic Growth (G) In summer. Haliotis fulgens had a greater increase in shell length than H. tubercidata. but there were no significant differ- ences (Table 1 ). At the end of the summer interval, H. tiihercidata lost weight as compared to H. fulgens. which showed a slight weight gain. The mean weight loss of H. tubercidata resulted in significant differences between species in this measurement of somatic growth. In winter, shell length and dry weight increase were significantly greater for H. tuberculata than H. fidgens. So- matic growth was greater for both species in winter than summer (Table 2). The specific growth rates (SGR), daily changes in DW, and shell length reflected the growth pattern seen from direct Abalone Bioenerghtics 661 TABLE I. Student's /-test results of somatic j;roHth, respiration, ammonia, and mucus production itf Hahdris fiilgens and H. tiihcnulala compared between species. Species (mean ± SD) Test statistic til Grovvth Season H. fulgens H. Itiberculata df P Significance Shell length mm/day Summer 0.02 ± 0.02 0.01 ±0,01 1 ,094 7 0.695 NS Winter 0.10 ±0.05 0.09 ± 0.02 0.397 7 0.291 NS Dry vveigln g/day Summer 0.01 ±0.001 -0.01 ±0.01 3.896 7 0.001 ** Winter 0.02 ±0.01 0.03 ±0.01 -2.847 7 0.001 ** Feed dry g/day Summer 0.02 ±0.01 0.04 ±0.01 -5.233 7 0.002 * Winter 0.02 ±0.01 0,03 ±0.01 -5.509 7 0.002 * E' % Summer 33 ± 2 22 ±1 5.711 7 0.001 * Winter 30 ± 2 44 + 2 -4.102 7 0.001 * V % Summer 76 ±2 87 ±5 -5.149 7 0.002 * Winter 80 ±3 64 ±3 4.065 7 0.001 * mg 0,/g DW/d Summer 3.9 ± 1.3 11 . 1 ± 4.5 -3.466 8 0.008 * Winter 2.1 ±0.9 2.3 ± 0.9 -0.545 12 0.596 NS Jig NHj+Zg DW/d Summer 17.1 ±0.6 67.2 ±32.0 -3.837 6 0.009 * Winter 14.0 ±5.6 15.5 ±6.5 -0.361 7 0.728 NS i* P < 0.05; ** P < 0.001; NS not significant; g DW = g dry weight; E' organic content of the feces. U' organic matter absorbed). growth measuremenls (Table 3). Energy tised for growth was greater in winter than in summer. Consumption (Cl Dry feed ingested/g DW abalone/day was significantly greater for H. liihcniiUira than H. fulgens during both seasons (Table 1 ). Neither species had significantly different feed ingestion rates when compared between seasons (Table 2). Gross food conversion efficiency was greatest for H. tuherciilata in the winter, followed by H. fulgens in the winter. H. fulgens in the summer, and least efficient in H. tuberculuta duritig summer months (Table 3). Algal protein content was greater in winter than summer (t = -2.5052. df = 12, P = 0.028, t = -8.483, df = 12. P = 0.005) for Ulva lacnica and Gracilaria conferta, respectively. Caloric values were not significantly different between seasons (Table 4). Organic absorption was significantly different between species in both seasons. H. tuherculahi had a higher food absorption rate than H. fulgens during sutnmer. and the reverse was true in winter. The organic content of both algal species was the same. 70%. and did not change with season (Table 2). Food absorption in H. tu- heivuUila in winter was the lowest found in this study. Feces (F) The organic content of the feces was significantly greater for H. fulgens in winter and for H. tuberculuta in summer. The organic content of the feces was significantly greater for H. luhenulata in winter than summer. Organic content of the feces had an inverse relationship with the organic absorption (Tables I and 2). Respiration (R) H. fulgens and H. tuberculuta had greater respiration I'ates in summer than winter. There were no sianificant differences in w in- TABLE 2. Student's Mest results of somatic growth, respiration, amniunia. and mucus production of Haliulis fulgens and H. luberculata compared within species between seasons. Species Test statistic (/I Df Significance Shell length (mm/day) Dry weight (g/day I Feed dry g/day E' (%) U' (%) mg O Jg DW/24 h (ig NHj+/g DW/24 h H. fulgens H. tiiherciilinu H. fulgens H. tnberculiila H. fulgens H. tuberculuta H. fulgens H. tuherculata H. fulgens H. tuherculata H. fulgens H. tuherculata H. fulgens H. tuherculata -4.465 -9.116 -4.196 -6.230 -1.276 -1.066 0.756 -8.341 -l.(.)59 5.677 3.030 4.687 0.140 3.236 7 <0.001 7 <0.()01 7 <0.001 7 <0.001 7 0.246 7 0.327 7 0.464 7 <0.001 7 0.312 7 <0.001 1 0.011 9 0.001 8 0.892 5 0.023 NS NS NS ** NS NS ' P < 0.05; ** P < 0.001; NS not siiznificant; g DW = a dry weight; E' organic content of the feces; U' organic matter absorbed. 662 McBride et al. TABLE 3. Specific grow th rates as % DVV/d and % SL/d and food conversion efficiencies ( FCE ) for Halioth fulgens and //. luberculata. H. fiilgens H. tuberculata Summer Winter Summer Winter SGR % DW/d SGR % SL/d Gross FCE— f/f . joule/g DW/d Net FEC— f/r, joule/g DW/day 0.19 0.03 6.6 S,7 0.29 0.06 9.0 11,2 -0.07 0.01 -0.5 -0.6 0.63 0.22 15.5 24.2 TABLE 4. Energy and protein content (mean ± SD( of Lira lactiica and Gracilaria confcrta. [Via lac liica Gracilaria conferta Summer Winter Summer Winter Gross energy (cal/g DW) Crude protein C/r/DW) 327S± 166 13.3 ±2.2 3299 ± 191 24.2 ±5.9 3657 ± 59 20.9 ± 4.6 3339 ± 170 29.3 ± 2.8 TABLE 5. Day and night respiration values (mgO,/g DW/h, mean ± SDl for Haliotis fiilgens and H. tuberculata. H. fiilgens H. tuberculata Day Night Day Night Summer Winter 0.2 ±0.1 0.1 ±0.1 0.2 ±0.1 0. 1 ± 0. 1 0.2 + 0. 1 0.1 ±0.1 0.3 ± 0.2 0.1 ±0.1 TABLE 6. Day and night ammonia excretion values (pMHNj*/g DW/h. mean ± SD) for Haliotis fiilgens and H. tuberculata. H. fiilgens H. tuberculata Day Night Day Night Summer Winter 0.3 + 0.1 0.6 ±0.2 1.0 ±0.3 0.7 ±0.2 1.3 ±0.3 0.6 ± 0.3 5.3 ±3.1 0.7 ± 0.4 TABLE 7. Energy budget per food ingested and scope for gro\\th per energy absorbed: units are joule/g DW/d unless otherwise indicated. H. fulgens Summer Winter H. tuberculata Summer Winter 509 427 198 273 161 34 125 29 68 154 8 6 116 164 156 41 30 10 -13 154 c A R U F M G Other % of total C Scope for growth 274 207 56 32 58 23 51 45 17 30 248 438 30 26 50 17 -13 14 6 49 AB ALONE BlOENERGETICS 663 ter, but H. tubercuUua had a greater rate than H. fiilgens in sum- mer. For both species, respiration was similar for day and night measurements, except for H. tiiherciilata in the siunmer. where night respiration was greater (Table 5). Ammonia iV) Ammonia excretion rate was significantly greater for H. liiher- ciilala than H. fiilgens in summer. There were no significant dif- ferences between species in w inter. When compared between sea- sons. H. tiibeniilala had a greater rate in summer than winter, and there were no significant differences in the ammonia excretion of H. fiilgens between seasons (Tables 1. 2). Ammonia excretion was similar during day and night measure- ments, except for H. tuberciilara. which showed a higher rate during the summer night samples (Table 6). Mucus (M) Mucus collected from all aquaria per species per season was pooled. Mucus secretion was 0.004. 0.001. 0.003. and 0.001 g DW/abalone/day for H. fulgens-summer. H. nihi'iriilata-summer. H. fiilgens-\\\n\er. and H. tiiheniilciui-winter. respectively. Energy Budget For both species and seasons, 83-94% of the assimilated en- ergy was accounted for. except H. tuhercidata in the summer, where 70% of the total energy intake was measured (Fig. I ). Be- cause of higher feed ingestion rates in summer, total energy in- gested was greater than in winter. In the summer. H. niberciiUiia responded to warm seawater conditions by high respiration rate and ammonia excretion. H. fiilgens had the highest mucus secre- tion rate overall, particularly during summer. The scope for growth was greater in winter than summer. Al- though H. tiiberculata would have been expected to show growth rates similar to H. fiilgens in the summer by the scope for growth calculation, this species lost weight and showed a slight increase in shell length. DISCUSSION The energy budget published for H. tiibcnidata (Peck et al. 1987) was done in its normal range and offers valuable insight to our results. H. fiilgens has very little basic biological information published, thus we were unable to make any direct comparisons of metabolic and nutritive characteristics. In general, the food ingestion increased in summer for both species. As expected, the energy expenditures in body processes of H. fiilgens and H. tiiberculata also intensified during the summer. Additional energy was available from the food ingested, but most of this was used for maintenance. The significant difference in responses of H. fiilgens and H. tiiberculata may result from the seawater temperatures tested here: 20-26°C. H. fiilgens is a relatively warm-water species and has shown accelerated growth for short periods at 28°C (Leighton 1981 ). H. tiiberculata has inaximal growth at temperatures near 20"C (Mgaya 1993. Shpigel et al. 1996a) and negative growth at 26°C (Shpigel et al. 1996b). On an annual basis. H. fiilgens had greater growth in shell length (22.1 vs. 1 8.5 mm/y ). and H. tiiberculata has greater growth in whole dry weight ( 10.6 vs. 9.0 g/y). The growth rates found here may not represent maximal performance of H. tiiberciihita. because hiaher growth rates were found between Hfulgens Summer □ respiration □ ammoma 17% 20% □ mucus □ growth □ feces Bother 23%\-:-:-: ^'•"'''-'■'i I -,0, 19% Hfulgens Winter 6% 12% 20% 46% // tiiberculata Summer 30% 10% 14% 2% 24% H tuherculala Winter 9% 8% 36% V Figure I. Knergy budgets for Haliotis fulgens and H. tuberculata for summer and winter. Values are percentage of total energy ingested; other is energy not accounted for. 664 McBride et al. March and May (Neori et al. 199S). a time of year we did not study. Fuller understanding of growth of these two species in a warm- water environment is shown in the scope for growth. The negative summer values indicate stress and reflect the high use of body reserves used for maintenance metabolism in both species. Other evidence of stress in H. liiberciilala is demonstrated by the nega- tive growth found in the energy budget. In contrast. H. fult;eus. continued growing at a lower rate in the higher seawater tempera- tures recorded in summer months. Mortalities observed during the decrease in seawater tempera- tures (Shpigel et al. 1996b) may result from the stressed condition of the H. tubercidata at this time. H. tubeicidala may deplete stored body reserves of glycogen (Bennett & Nakada 1968) during summer. This combined with the additional stress of temperature change may explain these mortalities. The negative scope for growth found for H. fidgens may approach the lethal limit but seems less severe for this species. Respiration is a useful and sensitive measure of daily energy expenditure. The stress of the higher seawater temperatures in- creased respiration rates. The solubility of oxygen changes little in the temperature and salinity range used in this study (Dejours 1988), but the ability of the abalone to cope with the seasonal environmental changes and maintain their metabolism was com- promised. The higher rates of respiration relative to food intake found at temperatures between 24 and 26 C limited growth of both species. This relationship is also shown in the gross and net food conversion efficiencies that are lower for both species in summer. The quality of the food is also important. Increased food in- gestion in summer was probably more related to temperature than the slightly lower protein content of the algae in summer. The energy content may be more important in determining feed inges- tion rates than protein (Shpigel et al. 1999). Energy content of U. lactuca and G. coufenci was not significantly different between seasons. The protein content of the algae fed here was consistent throughout the study and is higher than natural algal diets used in other studies. Cultured U. lactuca and G. conferta contains >1Q% protein, as compared to <1 19r found in natural populations (Peck et al. 1987. Barkai & Griffiths 1988. Shpigel et al. 1996b, Donovan et al. 1998). The elevated protein content seems to sustain growth in H. fiilgens in summer and winter and provides adequate nutri- tion for significant growth of H. tubercidata in winter. Organic absorption was significantly greater for H. tubercidata in winter than summer, again reflecting this species" ability for production in temperatures around 20°C. Seasonal variation in organic absorption is known for other invertebrates (McBride et al. 1998). The organic absorption of H. fulgens did not change sig- nificantly with season, again suggesting that this species is well adapted for production in seawater between 20-26°C. The gross feed conversion efficiency indicates the efficiency with which the abalone utilizes the ingested energy for growth. Gross and net food conversion efficiencies were greatest for H. lubercuhita during winter, because growth was high, and food ingestion and absorption were low, as compared to summer months, when growth was negative, and food ingestion was high. As vs'ith other measurements, the gross and net food conversion efficiency of H. fulgens were similar in summer and winter. The food conversion efficiencies were higher in winter for both .spe- cies, showing more energy was available for growth and less was used for maintenance. Mucus production was lower than other abalone energy bud- gets, between 1-9% of the total energy, as compared to 16-27% (Peck et al. 1987, Barkai & Griffiths 1988). except for H. ka- lutschatkana in winter, where mucus secretion contributed 4% of the total energy budget (Donovan et al. 1998). Two factors may contribute to the low mucus secretion values found here. First. H. fulgens and H. lubercuhita move very little in culture condition at the National Center for Mariculture (pers. observ.). This is similar to the low activity found in H. kauitschatkana in winter (Donovan et al. 1998). Second, other studies determined mucus secretion by subtraction, and we calculated it by filtration. The combination of low activity of abalone m our study and our method of collection may have contributed to this difference The scope for growth during winter months was similar for both species, reflecting their comparable growth rates. Although H. tubercidata lost weight during summer, their weight gain was greater on an annual basis than H. fulgens. Although not studied here, survival of H. fulgens seems to be greater than H. tubercidata because of seasonal seawater temperature changes (Shpigel et al. 1996b). adding more weight to the argument for pursuing culture of this species in a warm-water environment. In conclusion, the energy budgets of H. fulgens and H. iiiber- culata provide a good indication of the animals' response to the culture environment. The mean seawater temperatures experienced by each species in its natural range strongly influenced expendi- tures of energy, resulting in low growth and high maintenance during summer and high growth and low maintenance in w inter. In all factors measured, H. fulgens seems better adapted to the year- round culture conditions tested and would seem to be the most suitable species overall for culture in the Gulf of Aqaba. ACKNOWLEDGMENTS This study was supported in part by the Israeli Ministry for Energy and Infrastructure by a joint program of the European Economic Council and the Israeli Ministry for Science (Grant 4564192) to MS. This paper is funded in part by a grant from the National Sea Grant College Program, National Oceanographic and Atmospheric Administration. U.S. Department of Commerce, un- der Grant NOAA NA66RG()477. project number A/EA-1 through the California Sea Grant College System to SM. The views ex- pressed here are those of the authors and do not necessarily reflect the views of NOAA or any of its subagencies. The U.S. govern- ment is authorized to reproduce and distribute this paper for gov- ernmental purposes. LITERATURE CITED Barkai, R. & C. L. Griffiths. 1988. An energy budget for the South African abalone Halious inidae Linnaeus. / Mull. Stud. 54:43-31. Bayne. B. L.. J. Widdows & R. J. Thompson. 1975. Physiological inter- actions. In; B. L. Bayne. editor. Marine mussels, their ecology and physiology. London. UK: Cambridge University Press, pp. 261-292. Bennett. R. & H. Nakada. 1968. Comparative carbohydrate metabolism of marine mollusks. I. The intermediary metabolism of Mytdiis califomi- ciis and Haliotis nifesceihs. Coinp. Biochem. Physiol. 24:787-791. Conover. R. 1966. Assimilation of organic matter by /oopkmkton. Liniiinl. Oceanogr. 11:338-345. Dejours. P. 1988. Respiration in water and air. New York: Elsevier. 179 pp. Donovan. D. A. & T. H. Carefoot. 1998. Effect of activity on energy ABALONE BlOENERGETICS 665 allocation in the northern ahalone Haliotis kamtschatkami (Jonas). / Shell. Res. 17(3):729-736. Hahn, K. O. 1989. Handbook of culture ol ahalone and other marnie gastropods. Boca Raton, FL: CRC Press. 34S pp. Koike, v.. J. Flassch & J. Mazurier. 1979. Biological and ecological stud- ies on the propagation of the ornier Haliolis niheiriilaki Linnaeus II: Influence of food and density on growth of juveniles. Lti Mer 17: 43-52. Lawerence, J. M. 1975. On the relationship of manne plants and sea urchins. Oceanogr. Mar. Bid. .Ann. Rev. 13:213-286. Leighton, D. L., M. J. Byhower. C. J. Kelly, G. N. Hooker & D. E. Morse. 1981 . Acceleration and growth in young green abalone Haliotis fulgens using warmed effluent seawater. / Wartd Marie. Soe. 12:70-180. McBride, S. C, J. M. Lawrence, A. L. Lawrence & T. J. Mulligan. 1998. The effect of protein concentration in prepared feeds on growth, feed- ing rate, organic absorption, and gross assimilation efficiency of the sea urchin Strongylocenlronis franciseanus. J. Shellfish Res. 1 7(5): 1563- 1570. Mgaya, Y. D. 1995. Synopsis of biological data on the European abalone (ormer) Haliolis luberculata Linnaeus 1758 (Gastropoda: Haliotidae). FAO Fisheries Synopsis No. 156. Rome: United Nations Food and Agricultural Organization (FAO). 29 pp. Neori. A.. M. D. Krom, C. E. Boyd. D. Popper, R. Rabinowitch. P. J. Davidson. O. Dvir. D. Zuber. M. Ucko, D. Angel & H. Gordin. 1996. Seaweed biofilters as regulators of water quality in integrated fish- seaweed culture units. .Aquaenhnre 141:183-199, Neori. A.. N, L. Ragg & M. Shpigel. 1998. The integrated culture of seaweed, abalone, fish, and clams in modular intensive land-based systems: II. Performance and nitrogen partitioning within an abalone {Haliotis tiiberentaiu) and macroalgae culture system. .Aqna. Eng. 17: 21.5-2.^9. Peck. L. S., M. B. Culley & M. M. Helm, 1987. A laboratory energy budget for the ormer Haliotis tuherculuta L, / Exp. Mar. Biol. Ecol. 106: 103- 123. Shpigel. M.. A. Neon & .\. Marshall. 1996a. The suitability of several introduced species of abalone (Gastropoda: Haliotidae) for land-based culture with pond-grown seaweed in Israel. Israeli J. Aqua. Baniiilgeh. 48(4): 192-200. Shpigel. M.. A. Marshall. I. Lupatsch. J. P. Mercer & A. Neori. 1996b. Acclimation and propagation of the abalone Haliotis tuberculata in a land-based culture system in Israel. J. World Aqua. Soc. 27(4):435- 442. Shpigel. M.. N. L. Ragg. L Lupatsch & A. Neori. 1999. Protein content determines the nutritional value of the seaweed Ulva lactuca L. for the abalone Haliotis tuberculata L. and H. discus hannai Ino. J. Shellfi.swh Res. 18(l):227-233. Widdows. J. 1989. Physiological measurements. In: B. L. Bayne. D. A. Brown. K. Bums. D. R. Dixon. A. Ivanovici, D. R Livingstone. D. M. Lowe. M. N. Moore. A. R. D. Stebbing & J. Widdows, editors. The effects of stress and pollution on marine animals. New York: Praeger Special Studies, pp. 3-39. Mwnuil of Shellfish Ri:minh. Veil. 20. No. 2. fi67-672. 2001. SPAWNING INDUCTION OF YELLOWFOOT ABALONE, HALIOTIS AUSTRALIS USING CHEMICALS AND GANGLIONIC SUSPENSIONS NURUDDIN M. J. KABIR,' * MIKE F. BARKER,' PHILIP V. MLADENOV,' AND BRIAN E. NIVEN- ' Dcpartmcn! of Marine Science. University of Otago. Dunedin, New Zealaml: 'Department of Mathematics and Statistics. University of Otago. Dunedin, New Zealand ABSTRACT A two-slep experimental approach {in vitro and in vivo) using serotonin (5-hydroxytryptaniine). dopamine (3- hydroxytyraminel, prostaglandin E,. deionized water, filtered seawater. cerebral (CG) and pleural-pedal (PPG) ganglionic suspensions as inducers of the spawning of H. cnislrcilis were tested. A total of 88 laboratory conditioned abalone (« = 6-10 per treatment) with fully ripe gonads were used. Filtered seawater caused spawning in 1009f of females and 67% of males and 50% of females and 259!^ of males injected with serotonin ( 10"" M) spawned. Prostaglandin E, did not induce the release of gametes. Fifty percent of females injected with dopamine ( 10"' M) spawned a small number of eggs, while the males did nol respond. Males did not respond to injection of any kind of ganglia, and 20% of females spawned a few eggs in response to CG and PPG from females. Forty percent of females spawned in response to PPG from males. Injection of deionized water caused no weight changes; whereas, filtered seawater caused a reduction in weight. Pro.staglandin-treated animals gained weight on the second day but lost weight over subsequent days. Only females gained weight in the dopamine treated group, but both males and females gained weight in the serotonin-treated group. With ganglionic injections, males treated with CG or PPG from males gained more weight than their female counterparts, and the females treated with CG or PPG from females gained more weight than males. The increase in mean body weight of animals was followed by a swelling and softening of the ovaries, possibly because of increased water content in the ovaries. It seems likely that uptake of water in the ovary is a physiological precursor to spawning. KEY WORDS: Halioris aiistralis. spawning, ganglionic extracts, other neurotransmitters INTRODUCTION The control of spawning is essential for the etTicienl use of hatchery facilities and to the success of an aqtiactilture operation (Hahn 1989). Artificial induction of spawning allows the produc- tion of larvae over a longer period than would be possible relying solely on natural spontaneous spawning. Several methods of arti- ficial induction of spawning in abalone have been tried over the years, some have been successful and others of limited use. Com- monly used methods to induce broodstocks of abalone to spawn include immersion in ultraviolet (UV) irradiated or hydrogen per- o.xide (H^Ot) treated seawater for several hours (Kikuchi & Uki 1974, Morse et al. 1977). Both of these methods seem to induce spawning by a similar mechanism; that is, generation of hydrop- eroxy free radicals, HOC", or peroxy diradicals, °00°, in the seawater that activate the natural enzymatic synthesis of prostag- landins in the abalone (Morse et al. 1977). although the active molecule is produced differently (Hahn 1989). Prostaglandins, as well as some amines produced by nerve cells, are considered to play an important role in spawning in other molluscs as well (Khotimchenko & Deridovich 1991 ). The yellowfoot abalone. H. aiistralis. is the preferred species for New Zealand abalone farmers who want to produce cocktail size (60-70 mm) abalone for export, because the color of the foot makes them more acceptable on international markets (Moss 1998). The abalone reaches sexual maturity at about 70-90 mm (Poore 1972) and can be conditioned to ripeness in the hatchery by holding at constant temperature and ad libitum feeding (Moss 1998, Kabir et al. 1999). With the abalone aquaculture industry expanding worldwide, the demand for regular production of high- quality larvae is increasing. Existing methods for spawning brood- stock are often unreliable, species specific, and frequently result in *Corresponding author. the spawning of poor quality eggs leading to high larval mortality in cultures (Moss 1998, pers, observ.). Spawning induction of//. inistnilis by the H^O, method resulted in an overall spawning rate of 17% for males and 10% for females (Moss 1998). which is very poor. The present study aimed to investigate the effectiveness of different chemicals, neurotransmitters, and ganglionic extracts on the spawning induction in H. aiistralis. Another objective was to identify a suitable solvent for the preparation of solutions of chemicals or ganglionic homogenates, because there was evidence that the most commonly used solvent, seawater. itself caused spawning in other abalone species (Anon 1990). We also wanted to identify neurotransmitters, which may trigger egg maturation in this species. MATERIALS AND METHODS Abalone Collecliiin and Conditioning Yellowfoot abalone, H. niistralis. (length 84.44 ± 1.07 min. weight 1 1 1.32 ± 4.1.^ gm. mean ± SE) were collected from War- rington Beach in the Blueskin Bay area of the Otago coast of New Zealand and conditioned at 15°C for 150 days (effective accumu- lative temperature a l,400°C-days) in the holding facilities of a commercial abalone farm at Warrington. They were fed an artifi- cial diet. Makara (Proniak Technology, New Zealand), ad lihiiiim and exposed to a 1 2-h dark and 1 2-h light photoperiod. Throughout the conditioning period, gonadal development was evaluated sub- jectively and classified into four levels of gonad index from 0 (unripe) to 3 (gravid) (Ebert & Houk 1984). Most of the abalone reached the gravid stage after the conditioning and were spawnable. Trial I: Oocyte Maturation Assays Samples of gonad, haemocoel fluid, and cerebral (CG) and pleural-pedal (PPG) ganglia from ripe, unspawned, and spawned 667 668 Kabir et al. males and females of H. aitstndis were collected, frozen in liquid nitrogen, and stored at -80'C. Ganglia were warmed slowly and homogenized on ice in 1.0 mL of 0.2 |xiii filtered seawater per 10 ganglia using a glass-to-glass tissue homogeniser. Gonads were also homogenized, and the homogenates were centrifuged (8,000 rpm, 20 min), and the supernatant was used for the assay. One fully gravid female from the conditioned stock was dis- sected, gonad removed, and macerated. Small pieces of gonad, together with free oocytes were placed in wells of a tissue culture dish and exposed to the different treatments (different gonad e.x- tracts, hemocoel fluids, cerebral and pleural-pedal ganglionic ex- tracts, hydrogen peroxide, fresh seawater, and deionized water) with three replicates per treatment. Oocytes were monitored con- tinuously using an inverted microscope for any evidence of the occurrence of germinal vesicle breakdown (GVBD). extrusion of polar bodies, and change in shape. Trial 2: Effects of Freshwater and Seawater Injection on Induction of Spawning Twenty-four gravid abalone (3 males and 3 females per treat- ment) were kept individually in 2.0-L containers with individual water flow. One group was not treated and used as a control; whereas, other groups were: ( I ) sham Injected everyday for 3 days; (2) injected with filtered (0.2 (xm) seawater; and (3) injected with deionized water. The latter two treatments were injected once per day over 3 days ( I niL/abalone) into the hemocoel posterior to the head region using a TERUMO® needle 26G x i,:'" (0.45 x 1 3 mm). Trial 3: Effects of Serotonin, Dopamine, and Prostaglandin E, on Induction of Spawning Twenty-four gravid abalone (4 males and 4 females per treat- ment) were kept individually in 2.0-L containers with individual water flow and monitored as described above. Serotonin (5- hydroxytryptamine). dopamine (3-hydroxytyramine), and prosta- glandin E, (Sigma Chemicals. St. Louis, MO) at a concentration of 10"' M for serotonin and dopamine, and 10"^ M for prostaglandin En were injected once per day over 3 days ( 1 mL/abalone) into the hemocoel posterior to the head region. These doses were .selected based on the reports in other molluscs (Martinez et al. 1996). Solutions were made using deionized water as a solvent. Trial 4: Effects of Different Ganglionic Suspensions on Induction of Spawning Forty fully gravid abalone (5 males and 5 females for each treatment group) were used and maintained as described above. Each treatment group was injected with: { I ) cerebral ganglia from males (CGM); (2) cerebral ganglia from females (CGF); (3) pleu- ral-pedal ganglia from males (PPGM); or (4) pleural-pedal ganglia from females (PPGF). Ganglia were collected from the fully ripe wild yellowfoot abalone and were frozen in liquid nitrogen and stored at -80°C. Before use. the ganglion were thawed and ho- mogenized on ice with deionized water in a glass-to-glass tissue homogenizer at a concentration of one ganglia per 0.1 mL of solution. The homogenates were injected once a day over 3 days, at a dose of 0. 1 niL per abalone. This dose was selected after Yahala (1973). where he diluted one ganglion per 0.1 niL and injected into the abalone. Spawning activity was monitored continuously, and the weight of the abalone was measured at 24-h intervals. The response was recorded as positive when gametes were released. The quality of the gametes was monitored microscopically. Handling of the Abalone All the abalone used for the trials were subjected to gentle detachment from the tank, inversion on blotting paper to drain branchial water, weighing, and return to the respective containers. The weight for each abalone was recorded at the start and every 24 h up to the end of the experiment. The weight of the abalone was standardized considering the initial weight for the treatment group as 100%. Statistical Analyses The spawning results are treated as a logistic regression with the treatment and sex as factors using the regression model: pro- portion of spawning = exp (y) / |l-t- exp (y)] + binomial error, where y = constant -i- treatment effect + sex effect. It assumes that the probability of spawning for individual abalone is constant. The models ai-e found to fit the data adequately, and the results are considered significantly different if P < 0.05. RESULTS Trial I: Oocyte Maturation Assays The oocytes were irregular in shape when placed in the wells but became spherical within 20 min. Afterward, no change in shape, motile response, or release of polar bodies was observed for any of the treatments and replications. Trial 2: Effects of Freshwater and Seawater Injection on Induction of Spawning Filtered seawater caused spawning in all females and two out of three males. One female spawned in the control group. There was no spawning in other treatment groups (Fig. 1). Analysis of devi- ance (Table I ) indicated that treatments have a significant effect on spawning (P < 0.05) with the spawning by the filtered seawater differing from others. The sex had no significant effect [P > 0.05). Filtered seawater resulted in a loss of body weight; whereas, deion- ized water caused no weight change. Based on this result, deion- ized water was used as a solvent for further preparation of chemi- cals and ganglionic suspensions. Trial .?; Effects of Serotonin. Dopamine, and Prostaglandin £, on Induction of Spawning Serotonin-injected animals gained weight over time, and two females and one male spawned (Fig. 2). Two out of four females treated with dopamine spawned a small number of eggs and gained weight; whereas, no males responded and lost weight. Prostaglan- din E, did not induce the release of gametes, and the treated abalone gained weight on the second day but lost weight over subsequent days. Analysis of deviance (Table I ) indicated that treatments have a significant effect on the spawning (P < 0.05) with the spawning by serotonin differing from others. The sex had no significant effect (P > 0.05). Trial 4: Effects of Different Ganglionic Suspensions on Induction of Spawning Only one out of five females spawned a few eggs in response to CGF and PPGF. and two out of five females spawned with Spawning Induction of Yellowfoot Abalone 669 X o^ '53 ? >-. T3 O CD 106 104 102 100 9S 9[. - 94- 92- Control (tiol Ircaled) "'' 104- Control 102 - (sliam inj ectcd) 100 - €>«=-—. T ^..--S 9S - ^^~-^ --X~- —y' T 96- 94 - i —-6 — i 0 1 1 1 1 1 104 - Filtered seawalcr 102 - 100- T T X 98- ^ ^ ~^ — 4 96- 94- i =f=i 9? 1 1 1 1 u Dc-ioniscd water 1- 0 1 1 1 1 Days Figure 1. Changes in the mean body weight (± SE) (from a standardized initial mean weight of 100%) and cumulative spawning records of H. aiistralis «ith different treatments. PPGM. Abalone injected with CGM did not spawn (Fig. 3). Males treated with CGM and PPGM gained more weight than their fe- male counterparts, and females treated with CGF and PPGF gained more weight than males. Analysis of deviance (Table 1 ) indicated that treatments have no significant effect on the spawning (P > 0.05). but the sex had significant effect {P < 0.05), with females differing significantly from the males. DISCUSSION Water temperature is considered as the main exogenous factor that regulates the reproductive cycle in many marine gastropods (Webber 1977. Hahn 19S91. The process used to induce gonad ripeness and spawnability in adult abalone requires a number of exogenous and endogenous factors. The most common method to TABLE 1. Analysis of deviance for the spawning experiments in H. aiislralis using treatments and .sex as factors. Goodness of Fit Chi Square Test for Term Residual Residual Trial Number Source Deviance DF P Deviance DF P Trial2 Constant 19..V'>40 7 0.0071 +Treatment 3. 1 154 4 0.5289 16.1786 3 0.0010* +Sex O.OOO.'i 3 1.0000 3. 1 749 I 0.0748 Constant 7.4817 5 0.1872 +Treatment 0.5413 3 0.9097 6.9404 2 0.03 1 1 * Trial 3 +Sex Constant 0.0001 9.268.5 7 1 .0000 0.2339 0.5412 1 0.4619 +Treatmenl 6.27.14 4 0.1796 2.9951 3 0.3924 Trial4 4-Sex 0.0003 3 1 .0000 6.273 1 1 0.0123* Nonsignificant P value for ttie "goodness of fit" confirms the model fits adequately. "Chi square test for term" indicates the eftecl of treatments and sex on spawning and is considered significantly different if P < 0.05 (marked with asterisks). 670 Kabir et al. Prostaglandin E^ c o E 3 Dopamine A small number of eggs spawned from two females 08 - Serotonin T^ 1 W J 00 - 96 .: 1 1 Figure 2. Changes in the mean body weight (± SEl (from a standardized initial mean weight as 100%) and cumulative spawning records of H. aiislralis with different chemicals. bring abalone into a ripe condition for breeding is to provide adults "with good living conditions (tailored to the ecology of the par- ticular species) and ad Ubitiiin feeding" (Hahn 1989). The effect of temperature on abalone reproduction has been quantified in H. discus hannai (Uki & Kikuchi 1984) and H. australis (Moss 1998. Kabir et al. 1999). Our previous study (Kabir et al. 1999) showed that 21 wks of conditioning at 15'C (EAT > 1.400°C-days) with ad lihitiiin feeding induces gonad ripeness into a fully mature stage. Moss (1998) al.so reported 21 wks of conditioning at 15°C temperature was required for full gonad growth and initiation of spawning in H. aitstralis. but this time may be reduced to 14 wks if the animals are held at 18"C water temperature. For the present experiment, the abalone were reared at 15°C for 21 wks following the result of Kabir et al. (1999). After conditioning, the gonad was bulging across the shell with a rounded tip at the end of a conical appendage; the ovary was brown to blue-gray, and the testis was milky white to creamy in color. Oocytes treated with either ganglionic extract, hemocoel fluid, gonad extracts, or chemicals did not produce any motile responses or show any release of polar bodies. Saleuddin et al. ( 1983) re- ported that the freshly isolated oocytes from the ovotestis of Helix apsersa (Mollusca) showed amoeboid movement when treated in vitro with the extract of cerebral ganglia. They postulated that this form of induced motility might be associated with ovulation. On the other hand, Coggeshall (1972) suggested that ovulation in Aplysia is caused by the muscle contaction brought about by egg- laying hormone. Mature oocytes in the abalone do not undergo germinal vesicle breakdown until spawning is induced, and the oocytes are squeezed out of the gonad by the rapid contraction of the columnar muscle (Hahn I99(J). In the present study, the inac- tivity of the oocytes may be attributable to the failure of oocytes to have matured sufficiently and they were, therefore, not ready for ovulation, or the isolation procedure caused invisible damage (chemical) to apparently healthy looking oocytes (Saleuddin et al. 198.'!). Another possibility is that the donor H. australis (from where the gonad, ganglia, and hemocoel fluid were collected) were not sufficiently mature and lacked the required amount of hormone to induce the oocytes. The exact mechanism of ovulation in aba- lone requires further investigation. A preliminary study showed three out of five blackfoot aba- lone, Haliotis iris, spawned when a filtered seawater injection was used as a control. This prompted us to investigate the use of seawater as a control or its use as a solvent to homogenize different ganglia or other chemical solutions for the spawning experiment of H. australis. In the present experiment, injection of filtered sea- water into the hemocoel caused spawning that is significantly dif- ferent from other treatment groups {P < 0.05). Injection of 3.7% calcium chloride, sterilized or filtered seawater into the head and foot of H. discus hamuli has also been observed to cause spawning (Anon. 1990). Yahata (1973) did not observe any spawning when 0.6% NaCl solution was injected into Nordotis discus, but it re- sulted in weight loss in the abalone. Thus, it seems that seawater or other salt solutions often induce spawning in H. australis and other abalone perhaps by stimulating the nervous systems. There- fore, we must be cautious in selecting a solvent to homogenize ganglia or to prepare other chemical solutions for spawning ex- periments with abalone. The use of neurotransmitters and other chemicals in induction of spawning in molluscs (Velez et al. 1990. Ram et al. 1993, Martinez et al. 1996) and of abalone (Morse et al. 1977) has been reported. Serotonin has been shown to be one of the most effective spawning inducers in some molluscs (surf clam Hirai et al. 1988, scallop Matsotani & Nomura 1982, China clam Alcazar et al. 1987, and giant clam Braley 1985). Therefore, the release of se- rotonin may play a role in abalone spawning. Prostaglandins (PG) have been suggested to be involved in the release of gametes Spawning Induction of Yellowj-oot Abalone 671 an o : CG - Female KM - /. , T ■ ^J^ }II3I~T 100 - 0.05) Injections of crude homogenates of the cerebral ganglia into ripe females did not induce full spawning in ripe H. discus hannai (Yahata 1973). In their experiment, two out of five abalone in- jected with CG homogenates spaw ned a small amount of clustered eggs; however, there was a considerable gain in the mean body weight from an increase in water uptake without any noticeable change in the ovary. Hahn (1989) postulated that A-cells from cerebral ganglia of probably regulate vitellogenesis in females, but this ganglion does not contain a factor that is involved in the induction of spawning. Yahata ( 1973) also found that the injection of pleural-pedal and visceral ganglia crude homogenate into fe- male H. discus liaivnii caused a large swelling and softening of the ovary and spawnings after two injections. Therefore, these ganglia may produce and release factors that induce spawning in abalone. Histological examination of the PPG showed several cell types producing abundant neurosecretory material that seemed to vary in relation to the reproductive cycle (Hahn 1990). The handling stresses (such as detaching, injection, and weighing) of the abalone on spawning were not evaluated in the present experiment. A recent .study showed that handling for weighing inay produce 7 to 17% weight reduction in adults of H. iris (Ragg et al. 2000). An 672 Kabir et al. increase in body weiglit was observed in the present experiment witii ganglionic suspensions and other neurotransinitters. These increases in the body weight may be caused by the increased uptatce of water in the ovary, which indicates that the uptake of water by the gonad may be essential to bring about spawning (Yahala 1973). The weal< or almost no response of the ganglionic extracts in the spawning of H. austndis in the present experiment raises a number of issues. First, the maturity of the animals from where the ganglia were collected must be examined. The abalone were col- lected from Warrington Beach, and those with bulging gonad above or near the shell edge were taken. Abalone with such bulg- ing gonads are considered to be fully ripe (Ebert & Houk 1984). but this does not provide information on the oocyte maturation stage. It could be possible that in the wild population, abalone with fully developed gonads may not necessarily be spavvnable and lack factors in the ganglia that are responsible for spawning. Second. the abalone that were injected came from the conditioned stock ( 1.400°C-days). It has been shown histologically that after such conditioning period oocytes become fully grown, and the abalone are spaw liable (Kabir et al. 1999). Thus, the abalone used in this experiment were well gravid. Third, the dose of the ganglia (i.e., 1 ganglion per 0. 1 mL = I injection) was selected following the protocol of Yahata ( 1973). This dose may not be effective for this species. Therefore, it requires further investigation with the gan- glia to be collected from conditioned individuals and employed with different doses and combinations. ACKNOWLEDGMENTS We are grateful to Mr. Roger Bartlett, Executive Director of Abalone South Ltd., New Zealand for providing space on his farm and technical support. We thank two anonymous referees for con- structive review of the manuscript. Thanks to Mr. Taka Kai, who translated the Japanese papers. This research was undertaken as part of a Graduate Research in Industry Fellowship funded by the Foundation for Science, Research and Technology in New Zealand. LITERATURE CITED Anonymous. 1990. Training manual on artificial breeding of abalone {Hulkitis discus haimai) in Korea DPR. Training Manual 7. Shallow Seafarming Research Institute in Kosong, DPR, Korea. 107 pp. Alcazar, S. N.. E. P. Soils & A. C. Alcala. 1987. Serotonin-induced spawn- ing and larval rearing of the china clam. Hippopiis ponellaiuis Rose- water (Bivalvia: Tridacnidae). Aqiiacutlure 66:359-368. Braley, R. D. 1985. Serotonin-induced spawning in giant clams (Bivalvia: Tridacnidae). Aquacuhure 47:321-325. Coggeshall. R. E. 1972. The muscle cells of follicle of the ovotestis in Aplysia as the probable target organ for bag cell extract. Am. Zool. 12:521-523. Ebert, E. E. & J. L. Houk. 1484. Elements and innovations in the cultiva- tion of red abalone, Haliotis nifescens. Aqiiacultiire 39:375-392. Hahn. K. O. 1989. CRC handbook of culture of abalone and other marine gastropods. Boca Raton. FL: CRC Press. 348 pp. Hahn, K. O. 1990. Neurosecretory staining in the cerebral and pleural- pedal ganglia of Haliotis discus hannai and Trochus niloticiis. and its relationship to reproduction. Ph.D. Dissertation. University of Califor- nia at Davis, Hirai, S. T., T. Kishimoto. A. L. Kadam. H. Kanatani & S. S. Koide. 1988. Induction of spawning and oocyte maturation by 5-hydroxytryptamine in the surf clam. J. Exp. Zool. 245:318-321. Kabir. N. M. J., P. V. Mladenov & M. Barker. 1999. Effect of different temperatures on gonadal development of black and yellowfoot paua. Haliotis iris and H. australis. Book of Abstracts: World Aquaculture Society Conference, 26 April-2 May. 1999. Sydney, Australia, p. 853. Kikuchi. S. & N. Uki. 1974. Technical study on artificial spawning of abalone. genus Haliolis. II. Effect of irradiated sea water with ultra- violet rays on inducing to spawn. Bull. Tolniku Reg. Fish. Res. Luh. 33:79-86. Khotimchenko. Y. S. & I. I. Derodovich. 1991. Monoaminergic and cho- linergic mechanisms of reproduction control in marine bivalve mol- lusks and echinoderms: A review. Comp. Biochem. Physiol. 100C:3I 1- 317. Kunigelis. S. C. & A. S. M. Saleuddin. 1986. Reproduction in the fresh- water gastropod, Helisoma: Involvement of prostaglandin in egg pro- duction. Int. ./. Invert. Reprod. Develop. 10:159-167, Malsutani. T. & T. Nomura. 1982. Induction of spawning by serolonm in the scallop, Patinopeclen yessoensis (Jay). Mar. Biol. Lett. 3:353-358. Maninez, G., C. Garrote, L. Metlifogo, H. Perez & E. Uribe. 1996. Monoamines and prostaglandin E, as inducers of the scallop. Ar- gopecten piirpiiratiis Lamarck. J. Shellfish Res. 15:243-249. Morse. D. E.. H. Duncan. N. Hooker & A. Morse. 1977. Hydrogen per- oxide induces spawning in mollusks with activation of prostaglandin endoperoxide synthetase. Science 196:298-300. Moss, G. A. 1998. Effect of temperature on the breeding cycle and spawn- ing success of the New Zealand abalone, Haliotis australis. N.Z. J. Mar Fres. Res. 32:139-146. Poore. G. C. B. 1972. Ecology of New Zealand Abalone. Haliotis species; 2. Seasonal and diurnal movement. NZ. J. Mar. Fres. Res. 6:246-258. Ragg, N. L. C, H. H. Tylor & J. Behrens. 2000. Stress and weight loss associated. Haliotis iris. Abstracts. IV. International Abalone Sympo- sium. Cape Town. South Africa, February 2000. p.42— 43. Ram, J. L., G. W. Crawford, J. U. Walker. J. J. Mojares, N, Patel, P, P. Fong & K, Kyuzuka, 1993. Spawning in the zebra mussel (Dreissena polymorpha): Activation by internal or external application of seroto- nin. J. Exp. Zool. 265:587-598. Saleuddin. A. S. M.. C. L. Farrell & L. Gomot. 1983. Brain extracts cause amoeboid movement in vitro in oocyes in Helix aspersa (Mollusca). Int. J. Invert. Reprod. 6:31-31. Uki. N. & S. Kikuchi. 1984. Regulation of maturation and spawning of an abalone, Haliotis (Gastropoda) by external environmental factors. Aquaculture 39:247-261. Velez. A., A. Alifa & O. Azuaje. 1990. Induction of spawning by tem- perature and serotonin in the hermaphoditic tropical scallop. Pecten ziczac. Aquaculture 84:307-313. Webber. H. H. 1977. Gastropoda; Prosobranchia. In: A. C. Giese & J. S. Pearse, editors. Reproduction of marine invertebrates, vol. 4. San Di- ego, CA: Academic Press, pp. 1-97. Yahata, T. 1973. Induced spawning of abalone (Nordolis discus Reeve) injected with ganglionic suspensions. Bull. Jpn. Soc. Sci. Fish. 39: 1117-1122. Journal ol Shellfish Research. Vol. 20. No. 2, hiy-bll. 2001. REPRODUCTIVE PERFORMANCE INDICES BASED ON PHYSICAL CHARACTERISTICS OF FEMALE BLACKLIP ABALONE HALIOTIS RUBRA L. MAGDALENA LITAAY* AND SENA S. DE SILVA Scliool of Ecology & Enviroiinwiit. Dcuklii University. 3280 Victoria. Australia ABSTRACT Selection of abalone broodstock from the wild is often based on external appearances. The criteria used are size, color, and shape of the gonad. However, animals selected on external appearance are known to vary in egg fertilizability, hatching rates, and larval survival. The present study was instigated to develop easily useable indices, based on physical characteristics, for assessing the potential reproductive performance of female abalone. Analysis was conducted on wild-caught, artificially propagated blacklip abalone Haliotis rubra L.. obtained from Australian coastal waters i.\42''\y: 38°2rS). Females were spawned using a combination of the desiccation and the UV irradiated methods and shell characteristics: length (SL). width (SW). height (SH). and total weight (W) of the spawners were determined. The fertilized eggs from each spawn were incubated, hatched, and larvae were reared separately in a flow-through system. The criteria used for assessing reproductive performance were fecundity, fertilizability. hatchability. and pre- settlenient survival. The results show that fertilizability was positively correlated to hatchability and larval survival (P < 0.01). Furthermore, shell physical characteristics and weight can be used as predictors of reproductive performance. Accordingly, a number of polynomial regressions incorporating SL. SW. SH, W. and egg characteristics to fertilizability. and SL and W to fecundity, were developed. KEY WORDS: blacklip abalone. Haliotis rubra, broodstock assessment, shell characteristics, reproductive perfonnance INTRODUCTION Abalone {Haliinis spp.) is commercially exploited for its valu- able meal and shell. Long-temi exploitation of this group has resulted in the depletion of wild stocks in many areas of the world. To fulfill the high demand and because abalone take a long time to reach marketable size in the wild, culture of abalone has become a viable alternative to fishing wild stocks. Austi-alia is one of the main exporters of wild-caught abalone, contributing about 60% to world production. It is now in a prime position to become a world leader in the production and expoit of cultured abalone (Flemming & Hone 1999). In abalone culture, wild-caught broodstock are used exten- sively. Before this study, the suitability and selection of brood- stock from such wildstock has been based on the external appear- ances of mature animals. The main criteria used in such selection are size, color, and shape of the gonad. However, animals selected based on these criteria are known to vary widely in reproductive performance. The present study attempts to develop suitable indices for broodstock assessment based on the physical characteristics of female blacklip abalone. It is expected that this will provide a method for performance estimation and will contribute to knowl- edge on broodstock selection in abalone aquaculture. Accordingly, this paper is based on the results of the spawning of 8 and 1 1 female blacklip abalone. Haliotis rubra L. in the spawning season of 1997 to 1998 and 1998 to 1999. respectively. Furthermore, in the 1999 to 2000 spawning season, the results of five females spawned were used for validation of the models developed based on observations in the previous spawning seasons. MATERIALS AND METHODS Spawning Broodstock were obtained from coastal waters (l42'l.'i'E; 38°2rS) during the spawning seasons of 1997 to 1999 and 1998 to 1999. Spawning was conducted during the routine spawning ♦Corresponding author. E-mail; magdalen@deakin.edu.au operations of Southern Ocean Mariculture. Port Fairy. Victoria. Australia. To reduce stress, except for total v,'eight. all physical characteristics of females used in the present studies were recorded after spawning. Spawning was triggered by using a combination of desiccation and the UV iiradiated methods (Kikuchi & Uki 1974). Spawning was conducted using a flow-through system (velocity of 100 mL min^' ). In this study, to eliminate the influence of males, sperm from a number of males were pooled together, and females were spawned individually. Spawned eggs and sperm were immediately separated from the spawners with a siphon, washed with 1 jx of UV-treated seawater of 35% salinity, and the spawned eggs were concentrated into a known volume of clean seawater. The water containing the eggs was gently stirred to ensure a homogenous distribution of eggs through water column, and three subsamples were taken for esti- mating fecundity. Artificial fertilization was effected by mixing known numbers of newly released eggs with appropriate numbers of fiesh sperm at an approximate ratio of 5 sperm/egg. Incubated eggs and larvae from each female were kept separately. The hatch- ing stage was run using a static system, and hatched larvae were transferred into rearing containers soon after hatching. Larval rear- ing was done using a flow-through system (velocity of 80 mL min"'). for 3 days with the extent of presettled larvae being de- termined by formation of the third tubule on the cephalic tentacle (Hahn 1989). Hatchery procedures followed that described else- where (Hone et al. 1997). Parameters Such parameters as physical characteristics of adults, number of eggs (fecundity), diameter of eggs, fertilizability. hatchability, and presetllement survival were recorded for 19 female blacklip abalone. Physical features including shell length (SL), shell width (SW). shell height (SH). and total weight (W) of each female spawner were recorded. Shell measurements were inade with a Vernier caliper (to the nearest O.O.'i mm). Conventional measure- ments of the shell used in this study are shown in Figure 1. Weight was recorded using a scale (to the nearest 0.01 g), and egg diam- eter was measured using an eyepiece graticule with a compound 673 674 LlTAAY AND DB SlLVA Figure 1. Conventional shell measurement of length (L), width (VVl. and height (H). (Mgaya & Mercer 1994). microscope (to the nearest 0.5 [x). At least 150 eggs from each spawned female were measured, and means for individual females were used in regression analysis. Fertilizability was estimated by determining the number of fertilized eggs in a subsample of 1 mL taken after mixing eggs and sperm. For each spawner, five such subsamples were used, and the mean was determined. Similarly, hatchability and larval survival were estimated. Fecundity is the total number of eggs produced by an individual female. Fertilizability {%] Hatchabilitv C/r) Larval survival (%) Statistical Aiialvsis Total number of fertilized eggs Total number of eggs produced Total number of hatched larvae Total number of fertilized eggs Total number of survived larvae Total number of hatched larvae X 100 X 100 X 100 Scatter plots were used to observe the distribution of the data. Statistical relationships between physical characteristics and each of the reproductive performance indicators were explored using the software package Minitab version 12.1. The relationships then TABLE 1. Physical characteristics of female abalone spawned (19) and the mean (±SE) of fecundity, fertilizability. hatchabilitv. and larval survival from the spawning seasons 1997/98 and 1998/99. Parameters Mean (SEl Range .Shell length (nun) Shell width (mm) Shell height (mm) Total weight (g) Fecundity (iC*) Fertilizability (%) Hatchability (%) Larval survival {%) 128.97(1.70) 103.64(1.80) 40.35(1.52) 383.81 (19.51) 3.51 (0.43) 91.36(1.73) 87.54(1.21) 7941 (1.69) 120.05-141.50 90.9.5-121.00 29.65-52.35 274.37-598.52 0.54-5.92 70.47-97.30 75.73-96.87 66.00-88.41 tested for. between parameters, were linear, curvilinear and the second- and third-order polynomial. RESULTS Female Physical Fealiires and Reproductive Performance The legal size limit for wild-caught Victorian blacklip abalone is 120 mm. The mean (± SE). the range of the different physical features, and reproductive performance criteria of the I') female abalone used in the study are given in Table 1. Relationships of Shell Characteristics to Reproductive Criteria The potential statistical relationship of the shell characteristics to the female reproductive criteria; namely, fecundity, fertilizabil- ity, hatchability, and larval survival, were explored. The resulting relationship, in the form of a correlation matrix, is given in Table 2. It was evident that only shell width and total weight were positively coirelated to hatchability. None of the other shell char- acteristics, when individually considered, were correlated to any of the reproductive criteria in the study. Similarly, multiple regres- sion analysis incorporating shell characteristics did not result in a statistically significant relationship to performance predictors. It is also clear from Table 2 that fertilizability was positively correlated to hatchability (r = 0.64. P < 0.01 ) and to larval sur- vival (r = 0.65, P < O.OI). Also, hatchability was positively and linearly coiTelated to larval survival (r = 0.55. P < 0.05) (Table 2). TABLE 2. Correlation matrix of relationships of female blacklip abalone physical characteristics to the different indices of reproductive performances (.'V= 191. SL SW SH VV Fc Ft Ht Sv SL SW SH W Fc Ft Ht Sv 1.00 0.82* 1.00 0.50* 0.54* 1.00 0.81** 0.91** 0.47* 1.00 0.06 NS 0.06 NS 0.26 NS 0.04 NS 1.00 0.08 NS 0.44 NS 0. 1 7 NS 0.25 NS 0.52* 0.39 NS -0.28 NS 0.02 NS 0.09 NS 0.14 NS 0.46* 0.26 NS -0.15 NS -0.28 NS -0.03 NS 1.00 0.64** 0.65** 1.00 0.55* 1.00 hatchabilitv: SH = = shi II heiaht; Sv = = larval survival; Significance level is indicated as P < 0.01 (**). P < 0.05 ( ■ ) or not significant (NS). Abbreviations: SL = shell length; Ft = fertili/ability; W = total weight; SW = shell width; Ht Fc = fecundity. Reproductive Indices for Female Blacklip Abalone 675 TABLE 3. Correlation nuilrix. paiiiul cot.'l'tU'i(.'nts of llii' comhinations of physical characteri.stics to the indice.s of reproductive performance of female blackUp abalone Ui = 19). Larval Fecundity Kertilizubility Hatchubilit> Survival SL/SW 0.82*** 0.83*** 0.77*** 0.83*** SL/SH O..'i0* 0.55* 0.55* 0.50* SLAV 0.81*** 0.81*** 0.76*** 0.81*** SW/SH 0.54* 0.65** 0.62** 0.55* SWAV 0.91*** 0.91*** 0.88*** 0.90*** SHAV 0.47* 0.53* 0.51* 0.46* Significance level is indicated us P < 0.(K)1 (***i P < 0.01 (* For legends, see Table 2. 'lP<0.05i*). Hatchability (Ht) can be predicted by using performance indicator fertilizability (Ft), and its relationship is best described by the equation. Hi = 3.5.1 + 0.575 Ft (R- = 0.89. P < 0.001). The relationship between feitilizabiiity (Ft) and larval survival (Sv| is described by the equation Sv = 7.4 + 0.802 Ft (R- = 0.73. P < 0.001). Furthermore, hatchability (Ht) can also be applied to esti- mate survival (Sv), and their relationship is described by the equa- tion, Sv = -0.272 + 1.22 Ht (R- = 0.79. P < 0.001). The regression relationships between different combinations of shell characteristics and reproductive performance indicators are given in Table 3. Partial correlation coefficients indicated that combination of SL/SW and SWAV are strongly con-elated to fe- cundity, fertilizability. hatchability. and larval survival (in all cases r > 0.77, P < 0.001 ). A combination of SW/SH was also found to be significantly correlated to fertilizability (r = 0.65, P < 0.01) and to hatching rate (r = 0.62, P < 0.01 ). but was less significantly correlated to fecundity (r = 0.54. P < 0.05) and to larval survival (r = 0.55, P<0.05). However, none of the above relationships were linear. Shell characteristic; length, height, and width bear a quadratic relation- ship to fertilizability (Table 4). On the other hand, relationships between weight to feilili/ability are best described by a cubic function. The quantitative relationships between female physical characteristics (SL, SH, SW, W) and fertilizability, and the rela- tionship between SL and W and fecundity are shown in Table 4. The quantitative relationships between shell length and total weight to fecundity are illustrated in Figure 2. The fecundity (Fc) of blacklip abalone was also related quadratically to length (SL) and total weight (W). Their relationships are best described by the following equations. Fc = -622.238 + 9.492 SL - 0.0359 SL^R" = 0.59. P < 0.01 ) Fc = -26.628 + 0. 147 W - 1 .(i7E-04 W-(R- = 0.66, P < 0.01 ) Egg Quality and Reprudiiclive Performances Egg quality in terms of yolk diameter and the ratio between yolk and total diameter could also be used as a predictor of fer- tilization (Figs. 3, 4). The relationship between yolk diameter (YD) and fertilizablity (Ft) is best described by the equation Ft = -9028.77 + 97.557 YD- 0.261 YD" (R- = 0.92. P < 0.001). In addition, fertilizability (Ft) was correlated to the ratio of yolk diameter and total egg diameter (TED), the relationship being Ft = - 2862.67 -h 6786.67 TED-3891.62 TED= (R= = 0.77. P< 0.001). V alidation of Indices Validation of the present indices was done using the data from the 1999 to 2000 spawning seasons. The results of these particular spawnings, predicted and observed parameters are given in Table 5. The mean predicted fertilizability for the five females differed from the observed value by 5.7%. In all instances except in one female (-1.8'7r) the predicted fertilizability was higher (6.7 ± 2.2'/r). The predicted fecundity, however, was more variable, particularly with respect to the estimation based on one feinale (Table 5). DISCUSSION In abalone culture, selection of broodstock is mainly based on gonad appearance. The mantle and foot are manipulated to expose the gonad, which is markedly swollen, blunt and rounded at the tip indicating gravidity (Grant 1989). Abalone are judged to be ready for induced spawning by the amount of swelling of the gonad. Without a careful handling procedure, brood.stock selection ba.sed on this technique is bound to stress the animals. Therefore, a quick method of assessment with minimal disturbance is desirable. Present results demonstrate that female physical characteristics can be used for predicting the spawning performance. Regression analysis is one of the most widely used statistical TABLE 4. Results of regression analysis of shell and egg cbaracterlstlcs and total weigbt to fertilizability and fecundity ( 1st order equation, V bX; 2nd order equation, V = a -i- b,X + b,\- and 3rd order equation, Y = a + b,X + b,X" = b,X'). Independent Dependent Variable (X) Variable (Y) SL Ft SW Ft SH Ft W Ft YD Ft TED Ft Ratio YDATED Ft SL Fc W Fc b., -1344.890 -23.515 -355.011 7.858 36.609 3.482 923.485 -7.026 -9028.770 97.557 -5495.800 52.797 -2862.67 6786.67 -622.238 9.492 -26.628 0.147 -0.096 -0.034 -0.05 1 0.019 -0.261 -0.124 -3891.62 -0.035 -1.70E-04 -1.8E-05 12 0.86 0.000 11 0.66 0.013 12 0.87 0.000 14 0.81 0.000 11 0.92 0.000 12 0.71 0.003 13 0.77 0.000 15 0..59 0.004 14 0.66 0.002 YD = yolk diameter (|xm); TED = total egg diameter; Other legend, see Table 2. 676 LlTAAY AND DE SiLVA tools, because it provides a simple method for establishing a func- tional relationship between variables (Chatterjee & Price \9n). Linear relationship has been commonly used in performance stud- ies on other species (Shepherd et al. 1992, Estay et al. 1999). In the present study, a polynomial regression model successfully esti- mated the reproductive performance of female blacklip abalone. As indicated here, most of the relationships are best fitted with nonlinear models. Conelation analyses showed that some combinations of shell features correlated well with performance (Table 3). As indicated by correlation coefficients, we would expect bigger animals and ones with a flattened shell to be a good indication for performance estimation. The same trend was also observed in bigger and heavi- er animals. This conforms with other observations that larvae from blacklip abalone broodstock having bigger and wider sizes usually show a faster growth rate (SOM. pers. comm.). Fecundity is one important performance criterion. The relation- ship of fecundity to size (length or weight) has important conse- quences for both fisheries management and aquaculture. Determi- nations of the size limit, which allows the harvest of immature animals, in abalone fisheries, will affect egg production. In aqua- culture, however, fecundity plays a role in routine hatchery opera- tions. Estimations of egg production could allow hatchery opera- tions to be optimized. In blacklip abalone populations, fecundity can be appro.ximated by nonlinear relationships with length and has been described by a power function (Y = aX") by others (McShane 1990, Nash et al. 1994). Exponential increase of fecundity with increasing length has also been reported for other abalone species, HalUnis roei (Wells & Kessing 1989). H. midac (Newan 1967). H. tuheiculata (Hayashi 1980). and H. hicvii^aui (Wells & Mulvay 199.^). In CD N Yolk (pm) Figure 3. Quadratic model of mean yolk diameter (X) versus fertil- izabilitv (Y) (R' = 0.92, P < O.tKtl). addition, McShane (1990) and Nash et al. (1994) suggested that fecundity is size dependent rather than age dependent. In a separate study, on the other hand. Prince et al.( 1988) argued that fecundity of H. rubra was closely related to age rather than length. The present study indicates that fecundity varied widely, even among a small range of sizes (Table 1 ), and its relationship to length and weight was best described by a second-order polynomial function. As shown by the quadratic model (Fig. 2), fecundity increases with increasing size, but, decreased as animals grew beyond a certain size. This change may be brought about by a limited release of eggs by older and bigger animals, or simply because the number of eggs produced decreases as the animals grow beyond a certain size. It would be interesting to investigate whether or not there is a tendency for bigger animals to spawn early (or late) in the season and vice versa: a trend that may be of use in improving hatchery production. The egg yolk diameter and the ratio between yolk and total diameter seemed to be another method of predicting egg fertiliz- ability (Figs. 3. 4). For such nonfeeding larvae, as those of abalone, future spawn success is determined by egg quality (size and bio- chemical content). It has been widely known that egg biochemical content is size-dependent. Consequently, it is obvious that egg diameter is an important key in determining performance (Brooks et al. 1997). Investigations on other invertebrates have shown that small differences in egg size and content might have important biological consequences (George 1999). This author found that larval seastars from small eggs tend to grow and develop more Weight (g) Figure 2. Regression plot of shell length versus fecundity (.\) (R" = 0.59, P <0.0I( and weight versus fecundity (15) (R- = 0.66, P < 0.01) Yolk/Total Egg diameter Figure 4. Quadratic model of ratio of yolk and total egg-diameter (X) versus fertilizablity (V) (R" = 0.77, /' < 0.001). Reproductive Indices for Female Blacklip Abalone 677 TABLE 5. Shell chariictiTJstics and corrfspontlin;; pridlclttl \aliK^ for ri'rtill/abilit> and Itcundity. based on the equations in Table 4 and a comparison of the latter to those observed for live females spawned in the 1999/2000 season. For each female, the measurements given in order are shell length, shell width, shell height, and weight. Fecundity predictions are made using only length and weight. Weight Predicted Observed Shell Characteristics Fecundity Fertilizabilitv Fecundity F ertilizabilitv (SL;S\\ ;SH) (mm) (g) (10") t%) (10") ( 9, ) 121.65: 10.'i.95: 39 275 0.89; 0.94 95.0: 93.6; 94.5 U.65 92.3 127.7; 105..'i5; .34.8 285.95 4.13; 1.53 92.8: 93.4; 95.8 4.08 92.3 127; 94.65; 32.7 NA NA 93.7: 82.3; 95.7 NA 91.7 1-30.95: 98.5.^; 33 360.49 4.78; 4.27 88.2; 87.2; 95.8 4.75 86.1 136:37; 101; 34.3? NA NA 76.7; 89.7; 95.8 NA 74.0 NA = not avaiiahle. slowly than those from large eggs. AccoriJing to the present find- ings, bigger egg size di tagged. Treatment 1 Is the tonlrol. Water Quality DO, pH. ammonia, and temperature tor both pulse and chal- lenge exposures were recorded before any alteration to conditions in the tank (T = 0 h). pH and ammonia were then measured after 15-20 min (T = ]5 min). at 8 h (before restoring ambient con- ditions) and the morning following exposure (T = 24 h). Once DO had stabilized, it was checked on average every 30—10 min and flow of N-, adjusted as required. DO. temperature, and pH were recorded for all tanks every 2-3 days throughout the trial. Samples for ammonia analy sis were taken from two tanks on each treatment not being pulsed each week to provide ambient ammonia levels. DO was measured using hand- held meters (WTW Oxi96. TPS WP82-Y or Oxiguard Handy Gamma), which were calibrated daily and checked against saturated seawater and occasional Winkler titrations. The three meters gave similar readings when checked against each other. pH was recorded on a hand-held TPS meter and probe (WP 81 ), calibrated daily in fresh buffers (phosphate at pH 7.00. borate at pH 9.28. after Bruno & Svoronos (1989). Tempera- ture was recorded using the thermistor on the DO meter, which was checked against a calibrated mercury thermometer. All samples for ammonia analysis were collected in acid- washed glassware rinsed with deionised water. Samples were fil- tered through Whatman GF/C filters and frozen in polypropylene Growth Rale Data for Greenlips (mean + SE n=3) 0.8 - Specific °'' ■ [*1 I ' ^ ] r* . Growth Rate O'l 0,2 0 • J 1 1 1 J ■ Microns per Day Specific 0.1 Growth Rate 0.05 ■ Growth Rate Data for BlackJips (mean + SE,n=3) f { f ] -T 1 r- ' 1 ■ 5 ■ 0 ■ J i 1 1 J - . 20 15 10 Microns per Day ■ SGR - L □ SGR-W . Microns per Day Figure 2. Growth rate data from pulse exposures of two species of abalone (mean ± .SE, n = i). Treatments are given in Figure 1. S(;R-L = specific growth rate lor length, ii SGR-\V = specific growth rate for weight. bottles (acid washed, rinsed in deionised water) for subsequent analysis. Total ainmonia was measured within 4-5 wk of the samples being collected, by the method of Grasshoff (1989) but using the salicylate reagent of Bower and Holm-Han.sen (1980). UIA was calculated from pH and temperature using the equation in Bower and Bidwell (1978). Where necessary, samples were di- luted with the sand-filtered seawater such that TA was <0.5 mg L"'. Nitrite was analyzed by the diazotization method of Grasshoff (1989). Influent flow rates were recorded every 5-10 days (mean = ISOOmLmin"'). Challenge Exposure On day 45 of the growth trial, two tanks for each species on treatments 1. 2. 4. and 5 were exposed for 8 h to more .severe challenge conditions (Table 1), as described above, but using a stock solution of 1.33 g L^' of ammonia chloride. Statistical A ualysis The raw data were tested for homogeneity of variance by ex- amination of residual plots generated by JMP 3.2 (SAS Institute). Normality was assessed by observation of the distribution plotted by JMP 3.2. No transformations were required to meet the assump- tions of analysis of variance (ANOVA). as given in Underwood (1981). One-way ANOVA. using the replicate means for each treatment, was used to look for significant treatment effects, and Tukey's HSD was used to compare means. RESULTS Behavioral Observations Under ambient conditions, species-specific behavioral differ- ences were observed. Blacklips clustered in one or two spots in the tank, but greenlip abalone tended to gather in smaller groups dis- persed more evenly over the available surfaces. Although gener- ally inactive, small numbers of blacklip abalone (1-3) were ob- served to be active at any one time during the day; whereas, greenlips were rarely observed to be active during the day. For both species, inactive animals had all tentacles withdrawn and shells clamped tightly onto the substrate. During the pulse expo- sures, no change in behavior was observed until the end of the exposure period, when signs of stress were evident (shells starting to lift off the substrate, tentacles becoming actively extended). These behavioral effects were more pronounced during the fi- nal challenge exposure. In both species, the animals tended to separate from the clusters within 40-60 min of commencing the exposure, but once separated, did not continue moving. Rather, in both species, the front of the shell was lifted off the substrate, and epipodial tentacles were extended. Howe\'er, abalone in this posi- tion were not readily removed from the substrate, and if disturbed, clamped the shell back onto the substrate and withdrew the ten- tacles. Although food consumption was not directly determined by collecting uneaten food, ration was adjusted ad libitum for each tank, and no difference in weight of food fed between treatments was observed. However, in both species, consumption tended to decline slightly the night immediately after exposure to the pulse condition, and the abalone were also observed to be less active than normal. Consumption and foraging activity returned to nor- mal the following night (24-h postexposure). Following the chal- lenge exposure, this reduction in consumption and activity was 682 HiNDRUM ET AL. even more marked, but agiiiii returned to normal the following night. Mortality Over-all mortality was 0.8% for greenlips and I'^c for blacklips during the growth trial. No mortality was observed as a direct result of the pulse or challenge exposures. There was no clinical evidence (such as lack of vigor, lesions, or mortality) of any sec- ondary infection becoming established as a result of the pulse exposure regime. Two tanks of blacklips did show signs of bac- terial infection at different points during the growth trial, with some mortality, although no causative agent could be identified. The other replicate tank on this treatment showed no similar signs, and these were considered unrelated events. These mortalities were excluded from the mortality calctilations. Apart from these events, both species seemed generally healthy throughout the growth trial. Water Quality Table 1 shows the ambient w ater quality for days when DO and ammonia were not manipulated and the alterations achieved during pulses of ammonia and low DO. Table 1 also shows the water quality achieved during the final challenge exposure. Nitrite was always undetectable. The initial and final data collected for each pulse (at T = 0 h and T = 24 h) were always consistent with ambient conditions and. thus, were included in the ambient water quality. Total ammonia was elevated throughout both the pulse and challenge exposure periods (Table 1). pH and UIA declined over the 8-h period, and DO was kept relatively constant by adjusting the gas flow. Temperature increased, to an extent dependent on ambient air temperature, by 1-5°C during the exposure periods. During the highest temperature increases, mucus production was evident, and the water became noticeably cloudy. Growth Data The growth data are shown in Figure 2. For greenlip abalone. growth rates ( = 100 (j.m d*') were close to commercial growth rates. The blacklip abalone growth rates ( = 15 [o-m d"') were ap- proximately 10% of commercial growth rates. In terms of both length and weight, no significant difference was observed in growth of either species for any of the treatments (P < 0.05). For blacklip abalone. treatment 5 reduced growth, especially in terms of weight, although the difference was not significant. DISCUSSION The primary aim of tliis trial was to simulate a major systems failure, with an extended period of no influent water and little, if any. supplementary aeration, and the desired alterations in water quality were achieved within 5 niin. This rapid change was de- signed to exacerbate the severity of the exposures funher. because any similar changes during a real life event would take much longer to achieve. Although dissolved oxygen and ammonia were the primary water quality variables being manipulated, the lack of infiuent water flow also resulted in temperature increasing and pH declining over 8 h. which added to the reality of the simulation. Temperature changes retlected the ambient air temperature, which was beyond experimental control; whereas, the decline in pH was presumably attributable to respiration and metabolic excretions. Table 1 shows that substantially elevated TA levels were achieved over the baseline condition. Although TA levels did not change greatly during the exposure period, the decline in pH reduced the concentration of UIA. It was considered possible that the water currents generated by the submersible pumps may have modified the response of the abalone to the stress induced by the pulses of poor quality water, perhaps by improving water movement past the gills. This is cer- tainly consistent with known water-flow dynamics around and through abalone shells (Voltzow 1983). although animal behavior, in terms of both gross body positioning (Voltzow 1983) and un- derlying physiological modifications (see later), are also likely to be modifying factors. We are uncertain if this has been a signifi- cant factor in this trial, but note that at the end of the growth trial for Treatment 3. a pulse exposure was conducted without the sub- mersible pumps, and no behavioral differences were observed from previous exposures (data not shown). Based on observations of mortality and animal behavior, the pulse exposures and exposure regime used in this trial did not result in more than a transient stress. Coupled with the lack of significant effects on growth, this indicates that the two species used in this trial are well able to withstand periodic and relatively severe declines in water quality. Although a brief reduction in foraging activity was observed immediately following the pulse exposure, activity returned to normal the following night. The abalone seemed healthy throughout the growth trial, with the ex- ception of an isolated event, and mortality was similar to previous trials in this system (e.g.. Maguire et al. 1996). Conditions in the final challenge exposure seemed to be more stressful than the pulse exposures, based on observations of animal behavior, but were still not sufficient to produce mortality after 3 days. Consumption and activity patterns returned to normal within 24 h. as for the pulse exposures. The transient nature of the stress observed from the pulse and challenge exposures may have been at least partly at- tributable to optimal quality of the ambient conditions (Table 1 ). HaiTis et al. ( 1998. 1999a) reported that greenlip abalone were more sensitive than other invertebrate species and fish when chronically exposed either to elevated UIA or nitrite or decreased DO. For DO and UIA. growth declined through the experimental range in these chronic studies (25-188 p.g L"' UIA and 8.9-4.2 mg L"' DO (1 17-55% saturation)). However, these levels of UIA and DO were achieved in the pulse exposure from the cunent trial without producing a significant growth depression. Hindrum (un- published data) also found that growth of blacklip and greenlip abalone was significantly reduced as compared to controls when chronically exposed to similar combinations of DO and UIA as achieved in the pulse exposures. The chronic exposures outlined above were based on continu- ous exposure to altered water quality for several weeks. The pulse and challenge exposures in the cuirent trial were for 8 h. during the normally inactive daytime period. It is likely that gill activity is reduced during this quiescent period. Certainly, for mollusks in general, this is the case, with activity levels, heart rate, gill perfu- sion rates, gill ventilation rates and resultant oxygen uptake inti- mately related. This has been shown in sedentary species like Mytiliis ediilis (Bayne 1971) and in different gastropod species exhibiting various survival/feeding strategies (Newell & Roy 1973. Morton 1990). Indeed, for abalone. temperature and external oxygen levels have been clearly linked to such factors as heart rate, blood pressure, and resultant oxygen uptake (Nimura & Ya- niakawa 1989. Russell & Evans 1989). In the quiescent period, tissue exposure to such toxicants as ammonia may be reduced as a Simulated Systems Failure and Abalone Growth 683 result of lower transfer across the gills and lower tissue perfusion rates. Indeed, low DO and the expected reduction in heart rate as seen for other haliotids (Nakanishi 1978. Russell & Evans 1989) inay well have functioned in this way as a protective mechanism in our experiment. It is possible that a smiilar mechanism may also work in penaeids, as Allan and Maguire (1991) report that tiger prawns (Penaeus monodon) tolerated conditions for several hours that in chronic exposures resulted in significant growth reductions. It is possible that a greater impact on growth may have been observed if the pulse exposures had been conducted during the nocturnal foraging period, or over a longer period (e.g.. 12-24 h.). Potentially, the altered water quality would then have a bigger effect on the metabolic processes of the abalone. because they are active. However, the observed behavioral changes indicate that the altered water quality was still detected by the inactive abalone. The fact that exposing the abalone weekly to the pulse condition did not reduce growth or induce any clinical signs of stress, such as disease or morbidity, indicates that there was no additive effect. This is further confirmed by the lack of mortality and apparent recovery after the challenge exposure at the end of the growth trial. The pH after 8 h of either challenge or pulse exposure was in the range 7.91-7.62. Harris el al. (1999b) found that chronic ex- posure to pH 7.76 significantly reduced growth in both species of abalone as compared to controls at pH 8.27. Although tempera- tures in individual tanks reached 22-23 degrees during periods of high ambient air temperature, the temperatures recorded after the 8 h of either pulse or challenge exposures were still within the tolerance limits defined by Gilroy and Edwards (1998) for these species. The mucus production observed in the cuirent exposures on warmer days has been reported in abalone (Gilroy & Edwards 1998) and other bivalve niollusks (Bartsch et al. 2000) as a re- sponse to thermal stress. Studies of haliotids in the wild show that these animals tend to be found in large groups, commonly stacked around boulders and in crevices at various depths (Shepherd 1973. Douros 1987. Shep- herd & Partington 1995). H. rubra, in particular, occupies cave- like crevices during the quiescent daytime period in sufficient density that competition for space is a factor in their distribution (Shepherd 1973). In warmer weather and periods of low wave or tidal activity, low DO levels and elevated ammonia concentrations would be expected in these caves as a result of metabolic activity. This would be exacerbated when decaying organic matter accu- mulates (Wells et al. 1998). Not surprisingly, gastropod mollusks are physiologically and biochemically well suited to cope with episodes of hypoxia (Brix et al. 1979. Cade & Ellington 1983, Storey & Storey 1990). The tolerance shown in this trial is. there- fore, not unexpected. CONCLUSION Greenlip and blacklip abalone are resilient to periodic short- term exposure to low dissolved oxygen levels in conjunction with elevated ammonia levels. Transient effects on behavior and forag- ing activity were observed. Based on lack of mortality and on behavioral observations, exposure history had little effect on the response to a more severe challenge exposure. Evidence from the literature suggests that the low DO levels during the pulse and challenge exposures reduced tissue exposure to ammonia by re- ducing uptake across the gills. ACKNOWLEDGMENTS Funding for tins trial was provided by the Abalone Aquaculture Sub-Program of the Fisheries Research and Development Corpo- ration and the Cooperative Research Center for Aquaculture. Ma- rine Shellfish Hatcheries hosted the research, and Crow n Scientific provided backup oxygen meters. LITERATURE CITED Allan. G. L. & G. B. Maguire. 1 99 1 . Lethal levels of low dissolved oxygen and effects of short-term oxygen stress on subsequent growth of iuve- nile Penaeus monodon. Aquaculture 94:27-37. Bayne. B. L. 1971. Ventilation, the heart beat, and oxygen uptake by [Mylihis ediilis L.) in declining oxygen tension. Comp. Biiichem. Pins- /()/. 40:I065-I(W5. Bartsch. M.R.. D. L. Waller. W. G Cope, et al. 20()(). Emersion and thermal tolerances of Ihree species of un-ionid mussels: Survival and behavioral effects. J. Shellllsh Res. 19:233-240. Bower. C. E. & J. P. Bidwell. 1978. Ionization of ammonia in seawater: Effects temperature. pH. and salinity. J. Fish. Res. Board Can. 35:542- 575. Bower. C. E & T. Holm-Hansen. 1980. A salicylate-hypochlorite method for determining ammonia in seawater. Can. J. Fish. Aquat. Sci. 37: 794-798. Brix. C. G. Lykkeboe & K. Johansen. 1979. Reversed Bohr and Rool shifts in hemocyanin of the marine prosobranch Buccinum undatum: Adaptations to a periodically hypoxic habitat. / Comp. Physiol. 129: 97-103. Bruno. T. J. & P. D. N. Svoronos, editors. 1989. CRC handbook of basic tables for chemical analysis. Boca Raton. FL: CRC Press. Colt. J. E. & D. A. Armstrong. 1981. Nitrogen toxicity to crustaceans, fish. and mollusks. In: L. J. Allen & E. C. Kinney, editors. Proceedings of the Bio-Engineering Symposium for Fish Culture. Fish Culture Section of the American Fisheries Society. ECS Publ.l. pp. 34—47. Douros. W. J. 1987. Stacking behavior of an intenidal abalone: An adap- tive response or a consequence of space limitation. J. E.xp. Mar. Biol. Ecol. 108:1-14. Giide, G. & W. Ellington. 1983. The anaerobic molluscan heart: Adaptation to environmental anoxia. Comparison with energy metabolism in ver- tebrate hearts. Comp. Biochem. Physiol. 76:615-620. Gilroy. A. & S. Edwards. 1998. Optimum temperature for growth of Aus- tralian abalone: Preferred temperature and critical thermal maximum for blacklip abalone. Haliotis rubra (Leach), and greenlip abalone. Haliotis laevif-ata (Leach). Aqua. Res 29:481^85. Grasshoff. K. 1989. Methods of seawater analysis. New York: Velag Che- niie. pp. 134-137. Harris. J. O.. G. B Maguire. S. J. Edwards, et al. 1998. Effect of ammonia on the growth rate and oxygen consumption of juvenile greenlip aba- lone. Haliotis laevigata Donovan. Aquaculture 160:259-272. Harris, J. O.. G. B. Maguire, S. J. Edwards, et al. 1999a. Low dissolved oxygen reduces growth rate and oxygen consumption of juvenile green- lip abalone. Haliotis laevigata Donovan. Aquaculture 160:259-272. Harris. J. C. G. B. Maguire. S. J. Edwards, et al. 1999b. Effect of pH on growth rate, oxygen consumption rate, and histopathology of gill and kidney tissue for juvenile greenlip abalone {Haliotis laevigata. Dono- van) and blacklip abalone {Haliotis rubra. Leach). J. Shellfish Res. 18:611-619. Maguire. G. B.. S. M. Hiiidrum. D. R. Johns, et al. 1996. Effects of tank drainage frequency cm growth of juvenile greenlip abalone, Haliotis laevigata. In: P. Hone, editor. Proceedings of the 3rd Annual Abalone Aquaculture Workshop. Port Lincoln. .South .Australia. August: 1996. (South Australian Research and Development Institute. Adelaide.). Meade, J. W. 1985. Allowable ammonia for fish culture. Prog. Fish Cult. 47:135-145. Morton. B. 1990. The physiology and feeding behavior of two marine 684 HiNDRUM ET AL. scavenging gastropods in Hong Kong: The suhlidal [Biibylonia li{to.\a) (lamarck) and the interlidal {Na.ssariiis festivus) (Powys). / Moll. Sliid. 56:275-288. NalvUnishi. T. 1978. Studies on the effect of the environment on the heart rate of shellfishes 1 1. Effect of temperature, low salinity, and hypoxia on the heart rate of an abalone {Haliolis lNoidoric)disciiss hannai) Ino. Bull Hokkaido Reg. Fish. Res. Uib. 43:59-68. Newell, R. C. & A. Roy. 1973. A .statistical model relating the oxygen consumption of a mollusk (Linorina linorea) to activity, body si/e. and environmental conditions. Physiol. Zool. 46:253-275. Niniura. Y. & H. Yamakawa. 1989. Oxygen uptake rate and heart rate of small abalone {Sidciiliis siipeilexla) as related to the ambient oxygen concentration. Nippon Siiisan Gakkciishi 55:1869. Russell, C. W. & B. K. Evans. 1989. Cardiovascular anatomy and physi- ology of the black lip abalone {Haliolis ruber). J. Exp. Biol. 252:105- 117. Russo, R. C. & R. V. Thurston. 1991. Toxicity of ammonia, nitrite, and nitrate to fishes. In: D. Brune & J. Tomasso, editors. Aquaculture and water quality. Baton Rouge, AL: World Aquaculture Society, pp. 58- 89. Shepherd. S. A. 1973. Studies on Southern Australian abalone (genus Halunis) 1. Ecology of five sympatric species. .4».s7. ,/. Mi/r Freshwa- ter Res. 24:217-257. Shepherd, S, A. & D. Partmgton. 1995. Studies on Southern Australian abalone (genus Haliotis) XVI. Recruitment, habitat, and stock rela- tions. Aust. J. Mar. Freshwater Res. 46:669-680. Storey, K. B. & J. Storey. 1990. Metabolic rate of depression and bio- chemical adaptation in anaerobiosis, hibernation, and estivation. Quart. Rev. Biol. 65:145-174. Thurston, R.V & R. Russo. 198 1 . Ammonia toxicity to fishes. Effect of pH on the toxicity of un-ionized ammonia species. Env. Sci. Tech. 15:837- 840. Underwood, A. J. 1981. Techniques of analysis of variance in marine biology and ecology. Oceanogr. Mar. Biol. Ann. Rev. 19:513-605. Voltzow, J. 1983. Flow-through and around the abalone [Haliotis ka- intschatkana). Veliger. 26:18-21. Wajsbrot, N., A. Gasith, M. Krom & D. M. Popper. 1991. Acute toxicity of ammonia to juvenile gilthead seabream Sparus aurata under reduced oxygen levels. .Aquaculture 92:277-288. Wells, R. M. C, J. Baldwin. S. R. Speed & R. E. Weber. 1998. Haemocya- nin function in the New Zealand abalone Haliotis iri.\ and H. australis: Relationships between oxygen-binding properties, muscle metabolism, and habitat. Mar Fresh. Res. 49:143-149. .loiinuti iif Shellfish Rcstanh. VuL 20, No. 2, 685-687. 2001. ONTOGENETIC TRENDS OF MINERALOGY AND ELEMENTS IN THE SHELL OF ABALONE, HALIOTIS DISCUS HANNAI INO GEN HE AND KANGSEN MAI* Aquaciilttirc Rcscarcli Laboratory. College of Fislieric\. Ocean Uiiiversiix of Qingdao. Qiiii>clao, 2660(Ki People's Republic of China ABSTRACT Japanese abalone, Haliotis discus hawuii: with shell lengths 8. 14. 25. and 55 mm were used for mineralosy and element composition studies. Besides calcile and aragonite. dolomite (CaMgtCO,),) was detected tor the first time in abalone'shells. No calcite was delected in the 8-mm shells. The percentage of calcite increased steadily from 1.6% to 13.6% in shells from 14 to 55 mm, while that of aragonite decreased from 95.3% to 83.9%. Dolomite varied from 2.5% to 9.0%. The level of Mg, Mn, Fe, Na, and Al in 8 to 55 mm shells increased, whereas those of Zn and Cu declined. Possible reasons for these trends are di.scussed. A£) WORDS: ontogenetic trend, mineralogy, elements, uhalone, Haliotis discus hunmii INTRODUCTION The mineral composition of molluscan larval shells comprises exclusively of aragonite. regardless of postlarval or adult forms (Watabe 1988). In Haliotis cliseiis haimai. the larval shells consist of aragonite spherulites (Iwata 1980), whereas the prismatic layer of adult shell is composed of calcite and aragonite (Dauphin el al. 1989, Shepherd et al. 1995). The changing pattern of shell miner- alogy from larvae to adult remains unexplored. Many factors that can inllLience the mineralogy and chemistry of molluscan shells have been investigated, including water tem- perature and salinity (Wilbur 1972, Taylor & Reid 1990, Cohen & Branch 1992, Mann 1992), taxonomic differences (Turekian & Armstrong 1960, Dauphin et al. 1989). ontogetiy (Fuller 1985, Carriker et al. 1982, Carriker et al. 1982), and physiological con- ditions for shell formation (Simkiss 1976, Almeida et al. 1996). However, almost all of these studies investigated bivalves. Gas- tropods have been largely neglected. To our knowledge, there are no data on the ontogenetic trends of abalone shell mineralogy and chemistry. In this study, the shells of different sizes of abalone, Haliotis discus haititai. were analyzed to study the ontogenetic trend of shell mineralogy and element composition. MATERIALS AND METHODS Abalone Shells Live H. discus hauitai individuals were purchased from Rongcheng Abalone Farm, China, in October 1998. These were cultured under the same environmental and nutritional conditions in the hatchery. Shells 8 ± I mm, 14 ± I mm, 25 ± 1 mm. 55 ± I mm (mean shell length ± s.d.) were used for this study. Upon sacrificing the abalone. their soft tissue was immediately removed. .V-rav Powder Diffraction The mineralogical composition of the whole shell was investi- gated by X-ray powder diffraction. Ten to 20 shells were ground into a powder mixture for every replicate of each shell length. Three replicates of each shell length were used. A Philips counter diffractometer was employed, with a proportional counter and monochromatized copper target set at 40 kV and 80 mA. The ratio of polymorphs was calculated from the relative intensities of the principal reflection characteristic of each mineral through a cali- bration curve derived from mixtures of natural aragonite, calcite and dolomite (Neri et al. 1979). Element Analysis The abalone shells were ground into fine powder and carefully digested in double distilled nitric acid. Minor and trace elements in the digestion were analyzed by inductively coupled plasma atomic emission spectroscopy (ICP-AES) on a JOBIN (Model .lY 70PLUS). Statistical A nalysis All percentage data were square-root arcsine transformed prior to analysis. Tests of normality and homogeneity of variances con- firmed that the data sets met the requirements for one-way ANOVA (Zar 1984). When overall differences were significant at the 5% level, Tukey's test was used to compare the mean values between individual treatments. Correlation analysis was employed to evaluate the relationship between abalone size and mineralogi- cal or element composition. RESULTS Shell Mineralogy *Corresponding author. E-mail address: kniaiCs'mail.ouqd. edu.cn The percentages of different crystal polymorphs detected for different shell lengths are given in Table 1 and Fig. 1. Dolomite was detected for the first time in abalone shells. Dolomite and aragonite were found in all samples. However, although expected, calcite was not detected in 8-mm shells. The percentage of calcite increased steadily from 1.6% to 13.6% in 14-55 mm shells, whereas that of aragonite decreased from 95,3% to 83.9%. Corre- lation analysis showed that the percentages of both calcite and aragonite were significantly correlated with abalone sizes (calcite: r = 0.985, P < 0.0(J01: aragonite: r = -0.830, P < 0.001 ). The proportion of dolomite increased from 4.7 to 9.0% in 8-25 mm shells, but dropped to 2.5% in 55-mm shells. Element Composition The element composition of different sized shells is listed in Table 2. Ontogenetic changes of selected elements differed. The levels of Mg. Mn, Fe, Na, Al in shells increased from 8 to 55-mm shells, whereas those of Zn and Cu dropped. The positive corre- lation between abalone size and the level of Fe (r = 0.947. P < 0.00(11 ), Mg (r = 0.769. P = 0.003). Al (r = 0.787. P = 0.002). 685 686 He and Mai TABLE 1. The mean percentage of different crystal polymorphs in abalone, Haliotis discus liaiinai Ino. of different shell length (/i = 3, mean ± SDl. Shell length ( mm ) Calcite Aragonite Dolomite 8 ii.d. (O.O)-' 95.3 (0.9)" 4.7(0.9)" 14 1.6(0.1)-' 93.3 (0.4)'-' 5.1 (0.5)-' 2? 6.6(0.6)'- 84.5 (0.2)" 9.0(0.5)" 55 13.6(1.1)" 83.9(1.21-' 2.5(0.11)-' ANOVA F 212.09 113.91 17.23 P 0.000 0.000 0.009 n.d.: not detectable Means in the same row sharing the same superscript letter were not sig- nificantly different as determined by Tukej's test (/) > 0.05). or Na (r = 0.612. P = 0.034) was significant, as v\'ere the negative correlation between abalone size and Zn (r = -0.721, P = 0.008) and Cu concentrations (r = -0.646. P = 0.023). No significant coiTelation was established between abalone size and the level of Mn (r = 0.480. P = 0.114). DISCUSSION It has been reported that the shell mineralogy of adult and larval shells is quite different (Watabe 1988). although to date this had never been investigated in abalone. This is an inipoitant avenue of research, however, as the ontogenetic change of shell stiucture can determine the haidness of the shell, and thus affect the survival of a mollusk since its shell is usually its main souice of defense. In our experiments, clear ontogenetic trends of mineralogy and element composition were observed. The definite intrinsic mecha- nisms that determine the rnineral composition of shells during ontogenetic development remain unknown, although several fac- tors are known to be involved in the process. In in vitro tests, matrix proteins secreted by mantle epithelial cells determine the biomineral polymorphs of shells. Macromolecules extracted from the aragonite layers of some mollusk species induced aragonite formation in viiro. and macromolecules from calcite layers in- duced mainly calcite formation under the same conditions ( Belcher et al. 1996, Falini et al. 1996). It has beei-| reported that the con-i- position of matrix proteins of abalone shell varied in different developmental stages (Cariolou & Morse 1988). We obtained (^Mf^.M4'».-wl' 4,,JL ■1 /.' ■ i I L 55-mm Shell iicattmeim jJl 14- mm shell 23 '.Z 6a 26?(degrees) Figure 1. X-ray diffraction profiles of different Haliotis discus hannai shell sizes. Main reflections specific to aragonite (peak labeled '\'). calcite (peak labeled "C"), and dolomite (peak labeled 'D') are indi- cated. similar results in our laboratory (unpublished data). Therefore, we suggest that the matrix proteins can play a vital role in controlling the composition of crystal polymorphs in shells. Further work is needed to chai-acterize proteins in different shell layers during ontogenetic development and relate their composition to the crys- tal formation. On the other hand, some environmental factors, such as water temperature and salinity can also influence shell miner- alogy (Taylor & Reid 1990. Cohen & Branch 1992). However, given the constant en\ ironmental conditions (e.g. water tempera- ture, salinity and nutritional conditions) in our hatchery, the intfu- ences of these factors on our results can be largely excluded. Unexpectedly, besides calcite and aragonite. dolomite was de- tected in the cuirent study. The mechanism of its formation re- i-nains unknown. It is reported that the arrangement of atoms in aragonitic and calcitic lattice structures allows for the substitution of some divalents for calcium ions (Rosenberg 1990, Lingard et al. 1992). Under certain circumstances, this substitution can even re- sult in separate crystal polymorphs. Several isomorphs, such as vaterite (Wilbur & Watabe 1963), strontianite (SrCO,) (Odum 1951), dahllite (Watabe 1956) and barite (Fritz et al. 1990) have been reported. Pytkowicz ( 1965) points out that under conditions in which the Mg concentration is five times that of Ca. which is common in sea water and some bivalve extiapallial fluids (Cren- shaw 1972, Wada & Fujinuki 1976), collisions producing Mg-Ca- CO, aggregates are more frequent than those resulting in CaCO,. TABLE 2. Mean concentration (ppni) of certain elements in abalone, Haliotis discus hannai Ino. of different shell length tn = 3, mean ± SD). Shell length (mm) ANOVA Element 8 14 25 55 F P Zn 27.2 (0.0)" 53.1(6.0)'-' 6.9(0.2)° 0.2 (0.2)° 123.89 0.000 Mn 4.7 (0.0)" 1.7(0.1)° 16.4(1.4)" 10.5 (3.0)" 44.12 0.002 Fe 28.9 (2.0)° 29.6(3.5)° 133.8(28.4)" 211.9(1.3)' 76.08 0.006 Mg 140.1 (0.7)° 119.2(6.8)° 386.8(22.4)" 373.7 (0.5)" 305.28 0.000 Cu 6,2 (0.2) 9.1(4.0) 2.7(0.3) 2.5 (0.2) 3.63 0.122 Al 1498.8(10.6) 1373.3(30.3) 1344.4(44.5) 1 785.2 (.%.l) 4.93 0.079 Na 6883.6(11.7)- 9527.2(96.1)" 10439.1(24.6)" 10064.6(74.5)" 20.21 0.0(J7 Means in the same line sharing the same superscript letter were not significantly different as determined by Tukey's test (p > 0.05). Mineralogy and Elements in Abalone Shells 687 There is still no report reiiarding llie concentr;ilion of tliese ions in abalone extrapallial fluid (EPF). This study reports that the levels of Mg. Mn, Fe, Na. and Al in shells increased with shell size (age), whereas those of Zn and Cu dropped. This is similar to what is reported by Carriker et al. (1982) concerning the oyster Cms.uistrea virgiiiica (Gnielin). These changes can be attributed to many factors. For example, magnesium and manganese are enriched in calcite rather than ara- gonite (Dauphin et al. 1989). Mn can accumulate during the pro- cess of shell thickening (Yoshioka & Terai 1996). Hawkes et al. (1996) reported that the concentrations of elements in the shell were very different in the outer prismatic shell layer and the inner nacre layer. Therefore, the concentration changes of elements can be regarded as the effects of changes of shell mineralogy during ontogenetic development. ACKNOWLEDGMENT This investigation was supported by the Natural Science Foun- dation m China (NSFC. Grant No.: 39770589). LITERATURE CITED Almeida. M. J.. G. Moura. J. Machado. J. Coimbra. L. VUannlw. C. Ribeiro & P. Soares-da- Silva. 1996. Amino acid and metal content of Crassostrea gigas shell infested by Polydow sp. in the prismatic layer insoluble matrix and blister membrane. Aqmu. Living Resoiir. 9:179- 186. Belcher. A. M., .X. H. \Vu. R. J. Chnstensen. P. K. Hansma. G. D. Stucky & D. E. Morse. 1996. Control of crystal phase switching and orienta- tion by soluble mollusc-shell proteins. Nature 381:56-58. Cariolou, M. A. & D. E. Morse. 1988. Purification and characterization of calcium-binding conchiolin in shell peptides from the mollusc. Haliotis nifescens. as a function of developmenl. / Coiiip. Physiol. 157:717- 729. Carriker. M. R.. C. P. Swann & J. W. Ewart. 1982. An exploratory study with the proton microprobe of the ontogenetic distribution of 16 ele- ments in the shell of living oysters (Crassostrea virginica). Mar Biol. 69:235-246. Carriker. M. R.. C. P. Swann, R. S. Prezant & C. L. Counts. lU. 1991. Chemical elements in the aragonitic and calcitic microstructural groups of shell of oyster Crassostrea virginica: a proton probe study. Mar. Biol. 109:287-297. Cohen. .\. L. & G. M. Branch. 1992. Environmentally controlled variation in the structure and mineralogy of Patella graniilaris shells from the coast of southern Africa: implications for palaeotemperature assess- ments. Palaeogeogr. Palaeoclimatol. Palaeoecol. 91:49-57. Crenshaw. M. A. 1972. The inorganic composition of molluscan extrapal- lial fluid. Biol. Bull. 143:505-512. Dauphin. Y.. J. P. Cuif. H. Mutvei & A. Denis. 1989. Mineralogy, chem- istry and ultrastructure of the external shell-layer in ten species of Haliotis with reference to Haliotis tiiherciilata (Mollusca: Archaeogas- tropoda). Bull. Geol. Inst. Univ. Uppsala NS 15:7-17. Falini, G., S. Albeck. S. Weiner & L. Addadi. 1996. Control of aragonite or calcite polymorphism by mollusk shell macromolecules. Science 271:67-69. Fritz, L. W.. L. M. Ragone & R. A. Lutz. 1990. Biomineralization of barite in the shell of the freshwater Asiatic clam Corbuula tlununea (Mol- lusca: Bivalvia). Linmol. Oceanogr. 35:756-762. Fuller, C. S. 1985. Structure and mineralogy of larval and postlarval shells of Mytiliis edulis Linne and Ischadium recunnnn (Rafinesque). / Shell- fish Res. 5:36. Hawkes, G. P., R. W. Day, M. W. Wallace, K. W. Nugent. A. A. Bettiol. D. N. Jamieson & M. C. Williams. 1996. Analyzing the growth and form of mollusc shell layers, in situ, by Cathodoluminescence micros- copy and Raman spectroscopy. J. Shellfish Res. 15:659-666. Iwata. K. 1980. Mineralization and architecture of the larval shell of Hali- otis discus hannai Ino (Archaeogastropoda). J. Fac. Sci. Hokkaido Univ. Ser. IV 19:305-320. Linaard, S. M.. R. D. Evans & B. P. Bouraoin. 1992. Method for the estimation of organic bound and crystal-bound metal concentrations in bivalve shells. Bull. Environ. Contam. Toxicol. 48:179-184. Mann. K. O. 1992. Physiological, environmental, and mineralogical con- trols on Mg and Sr concentrations in Nautilus. J. Paleoniol. 66:620- 636. Neri. R.. G. Schifano & C. Papanicolaou. 1979. Effects of salinity on mineralogy and chemical composition of Cerastoderinu edule and Monodonta articulata shells. Mar. Geol. 30:233-241. Odum. H. T. 1951. Notes on the strontium content of seawater, Celestite radiolaria. and strontianite snail shells. Science 114:211-213. Pytkowicz. R. M. 1965. Rates of inorganic calcium carbonate nucleation. J. Geol. 73:196-199. Rosenberg. G. D. 1990. The 'vital effect' on skeletal trace element content as exemplified by magnesium. In: J. G. Carter & V. N. Reinhold, editors. Skeletal Biomineralization: Patterns, Processes and Evolution- ary Trends. Vol. 1. New York: Van Nostrum Reinhold. pp. 567-577. Shepherd. S. A., M. Avalos-Borja & M. O. Quintanilla. 1995. Toward a chronology of Haliotis fidgens, with a review of abalone shell micro- structure. Mar. Freshwater. Res. 46:607-615. Simkiss. K. 1976. Cellular aspects of calcification. In: N. Watabe & K. M. Wilbur, editors. The mechanisms of mineralization in the invertebrates and plants. Columbia: Univ. South Carolina Press, pp. 1-31. Taylor, J. D. & D. G. Reid. 1990. Shell microstructure and mineralogy of the Littorinidae: Ecological and evolutionary significance. Hydrohio- logia 193:199-215. Turekian. K. K. & R. L. Armstrong. 1960. Magnesium, strontium, and barium concentrations and calcite-aragonite ratios of some recent mol- luscan shells. J. Mar. Res. 18:133-151. Wada, K. & T. Fujinuki. 1976. Biomineralization in bivalve mollusca with emphasis on the chemical composition of extrapallial tluid. In: N. Watabe & K. M. Wilbur, editors. The Mechanisms of Mineralization in the Invertebrates and Plants. Columbia: Univ. of South Carolina Press. pp. 17,5-190. Watabe. N. 1956. Dahllite identified as a constituent of prodissoconch I of Pinctada nuirtensii (Dunker). Science 124:630. Watabe. N. 1988. Shell structure. In: E. R. Trueman & M. R. Clarke, editors. Mollusca. Vol. 1 1 . New York: Academic Press, pp. 69-74. Wilbur. K M. 1972. Shell formation in mollusks. In: M. Florkin & B. T. Scheer. editors. Chemical Zoology. Vol. VII. Mollusca. New York: Academic Press, pp. 103-145. Wilbur. K. M. & N. Watabe. 1963. Experimental studies on calcification in molluscs and the alga Coccolithus hu.xlevi. Ann. NY Acad. Sci 109:82- 112. Yoshioka, S. & M. Terai. 1996. Uptake of manganese by calcium carbon- ate skeletons in freshwater environment. In: D. Allemand & J. P. Cuif editors. Monoco: Musee Oceanographique. pp. 83-90. Zar, J. H. 1984. Biostatistical Analysis. 2nd edn. Englewood Cliffs: Pren- tice-Hall Inc. 718 pp. ./iiiirmil of Shellfish Research. Vol. 20, No. 2. 6S4-6y.^ 2001. GROWTH OF JUVENILE ABALONE, HALIOTIS FULGENS PHILIPPI. FED DIFFERENT DIETS ELISA SERVIERE-ZARAGOZA,' * ALEJANDRA MAZARIEGOS-VILLAREAL,' GERMAN PONCE-DIAZ.' - AND SILVIA MONTES MAGALLON' 'CeiiliD de InvcsligacioiH's Bioldgicus del Noraeste (CIBNOR), P.O. Box J28. Lu Paz, Bcija Culifoniia Siir, 23000, Mexico; 'Centra Interdi.fciplinario de Ciencias Marinas-IPN, La Paz. Baja California Siir. Mexico; ^SEMARNAP, Delegacioii en Baja California Siir. La Paz. Baja Cidifornia Snr. Mexico ABSTRACT Groulh rales ot juvenile Hatiotis fiilfiens (green ahalone) were evakiated with tour dilTerent diet,s over 106 days. Three diets were based on the algae palm kelp Eiseniii arburea. giant kelp Maeroeystis pyrifera. and Gelidinm rohiisium. and one on seagrass PhyllosiHidix torieyi. One artificial diet was used as a control. The best growth rates and specific growth rales were found in abalone fed M. pyrifera. which were significantly different from growth achieved on the other natural diets. The pattern of growth in juveniles fed an artificial diet was similar to juveniles fed M. pyrifera. The highest mortality (11%) was in juveniles fed the red algae G. roliHSliini- KEY WORDS: Haliolis fiil,i;ens. green ahalone, growth, algae, diets INTRODUCTION MATERIALS AND METHODS In recent years, dietary research in abalone has focused on the production of formulated feeds in countries with a history of aba- lone culture (Uki & Watanabe 1992. Wee el al. 1992. Viana et al. 1993, Viana et al. 1994. Britz 1996a. Britz 1996b. Fleming et al. 1996. Knaueret al. 1996. Britz et al. 1997. Clarke & Creese 1998. Lopez et al. 1998. Monje & Viana 1998) because of potential use of these feeds in commercial production. Information on natural diets is important in understanding the biology of abalone species. Abalone food preferences have been studied from both analysis of gut contents and feeding experiments in species such as Haliotis cracherodii Leach. H. discus hannai Ino. H. ful^iens Philippi. H. tuberculata (L.). and H. nifescens Swainson (Sakai 1962. Leighton & Boolootian 1963. Culley & Peck 1981. Uki 1981. Uki et al. 1986. Leighton 1966. Mercer et al. 1993, Corazani & Illanes 1998. Leighton & Peterson 1998, Simpson & Cook 1998). Shepherd and Steinberg ( 1992) reviewed the literature on the feeding biology of abalone. In Mexico, research into natural diets has been limited to the brown alga Maeroeystis pyrifera (L.) C. Ag. as a control in feeding trials (Viana et al. 1993. Viana et al. 1996). It has been assumed that abalone feed on the kelp alga M. pyrifera. Regional hatcheries use this species as a main source of natural food. M. pyrifera does not occur at the southern limit of the distribution of abalone species along the Baja California Peninsula. However, there are many subtidal algae along the coast that may be used as potential food. In Baja California Sur. the most common food items in gut content of adults of green abalone Haliatis fiilgens were the seagrass Phyllospadix torreyi S. Watson and the mac- roalgae Sargasuin sp.. Eisenia arhorea Aresh.. Cryptopleura erispa Kylin. and Rhodymeuia sp. (Serviere-Zaragoza et al. 1998). It is Important to develop feeding experiments oriented to test growth rates and feed conversion efficiencies on single species diets to evaluate local algal species as diets for abalone. Day and Fleming ( 1992) mentioned that although an alga may not support sustained growth when fed alone, it may be of great value when part of a mixed diet by providing essential nutrients to the diet. This study was designed to assess the growth of juvenile green abalone fed with common species in the benthic environments inhabited by abalone along the western coast of Baja California. ^Corresponding author. Fax: +1 1 2-5-47 1. S; E-mail: serviere@cibnor.mx Diets Maeroeystis pyrifera (MP) was selected as the primary alga species (the control) because it is the dominant algal species of southern California (Dawson et al. 1960) and it is believed by fishermen that Mexican abalone species feed extensively on this alga. The brown alga Eisenia arhorea (EA). red alga Gelidinm robusttiin (OR), and the sea grass Phyllospadix torreyi (PT) are thought to be important species in abalone communities as poten- tial food along Baja California (Guzman del Proo et al. 1972. Guzman del Piooetal. 1991. Serviere-Zaragoza et al. 1998). Natu- ral diets were harvested, dried, and stored at the beginning of the experiment, because chemical composition varies during the year. The grow th rate of abalone fed w ith algal diets is low and variable over time. Therefore, we decided to include an artificial diet as a control (AD). AD was manufactured by the nutrition group of Instituto de Investigaciones Oceanologicas. B.C. The dietary for- mulation is in Table 1. Proximate analyses (crude protein, ether extractables. crude fiber, and ash) were performed using the meth- ods of the Association of Official Agricultural Chemists (1995). Experimental I'rocidiire Animals used in this experiment were reared in a commercial hatchery in Erendira. Baja California, and transferred to the CIBNOR laboratory in La Paz, B.C.S. at age 8 mo. Feeding ex- periments commenced six months later. During this period, ani- mals were fed the brown alga Eisenia arborea. Six hundred ani- mals ( 17.3 (±2.2) mm shell length and 0,44 (±0.2) g body weight) were used to test the growth response of abalone fed four diets for 106 days. Experimental animals were held in 16-L fiber glass containers (50 x 30 x 35 cm. experimental units) with a concave bottom. Three replicates per treatment were made with 40 abalone per experimental unit (EU). Animals were marked with plastic tags attached to the shell. The EUs were supplied with temperature controlled (20 ± 1"C). fresh, filtered (10 (xm) water with a (low rate of 73 mL/min. The water was aerated vigorously. Salinity, oxygen. pH. nitrate, nitrite, ammonium content, and phosphate were monitored every week. Dead animals were removed and replaced to maintain densities. Shell length was measured with a vernier caliper, and body weight was measured with an electronic balance (to 0,001 g) at the beginning, and at 30. 60. and 106 days. 689 690 Serviere-Zaragoza et al. TABLE 1. Percent composition of the artificial diets tested in this study, given as percentage of dry matter. Ingredients g/lllO ! Fish meal Silage (dry basis) Soy bean meal Corn meal (whole) Vegetable meal Com starch Gelatin Vitamin mixture Mineral mixture Choline chloride Methionine BTH Sodium Benzoate 301) 2.0 10.0 12.0 15.0 19.37 6.0 1.7 3.3 0.11 0.23 0.086 0.23 Diets were given ml lihiluiii. Natural diets were fed every 42 hours in the afternoon and the artificial diet each night. The re- tiiaining food was carefully collected for drying and weighing; every other inoming for natural diets and each morning for the artificial diet. This was undertaken throughout the feeding experi- ment. Algae growing on the inside walls of the EU were removed twice a week with a soft brush. For diets, the amount of dry matter lost in seawater was estimated during the experiment by using EUs without abalone under the saine conditions as those of the growth experiments. Growth rates were calculated by the equation: GRsL = (SL, - SL,)/T and GRgw = (BW, - BW,)/T where SL, = mean final shell length. SL, = mean initial shell length. BW| = mean final weight, BW, = mean initial weight, and T = time in days. Specific growth rate (SGR % day"') was calculated for shell length and body weight by the equation (Britz 1996b): SGRsL = {(In SL, - In SL, )/T} x 100 and SGRbw = {(In BW, - In BW,)/T} x 100 where SGR^,, is percent shell length gain per day, SL, = mean final shell length, SL, = mean initial shell length, and T is time in days between measurements. SGRg^ is percent body weight gained per day. BW, = mean final weight, and BW, = inean initial weight. Consumption was calculated in terms of dry weight with the equation (Uki & Watanabe 1992): FC = (GS/IOO)-R where G is the weight of food offered per animal per day (in grams). S is the percentage of food recovered, obtaining a factor for each diet (from the controls without abalone). and R is the remaining food (in grams) after the abalone had fed. Food conversion efficiency ratio was calculated as (L'ki & Wa- tanabe 1992): FCE = wet weight gain (g)/di7 weight of tlxxi consumed (g) x KX) This measure of food utili/ation is related to conversion effi- ciency, but is in fact the ratio of animal live or wet weight gain to the amount of dry diet consumed (Monje & Viana 1998). Statistical Analyses Data for experimental replicates were pooled because no sig- nificant differences were found between them by one-way analysis of variance (ANOVA) at a significance level of P = 0.05. Data from each different treatment were analyzed by ANOVA test and Tukey test to determine differences of means (Sokal & Rohlf !99.'i). Statistical analysis was done with software STATISTICA 6.0 for PC. RESULTS The highest protein content was found in Gelidium robttstum (GR) followed by Phyllospadix torreyi (PT), Macrocystis pyrifera (MP), and Eiseiiia arborea (EA). The lipids were between 1.06% for G. riihiisriiin and 1.30% for M. pyrifera (Table 2). The daily average temperature was 20 ± 1°C. Water quality analyses showed the following averages (±s) for pH 8.03 (±0.06): oxygen 6.52 (±0.74) mg/L. salinity 40 (±1.29) ppm, nitrites 0.0052 (±0.0004) M-mol/L, nitrates 0.1253 (±0.0029) ixmol/L. ammonium 0.0226 (±0.00001) mg/L. and phosphate 0.00552 (±0.000017) [jimol/L throughout the experiment. Growth of Abalone A significant difference in means was shown between shell length and body weight of juvenile green abalone fed with the natural diets (P < 0.05). At 30 days of the experiment, the mean length, 19.15 mm ± 0.23 (SE). and weight. 0.59 g ± 0.02 (SE). of juveniles fed MP were statistically different from the means of juveniles fed EA, 18.28 mm ± 0.21 (SE) and 0.02 g ± 0.18 (SE). and GR, 18.24 mm ± 0.19 (SE) and 0.50 g ± 0.02 (SE). After 60 days, the differences increased and were significant. Both mean length and weight of ju\eniles fed MP were statistically different from the other diets; EA. GR. and PT. Differences between juve- niles fed these latter three diets were not detected. Mean shell length and body weight increased over time on all diets (Fig. I). The best growth in length and weight for green abalone was obtained with MP, 22.29 mm ± 0.33 (SE) and 1.0 g ± 0.06 (SE). The percent of survival was between 89% and 95% in natural diets. For AD, it was 97% (Table 3). The pattern of growth in juveniles fed the artificial diet. 22.01 mm ± 0.22 (SE) and 0.91 g ± 0.03 (SE), was similar to juveniles fed MP (Fig. 1). TABLE 2. Proximate analysis of the species and artificial diet used in experimental diets, EA, Eisenia arborea; MP, Macrocystis pyrifera: GR, Gelidium robustum: PT, Phyllospadix torreyi: and AD, artificial diet. Component Crude Crude Ether N-free Diet Protein Ash Fiber Extract Extract EA 7.60 27.13 6.44 1.15 57.68 MP 12.0 41.33 7.0 1.30 38.37 GR 17.61 21.26 10.19 1 .06 49.88 PT 13.94 31.88 13.45 1.28 37.44 AD 3.3.8."^ 10.70 5.85 7.09 40.5 1 Values are given as percent of dry matter. Growth of Juvenile Abalone 691 Days Days Figure 1. Mean growth of abalone fed with different diets. EA, Eisenia arborea; MP, Macrocyslis pyrifera; GR, Gelidium robiisliim; PT. Phyl- lospadix torreyi; and AD, artificial diet. weight growth rates were between 0.92 mg ±0.13 (SE) for GR during the first nionlh and 6.70 mg ± 1.29 (SE) for MP during the third month. The daily growth rates of juveniles fed artificial diet were simi- lar to growth rates of juveniles fed MP (Table 3). The values ranged between 31 |jim ± 2 (SE) in the third month to 57 (xm ± 2 (SE) in the .second month. Daily body weight rates were between 3.63 mg ± 0.18 (SE) in the first month and 6.13 mg ± 0.23 (SE) in the second month. SGR The ANOVA showed that there was no significant difference in mean SGR for any of the replicate diets. Using the Tukey test on the SGR (Tukey test P < 0.05) for green abalone. the exis- tence of a significant difference in mean was shown between the SGR shell length and body weight of the abalone fed with MP and the other diets (EA, GR. PT). The same analysis revealed there was no significant difference in the mean SGR for EA. GR. and PT. except for SGR body weight of juveniles fed PT, which was higher than EA and GR (Fig. 2). SGRs obtained from juveniles fed the artificial diet were similar to juveniles fed MP. The feed consumption rate of abalone ranged from 0.0033 g/day for EA to 0.0108 g/day for MP. Consumption did not differ significantly between EA and GR. and PT and MP (P > 0.05). Abalone consumed significantly greater amounts of PT and MP (Table 4). FCE ratios for the natural diets vary between 30% for PT and 63% for MP (Table 4). Differences between FCE ratios were not detected {P < 0.05). Growth Rales Significant differences occurred, between natural diets, in the daily growth rates of the shell length (SL) and body weight (BW) (P < 0.05). During the experimental period. 106 days, both mean SL and BW growth rates of animals fed MP were higher than mean growth rates of the juveniles fed the other natural diets (P < 0.05) (Table 3). The growth rates of the juveniles analyzed varied within the same diet during the experiment. In EA and PT a gradual decrease was observed during the experiment. For MP and GR. in the second month a gradual increase was measured, and at the end of the experiment the values had decreased. Juveniles fed EA showed the lowest shell length growth rates. 15 jjtm ± I (SE), during the third experimental month and the highest was in juve- niles fed MP. 53 p.m ± 2 (SE). during the second month. Body DISCUSSION Mean shell length and body weight increased over time on all diets. The best growth in length and weight for green abalone was obtained with Macrocyslis pyrifera. Feeding abalone on Eisenia arborea, Gelicliuin robuslwn. and Pliyllospadix torreyi diets re- sulted in lower growth and FCE. 46% to 81% of the values ob- tained with M. pyrifera. These results indicate that the dietary value of the common species along the coast of Baja California Sur, Eisenia arborea. Gelidium robustum. and Phyllospadi.x tor- rexi. was inferior to that of the dominant algal species of southern California M. pyrifera. This result may be related to the trends described by Guzman del Proo et al. (1976) about the size and weight means for Haliolis spp.. which decrease from north to south along the Baja California Peninsula. In southern California, the TABLE 3. Sur\ival, mean Initial size, mean growth gain, and mean growth rate of green abalone fed with different diets. Diets as defined in Table 2. Diet Survival % Mean Initial Size (mm) Mean (irowth Gain ( mm ) Mean Growth Rate (fim day"') Mean Initial Size (g) Mean Growth Gain ical and Chemical Sciences, Deakin University. Geelong, VIC 3217, Auslndia ABSTRACT Both prokaiyoles and eukaryotes express a set of highly conserveJ proteiiLs in response to e.xternal and internal stress. The stressors include tissue trauma, anoxia, heavy metal toxicity, infection, changed salinity, and the most characterized, heat shock. The resuU is an expression of stress proteins or heat shock proteins (HSP's) which lead to protection of protein integrity, and also to tolerance under continued heat stress conditions. The Australian blacklip abalone {Haliotis rubra) is found principally in southern coastal waters and also in estuarine/bay environments. Estuarine/hay environments have greater fluctuations in environmental condi- tions, especially those of salinity and water temperature, than are found along oceanic coasts. Abalone from estuarine/bay and oceanic coastal environments were subjected to either increased temperature (2°C/day for a total of 10°C) or hyposalinity (SOOi: seawater). Estuarine/bay abalone were less affected than the oceanic animals by temperature increase and also demonstrated the ability to volume regulate .i h after the initial salinity shock. SDS-PAGE and Western blotting technii|ues. together with dot blots of total protein, using HSP70 specific antibodies, were used to detect HSP70s in the foot muscle of the animals and indicated an expression of HSP70 in response to heat shock in abalone, hut not following hyposalinity shock. RT-PCR yielded a partial cDNA clone of HSP70 from the foot muscle. KEY WORDS: abalone. HaUolis rubra, heat shock proteins, H,SP70. stress, cDNA INTRODUCTION E.xtreine enviriuiiiiental conditions cause all organisms previ- ously studied, from mammals to bacteria, to e\oke a heat shock re.sponse (HSR) (Gehring & Wehner 1995). The HSR is a rapid reaction to environmental or internal stresses that can be elicited by factors such as tissue trauma, mutagens, anoxia, heavy metals, salinity and exposure to abnortnally high temperatures (Laursen et al. 1997). The HSR is more generally known as the stress response, because previous studies have shown that a wide range of stressful conditions induce its initiation. One exception to date has been established. Hydni oligaciis. a freshwater Cnidarian that lives in thermally stable environments (Bosch et al. 1988). Following stress, the HSR leads to significant alterations in expression of a small number of specific genes in the organism, designated as heat shock genes. These genes become actively tran- scribed under stressed conditions, which in turn leads to produc- tion of certain proteins, aptly named heat shock proteins (HSPs). The name is derived from the original stressor which led to the discovery of these proteins (Gehring & Wehner 1995). Upregula- tion in the transcription of one of these HSPs. HSP70. during stress is thought to be due to an elevation in the cellular levels of dena- tured proteins that activate a transcription factor termed the heat shock factor (HSF). which in turn binds to GC rich DNA in the promoter regions upstream to the HSP70 gene (Sanders 1993). Upon binding of HSF the upregulation of HSP70 expression is possibly 100 times basal levels (Wilkins & Lis 1997). The signifi- cance of the HSR is only partially understood, but there is suffi- cient evidence to indicate that production of HSPs leads to pro- tection or tolerance of the stress to which the organism is exposed (Woodet al. 1998). HSP70s are among the most highly conserved family of pro- teins known. There is approximately 50% homology in the amino acid (AA) sequence of every species characterised (Parsell & Lin- quist 1994). This high homology suggests there has been little phylogenic divergence in the protein, highlighting the importance ^Corresponding author. E-mail address: pjhta'deakin. edu.au of HSPs to survival. Most of the highly conserved regions in HSPs are found in the amino-terminus (N-term) of the protein, where an ATP binding domain is located. The carboxy-terminus (C-term) is the more divergent as it contains the substrate binding domains that recognize the vast an^ay of proteins. There are two distinct types of proteins in the 70kDa class, the inducible HSP70 proteins and non-inducible (constitutive) heat shock cognates. HSC70s (Wood et al. 1998). HSC70s are impor- tant in the normal function of a cell under non-stressed conditions and therefore are present at detectable low levels under such con- ditions. HSC70 proteins are itiiportant in the translocation of pro- teins within cells and especially in the transport of proteins into the mitochondrion (Stuart et al. 1994). HSC70s also prevent folding of polypeptide chains by cradling them when incorrectly folded or not completely assembled. Such a mechanism prevents the appear- ance of nonsense proteins and the aggregation of unfolded AA chains, which leads to cellular damage. Earlier observations on HSPs were interspecies comparisons in expression, or differences between HSP expression in control and stressed animals of the same species. These demonstrated varia- tions in the protein expression in both cases. For example, the lamprey Lampetm appendix has a HSP induction temperature (26"C). which is approximately 4"C higher than that of Pelni- myzon marinus, another lamprey species (Wood et al. 1998). Such differences are believed to be a consequence of evolutionary his- tories that ultimately led to genetic differences in the heat shock response. In another study. HSP expression was observed in two Californian mussel species. Myldiis Irossiiliis and Mytiliis gallo- proviiicicdis (Hofman & Somero 1996). Different latitudes sepa- rate these species. M. trasstdus being the more northern of the two. M. trossidns. was the more thermally sensitive, as measured by HSP7() standing-stocks and ubiquitin ctmjugates present in the gill tissues of animals tiiainlained al ]yC for 8 wk. In addition, in- duction temperatures for new HSP7() synthesis were lower in M. trossidiis than in M. <>aU()provincialis (23°C versus 25°C). suggesting protein damage at lower temperatures and greater tem- perature sensitivity in the species from the cooler climate. More recently, Chen and Chen (1999) have shown that juvenile abalone acclimated to higher temperatures survived heat shock at higher 695 696 Drew et al. temperatures. Interestingly, these abalone heat shock responses were negatively correlated with salinity. Salinity stress is another factor encountered hy coastal marine organisms. Eiintemoia ajfinis. a crustacean found in estuaries of North America and Europe, expresses a range of proteins when exposed to a salinity gradient, with particular groups of proteins being expressed at different salinities (Gonzalez & Bradley 1994). Molecular evidence for tolerance to heat has been characterised in a molluscan cell line from Bioinphalaiia glabrata. a fresh water snail (Laursen et al. 1997). Western blotting showed increased expression of H.SP70 resulting from a 10°C increase. Northern blots and RT-PCR led to the cloning of an induced HSP70 cDNA in this snail, which was sequenced and compared to that of other organisms. The highest sequence homology was to that of HSP70s of other aquatic animals, including HSP70a and HSP70b in Aply- sia californica. Abalone (genus Haliotis) is a marine mollusc v\hich inhabits coastal and bay waters of southern Australia. During summer, bay abalone have been observed to dramatically lose weight and de- teriorate in condition. Factors causing such deterioration could include the hotter seawaters (e.g. bay 10°C greater than ocean) and fluctuating salinit> during periods of high rainfall. Therefore, we started the cuixent study examining the HSR in commercially im- portant hlacklip abalone, Haliotis nibra (Leach) exposed to el- evated temperatures. In addition, we compared the ability of bay and oceanic populations to volume regulate after a hyposaline challenge. MATERIALS AND METHODS Animal Collection and Maintenance Haliotis rubra (Leach) were collected during Spring under per- mit number 97/R/()49A from a northern Port Phillip Bay site near Point Cook and from a site in ocean water near Barwon Heads in Victoria. Australia. The animals were transported to aerated hold- ing tanks with a fresh filtered flow-through seawater system where ambient water temperature ranged from 12-16'C. They were pro- vided an excess diet of mixed red macroalgae. The animals were held for 2^ wk before experimentation. Environmental Manipulation: Salinity Stress Two replicate hyposalinity tests took place in two 36-L tanks in which 80% sea water was produced from sea water taken from the holding tanks and distilled water. Water temperatures of the two replicate tanks were 16 ± 0.2°C for test 1 and 13.5 ± 0.2 C for test 2, being the ambient temperatures at the times of experimentation. Plastic plates of 300 mm x 300 mm were weighed wet, numbered, and placed vertically in the tanks. Six abalone were taken from each population, weighed, measured by shell length, and then placed on plates in the tanks (one population per tank). No water flow was permitted during the salinity trials and a carbon fiber filter and aerator were placed in each tank to provide sufficient oxygen and remove any organic matter the abalone might produce. The animals were weighed at 0.5 hour on the plates to reduce disturbance, and then weighed (on the same plate each time) at 1 h intervals for 6 h. Weights were also measured at 24 h, 48 h and 72 h. Abalone were returned to their respective holding tanks after the 72 h test period had finished. No animal was used more than once. Additional abalone were treated in the same manner for 1.5 h before taking tissues for molecular studies. Environmental Manipulation: Heat Stress Responses to heat stress were examined in a 9()-L glass tank in which 15 abalone were taken at random from the bay and ocean populations. The critical thermal maximum (CTM) was deter- mined using the point at which an abalone can no longer remain attached to the substrate, this being considered the temperature at which no long-tenn survival is considered likely. Each abalone was weighed, measured by shell length, and tagged with plastic discs, and then placed onto suspended vertical plates (600 mm x 300 mm) in the well-aerated trial tanks. No food was provided for the abalone during the tests. Ambient water temperature was 15 ± 0.2°C for test 1 and 12.5 ± 0.2°C for test 2. The tank was fitted with a 1000 W thermoregulator and the temperature was raised by 2 ± 0.2"C per h until it reached 20 C. after which it was then raised 1 ± 0.2°C per h until none of the animals could keep contact with the vertical plates. Continual visual monitoring of the trials was undertaken to record any behavioral responses of the abalone during the trials and to record the time and temperature that each abalone detached. Immediately after an abalone had detached, it was weighed and then placed into a well-aerated recovery tank 2 ± 0.5°C lower than the test tank. Two replicate trials were undertaken for each of the two popu- lations. Different animals were randomly selected for each of the tests. Differences between groups were tested using a 50% CTM calculated by linear regression. Only bay abalone were used for molecular studies. Before tis- sue collection, the water temperature was increased by 2°C daily until a temperature of 25°C was reached, representing a total rise of about 9°C. Animals were subsequently removed from the con- tainers and immediately killed by severing of cerebral ganglia. Tissues were quickly removed and snap frozen in liquid nitrogen prior to storage at -80°C. Di\A Extraction Approximately 0.5 g of frozen tissue from each foot muscle was minced (in a sterile petri dish using scalpel blades) until the tissue formed a coarse paste. Cells were then lysed by shaking the tissue in a 1.5 ml microfuge tube with 1ml of DNAzol (Life Technologies, genomic DNA isolation reagent), and the tube was then centrifuged ( 13000 xg. 10 min, 4°C) (MSE Microcentaur. microfuge). The supernatant was removed and the DNA was pre- pared by standard chloroform/isoamyl/alcohol precipitation. Ex- traction was confirmed using separation in agarose gels. Polymerase Chain Reaction (PCR) PCR was carried out on genomic DNA in order to gain se- quence information from the genes encoding HSP70 in H. rubra. PCR on cDNA was used to characterize expressed sequences (niRNA) of HSP70 genes within stressed animals. All reactions were perfomied using an automated thermocycler (FTS-320 Ther- mal Sequencer; Corbett Research) and all reactions were layered w ith one drop of mineral oil prior to cycling. Pritriers for PCR were designed from regions of high amino acid (AA) homology seen between aligned sequences of organisms previously cloned. Figure 1 is a general overall positioning of the primers (shown in italics and arrows) on the HSP70 niRNA tran- script, together with estimated sizes of PCR products. Identification of Expressed HSP's in Ab alone ATPase region 3' variable region 697 HSp-Of HSPdangf 3' HSI'-Ur AbHSPr [AA 175-238] 1 t)2bp AA 442-637 480bp A A 175-637 J 1400bp Figure 1. PCR primer binding sites on HSI'70 cDN A and expected PCR product sizes. Genomic DNA (HSP7() I92bp): 50 |xl reaclions contaiiiing 200 na H. rubra genomii: DNA. 5 jxl 10\ reaction butter (Life Technologies). 2— t niM MgCK (Life Technologies). 5 niM of dNTP's (Boehringer), 8 mM of both forward (HSP7()f; 5'- GNATHATHAAYGAASCANC-3') and reverse (HSP70r: 5'- GNATHATHAAYGAASCANC-3') primers. 2.5 units of Taq Polymerase (Life Technologies) and ddH-,0 were cycled through 36 cycles (1: 95X for 5 min: 2-35: 95X for 120 s. 52°C for 90 s. 72°C for 90 s: 36: 72°C for 5 min). Genomic DNA (HSP70 1400bp): 50 (jlI reactions containing 200 ng H. rubra genomic DNA. 5 (jil I Ox reaction buffer (Life Technologies), 2-4 mM MgCU (Life Technologies). 5 niM of dNTP's (Boehringer). 8 niM of both forward (HSP70f) and re\'erse (AbHSPr: 5'-YARTCNACYTCYTCNAC-3') primers. 2.5 units of Taq Polymerase (Life Technologies) and ddH^O were cycled through 31 cycles (I: 95°C for 5 min; 2-30: 95 =C for 120 s. 52°C for 90 s. 72' C for 120 s: 31: 72°C for 5 min). cDNA (HSP70 192bp): 25 p-l reactions containing 1 |jl1 cDNA, 2.5 [il lOx reaction buffer (Sigma). 2-4 mM MgCL (Life Tech- nologies), 5 inM of each dNTP's (Boehringer). 8 mM of both forward (HSP70f) and reverse (HSP70r) primers. 1.25 units of RedTaq Polymerase (Sigma) were cycled for 30 cycles (1: 95°C for 5 min; 2-29: 95°C for 90s. 52°C for 90s. 72°C for 60s: 30: 72''C for 5 min). cDNA (HSP70 480bp): 25 |jl1 reactions containing 1 p.1 cDNA. 2.5 \i.\ 10\ reaction buffer (Sigma). 2-4 niM MgCl, (Life Tech- nologies). 5 mM of each dNTP's (Boehringer). 8 mM of both forward (HSPdangF: 5'-GAYATHGAYGCNGAYGG-3') and re- verse (AbHSPr) primers. 1.25 units of RedTaq Polymerase (Sigma) and ddH,0 were cycled for 30 cycles ( 1 : 95°C for 5 min; 2-29: 95"C for 90 s. 50°C for 90 s. 72°C for 90 s; 30: 72°C for 5 min). cDNA (HSP70 I400bp): 25 |xl reactions containing 1 \x.\ cDNA. 2.5 p.1 lOx reaction buffer (Sigma). 2-4 niM MgCI, (Life Technologies). 5 mM of each dNTP's (Boehringer). 8 niM of both forward (HSP70f) and reverse (AbHSPr) primers. 1.25 units of RedTaq Polymerase (Sigma) and ddH^O were cycled for 30 cycles (1: 95°C for 5 min; 2-29: 95°C for 90 s. 50°C for 90 s. 72 C for 120 s; 30: 72°C for 5 min). Purification of PCR Products and Cloning Fragments amplified by PCR were separated on agarose gels, and. if PCR products were single products, ligations into plasmids were carried out directly from the PCR reaction mixture. Other PCR products were excised from agarose gels under UV light using sterile scalpel blades, and DNA extracted using QIAEX II gel extraction kits (QIAGEN) and QIAquick gel extraction kits (QIAGEN) following the manufacturer's instructions. Alterna- tively, gel pieces containing desired products, were cut from aga- rose under LIV light and incubated in 50 |jlI ddH,0 at 60°C over- night, then centrifuged at 13000 x i; for 15 min to obtain super- natants which were used in reamplification of product via PCR or used directly in ligation reactions. Ligations of PCR products were cauied out within 24 hours of PCR reactions. PCR products were ligated into pCR2.1 plasmid vectors using the Original TA cloning kit (Invilrogen) according to the manufacture's instructions. Plasmids were then transformed into INVaF' competent E. coli cells (Invitrogen) and transformed cells screened by standard blue/white screenuig procedures. White colonies were chosen as insert-containing clones and screened via PCR directly from colonies using the universal plasmid M13 (for- ward) and T7 (reverse) primers. Positive clones were selected and a QIAprep spin mini-prep kit (QIAGEN) used to prepare mini- preps of plasmids for subsequent sequencing on an Applied Bio- systems Automated Sequencer at Westmead Hospital. Sydney, us- ing the dye terminator method. RNA Extraction from Frozen Tissue and Production ofcDNA RNA was extracted from foot muscle tissue using a modifica- tion of the acid/phenol method (Choinczynski & Sacchi 1987). Frozen tissue (300 mg) was minced finely using scalpel blades and then homogenized with 3ml of TRIzol extraction reagent (Life Technologies), until a cloudy solution was obtained. Samples were centrifuged (13000 x g for 15 min at 4°C) in order to remove lipoproteins and fats, and the supernatant removed. RNA was then isolated using the standard chloroform/isopropanol technique, and final pellets washed with sterile 75% EtOH prior to air-drying pellets and storage at -20°C. First strand cDNA was made from freshly extracted RNA using a Superscript II kit (Life Technologies), according to the manu- facturer's instructions. cDNA was synthesised from total RNA and RNase H used to remove the remaining RNA. The cDNA was then stored at -20-'C. Western Blotting Frozen foot muscle samples (300 mg of tissue) were minced quickly into a coarse paste and the cells lysed by boiling for three 698 Drew et al. min in 2 ml of lysis buffer (32 mM Tris-HCl pH6.8, 2% SDS. 1 |jiM PMSF. 2 |xg/mL each of pepstatin. cfiymostatin and leupep- tin), followed by homogenization on ice with a teflon pestle (Cole- Parmer, U.S.A.) until all the tissue was suspended. The samples were boiled for a further 5 minutes, followed by centrifugation (I3000X g for 10 min at RT) and the supernatant collected for protein quantification using a BCA Protein Assay Reagent kit (Pierce), following manufacturer's instructions. Sample concentra- tions were then equalised by dilution and immediately separated by standard 10'7f SDS-PAGE. or stored at -20°C until used. Rep- licate gels were used for electro-transfer of proteins onto nitrocel- lulose (NC) membranes or Coomassie blue staining. Out-blots of Total Protein NC membranes and two pieces of Whatman chromatography paper were pre-soaked in Ix Western transfer buffer (200 ml MeOH. 5.24 g Tris, 17.6 g Glycine) for 15 minutes. The papers were placed under the NC membrane in a Bio-Dot apparatus (Bio- Rad) and the apparatus assembled. Proteins were prepared in a logarithmic series dilution ( 10-fold dilutions in lysis buffer) from a 100 |jLg stock solution (see Western blotting), to 0.01 |xg, and boiled for three minutes prior to being loaded and blotted under vacuum. NC membranes were blocked and visualized following probing with specific antibodies as described below. Iminiinostaiiiiiig NC membranes from Western transfers and dot-blots were blocked with 5% Blotto (5 g skim milk in 100 ml TBS-T), over- night with constant rocking. Membranes were then incubated with a primary (1°) goat anti-human HSP70 polyclonal antibody (Ab) (Santa-Cruz) in a 1:100 dilution with 5"^;^ Blotto for 2 h with constant rocking. This Ab was raised against the AA region 572- 591 of the C-term of human induced HSP70 and was non-cross reactive with HSC70. NC membranes were washed three times with TBS-T (2.42 g Tris, 5.85 g NaCl, 1 ml Tween-20 made up to 21 with dH^O). each for 15 minutes before being incubated with a secondary (2°) anti-goat horseradish peroxidase (HRP) conjugated Ab (Life Technologies) in a 1:2000 dilution with 57c Blotto for two h. Membranes were again washed three times each for 15 minutes with TBS-T. Bands were visualized w ith HRP developing solution (60 mg l-chloro-4-napthol, 20 ml MeOH, 100 ml TBS, 60 (xl I'O^c H,Oo) until bands appeared. Results were recorded under white light using an Eagle Eye II still video system (Stratagene). A preabsorption control was performed to confimi that the bands were specific. In brief. I °C Ab was preabsorbed overnight at 4°C with an excess of the antigen (Santa Cruz). RESULTS Responses to Reduced Salinity (Hyposalinity) All animals were allowed to acclimate in tanks for three days prior to experiments. The shell lengths (mm ± SD) of abalone used in the replicate tests were 1 15.67 ± 3.2 (test 1 ) and I 1 1 .0 ± I 1 .47 (test 2) for the ocean abalone. and 108.17 ± 8.4 (test 1 ) and 1 1 1.4 ± 6.02 (test 2) for the bay abalone. Exposure to hyposalinity stress caused swelling of tissue, particularly of the soft tissues behind the head. Only the data for test 1 are shown in Figure 2, but similar response curves were observed in both tests. Immediately after transfer into the lower salinity medium there was a distinct in- 01 O) c n ,c O *•' ,c o '5 § >> n o CD 2 0 3 0 Time (h) Figure 2. Percent bod> weight changes of H. rubra in response to a hypotonic salinit> stress. Abalones from oceanic (■) and estuarine/ba.\ (D) populations were subjected to 80'7f seawater. Bar lines represent SE of means. crease of weight in both abalone groups (ie. approximately 10'"/f at 0.5 h). Following this initial increase the bay abalone decreased their weight over the next 5 h to about that of pre-treatment (ie. 98.89r in test 1; 100.4% in test 2). In contrast, the ocean abalone of both groups did not decrease to the same extent as bay abalone (106.7% in test 1; 102.8% in test 2). The differences between the two groups was significant at 5 h in test 1 (;? = 0.012) but not in test 2 (p = 0.083). Overall, these studies on H. riihni populations suggest that the bay abalone are more able to volume regulate. Responses to Elevated Heat Shock The behavior of the aniiiials subjected to rising temperatures changed at approximately 19"C, when lamellapodia were erected and there was an observed lack of mobility in both ocean and bay abalone. Mucous production was evident in both abalone groups at temperatures above 20-2 TC. At about 23°C the abalone raised their shell from their foot by the extension of the adductor muscle and rotated from side to side around its own axis. Above 23.5°C for ocean abalone, and above 24.5°C for bay abalone, retraction of foot from the vertical plates started, and continued until the aba- lone lost their footing completely. One obvious sign of stress in the more stressed animals was a swelling of the soft tissue around the back of the head, causing it to protrude. Figure 3 shows the results of test 1 to establish the critical thermal maximum (CTM) for bay and ocean abalone. Replicate tests were performed but there was no significant difference at the p < 0.05 level between the groups. Starting from an acclimation temperature of 15 ± 0.2°C. the first ocean abalone in test I started to lose hold at 26°C. and bay abalone at 26.7°C. while the last to fall were 28.9°C. and 29°C. respectively. In test 2. starting with an acclimation temperature of 12.5 ± 0.2°C. the first ocean abalone started to drop off the plate at 25.3°C, and bay abalone at 25'C. whilst the last to fall were at 28°C. and 29°C, respectively. The temperature at which 50% remain attached (50% CTM) was cal- culated at 27.4 C for ocean abalone (r = 0.92). and 27.9°C (r = 0.93) for bay abalone in test 1. and 26.8°C for ocean abalone (r = 0.92). and 27.0°C for bay abalone (r = 0.95) in test 2, using linear regression on all points below 100%. The shell lengths (mm ± SD) for the ocean abalone were 1 16.27 ± 9.41 in test 1 and 1 15.73 ± 8.80 in test 2. whereas those for the bay abalone were 1 10.6 + 7.1 in test 1 and 108.67 ± 4.79 in test 2. A further test using bay abalone. was performed to obtain tis- sues from heat-shocked abalone and controls. In this test the water iDhNTIFICATlON OF EXPRESSED HSP'S IN ABALONE 699 90 - 1 1 80 - V. T3 70 - 60 - > •D 0) s 50 - 1 i < m (0 40 30 1 1 1 < n E 20 - V E 'E c < 10 - n ^ ^ < 24 25 26 27 28 29 Temperature (°C) 25 26 27 Temperature {°C) 28 29 Figure 3. Responses of //. rubra to heal stress. Scattergrams showing percent of animals remaining attached as temperature increases for oceanic abalone (■) and estuarine/hav abalone (D). temperature over the 7-day period for controls was 15.6 ± 2.36"C. The final temperature at which heat stress tissues were sampled was 25.4°C (i.e. an increase of 9.4°C). Western Blotting The polyclonal antibody (Ab) raised in goat against human heat inducible HSP7() bound to abalone proteins appro.ximately 70 kDa (Fig. 4). A more intense band was observed in muscle proteins from heat-shocked animals. It is possible that the reactions ob- served in the salinity stressed or control animals were due to un- degraded 70 kDa molecules which had been induced as a result of prior stress, such as transport to the laboratory. In additional West- em blots, preabsorption of the Ab with HSP70 protein produced no band, indicating that the HSP70 Ab was specific for abalone HSP7(). Dot-blots of Total Protein The results of dot-blots on total proteins isolated from heat shocked and control foot muscle of abalone are shown in Fig. 5. There was an increased expression of HSP70 in all 4 heat shocked animals compared with the 4 controls. Heat shocked animals 3 and • 4. both of which showed the least stress to heat (ie. attached to substrate and showing no swelling of tissues), also produced the greatest amounts of HSP70. with the detection limit occurring at 1 ixg of total protein. This increase in HSP70 was a 10 to 100-fold order of magnitude compared with controls (detection limit 10- 100 fjLg of protein). Total protein dot-blots were not performed on salinity stress animals, as HSP70 expression in the foot muscle of H. nihni did not appear to be upregulated by the stress (Fig. 4). Cloning of Induced Heat Shock Gene Regions from Abalone Genomic DMA DNA extraction and purification from H. rubra tissues was carried out and confirmed by the presence of a strong DNA band at approximately 23 kb in agarose gels (data not shown). There was almost no shared DNA. PCR of a 192 bp HSP70 fragment, using H.SP70f and HSP70r primers, was not successful. Upon cloning and sequencing of the fragments in three separate attempts, it was apparent 1 77 bp frag- ments had been cloned and was not from the HSP70 ATPase region using BLAST searches in Genebank. PCR of a 1400 bp HSP70 fragment from genomic DNA, using primers HSP70f and AbHSPr, was successful (Fig. 6A). The 1400 bp fragments were cloned and white colonies were obtained with plasmids having the correct size inserts (Fig. 6B). Nested PCR using primers (HSP70f and HSP70r) were used on the plasmid vj O * (» Figure 4. VVeslern blot using a HSP70 specific antibody on total pro- teins extracted from abalone foot muscles. Western blot showing 70kl)a products following membranes being probed with goat anti- human HSP70 polyclonal I antibody, then visualised via a HRP con- .jugated anti-goat 2 antibod>. Lanes: 1 and 5, protein MWM of 118, 85, 61. 50 and .^8 kDa (Life Technologies!: 2, heal shocked abalone proteins: 3, salinity shocked abalone proteins 4, control non-stressed abalone proteins. Figure 5. Dot-blots of total protein from heat shocked and control H. rubra. Membrane showing proteins evtracted from foot muscle of heat shocked and control abalone bound and probed with specific goat anti-human HSF70 polulonal I antibody visualised \ia a HRP con- jugated anti-goal 2 antibody. Lanes: 1-4 heat shocked animals: 5-8, control animals. Protein concentrations (top to bottom row l: 10(1 pg, 10 Mg. 1 Mg. 0.1 pg. 0.01 Mg- 700 12 3 Drew et al. 12 3 4 5 6 ♦ft- fcn*_ =J^ ~ Figure 6. Analysis of PCR products obtained In the cloning of a 1400 bp HSP70 fragment from H. rubra DNA using 1.5'7f agarose gel electrophoresis. A. PCR products amplified using primers AbHSPr and HSP70f. l.anes: 1, 1 kb marker (New England BioLab): 2, 1400 bp fragment plus non-specific bands, 3. negative control (no cDNA). B. PCR screening of bacterial colonies for a 1400 bp HSP70 insert using M13 and T7 vector primers. Lanes: 1. 100 bp MWM (Life Technologies), starting with a 200 bp marker (bottom); 2, negative (no insert. 167 bp of vector): 3 and 4, false positives (small inserts): 5, positive ( 1400 bp insert plus 167 bp of vector), 6, 1 kb MWM (New England BioLabs). C. Nested PCR on the vector containing the 1400 bp insert using HSP70f and HSP70r primers. Lanes: 1, 100 bp marker (MBI), 2, positive (192 bp fragment): 3, negative plasmid control. conI:jining the 1400 bp insert, and this gave a 192 bp product (Fig. 6C). These pniiier sets were subsequently used to identify induced abalone HSP70 sequences in cDNA. Cloning of Expressed Heal Shock Sequences from cD.VA High quality total RNA (182 ng/jil and a purity of 1.7) was extracted from tissues of abalone that had been heat shocked. This RNA showed characteristic 26s and 18s ribosomal RNA (rRNA) bands amidst other RNA's. The conversion of niRNA into cDNA was confirmed by successfully performing PCR on the cDNA using actin specific primers to produce a 474 bp fragment with >84% homology to three invertebrates (data not shown). PCR of a HSP70 192 bp fragment from heat shocked abalone foot muscle cDNA gave a product running at the appropriate size (Figure 7A). This fragment was purified and cloned to produce white colonies containing plasniids with the inserts (Fig. 7B). Se- quencing of a plasmid insert, along with similarity searches in GeneBank using BLAST, revealed that the clones were part of an abalone HSP70 sequence (Fig. 8). The sequence of 192 bp aligned with the same nucleotide sequence of other organisms with most homology to B. glabraia and Drosophila at 76% and 73%, respec- tively (Fig. 8A). A deduced amino acid sequence (AA). when aligned with known AA sequences, showed highest homologies to two other molluscs, B. ghihrcila and A. califoniica. exhibiting ho- mologies of 87% and 84%, respectively (Fig. 88). The 192 bp sequence was also compared with known consti- tutive heat shock proteins (HSC's) and relatively lower homolo- gies were observed in this region of H. rubra HSP70 (Fig. 8C). As well, these HSC70 sequences all had the GTC triplet (see Fig. 8A), encoding valine, absent. These data give a strong indication that the sequence obtained is in fact that from an induced HSP70. PCR of a 1400 bp and 480 bp fragments from abalone heat shock cDNA were successful. However, several attempts to re- amplify and clone these fragments were unsuccessful. DISCUSSION The blacklip abalone (W. rubra) is primarily found in southern Australian waters and is an economically important species to Australia, worth millions of dollars a year through exports alone. The condition of the animals in wild-catch and cultured stocks. including meat weight, can affect their economic value. Stress is especially seen in summer and during periods of high rainfall which affect water temperature and salinity. These conditions ap- ply particularly to those species in estuarine/bay environments. The heat shock response (HSR) is imperative to the survival of organisms and protects against protein aggregation and cellular damage, notably when an organisit) is under stressful conditions. These stressful conditions lead to the upregulation of HSPs, in- cluding HSP70. Our research indicates that the bay abalone are more physi- ologically adapted to respond to hypo-osmotic stress than the ocean abalone. Most marine molluscs, including presumably aba- lone, are osmoconformers, which means that the extracellular fluid is almost iso-osmotic with the environment. Usually, such animals A B Figure 7. .\nalvsis of PCR products obtained in the cloning of a 192 bp HSP70 fragment amplified from H. rubra cDN.\ using l.ttVc agarose gel electrophoresis. .\. PCR using HSP70f and H.SP7(lr primers. Lanes: 1, 100 bp ladder (MBI), starting with a 200 bp marker (bottom): 2, RNA negative control; 3, 192 bp product, which v*as subsequently isolated for T.\ cloning. B. PCR screening of white colonies for a 192 bp insert using primers M13 and T7. Lanes: 1, 100 bp ladder (MBI): 2. negative (167 bp of vector pCR2.1 with no insert): 3, 4 and 5. positive (359 bp including a 192 bp insert). Identification of Expressed HSP's in Abalone 701 H. rubra ( 1) B. glabrata ( 1) D. melanogaster { 1) H. rubra { 51) B. glabrata ( 51) D. melanogaster { 51) H. rubra (101) B. glabrata (101) D. melanogaster ( 101 ) H. rubra (151) B. glabrata (151) D. melanogaster ( 151 ) GGATTATCAACGAGCCGACGGAAGCTGCACTGGCGTATGGCTTGGACAAG .A. .A T C..A.CC C A C.T A . . . . C C. . . . CG . . G C. . C. . GO AATCTGAAAGGTGAGAAGAATGTATTGGTGTATGATCTGGGAGGCGGTAC GG . . AT A T . . . A . T . TC. . C T . . T . . C . . . . C . . CC . G . . G . . . CGC . . C . . GC. CA . C . TC. . CT . . . . C C. . CTTCGACGTCTCCGTCCTCACCATCGACGAAGGGTCGATGTTTGAAGTGA T AG.A..T.G T C. A .... G G . . A . . TC. . . . C. . G . . CC GATCAACAGCAGGGGATACCCATCTCGGAGGCGAGGACTTCG AGG .T..T..T..C..C G T..A T. .GG.C..C..C..A..C..GG.C..G..C..T B H. rubra ( 1) IINEPTEAALAYGLDKNLKGEKNVLVYDLGGGTFDVSVLTIDEGSMFEVR B. glabrata ( 1) A GH IF I K A. californica ( 1) A GQ....H..IF A I...K H. rubra (51) STAGDTHLGGEDF B. glabrata (51) A A. californica (51) A H. rubra (HSP) A. californica (HSC) C. griseus (HSC) D. rerio (HSC) 1 ) ATTATCAACGAGCCGACGGAAGCTGCACTGGCGTATGGCTTGGACAAGAA 1) ..C T..A..C..A.CC CA.T..C..C..TC 1) T..A..A..T.CT TA.T..T GC.A..T 1) ..C T..A..A..T.CT TA.T..T T A.. H. rubra (HSP) ( 51) A. californica (HSC) { 51) C. griseus (HSC) ( 51) D. rerio (HSC) { 51) H. rubra (HSP) (101) A. californica CHSCj (101) C. griseus (HSC) (101) D. rerio (HSC) (101) TCTGAAAGGTGAGAAGAATGTATTGGTGTATGATCTGGGAGGCGGTACCT ...G...C..TC.CA.C.T T..T..T .A.G GC.CA.T.T...CT T T. AG. TGGCAC. GG. TGG. .C. GG. TGGT.C GA CC.A.T.T T..T..T..C..T. TCGACGTCTCCGTCCTCACCATCGACGAAGGGTCGATGTTTGAAGTGAGA ....T..G...A T..G..C. . C. .C G....A. .T..T..G..TA T..T..G..T. . A. .T CA. .T..T..G..AA.T T..G..T. . C. .C..C..G..C.A. Figure 8. Sequence alignments of HSP7U ATPase region from a 192 bp PCR fragment of H. rubra heat shock cDNA compared with other invertebrates. A. Nucleotide sequence shows homologies of 7b9c and 73% with others HSP70s. Letters in bold print represent primer sequences. B. Deduced amino acid sequence shows homologies of 87% and 84%, respectively (AA residues 175-238 of HSP70s). C. HSP7() nucleotide sequence shows homologies of 69%. 69%^ and 68% with HSC70s. Region in italics and underlined indicates a region of low homology. Three nucleotides in bold (encoding valine! are absent in HSC70s. protect cells from volume perturbation, due to altered salinity, by modulating the concentration of free amino acids and other organic solutes in the intracellular compartment (Burton 1983). In these marine invertebrates exposed to low salinity, a rapid increase in fluid volume is compensated in a matter of hours (good volume regulators) or days (weak volume regulators) (Oglesby 198 1 . Dav- enport 1985). The success with which an osnioconfomiing species tolerates environmental dilution can vary between populations of the same species, depending on the variability of the salinity nor- mally encountered in the environment. For example, in a study on the bivalve, Geukensia deinissa. animals from an estuarine envi- ronment tolerated reduced salinities better than conspecifics from a higher salinity site (Garthwaite 1989). In the current study. v\e observed that the Port Phillip Bay population, which is exposed to a greater salinity range than oceanic populations, compensated for volume gain in reduced salinity faster and more effectively than the oceanic abalone (Fig. 1 ). Such a difference may be genetically determined because the bay abalone have a greater than 12% ge- netic difference compared to the oceanic populations (Huang et al. 2000). Studies on other molluscs have indicated that genetic varia- tion at the LAP locus is correlated with populations which are distributed along a salinity cline (see for example, Garthwaite 1989, Gardner & Palmer 1998). However, we were not able to correlate environmental salinity with changes in the expression of HSP 70. In our heat stress experiments, the upper lethal temperature for H. rubra was not determined, but the 50% CTM and the final temperature at which all animals had dropped off the plates were not significantly different. Similar 50% CTM's have been ob- served for H. midae at 27.9°C (Hecht 1994), and H. rubra at 26.9°C 702 Drew et al. and H. laevigata at 27.5°C (Gilroy & Edwards 1998). Hecht (1994) infers that differences in 50% CTM are considered to be of a greater zoogeographic significance than fixed physiological dif- ferences since they reflect adaptation to local environmental con- ditions. We had expected differences in 50% CTM as the bay populations can experience summer seawater temperatures of 30°C for short periods (unpublished), and there appears to be approximately a 12% genetic disergence between bay and oceanic populations (Huang et al. 2000). In the cuiTent study, attempts to characterise the HSR in H. rubra during stress by expression of HSP70 proteins, proved to be successful. Proteins were identified from the foot muscle of aba- lone through Western blotting and upregulation of HSP70 was observed in heat-stressed animals via dot-blots of total protein. All sample groups in the experiments showed HSP70 expression with heat-shocked animals displaying obvious upregulation over salin- ity-stressed and control animals. Control and salinity-stressed ani- mals did display some expression of the HSP70 protein. Given that antibodies were specific for an induced HSP70. it may be assumed that control animals in these experiments were under some form of minimal stress and were expressing induced HSP70. which may be basal level expression. All animals would also be expressing HSC70s. the non-induced form of the HSP70 family, but these proteins could not be monitored because of specific recognition of the antibody for the induced member. This assumed basal expres- sion of HSP70 might not necessarily be basal expression at all. but a response to one of many stresses the control animals experienced during their transport and maintenance, similar to that observed in a study of the HSR response in Mytilus califonuanus (Roberts et al. 1997). Therefore, it would be favorable to execute the experi- ments after one or two months of moving the animals to the new environment because, after initial expression, HSP70s are not de- graded for approximately two weeks, assuming no further stress is encountered (Sanders 199.'^). L'pregulation of HSP70 during heat shock, as seen by the 100- fold increase in expression of HSP70 proteins in H. rubra, has been previously well-documented in a variety of organisms, in- cluding other marine animals. A reported maximum 1000-fold increase of HSP7() has been observed after a heat shock in D. mekiiuigasrer (Parsell & Lindquist 1994). but levels resembling those of H. rubra have been monitored in oyster haemocytes (Ti- rard et al. 1995). cells originating from the fresh water snail B. gtahrala (Laursen et al. 1997). the mussel M. californiaiuis (Rob- erts et al. 1997), and the lamprey species Petroinyzon marinus and Lainpetra appendix (Wood et al. 1998). Genetic differences in the HSR between closely related species was reported in Collisella limpets. C. scabra and C. pelta (Sanders el al. 1991). This study monitored HSP expression under heat stress in both species, and found HSP70 expression was induced at lower temperatures, and at increased levels, in C. scabra. compared with C. pelta. An upregulation of HSP70 was not observed in salinity- stressed H. rubra. This lack of increased expression of HSP70 does not imply that the HSR was not engaged in H. rubra, as it could be due to the expression of proteins other than HSP70. This has been observed in Eunteniora affinis (Gonzalez & Bradley 1994). In this study, osmotic shock did not increase the expression in HSP70. However, new proteins in the ranges of 24-29 kDa, 45-50 kDa. 55-65 kDa and 75-85 kDa were expressed in response to changing salinities. Because the antibody used in our H. rubra study was against HSP70 specirically. it would not detect other types of expressed proteins. In contrast. HSP70 was induced by hyperosmotic salinity stress in Sahno salar (Atlantic salmon), to- gether with another stress protein, Osp54 (Smith et al. 1999). The increase in expression of proteins was accompanied by an overall decrease in other cellular proteins during salinity stress, which suggests that stress proteins were preferentially expressed. These findings present the first substantial evidence that HSP70, specifi- cally, is expressed under salinity stress and it is hypothesised to function in restoring osmotic homeostasis and renaturing proteins that have been destabilized during osmotic stress. The molecular cloning of HSP70 fragments using PCR of ge- nomic DNA and cDNA was initiated to characterise and identify genes of H. rubra that may be induced in response to stress. In particular, sequencing of the 3' variable end of an expressed niRNA would be useful in production of oligonucleotide probes for studying regulation of HSP70 in northern blots and for cDNA library screening. The cDNA sequence of the 192 bp region of HSP70 ATPase region had high homology to HSP70s of other species. This se- quence was unlikely to be constitutive HSC70 because there was less homology (<69%). In addition, a triplet at position 218 was absent from all the HSC70s. The HSC70s contained a region of low homology (encoding AA positions 188-191 in HSP70). It has been reported that position 190 in the HSP70 AA chain is a region not involved in tertiary structure formation in the overall func- tional HSP70 protein, and is therefore prone to mutation (McKay et al. 1994). The overall evidence indicates that the sequence ob- tained from H. rubra was in fact that from the heat-induced HSP70. The two larger fragments PCR amplified from cDNA (viz 1400 bp and 480 bp) show that we now have tools to further in\'estigate the nature of the HSR in abalone. Sequence data, particularly of the 3' end of an induced HSP70. would provide probes for future use in expression patterns of HSP70 niRNA and HSP70 proteins during stress. Such techniques have been employed previously using HSP70, where a nucleic acid HSP70-binding-protein was detected (Raynes & Guerriero, 1998). Overall, the current study has shown that abalone respond to heat stress with expression of HSP70. The identification of addi- tional molecules induced by heat stress, and the identification of a constitutive HSC70, are still to be carried out. In addition, the physiological factors responsible for the difference in volume regulation ability between bay and oceanic abalone remain to be elucidated. LITERATURE CITED Bosch. T. C, S. M. Kr>lovv. H. R. Bode & Steele RE. I9S8. Thermotol- erance and synthesis of heat shock proteins: These responses are present in Hydra alteniiala but are absent in Hydra olif^actis. Proc. Nat. Acad. Sci. USA 85:7927-7931. Burton, R. F. 1983. Ionic regulation and water balance. In: A. S. M. Saleuddin & K. M. Wilbur (eds.). The Mollusca. Vol 5 (part 2). Aca- demic Press. New York: pp. 291-352. Chen, J. C. & W. C. Chen. 1999. Temperature tolerance of Haiiotis diver- sicolor supertexta at different salinity and temperature levels. Camp. Bwchem. Pby.siol. 124:73-80. Identification of Expressed HSP's in Abalone 703 Chomczynski. P. & N. Sacchi. 1987. Single-step method ot RNA isolation by acid Guanidinium Thiocyanate-Phenol-Chloroform extraction. Ann- Ivl. Bioch. 162:156-159. Davenport, J. 1985. Osmotic control in marine animals. In: M. S. Laverack. editor. Physiological Adaptations of Marine Aniinals. Cambridge. UK: The Company of Biologists, pp. 207-244. Gardner. J. P. A. & N. L. Palmer. 1998. Size-dependant, spatial and tem- poral genetic variation at a leucine aminopeptidase (LAP) locus among blue mussel [Myiiliiis f^alloprovincialis) populations along a salinity gradient. Mar. Biol. 132:275-281. Garthwaite. R. L. 1989. Leucine aminopeptidase variation and fitness pa- rameters in the estuarine bi\al\e Geiikfiisui dt'ini.\sci. Mar. Biol. 10.^: 183-192. Gehring. W.J. & R. Wehner. 1995. Heat shock protein synthesis and thermotolerance in Caiaglyphis. an ant from the Sahara desert. Pnu: Nar .Acad. Sci. USA 92:2994-2998. Gilroy. A. & S.J. Edwards. 1998. Optimum temperature for growth of Australian abalone: preferred temperature and critical thermal maxi- mum for blacklip abalone. Halitotis rubra and greenlip abalone. Huli- lolis lae\ii;ara. .Aijuaculntre Res. 29:101-105. Gonzalez. C. R. M. & B. P. Bradley. 1994. Salinity stress protems in En- ryiemora aff'inis. Hydrobifllof>ia 292/293:461—168. Hecht. T. 1994. Behavioural thermoregulation of the abalone. Halioris midae. and the implication for intensive culture. ALjuucidture 126:171- 181. Hofmann. G. E. & G. N. Somero. 1996. Interspecific variation in thermal denaturation of proteins in the congeneric mussels Mytilus irossulus and M. f;alloprorincialis: evidence from the heat shock response and protein ubiquitination. Mar. Biol. 126:65-75. Huang. B. X.. R. Peakall & P. J. Hanna. 2000. Analysis of population genetic structure of blacklip abalone ifialiotis rubra) using R.APD. minisatellite and microsatellite markers. Mar. Biol. 136:207-216. Laursen, J. R.. di Liu Hong. Xiao-Jun Wu & T. P. Yoshino. 1997. Heat- shock response in a molluscan cell line: Characterisation of the re- sponse and cloning of an inducible HSP70 cDNA. / Invert. Pathol. 70:226-233. McKay, D. B., S. M. Wilbanks, K. M. Flaherty. J. H. Ha, M. C. O'Brien &. L. L. Shirvanee. 1994. Stress-70 proteins and their interaction with nucleotides, pp. 153-179. In: R. I. Morimoto. A. Tissieres & C. Geor- gopoulos. (eds.). The Biology of Heat Shock Proteins and Molecular Chaperones. Cold Spring Harbor Laboratory Press. Cold Spring Har- bor. Oglesby, L. C. 1981. Volume regulation In aquatic invertebrates. J. Exp. Zool. 215:289-301. Parsell, D. A. & S. Lindquisi. 1994. Heat shock proteins and stress toler- ance, pp. 459—^94. In: R. I. Morimoto, A. Tissieres and C. Georgopo- ulos, (eds.). The Biology of Heat Shock Proteins and Molecular Chap- erones. Cold Spring Harbor. NY: Cold Spring Harbor Laboratory Press. Raynes. D. A. & V.. Jr. Guen-iero. 1998. Inhibition of HSP70 ATPase activity and protein renaturation by a novel HSP70-binding protein. J. Biol. Chem. 273:32883-32888. Roberts, D. A., G. E. Hofmann & G. N. Somero. 1997. Heat shock protein expression in Mytilus califoniianus: Acclimatization (seasonal and tidal-height comparisons) and acclimation effects. Biol. Bull. 192:309- 320. Sanders, B. M. 1993. Stress proteins in aquatic organisms: An environ- mental peipesctive. Crit. Rev. To.xicol. 23:49-75. Sanders, B. M., C. Hope, V. M. Pascoe & L. S. Martin. 1991. Character- ization of the stress protein response in two species of Collisella Lim- pets «ith different temperature tolerances. Physiol. Zool. 64:1471- 1489. Smith. T. R.. G. C. Tremblay & T. M. Bradley. 1999. Hsp7l) and a 54kDa protein (Osp54) are induced in salmon {Salmo salar) in response to hyperosmotic stress. J. Exp. Zool. 284:286-298. Stuart. R. A., Cyr, D. M. & W. Neupert. 1994. Hsp70 in mitochondrial biogenesis: From chaperoning nascent peptide chains to facilitation ot protein degradation. Experientia 50:1002-101 1. Tirard. C, T., R. M. Grossfeld, J. F. Levine & S. Kennedy-Stoskopf. 1995. Effect of hyperthermia in vitro on stress protein synthesis and accu- mulation in oyster haemocytes. Fish Shellfish Imnuinnl. 5:9-25. Wilkins, R. C. & J. T. Lis. 1997. Dynamics of potentiation and activation: GAGA factor and its role in heat shock gene regulation. .Wucl. Acids Res. 25:3963-3968. Wood. L. A.. I. R. Brown & J. H. Youson. 1998. Characterization of the heat shock response in the gills of sea lampreys and a brook lamprey at different intervals of their life cycles. Comp. Biochem. Physiol. Part A 120:509-518. .Iinirihil i>f Shclll'isli Rcstanh. Vol. 20. No. 2. 703-710. 2(101 LOCATION OF EGG-LAYING HORMONE IN REPRODUCTIVE STRUCTURES AND NEURONS OF HALIOTIS RUBRA (LEACH) USING ANTIBODIES RAISED AGAINST RECOMBINANT FUSION PROTEINS SCOTT CUMMINS.' AMPORN THONGKUKIATKUL.- AND PETER J. HANNA' * School (if Biological & Chemical Sciences, Deakin University. Geelong. VIC 3217. Australia; -Department of Biology. Burapha University. Bangsaen. Chonhiiri. Thailand .\BSTK\CT Recomhinant abaUine egg-laying hormone was produced using a bacterial expression vector. This required directional clonnig of a lOS-bp abalone egg-laying hormone (aELH) gene sequence, using PCR of genomic DNA with primers incorporating restnction enzyme sites, into a pGEX-2T vector. Following transformation of the recombinant vector into Escherichia coll, a GST:aELH fusion protein was produced in large quantities and then purified. This fusion peptide was used to immunize mice for the production of polyclonal and monoclonal antibodies, which were subsequently tested for specificity using ELISA's and Western blots, .-kntisera and two IgM monoclonal antibodies were shown to react with aELH. The.se antibodies were used in immunocytochemistry studies of neural and gonad tissues of sexually mature female abalone. The aELH was found to be located in neurosecretory cells of cerebral and pleuro-pedal ganglia, slatocysls. and Irabeculac in female gonads. KEY WORDS: egg-laying hormone. aELH. fusion protein, monoclonal antibodies, immunocytochemistry INTRODUCTION Iniiiuinocytochemistry has previously been used to characterize central neurons which react to antibodies raised against neuropep- tides, including the a-caudodorsal cell peptide (t«-CDCP) and cau- dodorsal cell hormone (CDCH) in Lymnaea stagnalis, and corre- sponding egg-laying hormone (ELH) in Aplysia californica (Van Minnen et al. 1989, Nambu & Scheller 1986). These antibodies allowed the identification of the neurosecretory cells controlling egg laying. In Lymnaea. the neurosecretory cells are located w ithin the caudo dorsal cells (CDC) of the cerebral ganglia. The CDCs are arranged in two clusters in the left and right cerebral ganglia. In Aplysia, bag cells of the abdominal ganglion appear immuno- reactive for ELH. The bag cells consist of two clusters of 250 to 400 cells located in the abdominal ganglion, near the origin of the pleurovisceral connectives. Further investigations have indicated that the egg-laying pre- prohormone is relatively conserved across a wide range of mol- luscan classes (Nambu & Scheller I9S6). resulting in the identi- fication of additional peptides associated with neurons, which pos- sibly control egg-laying and egg-laying behavior. Van Minnen et al (1992) have found that a number of gastropod molluscs, includ- ing Helix aspersa. Biomphalaria glahrara and Liiiuix luaxiiimis. contain immunoreactive proteins using antibodies raised against a-CDCP, a short peptide within the CDCH preprohormone. Using these antibodies, as well as antibodies against CDCH in cross- reaction studies, it has also been shown that neurons in the bivalves Mytihis. Mya and Placopecten also contain a similar vitellogenic factor (Croll et al. 1993). These selective immunological markers, therefore, suggest that related peptides may be involved in the egg laying of a wide range of gastropod and bivalve molluscs. In addition, peptides immunoreactive to antisera specifically directed against CDCH, a-CDCH and p-CDCH (another short peptide in the CDCH preprohormone), have been detected in the central nervous system of the rhyonobdellid leech, Theromxzon tessiilatum (Salzet et al. 1997). As well, it has been shown that polyclonal antisera directed against the same three peptides results in positive immunoreaction in Sarcophaga bidlata (Diptera), Lep- 'Corresponding author. E-mail address: pjh@deakin.edu.au tinalarsa dcccmlincata (Coleoptera), Locusia migratoria and Periplaneta americana (Orthoptera) (Theunis et al. 1990). Abalones are classified as primitive gastropods and have a nervous system containing cerebral ganglia rather than a defined brain. Nerve cords and connectives lead from the cerebral ganglia, two of which join to a single pleuro-pedal ganglion. This ganglion is composed of fused pleural and pedal ganglia, both of which are elongated and flattened (Crofts 1929). A common feature of all the.se ganglia is the pos.session of an outer cortex of ganglion neurons and glial cells surrounding a central neuropril. The small neuronal cells and glial cells lie below the surface layer but do not extend into the neuropril. Neurosecretion in the cerebral ganglia was investigated in H. discus lumnai, in order to determine the role of hormones in the regulation of reproduction (Hahn 1994). The study showed that four types of cells exist in the cerebral ganglia, and these were defined as cell types A, B, C and D, of which only cell types A and B appear to be neurosecretory. Cell type A shows a correlation with vitellogenesis and gametogenesis in the ovary (Hahn 1994). More recently, similar neurosecretory cells of cerebral ganglia have been identified in H. asinina (Upatham et al. 1998). In other experiments, Yahata (1973) induced spawning of abalone by in- jection of homogeni/ed pleuro-pedal and visceral ganglia of ma- ture females, indicating the existence of an egg-laying hormone. However, injections of homogenized cerebral ganglia produced no notable change in the ovaries. Despite these cytological and physi- ological studies indicating the existence of sets of neurosecretory cells in the ganglia, there is no definitive evidence that peptides produced in these cells are regulating reproduction and growth. The nucleotide sequence of an abalone ELH (aELH) has been recently obtained for H. rubra (Wang & Hanna 1998). and other abalone sequences (Hanna el al. 20(X)). all of which show high homology. Fig. I shows the egg-laying precursor peptide arrange- ments of L. stagnalis and A. californica. together with the location of the sequence encoding the aELH of H. rubra. The nucleotide homology of the aELH is 94^!^ compared with the CDCH of L. stagnalis and 369r compared with the ELH of A. californica. Despite this, very little is known about the regulation of egg- laying hormones during reproductive cycles in the genus Haliolis. and the sites at which aELH is expressed. Therefore, research was 705 706 Cummins et al. L'/mn3e2 P P P CiFi a: CDCH WH2 AP III! II I C H=l Ab alone Aplysis P y E TTTT X] aELH ELH KH2 AP II II Ab alone L XJ iELH Figure I. Comparison of Lymmtea CDCH (A, Vreugdenhil et al. 1988) and Aplysia (B. Scheller el al. 1982) egg-lavins preprohormone sequences Hilh Ihe known abalone sequence. NH2 represents a potential aniidation signal. Percentage nucleotide homology between corresponding regions are indicated. Positions of potential peptides are indicated b> p, ,-CDCP/p-BCP, 7-BCP. a-CDCP/aBCP. o-BCP. calfluxin (CaKl) and an acidic peptide (.\P). Known or potential clea>age sites are shown with black bars, and the larger black regions (left) represents signal sequences at the 5' region. carried out to develop antibodies for subsequent use in immuno- logical techniques to locate cells producing egg-laying hormone within abalone neural and reproductive organs. MATERIALS AND METHODS Recombinant Protein Expression and Purification Recombinant aELH was expressed in large quantities and pu- rified by the use of a pGEX-2T (Amersham) expression vector in which the 108-bp egg-laying hormone sequence of H rubra had been inserted (Wang & Hanna 1998). This recombinant plasmid was transformed into E. coli competent cells and colonies selected on LB medium containing ampicillin. A lO-mL bottle of pre- warmed 2x LB/Amp medium was loop inoculated with cells from one colony, and incubated at 37°C with shaking overnight. This culture was transferred into 1 L of the same medium and grown at 37°C with shaking until A,,,,,, reached 0.5-0.2. Bacterial fusion proteins were induced by adding 0.5 niL of lOOmM IPTG with incubation for another three to five hours. The culture was centri- fuged at 8.000 rpm for 3 min at 4°C (Beckman. JA-20) in 40-mL conical flasks. The supeniatants were discarded and the tubes placed on ice. The pellets were resuspended in ice-cold PBS and the cells sonicated for five 30-second intervals on ice. ensuring that no frothing occurred. Immediately following this, 10% Triton X-100 was added to make final concentrations of 1% and the solutions were then incubated for 30 minutes with gentle mixing on ice to aid in solubilization of fusion protein. The suspensions were centrifuged at 10.000 rpm for 10 min at 4°C and supematants transferred to 50-mL centrifuge tubes. Glutathione .sepharose 4B beads were mixed with each 100 mL of sonicate and the contents left to stand for one hour. The glutathione sepharose 4B beads with fusion protein were sedunented by centrifugation and the super- natant collected for later reference. Ice-cold PBS was added and beads washed three times by repeated centrifugations. The fusion protein was then eluted with glutathione elution buffer, and analy- sis of fusion protein was performed by SDS-PAGE and Coomassie blue staining. Production of Antibodies Antigens used in primary immunizations were prepared by mixing 50 p-g of the fusion protein with an equal volume of Freund's complete adjuvant (Sigma). BALB/c mice were then im- munized intraperitoneally. Subsequent immunizations were per- formed each wk. for 3 wks. using the same dose of fusion protein but mixed with Freund's incomplete adjuvant (Sigma). Polyclonal antisera were obtained by tail-bleeding mice and stored for use when ELISA end-point litres of > 10,000 were ob- tained. For immunocytochemical studies, antisera were diluted 1/500 to 1/5000. Monoclonal antibodies were prepared four days after the last injection of immune mice, using standard techniques (Coding 1987). The mice were killed by cervical dislocation and the spleens removed aseptically. Spleen cells were fused with Sp2/ O-Ag-14 myeloma cells using PEG 4000. The supematants of the hybridomas surviving the selective medium consisting of RPMI 1640, bovine fetal calf serum and HAT, were tested for the pres- ence of specific antibodies by means of ELISA. Antibodies se- creted from positi\e clones were isotyped and the cells re-cloned twice to ensure monoclonal ity. Quantification and Specificity of Antibodies in ELISA ELISAs were undertaken to detennine the titre of anti-aELH sera obtained from mice and for screening hybridomas and speci- ficity of reaction with antigen. Antigen concentrations, which were calibrated to 20 p.g/ml. consisted of expressed GST-aELH fusion protein or expressed GST. The wells of a microtitre plate (Bio-Services) were coated with 50 p-L of antigen solution, which was allowed to adsorb to the plate at 4 'C overnight. Unoccupied binding sites on the plates were blocked with 1 9f BSA in coating buffer for one hour at 37'"C. The plates were then washed three times in washing buffer and incubated with 50 |jlL of polyclonal antiserum diluted to 1:50, 1:100, 1:500, 1:1000, 1:5000 and 1:10000, in incubation buffer at 37°C for one hour. Supematants of hybridomas were substituted for the antisera in additional tests. For the detection of bound antibody, plates were washed three times in washing buffer and incubated for one hour at 37°C with 50 |jlL per well of a secondary alkaline phosphatase labeled antibody (Sigma) diluted 1000-fold in incubation buffer. Plates wei-e then washed three times in wash- ing buffer and reacted with a 50 p,g/mL solution of the enzyme substrate. After 100 min of color development, the plates were read on an automated ELISA plate reader (Titertek Multiskan MCC/340) at 405 nm. Dissection of Abalone Tissues Live female H. rubra (Leach) with mature gonads were col- lected under a research permit (97/R/049A). during the months Egg-Laying Hormone, Hauotis Rubra 707 September to November. If not used immediately they were trans- ferred to an abalone aquaculture facility where they were kept for no more than 5 days under a 12-h Ii2ht/l2-h dark cycle. Animals were removed from the shell prior to dissection. To remove the left and right cerebral ganglia (cgj. a blunt probe was used to separate the tissues from each side of the head. The eg with two nerves protruding was then detached. Next, the pleuro- pedal ganglia (ppg) and statocysts were exposed by removing all organs in the head cavity with a scalpel and then removed from within the muscle tissue. A piece of female gonad was also re- moved from the conical end of the organ. They were rinsed three times in PBS. and then incubated for 1 h at room temperature in primary antibody (i.e. polyclonal or mono- clonal). After three rinses in PBS. they were incubated in FITC- labeled goat anti-mouse secondary antibody (Silenus) in the dark at RT for 1 h, and then rinsed three more times in PBS. before mounting in FITC mounting solution. Sections were viewed under blue light with a Zeiss Axioskop MC 80 microscope and images recorded. Negative control procedures included the omission of primary antibodies and preadsorbing the anti-aELH antibody with recom- binant aELH. Fixation and Siclioiiiiif; of Tissues After dissecting out the ganglia and pieces of gonad, they were fixed in freshly prepared 4'^'f paraformaldeh\-de at 4°C for 24 h. They were then transferred to phosphate buffer, and stored at 4''C before further processing. Prior to cryosectioning. tissues were transferred to 30% sucrose in phosphate buffer and stored overnight at 4°C. The preparations were then frozen in TissueTek (Bayer Diagnostics) in the cryostat at -20°C and subsequently sectioned at 6 p-m. Alternate sections were placed on separate slides so that different antibodies or con- trol procedures could be compared. Slides were then stored at -80°C, until required. Western Blots Standard SDS-PAGE of expressed GST and GST-aELH were carried out prior to Western transfer of the proteins onto 0.45 p.m nitrocellulose membranes. The membranes were then reacted with antisera or mAbs to determine the sizes of antigenic proteins. Fresh tissues of eg, ppg and gonad were used to prepare total protein extracts using TriPure isolation reagent (Roche), according to the manufacturer's instructions. These extracts were separated on SDS-PAGE. Proteins were transferred, as before, and probed with antisera or mAbs. Immunocytochemistry A DAKO LSAB2 (streptavidin-biotin labeled) kit was utilized to identify the location of aELH in sections. Initially, sections were heated in buffered citrate (pH 6.0) to enable antigen retrieval. To prevent endogenous peroxidase activity, the slides were treated with 3% hydrogen peroxide, after which, primary antibody (di- luted polyclonal or monoclonal in ascites fluid) was added and the slides incubated for 1 h at RT. After washing with PBS. biotiny- lated anti-mouse immunoglobulins were added and the slides in- cubated for 30 min at RT. After a brief washing, two drops of streptavidin peroxidase were added and another incubation carried out at RT for 20 min. During this step, a fresh solution of 3- amino-9-ethylcarbazole (AEC) chromogen was prepared. The ehromogen was added after another washing and a further incu- bation at RT for 3 to 5 min performed, before being examined for color development. When appropriate, color development was stopped by washing the slides in distilled water, a counterstain in Mayer's haematoxylin performed, and the slides washed again and before mounting in Faramount aqueous mounting solution (DAKO). Digital images were then taken using a Zeiss Axioskop MC 80. ^ " Immunofluorescence was performed on stored frozen sections. The sections were allowed to thaw at RT for 30 min before block- ing, first in 1% glycine for 30 min, then in 4% BSA for 30 mm. RESULTS Monoclonal Antibodies Table 1 shows the properties of five monoclonal antibodies (mAbs) that were produced with medium to strong reactions in ELISA and immunofluorescence testing. MAbs with weak reac- tions were discarded. Three mAbs (i.e. F61P3E3, F62P1A1 and F62P6D4) reacted with the GST component of expressed GST- aELH fusion protein as well as expressed GST. The other two mAbs (i.e. F61P1A5 and F62P4C1) reacted with the aELH com- ponent of expressed GST-aELH fusion protein. It was unusual that most of the mAbs were IgM as a period of four days from the last immunization of peptide to the production of niAbs would have allowed for class-switching of cells from IgM to IgG's. Specificities of mAbs were also confirmed using Western blots of expressed GST and GST-aELH fusion proteins (data not shown). Western Blots of aELH in Abalone Tissues The results of Western blots of eg, ppg and mature female gonad proteins are shown in Fig. 2. There were no bands at ap- proximately 4 kDa, the estimated size of aELH (36 amino acid residues deduced from a 108 nucleotide sequence), in any of the four experimental lanes. This absence may be attributed to the small size of the aELH (i.e. 4 kDa) and relatively small amounts present, compared with the total protein present. However, the eg showed 6 bands of sizes ranging from 24 to >77 kDa, the ppg 3 bands of which two were between 34 and 48 kDa and the other >77 kDa, and the gonad showed two bands between 34 and 40 kDa. The presence of these multi-bands indicates the different process- ing of preprohormones in the tissues and/or the presence of multi- gene products. TABLE L Monoclonal antibodies produced and their properties Monoclonal Specificity Reactivity using .\ntibody Isotjpe in ELISA Immunofluorescence F6IP3E3 IgM GST +++ F62P1A1 IgM GST ++ F62P6D4 IgG, GST +-H- F61P1A5 IgM aELH -!•+ F62P4C1 IgM aELH ++ -^++. strong reaction; ++. medium reaction; +. weak reaction (inAb dis- carded) 708 Cummins et al. 77.000 48,000 34,200 28,400 20.500 Figure 2. Western blots of proteins from H. rubra tissues usin;; anti- aELH antibodies. I.anes: 1, pre-stained molecular weight markers (Bio-Rad); 2 & 3 mature female gonad; 4, pleuro-pedal ganglia; 5, cerebral ganglia. Immunocytochemistry A brick-red precipitate of the immunological test, together with a blue haemotoxylin counter stain, gave clear indications of anti- gen location. We had initially used 3.3' diamino-benzidine tetra- hydrochloride (DAB) in the final reaction, to gi\e a brov\n pre- cipitate, but found that there were endogenous brown pigments in control tissues (data not shown). This made it difficult to determine the location of the immune reaction, so this method was substituted by the AEC substrate to give a clearer result. Consequently, immunopositive reactions were observed in eg. ppg. and mature female gonads (Fig. 3). However, there were marked differences in each of the structures. The ceils reacting in the eg were equivalent to the NS 1 neurosecretory cells observed by Upatham et al. (1998). In the ppg. peripheral neurosecretory cells have also been identified by the same group (unpublished), and our current work showed these cells immunoreacli\e. Immunoreactive material was also observed around the statocysts. In the gonads, a small number of cells spread throughout the trabeculae showed strong immunoreactions. The results obtained using immunofluorescence location of aELH were the same as for immunoenzyme staining (Fig. 3). DISCUSSION Production of immunocytochemical probes has relied exten- sively on the use of native protein to provide specific polyclonal and monoclonal antibodies. However, there has been a transgres- sion toward the use of synthetic or recombinantly expressed pro- teins, to facilitate antibody probe production. The advantage of these proteins is three-fold: large amounts of protein can be pro- duced; only partial sequence data needs to be obtained; and labo- rious protein purification techniques can be avoided. In this study, the aELH sequence data obtained (Wang & Hanna 1998) was used for the in vitro expression and subsequent immunization of experi- mental animals. Antibodies were then produced that with immu- noreactivity toward the recombinant protein in ELISA. However, further analyses were required to determine immunoreactivity to native protein in vivo, and this was ratified by immunocytochem- ical studies. This was not unexpected as previous studies have shown that antibodies raised against native ELH protein of Aplvsia is immunoreactive against a synthetic ELH peptide. In addition to ELISA, further specificity of the monoclonal and polyclonal antibody immunoreacti\ity to aELH was achie\'ed by Western blots. These confirmed that the antibody probes were reactive to a protein extracts from the central nervous system and the gonad. The antibodies against aELH recognized proteins of various sizes within gonad, cerebral and pleuro-pedal ganglia ex- tracts. This indicates that there are a number of polypeptides of different sizes, each containing the aELH peptide component. These bands probably represent differential cell processing of the polypeptides, typical of that found in A. californica (Fisher et al. 1988). They showed by Western blot analysis that antibodies rec- ognized small final-product peptides, intermediates in the process- ing pathway which contain the sequence used as the immunogen, and also the large prohormone. However, the multi-band data may also be the result of the presence of a multigene aELH family, the genes of which are expressed in different tissues. We have pre- liminary data (unpublished) indicating that abalone have several genes encoding aELH. Our data is consistent with the findings of Nambu and Scheller (1986), in their studies of Alysia genomes. They used Southern blotting, gene cloning and immunocytochem- ical techniques to identify and characterize ELH related genes. Overall, there is a need for a substantial investigation into the number of aELH related genes in the abalone genome, the cellular expression of these genes, and the processing of preprohormones in each tissue to give immunoreactive proteins of various sizes. In vivo expression of the female H. rubra aELH was shown in our studies of sections of abalone tissues, using monoclonal and polyclonal antibodies. The aELH appeared to be localized to the NSl cells within the cerebral, and were also found in pleuro-pedal ganglia, statocysts, and to the trabeculae of the gonad. Our data supports the previous research by Hahn (1994) and Upathain et al. ( 1998), in which they showed that differential staining could iden- tify neurosecretory cell types in the cerebral ganglia. In addition, the cerebral ganglia of other molluscan species are known to be a major site of ELH-like expression (Croll et al. 1993. Van Minnen et al. 1992). Immunoreactivity within the pleuro-pedal ganglia was not only localized to the cells of the periphery, but also the statocyst mar- gins and interior. The statocyst is a chambered sense organ con- taining granules to sense the direction of gravity. Despite extensive research being conducted on neuropeptide immunoreactivity within the central nervous system, little research has focused on the reproductive tissue. Our study, however, shows aELH located in the trabeculae of the mature pre-spawned female gonad. It was only located within certain cells, which may function to store aELH with the maturing gonad. It is postulated that a stimulus may act to release the hormone that acts directly on the smooth muscle of the gonad to release ova. Evidence from Droso- pltila shows that when a male mates and releases sperm, he also deposits an egg-laying hormone that induces egg-laying in the female (Park & Wolfner 1995). This hormone is related to the ones found in gastropods and other invertebrates. The large amount of aELH expressed within the reproductively pre-spawned mature female may be a function of the gastropod requirement to release high concentrations of hormone into the hemolymph. It is known in molluscs that hormone release areas are high in number (Joosse 1979). The antibodies we have produced can now be extended to further studies. They could be used to determine site-specific ex- pression of aELH during the reproductive cycle of H. rubra, and other abalone species. Preliminary tests have indicated that they do react with the aELH in H. asinina. It would be interesting to determine if bioassays of gonads could quantify the presence aELH sufficiently to indicate spawning readiness. We are already doing bioassays with the aELH to induce spawning. Egg-Laying Hormone, Hauot/s RuB/iA 709 A r r ^ -n '*^\^|| D ^- ^^.^ Figure 3. Localization ol'aELH in tissues of //. rubra using immunocytocheniistry staining. Left: Tissues incubated with murine anti-aELH prior to detection using a DAKO LSAB2 Ivil to produce bricli-red precipitate. Right: Tissues incubated uith murine anti-aKLH prior to detection using secondary incubation with FITC-conjugated goat anti-murine Ig to produce a green immunofluorescence under IV. A, immunorcacti\ity of aFLH in cerebral ganglia: B. imnumoreactivity of aELH in cells of a pleuro-pedal ganglia: C. immunoposilive material around statocysts: D. imnuniop()siti>e cells in the trabeculae of mature female gonads. LITERATURE CITED Crofts, D., R. 1929. Haliotis. Liverjiool Mar. Biol. Commiitce Memoirs 29:1-74. Croll. R., J. Nason & J. Van Minnen. 1993. Characterization of central neurons in bivalves using antibodies raised against neuropeptides in- volved in gastropod egg-laying behavior. Imeri. Rcprod. Develop. 24: 161-168. Fisher, J. M., W. Sossin & R. Newcomb. 19SS. Multiple neuropeptides derived from a ci)nimon precursor are differentially packaged and transported. Cell 54:813-822. Coding, J. W. 1987. Monoclonal Anlihoilies: Principles and Practice. London: Academic Press. Hahn. K. O. 1994. The neurosecretory staining in the cerebral ganglia of the Japanese abalone le/oawabi), Haliotis discus haniiai. and its rela- tionship to reproduction. Gen. Coinpar. Endocrinol. 93:295-303. 710 Cummins et al. Hanna. P. J.. Z. Chai. S. F. Cummins & B. Huang. 2000. Applications of molecular biology to abalone fisheries and aquaculture. J. Med Appl. Malacol. (in press). Joosse, J. 1979. Evolutionary aspects of the endocrine system and of the hormonal control of reproduction of molluscs. New York: Academic Press, pp. 119-157. Namhu. J. R. & R. H. Scheller. 1986. Egg-laying hormone genes oi Aply- sia: evolution of the ELH gene family. J. Neiirosci. 6:2026-2036. Park. M. & M. F. Wolfner. 1995. Male and Female Cooperate in the Prohormone-like Processing of a Drosophila melanogaster Seminal Fluid Protein. Dev. Biol. 171:694-702. Salzet. M., M. Verger-Bocquet. F. Vandenbulcke & J. Van Minnen. 1997. Leech egg-laying-like hormone: structure, neuronal distribution and phyiogeny. Mol. Brain Res. 49:211-221. Scheller, R. H., J. F. Jackson, L. McAllister, J. Schwanz, E. R. Kandel & R. Axel. 1982. A family of genes that codes for ELH, a neuropeptide eliciting a stereotyped pattern of behavior inAplysici. Cell 28:707-719. Theunis, W., J. Van Minnen & A. De Loof. 1990. Imniunocytochemical localization in the central nervous system of four different insect spe- cies of molecules immunoreactive against peptides present in the cau- dodorsal cells of Lymnaea skignalis. Gen. Comp. Endo. 79:415—422. Upatham, E. S., A. Thongkukiatkul, M. Krutrachue. C. Wanichanon. Y. P. Chitramvong, S, Sahavacharin & P. Sobhon. 1998. Classification of neurosecretory cell, neuron, and neuroglia in the cerebral ganglia of HuUotis asinina Linnaeus by light microscopy. J. Shellfish Res. 17: 737-742. Van Minnen, J., R. W. Dirks, E. Vreugdenhil & J. Van Diepen. 1989. Expression of the egg laying hormone genes in peripheral neurones and exocrine cells in the reproductive tract of the mollusc Lymnaea siag- nalis. Neiiroscience 33:35^6. Van Minnen, J., H. D. F, H. Schallig & M. D. Ramkema. 1992. Identifi- cation of putative egg-laying hormone containing neural systems in gastropod molluscs. Gen. Comp. Endocrinol. 86:96-102. Vreugdenhil. E., J. F. Jackson, T. Bouwmeester, A. B. Smit, J. Van Min- nen, H. Van Heerikhuizen, J. Klootwijk & J. Joose. 1988. Isolation, characterization, and evolutionary aspects of a cDNA clone encoding multiple neuropeptides involved in the stereotyped egg-laying behav- iour of the freshwater snail Lynimiea siagnali.\. J. Neurosci. 8:4184- 4191. Wang, L. & P. J. Hanna. 1998. Isolation, cloning and expression of a gene encoding an egg-laying hormone of the blacklip abalone {Halunis ru- bra Leach). J. Shellfish Res. 17:785-793. Yahata, T. 1973. Induced spawning of abalone {Nordolis discus Reeve) injected with ganglional suspensions. Bull. Jap. Soc. Sci. Fish. 39: 1117-1122. Jininiul ,if Shi'llftsh Research. Vol. 20. No. 2. 711-716. 2001. MORPHOFUNCTIONAL STUDY OF THE HEMOCYTES OF HALIOTIS ASININA S. SAHAPHONG,' * V. LINTHONG." C. WANICHANON," S. RIENGROJPITAK,' N. KANGWANRANGSAN,' V. VIVANANT/ E. S. UPATHAM,' ^ T. PUMTHONG,' N. CHANSUE," AND P. SOBHON" 'Depaifiiiait of Pathobiology, Mahidol University, Bangkok, 10400, Thailand; 'Department of Anatomy. Maliiclol Uniren^ity. Bangkok 10400. Thailand: ' Department of Biology. Mahidol University, Bangkok 10400, Thailand: '^Department of Biology. Burapha University, Chonliuri. Thailand: ^Coastal Aquaculture Development Center. Department of Fishery, Ministry of Agriculture and Cooperatives. Prachuapkhirikhun 77000. Thailand: ^Veterinary Medical Aquatic Animal Research Center. Department of Veterinary Medicine. Faculty of Veterinary Science, Chulalongkorn University, Bangkok 10400, Thailaitd ABSTR.\CT The henlocyte^ of abalone iHalio!is asiiiiiw) were studied hy light and electron microscopy in order to describe their main morphological features and to relate these to their role in immune defense. The cells are comprised of two differentiated types: agranulocyte or hyalinocyte and granulocyte. The hyalinocyte is characterized by the presence of several filopodia. a large nucleus with dense chromatin, a moderate amount of cytoplasm, microfilaments, oval and round-shaped mitochondria with rather dense matrix, a considerable amount of rough endoplasmic reticulum, few cytoplasmic granules, coated pits and vesicles, phagocytic vacuoles, and numerous large and small vacuoles. Like the hyalinocyte, the granulocyte possesses similar cytoplasmic organelles but in fewer number, and a peripheral organelle-free zone containing numerous dense granules of various types. The shape of the granules varies from round to oval to elongated forms. Several dense granules e.\hibit a crystalloid substructure that show close relationship to the plasma membrane. The average size of granulocytes is 9.68 ±1.12 p.m and hyalinocytes is 8.65 ± 0.77 |j.m. KEY WORDS: HalmUs ashiina. hemocytes INTRODUCTION Approximately 75 species of abalone have been reported in the world. Thailand harbors three species, namely, Haliotis as- iniiui, H. ovina, and H. varia (Nateewatana and Hylleberge 1986, Tookvinart et al. 1986. Nateewatana and Bussarawil 1988). Among the three species, H. asinina is the biggest and has the most economic potential because of its large proportion of fiesh and its good taste (Singhakriwan & Doi 1993). Due to the eco- nomic potential of this abalone species, a thorough understanding of the biology of H. asinina including its immune response is needed. In molluscs, hemocytes are involved in a variety of physi- ological and pathological functions including nutrient trans- port and digestion, wound and shell repair, internal defense, and exogenous and endogenous tnaterial excretion (Cheng 1981. Bayne 1983, Fisher 1986). In the literature, most of the morpho- logical studies of molluscan hemocytes reported are of bivalve molluscs; only a few exist on abalone. There has been no study on the hemocytes of H. asinina. In this report, we studied the mor- phology of H. asinina hemocytes by using light and electron tiii- croscopy. MATERIALS AND METHODS Hemolymph was withdrawn from the cephalic arterial sinus of individual abalone atid pooled. Pooled hemolymph was itnmedi- ately poured into cold 2% glutaraldehyde in 0.1 M sodium caco- dylate buffer pH 7.4. at 4°C. overnight. The hemocytes were cen- trifuged at 800 x g for 10 min at 25°C. The hemocyte pellets were post-fixed in \9r osmium tetroxide in 0.1 M sodium cacodylate t I ft^^ M> *Corresponding author. E-mail address; sessh@mahidol.ac.th * J50u Figure 1. Semi-thin section of hemocytes of//, asinina, methylene blue stain. .\l this magnification, granulocytes (G) and hyalinocytes (H) are clearly differenliated. 711 712 Sahaphong et al. buffer, at 4°C. for an additional 2 h. They were then washed with sodium cacodylate buffer, dehydrated in a graded series of alcohol, cleared in propylene oxide and embedded in Aradite 502 resin. Blocks were sectioned at one-micron thickness by an ultramicro- tome and stained with methylene bhie for light microscopic ob- servation. Ultrathin sections were cut and stained with uranyl ac- etate and lead citrate and viewed by Hitachi TEM H-300 at 75 kV. RESULTS Light Microscopy Micrographs of the semithin sections of the H. asiniiui hemocytes are shown in Figures 1 and 2. Two populations of hemocytes. granulocytes and agranulocytes or hyalinocytes. were observed based on the presence and the absence of cytoplasmic granules. The proportions of granulocytes and hyalinocytes are 11.43% and 88.57% in this present study, respectively. Granulocytes are characterized by the presence of numerous cytoplasmic granules. The granules tend to be arranged at the periphery of the cell, thus leaving the clear zone around the 0''. a: ^'% ft® ' ~ H — mi c> * ^ ® lOu % ^ ^ "f If , - *i !i B si--. ^lOu Figure 2. A-B. M higher magnincation, more details are seen. Note the peripheral distribution of the granules and perinuclear clear area in the granulocytes, while large cytoplasmic clear zones (CZ) are seen in the hyalinocytes. Different shape, size and location of the nucleus in hoth granulocytes and hyalinocytes can be ob.served. Note also the niopodia (arrow) in both cell types. nucleus. Most of the cells are spherical or slightly oval, the largest measured 13. 1 5 |jim (9.68 ±1.12 p.m in diameter on average) with some filopodia extending from the plasma membrane. The nucleus is round to oval shaped and centric with a rather small nucleus/ cytoplasmic ratio. The maximum nuclear size is 4.70 p.m with an average of 3.70 ± 0.59 p.m. Some nuclei are bilobed. elongated and indented. The agranulocytes or hyalinocytes are spherical or oval. They are characterized by large nuclei and contain one or two prominent clear cytoplasmic zones. The cells are only slightly smaller than the granulocytes. The largest cell measured 1 1.22 p,m in diameter and the average size is 8.65 ± 0.77 p,m. Similar to the granulo- cytes, the nuclear profiles were round or oval, with some bilobed and indented. Nuclei were both eccentrically and centrically placed. The maximum nuclear size of the hyalinocyte is 5.98 (xm with an average of 4.70 ± 0.60 jjtm. The nucleus/cytoplasmic ratio was relatively larger than that of the granulocyte. Filopodia were clearly visible extending from the cell. V ^ Gly f^~.. - f- 1u Figure 3. Electron micrographs of a hyalinocyte. .\. Hyalinocyte show- ing bilobed nucleus (Nl with moderate amount of heterochromatin. A lake of glycogen l(;i> ) is prominent. Few mitochondria (.Ml are seen. F. niopodia. B. In contrast to figure .\, this hyalinocyte does not contain as prominent a glycogen area. There are few small round electron dense granules (G), rough endoplasmic reticulum (RERl and mito- chondria (M). Note the presence of small vesicular bodies (arrow). MORPHOFUNCTIONAL STUDY (IF HaLIOTIS ASIN/A 713 Go \ Gly PER 0.5u B 0.5u 0.5u D ■■ ^- ■ 'i 4 -^ \ - #■ '-* \ Mt. \ Mt P D 0.5u M W :%■?. ^&i/ •~?"^ N dREB ^ dRER dRER Figure 4. A series of electron miirDnriiplis vliowinj; ditails iil'some liyalinocyte organelles. A. Hyalinocyte with spherical nucleus iNl containing a moderate amount of chromatin and a prominent nucleolus (Nul. Note the presence of a glycogen area (Glyl. Colgi apparatus (Gol, and rough endoplasmic reticulum (RER). I'M, plasma membrane. B. Hyalinocyte showing profiles of oval shape mitochondria (M) with rather dense matrix. Several small vesicles (SV) are also seen. M, mitochondria. N, nucleus. RER. rough endoplasmic reticulum. C. Well developed Colgi apparatus (Gol and residual bodies (RB) are seen here in this electron micrograph. Go, (;olgi apparatus. N, nucleus. RB. residual body. D. Microtubules (Mt) and small \esicles (SX I are seen in the cytoplasm of the h>alinoc>te. Mt. microtubule. SV. small vesicle. E. Mitochondria (M) with rather dense matrix are shown here with rough endoplasmic reticulum (RER). N, Nucleus. F. Hyalinocyte showing stack of rough endoplasmic reticulum (RER) and a few dilated endoplasmic reticulum (dRER) with light density amorphous material within cisternae. N, nucleus, RER, rough endoplasmic reticulum. 714 Sahaphong et al. Electron Microscopy The hyalinocyte is characterized by the presence of several filopodia, a large nucleus with dense chromatin and a moderate amount of cytoplasm Fig. 3. The mitochondria are round or oval in shape with rather dense matrix. A considerable amount of rough endoplasmic reticulum. Golgi apparatus, few cytoplasmic granules mostly of round shape, coated pits and vesicles, phagocytotic vacuoles, numerous large and small vacuoles including microfila- ments were observed Fig. 4. When viewed by electron microscopy, the clear cytoplasmic zones observed by light microscopy corre- sponded to the areas composed mostly of unstained glycogen with some visible glycogen particles within. In some cells the pools of glycogen are quite large and seem to push the nuclei to the pe- riphery. Like hyalinocytes, the granulocyte possesses similar cyto- plasmic organelles but in fewer numbers and a peripheral organelle free zone containing cytoplasmic granules of various shapes and sizes and various densities Fig. 5. The shapes of the granules vary from round, to oval to polyhedral elongated forms. The maxi- mum size of the granule is 0.78 p.m and the average size is 0.34 ± 0. 1 1 |xm. However, most of them are polyhedral elongated and only a few are spherical or oval. The close relationship of the granules and plasma membrane is noted. Some of the gran- ules protruded from the plasma membrane Fig 6, while others coa- lesced or fused with the cell membrane. The nucleus is rela- tively small in size, round and located either eccentrically or cen- trically. DISCUSSION In this preliminary study on the hemocytes of the H. asininu. were composed of two differentiated cell types, the granular and the agranular or hyalinocytes. Cheng (1981) suggested that hemocytes should be designated as granulocytes and hyalino- cytes. By using different techniques such as phase contrast mi- croscopy, several fixatives and stains including electron micros- copy, Foley and Cheng (1972), Foley and Cheng (1974), Cheng and Foley (1975), identified another cell type in the hemolymph of the bivalve mollusc, the fibrocyte. Fibrocytes were also de- scribed in the black abalone, Halioris cracherodii (Shields et al. 1997). In this study, we have shown that hyalinocytes contained very prominent areas or aggregates of glycogen. Cheng and Call (1974) however reported that the granulocytes not the hyalinocytes con- tained a large aggregate of glycogen in the cytoplasm. This is in contrast to our findings in which we found glycogen primarily in the hyalinocytes Fig. 3B. Of interest, is the observation of the close relationship of the granules of the granulocyte to the plasma membrane. Several gran- ules protruded from the membrane and some fused with the mem- brane. This represents the process of releasing the content of the granules or most likely the lysosomes into the serum. The migra- tion of the lysosome to the cell membrane and the extrusion from the granulocytes in Mercenaria mercenaria as evidenced by scan- ning and transmission electron microscopy were reported by Mo- handas et al. (1985) and Mohandas and Cheng (1985). Fewer organelles observed in granulocytes indicate that when a cell be- Figure 5. Electron micrographs of a granulocyte. A. Granulocyte showing different shapes and sizes of its granules. Note the peripheral distribution of the granules ((;i, several of them in close relationship with the plasma membrane (PM). As in hyalinocytes. the oval shaped nucleus contains moderate amount of heterochromalin. This cell is almost devoid of recognizable organelles. F, filopodia, N, nucleus. B. Variation of intragranular densities are illustrated in this micrograph from marked to light electron dense. The light granules (LG) contain intragranular inclusion bodies. F, fdopodia, N. nucleus, PM, plasma membrane. comes fully differentiated, reduction in the number of organelles accompanies this process. This present study shows for the first time, the presence of two main types of hemocytes in H. asinina: hyalinocytes and granu- locytes. We believe that the granules are the lysosomes based on the marked morphological resemblance to the lysosomal granules observed in the eosinophil. The particular function of each popu- lation in the defence mechanism in this species of abalone will be further studied. ACKNOWLEDGMENT This work was supported by the Thailand Research Fund (Se- nior Research Scholar Fellowship to P. Sobhon and grant No. BRG4080004). MORPHOFUNrTIONAI, STunv OF Haiioiis asinia 715 Figure 6. Variations in sliape, size, density and tlie relationsliip to the plasma nieniiirane of the cvtopiasniit granules in the granulocyte of//. asinina are illustrated in details in this series of electron micrographs (A-F). A. Klongatcd polyhedral forms. F, filopodia. (!, granule. PM, plasma membrane. B. Another yie« of the cytoplasmic granules of the granulocyte. G, granule. N. nucleus. PM. plasma membrane. C. Fusion of one of the granules It;) with the plasma membrane (PM). D. One large and one small spherical granule (Sp(;i and .Sp(;2) is seen. Note the heterogenicity of the elongated form granules, some of which contain intragranular inclusions. K. Picture frame like granule with hollow core is seen here (arrow). (;, granule, PM, plasma membrane. F. Protrusion of one of the granules iseyident here. Note the limiting membrane (arroyv) around the small round shaped granule (G). N, nucleus. PM, plasma membrane. 716 Sahaphong et al. LITERATURE CITED Bayne, C. J. 1983. Molluscan immunobiology. In: A. S. M. Saleuddin & K. M. Wilbur, editors. The Mollusca. 5. Physiology. Part 2. London: Aca- demic Press, pp. 407-486. Cheng. T. C. 1981. Bivalves. In: N. A. Ratcliffe & A. F. Rowley, editors. Invertebrate blood cells. London. New York: Academic Press, pp. 231- 300. Cheng. T. C. & A. Cali. 1974. An electron microscope study of the fate of bacteria phagocytized by granulocytes of Cra.ssoslrea virginica. Con- temp. Topics Immunobiol. 4:25-35. Cheng. T. C. & D. A. Foley. 1975. Hemolymph cells of the bivalve mol- lusc. Mercenaria mercenaria: an electron microscopical study. J. In- veitbi: Pathol. 26:341-351. Fisher. W. S. 1986. Structure and function of oyster hemocytes. In: M. Brehelin. editor. Immunity in invertebrates. Berlin. Heidelberg: Springer-Verlag. pp. 25-35. Foley. D. A. & T. C. Cheng. 1972. Interaction of molluscs and foreign substances: the morphology and behavior of hemolymph cells of the American oyster, Cms.si>. elope and in the center. The c>toplasm contains large rughj -shaped granules about 270 x 550 nni in size, «ith an electron-dense matrix (II. In B. branches of granulated cells (1) are dispersed at intervals along the basal lamina (Ba). (C-F) The nerve bundles in trabeculae consist of two types of axons that contain dense (2) and light (3) spherical granules (about 165 and 150 nm in diameter). Processes of granulated cells ( I ) are closely associated with the periphery of each nerve bundle (in D, Kl. 724 Apisawetakan et al. DISCUSSION The general histology of the gonad and classification of various types of germ cells in many species of abalone. such as. H. tiiher- culata (Stephenson 1924. Croft 1929), H. discus hannai (Toniita 1967. 1968). H. cracherodii (Webber & Giese 1969). H. nifcscens (YoLuig & DeMartini 1970. Martin et al. 1983). H. diversicolor divcTsicolor (Takashima et al. 1978). H. asinina (Apisawetakan et al. 1997. Sobhon et al. 1999) have been reported. However, most studies neglect the cellular compositions and detailed structure of the connective tissue scaffold of the gonad, apart from mentioning casually that the gonadal capsules and traheculae are made of fibro-muscular tissues. In our detailed studies using both light and electron micros- copy, we found that the cellular compositions of connective tissue scaffold are more complex than previously thought. Within these connective tissue frameworks, which may be termed trabecular- capsular compartments, there are many types of cells that may be involved in the physiology of the gonad, including the production of reproductive hormone, and the release of mature gametes. These cells are muscle cells, fibroblasts and granulated cells. The most striking feature is the presence of a large number of granulated cells, with large-endocrine like granules, in the trahec- ulae and capsules of the gonads. It is remarkable that these types of cells and their branches form an extensive network within the connective tissue scaffold of the gonad. Immunolocalization stud- ies of abalone egg-laying hormone (aELH) performed by our group demonstrated that these cells could be one of the primary producers and/or storage sites of aELH (Chanpoo et al. 2000). Even more remarkable is that there are numerous nerve processes, which consist primarily of axons containing neurochemical vesicles, coming into close contact with branches of granulated cells. These bundles of nerve and granulated cell processes are mostly observed within the gonadal capsules. Judging from this appearance, gonads of H. asinina are highly innervated organs. In contrast to vertebrates, it is possible that the nervous system still play a more direct role in controlling the physiology of the gonad. The muscle cells show typical features as present in most in- vertebrate species, and these characteristics are intermediate be- tween skeletal and smooth muscle cells of vertebrates (Bennett & Threadgold 1973. Stitt et al. 1992). The thick filaments are short bundles consisting primarily of paramyosin protein (Ishii & Sano 1980) and each is surrounded by up to 12 to 13 thin filaments. This indicates that the muscle cells may be able to generate contractile force greater than that of vertebrate smooth muscle. It was found that branches of granulated cells, the putative egg-laying hormone producer, are lying close and frequently tightly adhered to the muscle cells. Immunolocalization study by our group demon- strated that aELH also binds to muscle cells in the capsules and cores of trabeculae (Chanpoo et al. 2000). It is. therefore, possible that aELH. upon being released from the granulated cells, may stimulate the contraction of muscles in the trabeculae and capsules of the gonad, and results in the release of mature gamete cells from the abalone at the time of spawning. Fibroblasts and collagen fibrils appear to be similar in most respects to their counterparts in vertebrates' connective tissues. Thus, their functions could be primarily supportive. ACKNOWLEDGMENT This investigation was supported by the Thailand Research Fund (.Senior Research Scholar Fellowship to P. Sobhon). LITERATURE CITED Apisawetakan. S., A. Thongkukiatkul. C. Wanichanon. V. Linthong. M. Kruatrachue. E. S. Upatham. T. Pumthong & P. Sobhon. 1997. The gametogenic processes in a tropical abalone. Haliotis asinina Linnaeus. J. Sci. Soc. Tltailand 23:225-240. Bennett. C. E. & L. T. Threadgold. 1973. Electron microscope studies of Fasciola lieparica. XIII. Fine structure of newly excystedjuvenile. Exp. Parasiiol. 34:85-99. Bern. H. A. & I. R. Hagadom. 1965. Neurosecretion. In: T. H. Bullock & G. A. Horridge, editors. Structure and function in the nervous system of invertebrate. San Francisco: Freeman, pp. 353-429. Chanpoo, M.. S. Apisawetakan, A. Thongkukiatkul. C. Wanichanon. V. Linthong. M. Kruatrachue. E. S. Upatham. T. Pumthong & P. Sobhon. 2000. Localization of egg-laying hormone (ELH) in the gonads of a tropi- cal abalone. Haliotis asinina Linnaeus. / Sliellfisli Res. 20:725-73 1 . Croft. D. R. 1929. Haliotis. Liverpool Mar. Biol. Coinni. Mem. 29:1-174. Dorsett. D. A. 1986. Brains to cells: The neuroanatomy of selected gas- tropod species. In: A. O. D. Willows, editor. The mollusca. Vol. 9. New York: Academic Press, pp. 101-87. Hahn. K. O. 1994. The neurosecretory staining in the cerebral ganglia of the Japanese abalone (ezoawabi), Haliotis di.'icus liannai. and its rela- tionship to reproduction. Gen. Comp. Endocrinol. 93:295-303. Ishii. A. I. & M. Sano. 1980. Isolation and identification of paramyosin from liver tluke muscle layer. Com. Biocliein. Plivsio. 65B:537-541. Joose, J. 1979. Evolutionary aspects of the endocrine 'system and of the hor- monal control of reproduction of molluscs. In: E. J. W. Harrington, editor. Hormones and evolution. Vol. 1. New York: Academic Press, pp. 1 19-57. Joose, J. 1988 The hormones of molluscs. In; H. Laufer & R. G. H. Downer, editors. Endocrinology of selected invertebrate types. New York: A. R. Liss, pp. 89-140. Martin. G. G.. K. Romero & C. Miller- Walker. 1983. Fine structure of the ovary in the red abalone Haliotis rufescens (Mollusca: Gastropoda). Zoomorphol. 103:89-102. Singhagraiwan. T. & M. Doi. 1993. Seed production and culture of a tropical abalone. Haliotis asiiiin{i Linnaeus. Thailand: The Eastern Ma- rme Fisheries Development Center. Department of Fi^herie^. Ministry of Agriculture and Cooperatives. 33 pp. Sobhon. P.. S. Apisawetakan. M. Chanpoo. C. Wanichanon. V. Linthong. A. Tliongkuklatkul. P. Jarayabhand. M. Kruatrachue. E. S. Upatham & T. Pum- thong. 1 999. Classification of gemi cells, reproductive cycle and maturation of the gonad in Haliotis asinina Linnaeus. Science Asia 25:3-2 1 . Stephenson. T. A. 1924. Notes on Haliotis tulwrculala. J. Mar. Biol. Assoc. UK. 13: 480^95. Stitt, A. W., I. Fairweather. A. G. Trudgett, C. F. Johnston & S. M. L. Anderson. 1992. Localisation of actin in the liver fluke. Fasiola lie- patica. Parasitol. Res. 78:96-102. Takashima, F.. M. Okuno, K. Nichimura & M. Nomura. 1978. Gameto- genesis and reproductive cycle in Haliotis diversicolor diversicolor Reeve. / Tokyo Univ. Fisli. 1:1-8. Tomita. K. 1967. The maturation of the ovaries of the abalone. Haliotis discus liannai Ino, in Rebun Island. Hokkaido. Japan. .Sci. Rep. Hok- kaido Fish. Exp. Sta. 7:1-7. Tomita, K. 1968. The testis maturation of the abalone. Haliotis discus hannai Ino, in Rebun Island. Hokkaido. Japan. 5c/. Rep. Hokkaido Fish. E.xp. Sm 9:56-61. Webber. H. H. & A. C. Giese. 1969. Reproductive cycle and gametogen- esis in the black abalone Haliotis cracherodii (Gastropoda: Prosobran- chiata). Mar Biol. 4:152-159. Young. J. S. & J. D. DeMartini. 1970. The reproductive cycle, gonadal histology and gametogenesis of the red abalone. Haliotis rnfescens (Swainson). Calif. Fish Game 56:298-309. Jniinial I'l Slwllfisli Kcscanh. Veil. 21). No. 2. 72.^-7.^1. 2001. LOCALIZATION OF EGG-LAYING HORMONE IN THE GONADS OF A TROPICAL ABALONE, HALIOTIS ASININA LINNAEUS M. CHANPOO,' * S. APISAWETAKAN,' A. THONGKUKIATKUL,' C. WANICHANON,' V. LINTHONG,' M. KRUATRACHUE,' E. S. UPATHAM,' ^ T. PUMTHONG," P. J. HANNA," AND P. SOBHON' Dcpartnicnis aj 'Aiuitomy and ~Bi(>l(>i;y. Fciciiliy of Science. Maliiclol University. Bangl^' ^ *-, LSc 1 '^^B I0|im I -Sz -10tllti%* Figure 1. (A. Bl Liyhl miironraphs ot I'AS-heniatoxylin stained parafrin sections (if //. asiniim ovary (Al and testis (B) in the mature (prespawn- ingl phase, where there are ahundant mature stage \ oocytes (Oc^) with fully formed jelly coats (jc) in the ovary (in A I, and numerous spermatozoa (S,) with dense nuclei in the testis (in Bl. The connective tissue scaffolds of the gonads in hoth seves consist of the outer capsule (Ocp), the Iraheculae (tr) and the inner capsule (not shown). In the core of each trahecula. there is a small hemolymph capillary (Ca). Sc = spermatocytes, SI = spermatids. (C, I)) High magnification micrographs showing the granulated cells containing PA.S positive granules (gr-in t) in the outer capsule of the ovary (Ocp) and in a trahecula (tr) of the testis (gr-in I)). F = flhrohlasts. I.Sc = leptotene spermatocyte, mu = muscle cells, S/ = spermatozoa. (K, F) The semi-thin plastic secliims of the ovarian inner capsule (Icp- in F) and the testicular outer capsule (Ocp-in F) showing the granulated cells with their ellipsoid nuclei (gr-in F), and their attenuated processes ccmtaining dense granules (gr-in F). Ca = haemolymph capillary, ept = epithelium, F = fibrohlast, mu = muscle cells, Stj = spermatid stage 4. 728 Chanpoo et al. Ocp \ / Oc- -tr- gr gr- % kf I Icp ii ^ HP Icp 2 5|im HP Jfe tr •5 0Hm B tr. Oci tr Icp- HP I. -gr 0C2 OCo^ 1 -Icp HP f*' . -gr 2 5)Im 10|Im Figure 2. Light micrographs of the ovary in the proliferate phase, stained with anti-aELH by the inimunoperoxidase method. ( Al Control section shows no staining in the ovarian tissue. (B-D) Sections stained with anti-aELH, showing intense staining in granulated cells (gr) within the inner (Icp) and the outer capsules (Ocp), while moderate staining is seen in the connective tissue proper of trabecula (tr), both capsules (Ocp & Icp), and the cytoplasm of immature oocytes (Oc„ Oc2, Ocj). HP = hepatopancreas. Localization of Abalone Egg-Laying Hormone 729 Ocp I tr_ -qr Oc, tr tr I V HP -Icp 2 50|am ri. .J :^ 2 5|im --^ OC5 B / ^.*^---.*. mu I . > ^-gr Vx-^ '>v::"*V ■?' OCc •i5' ^k tr -.^>="-' '1 0 0|am' tr-.-' ■Vl -mu ^^^ >"\'- vv^. »\ V ^,mu \ .2 5|am Figure 3. Lisht micrographs of the ovary in the mature phase, stained with anIi-aEKH hy the immunoperoxidase method. I A, B) In A. anti-aELH exhibits staining in the traheculae (Irl and hoth capsules (Ocp. Icpl of the gonad. In B, granulated cells (gr) in a trabecula are intensely stained. The cytoplasm of late stage oocytes l()c,l surrounded by a thick jelly coat (Jcl is not stained. (C. I)) Sections stained with antl-aKI.H and counter-stained with hematoxylin. In C. the connective tissue of the traheculae (tr) is positively stained in comparison to oocytes (Oc,). In D, which is higher magnification of an area from the traheculae in C, granulated cells (gr) and muscle cells (mu) are intensely stained. 730 Chanpoo et al. Ocp^ V Ocp '-gr ^ Sz .'9r . gr tr~ \ _ ^ >r HP <^ -gr tr Sz ' Sz Icp , t,r gr Icp ' 'l 100U " SOU HP -tr-^ 1 gr gr-y Sz ■*^^ "Ocp Sz '-'2«|Pr ""^ «^ V gr -gr 1 ou gr Sz 25U 25 U Figure 4. Light micrograplis of the testis in the mature phase, stained «ith anti-aELH hy the immunoperoxidase method. (A-D) Sections stained with anti-aELH, showing strong staining of the tral)eculae (tr), hoth the outer and inner capsules (Ocp, Up). In C and I) inset, which are high magnifications of a trabecula and the outer capsule, granulated cells (gr) with large granules in their cytoplasm are intensely stained, while spermatocytes and spermatozoa (Sz) are not stained. Localization of Abalone Egg-Laying Hormone 731 is transported through the axon and released by exocytosis to the hemolymph (Dirks et al. 1989. Dirks et al. 1993. Van Minnen et al. 1988). Expression of CDCH v\as not restricted to the CNS alone, but was also found in the reproductive tract, including the oothecal gland, muciparous gland, and pars contorta. which are feinale ac- cessory sex glands in Lyiunaea sraginilis. In these glands the pro- cesses of positively labeled neurons terminated on the secretory cells, suggesting that they controlled the activities of these tissues (Van Minnen et al. 1988). CDCH immunoreactive material has also been found in secretory cells of the prostate gland and sperm duct (Van Minnen et al. 1989). In contrast to CDCH little is known about the origin of egg-laying hormone in abalone. Histological studies in the Japanese abalone. Haliotis discus luiiiiiiii. showed that the number of neurosecretory cells, especially type 1 and 7, in pleural-pedal ganglia was correlated with the induction of spawn- ing (Hahn 1992). Injections of pleural-pedal and visceral ganglion crude homogenates. or the combination of both, caused female H. dixciis liannai to spawn (Yahata 1973). The quantity of eggs being spawned was significantly greater with the injections of homoge- nates from visceral ganglion or the combination of pleural-pedal and visceral ganglion, when compared with the injection of pleu- ral-pedal ganglion alone (Yahata 1973). In our preliminary study of H. asinina visceral ganglia, neurosecretory cells type 1 were also positively stained with anti-aELH (unpublished observation). Hence, existing evidence implies that abalone egg-laying hormone is mostly produced by neurosecretory cells of the nerve ganglia, particularly plcuropedal and visceral ganglia. In the present study, we found that unti-aELH from H iiihni showed strong cross reaction with H. iisinina gonadal connective tissues, and this implied that aELH may also be produced and stored in the granulated cells within the trabeculae and capsules of the gonad. Similarly, muscle cells in these connective tissue scaf- folds were also stained with the anti-aELH, which suggested that this group of cells also bind aELH. Coggeshall (1972) suggested that, in Aplysia. ELH acted directly on muscle cells to induce their contraction, which caused the expulsion of ripe oocytes from the ovary. From the evidence gathered in the present study we, there- fore, would like to suggest that the granulated cells in the trabec- ulae and the capsules of gonads in both sexes of abalone can synthesize aELH. After being released from the granulated cells, this hormone could bind to muscle cells in trabeculae and capsules and cause them to contract, which results in the expelling of ripe oocytes or spermatozoa from the gonads. The significance of the binding of anti-aELH to early stage oocytes is not known, but this hormone may also participate in the developmental process of germ cells. In contrast aELH did not bind to male germ cells thus its role in the male abalone may be limited to controlling the release of spermatozoa as has recently been demonstrated by our group (unpublished observation). ACKNOWLEDGMENTS We sincerely thank Associate Professor Peter J. Hanna. School of Biological & Chemical Sciences. Deakin University. Geelong. Australia, who kindly supplied mouse polyclonal antibody against aELH of H. rubra. This work was supported by the Thailand Re- search Fund (Senior Research Scholar Fellowship to P. Sobhon). LITERATURE CITED Arch. S. 1976. Neuroendocrine regulation of egg-laying in Aplysui cuti- fornica. Am. Zool. 16:167-175. Chiu. A. Y.. M. W. Hunkapiller. E. Heller. D. K. Stuart. L. E. Hood & P. Struniwasser. 1979. Purification and primary structure of the neuro- peptide egg-laying hormone of Aplysia califonucii. Proc. Nail. .Acciil. Sci. USA 76:6656-666(1 Choate, J. V. A.. T. E. Kruger. M.-A. Micci & J. E. Blankenship. 1993. Isolation of an egg-laying hormone-binding protein from the gonad of Aplysia catifornicu and its localization in oocytes. J. Coinp. Plivsiol. A. 173:475-^83. Coggeshall. R. E. 1972. The muscle cells of the follicle of the ovoteslis in Aplysia as the probable target organ for bag cell extracts. Am. Zeal 12:521-523. Dirks. R. W,. A. K. Raap. J. Van Minnen. E. Vreugdenhil. A. B. .Smil & M Van der Ploeg. 1989. Detection of niRNA molecular coding for neuropeptide hormones of the pond snail Lynmaea stagnalis by radio- active and non-radioactive in situ hybridization: a model study for niRNA detection. J. Histocliem. Cytochem. 37:7-14. Dirks. R. W.. A. G. M. Van Dorp, J. Van Minnen. J. A. M. Fransen. .M. Van der Ploeg & A. K. Raap. 1993. Ultrastructural evidence for the axonal localization of caudodorsal cell hormone mRNA in the central nervous system of the mollusc Lymnaea sktgnalis. MicrcK. Res. 25: 1 2- 1 8. Ebberink. R. H. M.. H. Van Loenhout. W. P. M. Geraerts & J. Joosse. 1985. Purincation and amino acid sequence of the ovulation neurohor- mone ol Lymmwii stagnalis. Proc. Nat. .Aciid. Sci. USA 82:7767-777 1 . Geraerts, W. P. M. & S. Bohlken. 1976. The control of ovulation in the hermaphrodite freshwater snail Lymnaea stugiwlis by the neurohor- mone of the caudodorsal cells. Cen. Comp. Endocrinol. 28:350-357. Hahn. K. O. 1992. Review of endocrine regulation of reproduction in abalone Haliotis spp. In: Abalone of the World. Fishing News Books, pp. 49-56. Kuptermann. I. 1967. Stimulation of egg-laying: possible neuroendocrine functions of bag cells of abdominal ganglion of Aplysiii californica. Nature 216:814-815. McAllister. L. B.. R. H. Scheller. E. R. Kandel & R. Axel. 1983. In situ hybridization to study the origin and fate of identified neurons. Science 222:800-808. Scheller. R. H., J. F. Jackson. L. B. McAllister. B. S. Rothman. E. Mayeri & R. Axel. 1983. A single gene encodes multiple neuropeptides me- diating a stereotyped behaviour. Cell 32:7-22. Singhagraiwan. T. & M. Doi. 1993. Seed production and culture of a Q-opical abalone. Haliotis usiniuu Linnaeus. The Research Project of Fishery Re- source Development in the Kingdom of Thailand. Thailand: Department of Fisheries. Ministry of Argriculture and Cooperatives. 32 pp. Van Minnen. J.. C. Van der Haar. A. K. Raap & E. Vreugdenhil. 1988. Localization of ovulation hormone-like neuropeptide in the central ner- vous system of the snail. Lymnaea stagnalis by means of immunocy- tochemistry and in situ hybridization. Cell. Tiss. Res. 251:477-484. Van Minnen. J.. R. W. Dirks. E. Vreugdenhil & J. Van Diepen. 1989. Expression of the egg-laying hormone genes in peripheral neurons and exocrine cells in the reproductive tract of the mollusc. Lymnaea stag- nalis. Neuroscience 33:35— f6. Vreugdenhil. E.. J. F. Jackson. T. Bouwmeester. A. B. Smit. J. Van Min- nen. H. Van Heerikhui/en. J. Kloolwijk & J. Joosse. 1988. Isolation, characterization and evolutionary aspects of a cDNA clone encoding multiple neuropeptides involved in the stereotyped egg-laying behavior of the freshwater snail Lymnaea stagnalis. J. Neurosci. 8:4184—4191. Wang. L. & P. J. Hanna. 1998. Isolation, cloning and expression of a DNA sequence encoding an egg-laying hormone of the blacklip abalone i Haliotis Rubra Leach). / Shellfish Res 17:789-793. Yahata, T. 1973. Induced spawning of abalone {Nordotis discus Reeve) injected with ganglional suspensions. Bidl. Jap. Soc. Sci. Fish 39: 1 1 17-1 122. .loiiinal of Shellfish Research. Vol, 2(1. No. 2. 7.^.V74I. 2001. IILTRASTRUCTURE OF NEUROSECRETORY CELLS IN THE CEREBRAL AND PLEUROPEDAL GANGLIA OF HALIOTIS ASININA LINNAEUS A. THONGKUKIATKUL,' ' P. SOBHON," E. S. UPATHAM,'^ M. KRUATRACHUE/ C. WANICHANON,- Y. P. CHITRAMYONG' AND T. PUMTHONG' ^Department of Bioloiiy. Faculty of Science. Biirapha Unlver.siry. Chonhitri 201 31 . Tliailaiul: -Department of Anatomy. Faculty of Science. Mahidol University. Baiii^kok 10400. Thaihuul: ^Department of Biology. Faculty of Science. Mahidol University. Bangkok 10400. Tliailaml; Faculty of Science. Burapha University. Chonhuri 20131. Thailand: ^Coastal Aquaculture Development Center, Prachuap Khiri Khan 77000, Thailand ABSTRACT The ultrastructure of all three types of neurosecretory cells (NS,, NS, and NS,) in the cerebral and pleuropedal ganglia of Hulkitis asiiiimi was studied. NS, cells contained a euchromatic nucleus, and the cytoplasm contained RER, Golgi complexes, mitochondria and rihosomes. There were two types of neurosecretory granules in the NS, of cerebral ganglia: type 1 were large osmiophilic membrane-bound granules and type 2 were small electron-dense spherical granules. The NS, cells of pleuropedal ganglia only had one type of round cytoplasinic granule with a moderate to strong electron-dense matrix. NS, cells contained blocks of heterochromatin in the nucleus. The cytoplasm of the NS, of cerebral ganglia contamed the usual organelles similar to those of NS, cells and large membrane-bound granules containing crystalline structures einbedded in a inoderately dense osmiophilic matrix. The NS, cells of pleuropedal ganglia contained one type of granule that had a dense matrix. NS, cells were smaller than NS, and NS,. The nucleus contained thick heterochromatin strands. The organelles in the cytoplasm appeared to be fewer than those of NS , and NS,. The secretory granules of NS, of both cerebral and pleuropedal ganglia were composed of aggregates of dense osmiophilic globules of various sizes. KEY ^yORDS: Haliinis asiiiiiiti. neurosecretory cells, cerebral ganglia, pleuropedal ganglia, ultrastructure INTRODUCTION Neurosecretory cells present in the cerebral iziinglia of proso- branchs have not been extensively studied, and consequently little is known about them. The neurosecretory cells in the cerebral ganglia of Bilhynia tentaculata Linnaeus were stained with phlox- ine (Andrews 1968). They were found to be unipolar and their nuclei were usually concave on one side. Neurosecretory material accumulated in the periphery of the cytoplasm and the axon hillock (Andrews 1968). In Haliotis discus hamuli Ino and Nordolis discus Reeve, two cell types in the cerebral ganglia were identified as neurosecretory cells. They were large and medium sized cells with euchromatic nuclei and contained neurosecretory granules in the cytoplasm (Yahata 1971. Hahn 1994). More recently, in the cerebral ganglia of Haliotis asinina Lin- naeus, two types of neurosecretory cells were found (Upatham et al. 1997). These cells are either large or medium in size and stained positively with chrome-hematoxylin-phloxine and paraldehyde- fuehsin. The large sized cells contain a round nucleus with eu- chromatin and a distinct nucleolus. The medium sized cells also contain a round nucleus with patches of heterochrotnatin. Neurosecre- tory granules are present in both cell types (Upatham et al. 1997). Most studies on the ultrastructure of neurosecretory cells in gastropods have concentrated on pulmonates and opisthobranchs with only a few on prosobranchs. In the cerebral ganglia of Lym- naea slaf;nalis (Linnaeus), two groups of neurosecretory cells have been described (Joos.se 1964. Boer 1965. Boer et al. 1968). The cytoplasm of these cells contain electron-dense granules, which had a mean diatneter of 20 nm. extremely elongated mitochondria, rough endoplasmic reticulum, free ribosomes. polyribosomes, Golgi complexes, multivesicular bodies, neurotubules and cyto- somes. Bonga ( 1970). using the alcian blue-alcian yellow staining 'Corresponding author. E-mail address: aniporn(S'bucc4.buu.ac.th. technique, reported that there was only one type of neurosecretory cells /.('. dark green cells in the pleuropedal ganglia of L. stagindis. At the electron microscopic level, the dark green cells appear to contain a large quantity of elementary granules with a mean di- ameter of 20 nm. Numerous Golgi complexes were found to be evenly distributed in the cytoplasm, and there was extensive rough endoplasmic reticulum. A low number of cytosomes were present. In Achatina fulica (Bowdich). neurosecretory cells in the cerebral ganglia contain a round shaped nucleus with patches of hetero- chromatin and a conspicuous single large vacuole in the cyto- plasm. In addition, electron-dense granules with a mean diameter of 16nm are associated with extensive Golgi complexes and rough endoplastiiic reticulum (Kruatrachue et al. 1994). In the prosobranchs, the ultrastructure of neurosecretory cells and neurons has been described in B. tentaculata (Andrews 1971) and Haliotis rufescens Swainson (Miller et al. 1973). In B. ten- taculata. there are three types of neurosecretory cells, viz, SI, S2 and S3. In the cytoplasm of these cells, there are well-developed rough endoplasmic reticulum and Golgi complexes, mitochondria, lysosomes. glycogen granules, neurofibrils, and neurosecretory granules (Andrews 1971 ). Miller et al. (1973) describe that most of the neurons of H. rufescens contain the usual cytoplasmic or- ganelles along with large membrane-bound inclusions. A few neu- rons contain small dense granules, which are similar in appearance to typical neurosecretory granules. It was. therefore, apparent that an extensive investigation of the ultrastructure of neurosecretory cells in Haliotis was needed. Hence, the aim of the present study was to describe the ultrastruc- ture of different types of neurosecretory cells in the cerebral and pleuropedal ganglia of H. asinina. MATERIALS AND METHODS Cerebral and pleuropedal ganglia from matuie H. asinina were fixed in a mixture of 4% v/v glutaraldehyde and 2% v/v paraform- 733 734 Thongkukiatkul et al. aldehyde inO. IM Millonig buffer (pH 7.8) at 4°C for 24 h, washed several times with the same buffer, and postfixed in 1% OsOj in O.IM Millonig buffer. The specimens were dehydrated in a graded series of ethanol, infiltrated in acetone and embed- ded in Araldite 502-epoxy resin. Sections were cut on a Sorvall MT2 ultramicrotonie. stained with sattirated uranyl acetate and lead citrate, and viewed with a Hitachi H-3()0 TEM, operating at 75 KV. RESULTS Based on ultrastructural characteristics, there are three types of neurosecretory cells in the cerebral and pleuropedal ganglia of//. asiiiiiui. These are: type 1 neurosecretory cell (NS,), type 2 neu- rosecretory cell (NSt). and type 3 neurosecretory cell (NS,). NS| cells are round to oval (15-20 |xm in dimension) and contain round nuclei (6-S |j.m in diameter) (Fig. 1. Fig. 2). The m^^^z^'ms:-:: **:*». NSh Nu -Gc /Gr^ 1 um rer 'Gr^ Gc O.Sum No -Gc Figure 1. TEM micrographs of NS, cells of cerebral ganglia. (A, C) Medium power micrographs of the NS, shoHing the round nucleus (Nu) which contains a thin rim of heterochromatin (He) near the nuclear envelope. The nucleolus (No) is round, large and very distinct. There are abundant secretory granules in the cytoplasm. NS,, type I neurosecretory cell; NS„ type 3 neurosecretory cell; Gc, Golgi complex; Or, type 1 granule. (B, I), El High magnifications of the cytoplasm of the NS, demonstrating abundant rough endoplasmic reticulum (rer), Golgi complexes ((;c) and secretory granules. Type 1 granules (Gr,) are large spherical membrane hound granules whereas type 2 granules (Or,) are small and round, containing electron-dense cores. Neurosecretory Cells of Abalone 735 Figure 2. TEM mkriijjniplis ol NS, cells (it pleunipedal nanylia. ( \. Hi I,al nucleus (Nu) which contains thick strands of heterochromatin (He) near the nuclear envelope and in the central region. The cytoplasm contains aggregated granules (Gr). 740 Thongkukiatkul et al. thick rim of iieterodiromatin near the nuclear envelope, and a thick strand of heterochromatin in the center (Figs. 5C, E. Fig. 6D). In the NS, cells t)f the cerebral ganglia, the cytoplasm contains rough endoplasmic reticulum, mitochondria, and Golgi complexes, but these organelles are less abundant, and of smaller sizes, than those in NS| and NS, cells. Much cytoplasm is filled with secretory granules (Figs. 5A. D). The.se granules are large (2000-7000 nm in diameter), membrane-bound, and contain dense, osmiophilic glob- ules of various sizes that are aggregated together. When examined in detail, these granules are divided into two subtypes. Type 1 secretory granules are composed of strongly osmiophilic crystal- line material in a clear ground substance (Figs. ?A, B). and type 2 secretory granules are composed of highly condensed osmiophilic globules aggregated together (Figs. 5D, E. F). It is possible that type 2 secretory granules are developed from type 1 secretory granules. In the NS, cells of pleuropedal ganglia, the cytoplasm contains less rough endoplasmic reticulum, fewer mitochondria, and fewer Golgi complexes, than those of NS, and NS,. NS, cells are divided into two subtypes. Subtype 1 cytoplasm contains few, but large secretory granules, with a strong osmiophilic substance within a clear ground matrix (Figs. 6A. B). and subtype 2 cytoplasm con- tains numerous large secretory granules, each composed of strong osmiophilic globules aggregated together in a homogeneous ground substance (Figs. 6C, D). It is possible that neurosecretory cells of subtype 1 develop from subtype 2 through the condensa- tion and dehydration of the osmiophilic substance. DISCUSSION The ultrastructural study of the neurosecretory cells in the ce- rebral ganglia of H. asiiiiini revealed that there are three types of neurosecretory cells (i.e. NS,, NS,, NS,). This is in contrast to only two types reported by Upatham et al. (1997) using light microscopy. The pleuropedal ganglia also contain three types of neurosecretory cells. The details of the heterochromatin and eu- chromatin in the nuclei of cerebral and pleuropedal ganglia neu- rosecretory cells were revealed by light microscopy and confinned by the extra resolution of TEM. The NS, nucleus contains mostly euchromatin. while large amounts of heterochromatin were present in NS-, and NS, cells. In general, the cytoplasm of these cells resembles that of neurosecretory cells described in other gastro- pods, such as B. tentaculata (Andrews 1971 ). H. mfescens (Miller et al. 1973), A. fiilica (Kraatrachue et al. 1994) and L. stai;inilis (Boer et al. 1968). The main differences between the neurosecretory cells of the cerebral ganglion and those of the pleuropedal ganglion are the type and size of neurosecretory granules. In the cerebral ganglia, the NS, cell contains 2 types of secretory granules (large and small) while the NS, cell of the pleuropedal ganglion contains only one type (small granules). In addition, the NS, cells of the cerebral and pleuropedal ganglia both contain one type of secretory gran- ule. However, they are different both in size and content. The NS, of both ganglia appear to contain one type of secretory granule that contains aggregates of dense globules. In the cerebral ganglia, the NS, cytoplasm is composed of cell organelles, including numerous mitochondria, rough endoplasmic reticulum. Golgi complexes and secretory granules, which reflects a highly active secretory function. Golgi complexes are extremely large and there may be several present in a cell. Small electron- dense secretory granules are associated with the maturing face of Golgi complexes. These granules were later widely distributed throughout the cytoplasm. Hence, this indicates that the Golgi complexes in the neurosecretory cells of H. asinina have a similar role in packing of electron-dense material, to those of neurosecre- tory cells reported in other gastropods (Boer et al. 1968. Kai-Kai & Kerkut 1979). The ultrastructural characteristics of the NS, of the cerebral ganglia indicate that it is a highly active synthetic cell. In com- parison, the cytoplasm of NS, and NS, contains only one type of granule, which are large and round in NS, and polymorphic in NS,. These granules bear crystalline structures in NS, and in NS, and are composed of aggregates of dense osmiophilic substances. The similarity between granules in NS, and NS, lends to indicate that the two cells could be of the same group. While the NS, appears to be in a more active secretory phase, the NS, has reached the fully differentiated state, in which hormonal product is already produced in abundance, stored, and ready for release. In the present study, the cytoplasms of NS , and NS, cells in the pleuropedal ganglia of H. usinina exhibit characteristics which imply that they have actively synthetic features. These are the numerous mitochondria, rough endoplasmic reticulum, Golgi com- plexes, multivesicular bodies and secretory granules, and numer- ous small clear vesicles that may be transport vesicles in the cy- toplasin. Rough endoplasmic reticulum and Golgi complexes are well de\eloped. The electron-dense secretory granules are prob- ably tbrmed from the Golgi complex. This process is similar to that described in neurosecretory cells in the pleural ganglia of Z.. stag- inilis (Bonga 1970). ACKNOWLEDGMENT This study received financial support from the Thailand Re- search Fund #BRG 408 0004 and PG 2/015/2539. LITERATLl Andrews. E. 1968. An anatoniicd and histological study of the ner\'Ous system of Birhyir ill tentaculata (Prosobranchia) with special reference to the pos- sible neurosecretory activity. Prix. Malac. Soc. Land. 38;213-232. Andrews. E. 1971. The fine structure ot the nervous system of Bilhvniu tentaculata (Prosobranchia) in relation to possible neuro.secretory ac- tivity. Veliger 14:13-23. Boer. H. H. 1965. A cytologicul and cytochemical study of neurosecretory cells in Basommatophora. with particular reference to Lyinnaea stag- nctlis (L.). Arch. Neerl. Zool. 16:313-386. Boer. H. H.. E. Dounia & J. M. A. Koksnia. \9bH. Electron microscope RE CITED study of neurosecretory cell and ncurohaemal organs in the pond snail Lymnaea stagnalis. Symp. Zool. Soc. Land. 22:237-256. Hahn. K. O. 1994. The neurosecretory staining in the cerebral ganglia of the Japanese abalone (ezoawabi), Haliotis discus hannai, and its rela- tionship to reproduction. Gen. Comp. Endocrinol. 93:295-303. Joosse. J. 1964. Dorsal bodies and dorsal neurosecretory cells of the ce- rebral ganglia of Lymnaea stagnalis (L.). Arch. Neerl. Zool 16:1-103. Kai-Kai. M. A. & G. A. Kerkut. 1979. Mapping and ultrastructure of neurosecretory cells in the brain of Helix aspersa. Comp. Biochcin. Physiol. 64A:97-107. Neurosecretory Cells oh Abalone 741 Kruatrachue. M.. V. Seehabutr. J. Chavadej. P. Sretarugsa. E. S. Upathani & P. Sobhon. 1994. Development and histological characteristics iil neurosecretory cells in the cerebral ganglia o( Acluiiiiui ftdica (Bowd- ich). Moll. Res. 15:29-37. Miller, W.. R. S. Nishioka & H. A. Bern. 1973. The "juxtaganglionic" tissue and the brain of the abalone. Haliotis nifcscens Swainson. Ve- liger 16:125-129, Upatham. E. S.. A. Thongkukialkiil. M. Kruatrachue. C. Wanichanon. Y. P. Chitramvong. S. Sahavacharin & P. Sobhon. 1997. Classification of neurosecretory cells, neurons, and neuroglia in the cerebral ganglia of Hiilioris iLsiitiiHi Linnaeus by light microscopy. J. Shellfish Res. 17: 737-742. Wendelaar Bonga, S. E. 1970. Ultrastructure and histochemistry of neu- rosecretory cells and neurohaemal areas in the pond snail Lymncuii stagnalis. Z. Zellforsch. Mikrosk. Anat. 108:190-224. Yahata, T. 1971. Demon.stration of neurosecretory cells in the cerebral ganglion of the abalone. Nonlotis discus Reeve. Bull. Fac. Fish. Hok- kaido Univ. 22:207-214. Joiiimil of Shellfish Resvairh. Vol. 20. No. 2. 743-75.1. 2001. INTERACTIONS AMONG RED ABALONES AND SEA URCHINS IN FISHED AND RESERVE SITES OF NORTHERN CALIFORNIA: IMPLICATIONS OF COMPETITION TO MANAGEMENT KONSTANTIN A. KARPOV,' * MIA J. TEGNER,' t LAURA ROGERS-BENNETT," PETER E. KALVASS,' AND IAN K. TANIGUCHl' 'California Departiueni of Fish and Game. Fori Brafifi. California 95437: 'California Departmeni of Fish and Game. Bodef^a Marine Laboratory. Bodega Bay, California 94923: 'California Departmeni of Fish and Game. Los Alamltos. California 90720 ABSTRACT Red abaloiies {Haluilis rufcsccns). red sea iiivhins (SlrongylDceiUnitus fnincisciiiuis). and purple sea urchins (S. purpuratiis) share similar food and habitat requirements in northern California. Red abalones and red sea urchins also support important fisheries. Here we explore spatial interactions and apparent competitive effects among these species at an area where fishing has large impacts on both taxa. and at unfished reserve sites in which invertebrate density and food availability differ. There was an inverse correlation between adult red abalone and red sea urchin abundance at the scale of our transects when density of either or both species was high. In the poorest habitat for macroalgae. red abalones seldom occurred on the same transects with red urchins. The results suggest that differences in density, depth, and food availability play an important role in the observed spatial patterns of red abalones and red sea urchins. Purple sea urchins were not cortelated to either of the other two species' distributions. An intense fishery for red sea urchins appears to have had a positive effect on kelp availability, and abalone growth and abundance. Aerial photographs during the period of intense urchin fishing (froin 1982 to 1989). showed a dramatic increa.se in the surface canopy. Siinilariy. during this period, size frequency distributions of fi.shed red abalones show an increase in the number of individuals in larger size classes. Modal progression in abalone size frequency distributions suggests a faster growth rate during this period when compared with a growth study, at the same location, conducted during the pre-urchin fishery years. Ultimately, red sea urchin removal apparently led to an increase in red abalone abundance even at a site that was heavily fished by recreational abalone fishers. Meanwhile, at a nearby reserve site where kelp populations are lower, red abalones have declined in abundance as red sea urchins increased. Our results suggest the need for multi-species ecosystem-based approaches to management of these valuable resources. KEY WORDS: reserves. Haliotis nifesiens. Strongyloceiurotits fnmciscaiuis. Stinngylocenrronis piirpuninis. competition, spatial exclusion, ecosystem-based management INTRODUCTION Red abalones {Haliotis rufescens). red sea urchins (Strongylo- centrotiis franciseaniis) and purple sea urchins (S. purpuratus) share similar food and habitat preferences in kelp forest commu- nities along the California coast. In noithern California, red aba- lones are found in rocky intertidal and shallow subtidal areas in high abundance at 7-8-ni depths, but also occur down to 25 m in areas where drift algae accumulate in surge channels. Red and purple urchins are found tVom mid- to low- intertidal zones to depths in excess of 50 m. Both species prefer rocky substrates, particularly ledges, crevices and surge channels, and avoid sand and mud (Schroeter 1978, Kato & Schroeter 1985). In areas of high predation, red abalones show a preference for crevice habitat (Hines & Pearse 1982). Red abalones and red and purple urchins feed priinarily on the same species of macroalgae iLeighton & Boolootian 1963). and have been described as potential competi- tors for food and space (Leighton 1968, Tegner & Levin 1982). Both urchins and abalones feed primarily on drift kelp, but sea urchins are well known for their destructive grazing on attached plants when drift becomes limiting. Schroeter (1978) presented evidence that red urchins out compete purple urchins for food and habitat, suggesting that the red urchin fishery could lead to an increase in populations of its smaller congener. .Several authors have considered the potential of sea urchin populations being re- leased from competition as abalones were fished down in southern California (North & Pearse 1970, Teener 1980, Teener & Levin *Cortesponding author. E-mail address: kkarpov@mcn.org tDeceased. 1982, Tegner & Dayton 2000). No one has examined the effects of red urchin removal in areas where red abalones are still abundant. Red abalones and red sea urchins are both fished intensively in northern California. Red abalone take is restricted to recreational fishers who are prohibited from using SCUBA. Karpov et al. ( 1998) reported that this results in a "defacto" refuge for red aba- lone at depths greater than 8.4 m. Since 1985, Red urchins have been subjected to an intense commercial fishery at all depths (Kal- vass & Hendrix 1997). Purple sea urchins are essentially unfished. comprising less than X'^i of the total urchin landings. The red abalone, the largest member of the genus, attains sizes of up to .312 mm in northern California (Department of Fish and Game -DFG- unpublished data). Legal minimum size for the rec- reational-only abalone fishery is 178-mm shell diameter. Red sea urchins, characterized by long spines in relation to test diameter (TD), attain sizes of 140 mm TD, and were first protected by a minimum size limit of 89-mm TD in 1991. Purple sea urchins have short spines relative to their TD, and maximum TD is about 85 mm. In northern California macroalgae are highly seasonal as a food source (Tegner et al. 1992). Wave energy is markedly higher in northern compared to southern California, and thus is likely to be a more important factor affecting the distribution of abalone and urchins (Deacon 19731. Fishing effects, both direct and indirect, make field studies of competition difficult to conduct. Here we take advantage of three locations in northern California, one fished and two reserves closed to sport and commercial fishing, to examine biological and fishery interactions between red abalones, red sea urchins, and purple sea urchins. Differences in fishing regimes constitute a natural experiment, which offers an opportunity to examine the factors structuring this nearshore community, the potential produc- 743 744 Karpov et al. tivity of the three grazers, and the ecosystem effects of fishing. We compare observations of densities, size-frequency distributions. and spatial distributions of the three grazers and kelp populations over time. Two major questions are asked in this study: ( 1 ) Has intense red urchin removal by the fishery had an affect on the abalone resource, and (2) are there significant differences by habi- tat, density, or depth that indicate spatial exclusion between sea urchins and abalones. Materials and Methods Our study is focused in noithern California on a fished area. Van Damme State Park (VDSP). and two unfished "control" areas: Bodega Marine Life Refuge (BMLR) and Point Cabrillo Marine Reserve (PCMRxFig. 1). VDSP is highly impacted by both the sport only abalone fishery and the commercial red sea urchin fish- ery. In this study, we used the same 1 20 dive stations examined for changes in emergent abalone abundance and size distributions dur- ing 1986. and from 1989-1992 by Karpov et al. I 1998). In 1999. we added 34 more stations that were at comparable locations to those surveyed during earlier years. Red urchins were surveyed throughout these periods: purple sea urchins were added in 1989. As in the previous study (Karpov et al. 1998). we stratified our sampling into two strata, "shallow" and "deep", using 8.4 m as the dividing depth because free divers seldom dive deeper than 8.4 m to collect abalone. Thus, the VDSP study area represents a treat- ment of large-scale continued removal of red urchins at all depths and red abalones from shallow water. Two no-take reserves were also surveyed. PCMR and BMLR (Fig. 1 ). Parker et al. (1988) first surveyed 30 stations at PCMR in 1986. PCMR was re-sampled in 1988 for sea urchins at 14 stations and again in 1999 at 30 stations. About half of the 1999 stations were in close proximity to loca- tions sampled in 1986. Sampling was normally conducted during late summer months. Station depths throughout the study period ranged from 2 to 18 m. Sampling was conducted along 2 x 30 m transects randomly placed on rocky habitat. Transects were not sampled if they were placed on substrate with more than 50% sand. PCMR. fiist established as a "no-take" reserve in 1975. is lo- cated 9 km north of VDSP. PCMR provides both a site for com- paring spatial conelation for high densities of red and purple sea urchins to high densities of red abalone, as well as a nearby un- fished control for comparison with the fished VDSP site. BMLR, located 130 km south of VDSP. was established as a no take reserve in 1966 (Fig. 1). In 1999. 33 stations at BMLR were sampled to examine spatial correlation in an area with no surface canopy of kelp in more marginal abalone and urchin habitat. We also examined this area on a smaller spatial scale. Stations were grouped into three sites of distinct habitat, designated Horseshoe Cove, Cave, and Points (North and South Point combined) (Fig. 2). To compai-e the relative availability of food at each study area, percent cover of algae was estimated by divers on each quadrant and averaged in all three of the study sites during 1999. Algae and cover were classified in five categories: encrusting, coralline algae turf, foliose. underslory. or canopy kelp. Understory kelp included brown algae such as Pterygopheia ccilifonilca and Laniinaria den- tiijera. The primary canopy species in northern California was the annual kelp, Nereocystis luetkeana. Percentage cover at times ex- ceeded 100%. Size frequency distributions, weighted for density, were exam- ined for patterns of recruitment of red sea urchins and red abalones at VDSP and PCMR. Size frequency distributions of red urchin catches were obtained from commercial samples. Estimates of -■ ■ / ■ ■ ■ 1 ■ 1 . ) Humboldt County Xrea lot "■""^auforniaV \ -^^^^ \ \ ■ Pt. Cabnilo Re«^e -r^^te,Ktoc*» Van Damme -\_ /*y«»,#w State Park ^^...Cow^ • * *\ L V ^^ 3 \ ««. . 39* ) '^^..^^ X,_r— ^ A ^'"^S' N \ Sonoma V Cotmty \ 16 km T ^ ^ . ; 0 10 Bodega Marine Life W->. Refuge yVv^~^ V:: - 38- J^^^ly'*^'^=K8,4 m) stations. Red sea urchin densities were low in Horseshoe Cove (0.6 red urchin m"" SE = 0.2) but higher at both the Cave ( 1 .2 m'- SE = 0.3) and Points (1.1 m"" SE = 0.2). Purple sea urchins were found in very low abundance throughout this reserve. No signifi- cant correlation in counts was found between purple urchins and either red abalones or red urchins at any of the three BMLR sites. Red urchin and red abalone counts, however, were negatively cor- related on transects at each of the three sites (Fig. 3). Plots of the distribution of counts by station showed marked segregation by species at almost all the stations, with few containing both species. Patterns of segregation were similar among the three locations in the reserve irrespective of relative density decrease for abalones or increase for sea urchins. PCMR While generally higher than BMLR, red abalone densities at PCMR declined in both shallow water (from 1.2 SE = 0.2 to 0.6 m"- SE = 0.1 ) and in deep depths ( 1.2 SE = 0.2 to 0.1 m"- SE = 0.1) during this 13-year interval (Fig. 4, Fig. 5). ANOVA comparison showed the difference in density to be significant by TABLE L Percent cover of encrusting organisms and macroalgae at Van Damme State Parke (VDSP), Point Cabrillo Marine Reserve (PCMR), and Bodega Marine Life Refuge (BMLR) for 1999. I'nder- N Encrust, Turf Foliose story Canopy VDSP Shalk.u 21 62 39 41 53 65 Deep 11 44 IS 31 30 79 PCMR Shallow 17 72 35 49 40 44 Deep 12 68 7 12 9 21 BMLR Horseshoe Cove 13 36 56 41 12 0 Cave 6 44 33 55 0 0 Points (NSl 12 41 24 33 3 0 746 Karpov et al. Densities at BIVILR E •1 1.0 1 1 I II ♦ Red Abalone ■ Red Sea Urchin A Purple Sea Urchin L II i i CAVE Study Areas (S 100 '.' -0.84 p = 0.03 N = 6 Figure 3. Red abalone. red sea urchin, and purple >.ea urchin densities (± sel by study area of BMLR with paired comparison of counts by species at each station with Spearman Rank correlation coefficients Ir..!, statistical signiticance, and number of stations. year, depth, and interaction (Table 2). Red sea urcfiin densities were significantly greater at depth and in 1999 (interaction was not significant). Size distributions re\eal very little differences with large proportions of animals al both depths at or above legal sizes for both species (Fig. 6). Deep water red urchins showed evidence of recruitment in both years w ith bimodal distributions of YOY (<.^Omm) or small juveniles and large adults. Purple sea urchins, not surveyed in 1986, were at greatest abundance at shallow depths in 1999 (3.2 m"-. SE = 1.0. Fig. 4 and Fig. 5). Purple sea urchins at PCMR were not significantly spatially correlated to the other two species at the scale of our transects. Correlation comparisons between red abalones and red sea urchins were significantly negative in both 1986 and 1999. when shallow and deep stations were combined. Comparisons at shallow depths were significantly negative in 1986 and 1999 (Fig. 4). Graphical comparison of correlation revealed more overlap or concurrence of both species on the same transects than observed at BMLR. At deep stations in 1986 the correlation was negative but not signifi- cant (Fig. 5). PCMR was the only area where high densities of both red abalones and red urchins were found in our study. Under these high densities both species were significantly negatively correlated (r„ = -0.52 P < 0.0001. Table 3). Shallow ( 2.0 - 8.3 m ) 8.0 E 6.0 d- (0 Q. 4.0 2.0 0.0 1973 ♦ Red Abalone DRed Sea Urchin A Purple Sea Urchin ft * 1986 1986 1999 Year 800 600 400 200 0 r 20 40 Red Abalone 60 1999 80 100 r. = -0.50 P = 0.02 N = 20 400 [%t ^v ^^^-^ i ► 50 100 Red Abalone r, = - 0.45 p = 0.05 N = 18 Figure 4. Red abalone. red sea urchin, and purple sea urchin densities (± se) at PCMR at shallow depth stations with paired comparison of counts by species at each station with Spearman Rank correlation coefficients (r,), statistical significance, and number of stations. Abalone and Urchin Interactions Deep( 8.4-18 m) 747 (0 V Q. 8.0 6.0 4.0 2.0 0.0 ♦ Red Abalone ■ Red Sea Urchin A Purple Sea Urchin 1 i i 1986 1999 Year Figure 5. Red abalone. red sea urchin, and purple sea urchin densities (± se) at PCMR at deep depth stations with paired comparison of counts b> species at each station with Spearman Rank correlation coefficients (rj, statistical signiflcance, and number of stations. VDSP In 1986 red abalone densities at VDSP were low in deep (0.1 m"- SE = 0. 1 ) and shallow depths (0.4 m"" SE = 0. 1 ), while red sea urchin densities were low at shallow (0.4 m"" SE = 0.1) and high at deep stations (1.7 m~- SE = 0.3) (Fig. 7, Fig. 8). Abalone densities had increased significantly to 0.8 (SE = 0. 1 ) at shallow stations, and to 0.9 m'" (SE = 0.2) at deep stations by 1992. Karpov et al. (1998) examined this difference using ANOVA and found the increase to be significant between years and depths (Table 2). During the same period, red urchin densities underwent significant decline by year and depth {P = 0.04 and 0.0001 re- spectively). By 1999. deepwater red urchin abundance had recov- ered to 2.4 m"- (SE = 0.5). while red abalones had declined to 1986 levels (0.1 m"- SE = 0.1). Repeating the ANOVA compari- son revealed that the change in red urchin abundance was no longer significant by year, but still significant by depth. Purple sea urchins, first examined in 1989, underwent significant increases by depth and by year, increasing primarily at deep stations to 1.8 m~" (SE = 0.7) in 1992 (Fig. 7. Fig. 8. Table 2). VDSP data allowed comparison of species correlation at vari- ous combinations of density for each of the species (Table 3). Except at low densities of each, red abalones and red sea urchins were significantly inversely correlated iP < 0.04). Again, as in the other two study areas, no significant correlation was observed for purple sea urchins when counts were paired to either of the other two species. Size distributions of both red abalones and red sea urchins at VDSP weighted for abundance were compared for evidence of recruitment by depth (Fig. 9). Red abalone size frequencies showed a clear pulse of juveniles at shallow depths in 1989. The mode in deep water appeared later, and grew into sizes larger than the sport size limit by 1992. In 1999. the deep water abalones remaining included few juveniles and resembled the 1986 and 1989 distribution. In shallow water, the distributions showed a large buildup of large adults over all previous years, with less evidence of recent juvenile recruitment than had been apparent in 1989 through 1992. Size distributions of red sea urchins, first measured in 1990. showed a deep water peak at the minimum commercial legal size TABLE 2. Red abalone and red sea urchin two-way ANOV.A probability values for log transformed density comparison by year, depth, and year*depth interaction at Van Damme State Park (VDSP) and Point Cabrillo Marine Reserve (PCMR). The first comparison at VDSP excludes 1999 data. Comparison Class Variable Red Abalone Red Sea Urchin Purple Sea Urchin VDSP (1 986. 1989- (Shallow, Deep) 1992) VDSP(19S6. 1982-1992. 1999) (Shallow. Deep) PCMR (1986. 1999) (Shallow. Deep) Year Depth Year*Depth Year Depth Year* Depth Year Depth Year'Depth 0.004 0.005 n.s. 0.0008 <0.0001 n.s. <0.0001 0.029 0.034 0.04 0.0001 n.s. n.s. <0.0001 n.s. 0.008 0.001 n.s. 0.03 0.0005 n.s. 0.03 0.0001 n.s. 748 Karpov et al. RED ABALONE SIZES PCMR RED SEA URCHIN SIZES PCMR UJ a. < o u s a: m s Z3 Figure 6. abalunes (89 mm). 0 25 50 75 100 125 150 175 200 225 250 0 20 40 60 80 100 120 140 160 180 SHELL LENGTH (MM) TEST DIAMETER (MM) SHALLOW DEEP _ SPORT LEGAL SIZE LIMIT COMMERCIAL LEGAL SIZE LIMIT Size density of emergent red abalone and red sea urchin transects at PCMR at both shallow and deep depths. Years compared for red are 1986 to 1999 and for red urchins 1988 to 1999. \ ertical Lines represent minimum legal size for red abalone ( 178 mm) and red urchin limit of 76 mm. first imposed in 1989 (Fig. 9). That mode declined through 1992 with little evidence of recruitment at either depth during that period. In 1999. a large cohort of red urchins, analo- gous to the earlier event for shallow water red abalones. had be- come established at deep depths with the peak, knife-edged, at the new 89 mm legal size limit, first imposed in June 1990. Apparent Competitive Effects Commercial landings of red sea urchins were first reported at the Van Damme area in 1985 (Fig. 10). Landings peaked at 800 mt in 1987. and declined along with CPUE (kg per diver day) through 1993. Size distributions of red urchins remained essentially fiat from 1987 through 1990. becoming increasingly positively skewed by 1 994 with few animals taken below the newly imposed size in 1990. and a decreasing number of animals larger than 120 mm taken (Fig. 1 1 1. Suiface kelp canopy at the VDSP area increased dramatically from 3 to 45 ha between 1982 and 1987. This coincided with increased red sea urchin renioxal (Fig. 10). Nereocysris tuetkeana was the dominant kelp but the canopy also included Macracystis integrifolia. While the level of sport take was not monitored at VDSP. telephone surveys conducted from 1986 to 1989 estimated that there were 30.000 to 40.000 sportsmen in the northern red abalone fishery (CDFG unpublished). In 1998. a $12 stamp was instituted for the abalone sport fishery. 32.000 were sold in 1998. and 35.000 in 1999. VDSP is among the most highly utilized sport fishing sites in northern California. Creel surveys conducted at this site duiing spring and early summer since 1977 reveal continued high use with no decline in CPLIE throughout our study period. Red abalone size distributions from diver creels at VDSP can be segregated into pre- and post-onset of the sea urchin fishery. The first period, from 1977 through 1985. shows little change (Fig. 1 1 ). Size distributions during this period were highly skewed to the right, with the mode adjacent the sport legal size of 178 mm. Two of these years. 1983 and 1984. had the most positively skewed distributions with the lowest mean size of 188 mm. The second period (1986 to 1994) is characterized by dispersal into a fiat distribution and an apparent modal progression from 182 mm in 1987 throuah 206 mm in 1992. DISCUSSION Spatial Interactions There was a significant negative correlation between red aba- lones and red sea urchins at all sites and depths with the exception TABLE 3. Density matrix summarizing Spearman Rank Correlation test for red sea urchins compared to red abalones at low and high densities of each at \ an Damme State Park (\'DSPl and Point Cabrillo Marine Reserve (PCMR). Species Red .4balones Density Low High (>l).4 per sq. m) Red Sea Urchins lOtt High Ol.O per sq. m) VDSP 1986 Shallow and 1989 Deep N = 29 n.s. VDSP 1986 and 1999 Deep N = IS r, = -0.63 p = 0.00.'; VDSP 1989-1992. 1999 Shallow and 1990-1992 Deep N = 101 r, = -0.22 p = 0.003 PCMR 1986 and 1999 N = 60 r. = -0.52 p < 0.0001 Abalone and Urchin Interactions 749 VDSP Shallow (2.0 -8.3 m) E 6 0.5 z — ♦ — Red Abalone ■ Red Sea Urchin A Purple Sea Urchin u 1985 1987 1989 1991 1993 1995 1997 1999 Year Figure 7. Red abalone. red sea urchin, and purple sea urchin densities (± se) at VDSP shallow depth stations in 1986. I9S9 to 1992, and in 1999. of deep-water stations at PCMR in 1999. This positive coiTelation is an artifact of data from only two stations; abalone were absent from 1 0 of 12 deep stations. The consistent negative correlation for all other comparisons suggests spatial interspecific exclusion be- tween these two species. Deacon (1973) found a similar negative correlation between red abalones and red urchins in Sea Ranch, a northern California location between BMLR and VDSP. In 2000 we completed an additional 24 einergent transects. 13 deep and I 1 shallow, at Salt Point State Park, located north of BMLR and south of VDSP (Bennett unpublished). Red urchins and red abalones were again negatively correlated in an area of high red abalone (0.84 m"^) and low red urchin (0.90 m"") densities. Once more, red urchins were most abundant at deep and red abalones at shal- low stations. Our finding of significant negative correlation at all but low densities of both red abalones and red urchins suggests that the observed spatial exclusion is density related. The increase in abalone abundance occurred during a period of increased surface VDSP Deep (8.4 -18 m ) 3.0 2.5 E 2.0 (A I- 1.5 » a i 1.0 0.5 ^ Red Abalone _»- Red Sea UrcNn -A_ Purple Sea UrcNn 1 • 1 \ 1 * ^ 1991 1993 Year Figure 8. Red alialone, red sea urchin, and purple sea urchin densities (± se) at \ l)SI' deep depth stations in 1986. 1989 to 1992. and 1999 with paired comparison of counts hj species at each station with Spearman Rank correlation coeflleienis (r,), statistical signitlcance, and number of stations in 1986 and 1999. canopy and therefore increa.sed food abundance. Karpov et al. ( 1998) reported on the increase in abalone populations at VDSP at both depths from 1986 to 1992. Mechanisms responsible for this increase could include large-scale lemoval of potential competitors by the urchin fishery and positive effects on abalone growth due to increased kelp production. The most profound lack of co-occurrence between red abalones and red sea urchins was at BMLR, the site with the least available macroalgae. This suggests competitive exclusion is greatest in habitat where food is limiting. It is likely that high silt loads from the Russian River (8 km to the north) and urchin grazing may limit kelp productivity in this area. Tegner et al. (1992) described areas adjacent to San Francisco Bay as marginal habitat, largely influ- enced by sediment loads from the bay. et al. ( 1997) reported com- parable habitat to BMLR, with an absence of canopy and lack of understory algae below 1 3-m depth, off Fitzgerald Marine Reserve (another area near San Francisco Bay) in 1993. Canopy and un- derstory kelps are the primary source of drift needed to support abalone as well as sea urchin populations (Deacon 1973, Tegner & Dayton 1991, Tegner et al. 1992). The lack of correlation (positive or negative) between purple sea urchins and the other two species, at the scale of our transects, suggests that the spatial distribution of this species is not strongly conelated to the other two species. We note, however, that Schroe- ter (1978) found that the longer-spined red urchins exclude purple urchins from the most desirable habitat (within the nr scale) by spine fencing. The mechanism responsible for the increase in purple urchin abundance at both shallow and deep stations in VDSP may have been a release from competition for space as red sea urchin stocks declined due to fishing, or a strong coincident recruitment event. Off Sea Ranch, northern California, Deacon ( 1973) found a significant positive coiTelation between red urchins and purple urchins but no significant correlation between purple urchins and red abalones. Tegner et al (1989) observed significant increases in purple urchins at Johnson's Lee. Santa Rosa Island, an area that had also been subjected to intense fisheries for both red urchins and red abalones. Lafferty and Kushner (pers. comm.) reported periodic rapid increase and collapses of purple urchin populations off the northern Channel Islands in southern California from 1983 to 1998. with decreases due to starvation and disease following warming events. Our results and those of Tegner et al (1992) and Ebert and Russell (1988) are consistent with strong episodic recruitment for this species but do not clarify whether this follows release from competition for space. Sea urchins are more adapted to periods of starvation than abalones, persisting for long periods on drift and microflora after macroalgae have disappeared from an area (Leighton 1968, Shep- herd 1973, Tegner & Levin 1982). In spite of the overlap in their diet, red sea urchins have a more generalist feeding strategy than abalone (Leighton 1966). Abalones are essentially drift feeders, while sea urchins can act as drift feeders or can form feeding fronts that actively graze attached kelps when food is extremely limited (Han-old & Reed 1985). At both PCMR and VDSP, we found understory algae to be least abundant at deep depths, suggesting that food is more limiting at deeper depths. More drift algae at shallower depths have been observed at other sites in northern California (Rogers-Bennett et al. 1995). The significant difference in red abalone and red sea urchin abundance by depth at VDSP and PCMR suggests differences in habitat preference between the two species. Greater numbers of red urchins were found at deeper stations and conversely greater num- 750 Karpov et al. RED ABALONE DENSITY RED SEA URCHIN DENSITY SHALLOW DEEP SPORT LEGAL SIZE LIMIT COMMERCIAL LEGAL SIZE LIMIT NO DATA- 0 08 004 : 1989 I 20 40 60 80 100 120 140 160 18 TEST DIAMETER (MM) 0 26 50 75 100 125 150 175 200 225 250 SHELL LENGTH (MM) Figure •>. Red ahalone and red sea urchin size densit) from emergent dive transects at \'DSP at botli shallow and deep depths. Years compared for red abalones are 1486, 1989-1992, and 1999; and for red urchins 1980-1992 and 1999. Vertical Lines represent minimum legal size for red abalone (178 mml and red urchin (89 mm). bers of red abalones at shallower depths. The absence of red aba- lones was most pronounced at deep depths off BMLR where food is most limiting, but was also apparent for abalones at VDSP. in spite of intense fishing pressure for abalone at shallow depths. The increased red abalone densities observed in deep stations at VDSP during the 1989 to 1992 period did not persist into 1999 and declined significantly following an increase in abundance of red urchins. A greater abundance of abalones at shallow depths in northern California may reflect their morphological adaptation to the high surge conditions found in shallow waters (Cox 1962. Deacon 1973). While both species prefer crevice habitats, espe- cially in areas of high predation such as the sea otter range (Hines & Pearse 1982). urchins are more easily dislodged by wave action (Dayton 1973). and would be most vulnerable outside of crevices at shallow depths. In New Zealand, manipulative experiments re- ducing sea urchin densities enhanced the recruitment, survival, and growth of abalones (Andrew et al. 1998). S < 1 600 — kelp cover CATCH / 1 o o CPUE 600 ^ / ^ ,o 400 - f \ 200 - 1 1 i \ o o ° o o o 0 - ■1 1 1 V-— -^ '*\ ■^ 90 YEAR Figure 10. Commercial landings of red sea urchins, catch per unit effort (CPUE), and kelp canopy cover in the VDSP area. Abalone and Urchin Interactions 751 >- o 2 lU o HI o cc. m a. RECREATIONAL RED ABALONE CATCH 1987 N=1440 mean=191 COMMERCIAL RED SEA URCHIN LANDINC 1 1987 1988 N=1670 mean=194 1989 N=1887 me3n=196 I Hnn 1990 N=1472 mean=198 miuu. 1991 N=1312 mean=200 \SU^ fl 11 1992 N=594 mean=197 n 1993 N=1518 mean=194 llflflnnn^- . 1994 N=1287 mean=199 170 182 194 206 218 230 242 250 SHELL DIAMETER 80 100 120 140 TEST DIAMETER 170 182 194 206 218 230 242 250 SHELL DIAMETER Figure 11. Size frequency distribution of commercial red sea urcliins taken from the VDSP area from 1987 (o 1W4 «itli size of sport-diver- landed red abalones from 1977 through 1994. Legal minimum sizes are 178 mm for red abalones and 89 mm for red sea urchins. Apparent Competitive Effects and Fishery Changes Karpov et al. ( 1998) reported on the strong recovery in abalone populations in both depths at VDSP from 1986 to 1992. Both red sea urchin removal and enhanced abalone growth with increases in algal food are likely to have contributed to this increase in abalone abundance. Low kelp abundance observed at VDSP in 1982 and 1983 probably reflected the effect of grazing by an unfished popu- lation of red urchins. The rapid increase in kelp cover concurrent with high levels of urchin removal suggests kelp productivity and abundance increased along with removal of the sea urchins. Un- fortunately, surface canopy aerial photographs were discontinued following 1989. Continued low levels of CPUE and increasingly skewed urchin size distributions in the fishery show (hat intense fishing continued throughout our study period. Red abalone growth appears to be somewhat plastic and able to respond to increased food availability (Haaker et al. 1998). Modal progressions observed in size frequency distributions of sport- caught red abalones in the VDSP vicinity between 1987 and 1993 imply a growth rate of about 4 mm per year for abalone greater than about 180 mm shell length (SL) (Fig. 1 1 ). This growth rate is greater than observed in previous tag and recapture studies in northern California (Tegner et al. 1992) and southern California (Haaker et al. 1998). A DFG (unpublished) study at Van Damme from 1973 to 1977 based on 275 recaptured tagged red abalone showed an annual growth rate of <1 mm at sizes above 178 mm SL. While this is an area for more rigorous research, we propose a mechanism for this empirical observation of increased abalone growth at VDSP. The importance of sea urchin grazing to the structure of kelp communities is well known (Nicholson 1970. Lawrence 1975, Harrold & Pearse 1987). Removal of an estimated 2.878 mt of red sea urchins from the VDSP area (appro.x. 3.5 km of coastline) from 1985 to 1993 coincided with an explosive increase in kelp canopy from 12.5 ha in 1985 to 44.2 ha in 1987 (Fig. 10). Concomitant with these events was a reduction in the larger size classes of red urchins. Red urchins Si 20 mm TD comprised 24.9% of the Men- docino County catch in 1992 (Mendocino County is used as a 752 Karpov et al. proxy for the VDSP sea urchin catch, because of inadequate sam- phng at that location). By 1998. these sizes made up only 9.5% of the catch. This suggests a mechanism for the observed abalone growth rate increase, whereby removal of significant amounts of a primary niacro-benthic grazer like sea urchins, as well as a con- tinuing vigorous fishery for the abalones themselves, allowed re- lease of the kelp canopy as well as subsurface kelps from grazing. This led to increased food availability for the remaining kelp graz- ers. Red sea urchin size distributions by density at deep stations in 1999 showed peaks at 85-99 mm TD at PCMR and 80-94 mm TD at VDSP (Figs. 6 & 9). Based upon a combination of observational and experimental evidence, these peaks appear to be comprised primarily of a cohort that settled during the 1992-93 El Nino. Significant settlement events of red urchins were noted on artificial substrates monitored since 1990 at shallow subtidal stations at PCMR and Westport (about 30 km north of PCMR) in late spring and early summer 1992 and spring 1993 (Ebert et al. 1994). Sub- tidal surveys in the vicinity of PCMR in fall 1994 noted a strong cohort averaging 18 mm TD. 16 months following the spring 1993 artificial substrate settlement event (Kalvass & Hendrix 1997). Ebert (1997) developed a growth transition matrix for California red urchins with annual probabilities of transfer between 10 mm size groups. We applied these probabilities to the observed 1994 10-20 mm cohort and estimated that 63'/(- would grow to the 80-100 mm size groups by the time we observed them in 1999 at both PCMR and VDSP. Growth of the observed cohort in 1994. from significant settlement events in 1992 and 1993, might ac- count for much of the peak near the commercial red urchin mini- mum size limit of 89 mm TD observed in 1999 (Fig. 9). Mechanisms Our finding that red abalones are displaced from deep but not shallow depths at VDSP. combined with the negative coiTelation observed in abalones and red sea urchins at comparable densities in 1986, strongly suggests that competitive exclusion of red aba- lones has occurred. Disappearance of red abalone in deep water could have resulted from other mechanisms including illegal take (Daniels & Floren 1998), a major mortality event from disease (Lafferty & Kuris 1993), or movement to shallow depths during a period of successful red urchin recruitment. Temporal changes in density at PCMR support the trend of exclusion of red abalones in deeper water but need to be inter- preted cautiously. Densities where compared between two years bracketing a thirteen year interval and not a time series as sampled at VDSP. In addition, since only half of the same locations were resampled in 1999, spatial variations in habitat could have exag- gerated (or underestimaled) actual temporal differences in abun- dance (Thrush et al. 1994). Red sea urchins in northern California appear to be much less mobile than red abalones. Ebert et al. ( 1999) tagged thousands of red urchins at 20 locations between southern California and south- east Alaska, for growth analysis. Recovery rates at these sites after approximately one year ranged from 13% to 76% for internally tagged urchins. These high recovery rates indicate that red urchins are relatively sessile. Conversely, red abalone tugging studies in- dicated abalone can move considerable distances. In a study at PCMR, Ault and DeMartini (1987) found that in an area of high abundance 11% of the tagged population moved over 90 m, with one animal moving 0.6 km. Similarly, high numbers of tagged red urchins ( 16-38%) were recovered after one year from sites in Salt Point in northern California. Twice as many urchins were recov- ered from a shallow site compared with a deep site, suggesting urchins in shallow water are more sedentary (Rogers-Bennett 1994). Deacon (1973) also found that the vast majority of red urchins moved far less than red abalone, during movement experi- ments at Sea Ranch. In the spring of 1998 starving red abalone were reported by divers south of Noyo Bay. 14 km north of VDSP (Haaker & Karpov, unpub. obs.). They observed a marked absence of understory kelps with evidence of abalone undergoing unusual movements over foliose algae in search of food. One possible explanation for the decline in deep-water abalones off VDSP could be recent movement to shallow depths in search of food. Continued low numbers of red sea urchins at shallow depths in Van Damme are expected, given the high density of red abalone remaining at shallow stations in 1999. Urchins could only survive at shallow depth in crevice areas protected from extreme sea con- ditions (Deacon 1973). and it is unlikely that red urchins could displace red abalones from such habitat once occupied. While we found that the removal of red sea urchins at shallow depths in VDSP appeared to enhance the abundance of adult red abalones. juvenile abalones are known to shelter under the spine canopy of red sea urchins, complicating the relationship between red abalo- nes and red sea urchins (Tegner & Dayton 1977. Rogers-Bennett & Pearse 2001). CONCLUSIONS Our results suggest that spatial competition between red aba- lones and red sea urchins is density related and appears to be most pronounced in habitats where macroalgae are scarce. In kelp beds these species co-occur on transects more frequently but are nega- tively correlated when either species is in high abundance. Intense red urchin fishing at VDSP coupled with a major recruitinent event of red abalones appears to have enhanced abalone density in both shallow and deep depths through 1992. This release from inter and intraspecific competition for space and food resulted in a large surplus abalone population at refuge depths through the early 1990s. However, significantly lower densities of abalone are now apparent at depth. A major recruitment event for red sea urchins appears to have depressed abalone populations at depth. We cannot rule out the possibility of movement, mortality, or poaching of deep water red abalones as causes for the decline. In the absence of fishing for both species, densities of red abalones at an adjacent PCMR reserve site are now below those found at the intensely fished area, while sea urchins have increased to greater numbers than fished abundance. With its high urchin abundance, PCMR kelp populations are also lower than at nearby VDSP. Together this suggests that fishing of both red urchins and red abalones at VDSP enhanced abalone densities. We therefore suggest that an ecosystem approach that takes into account multispecies interac- tions, should be an important consideration in managing these fisheries in northern California. ACKNOWLEDGMENTS We especially wish to thank Neville Sweijd and Peter Cook for making presentation of our work possible. We thank Carolyn Friedman. Jerry Kashiwada, Mike Prall, and Kristen Riser for their contributions to the analysis. We also wish to thank the many research divers and biologists who over many years have contrib- uted their time and efforts in field collections used in our study. Thanks to Sonoma County Fish and Wildlife Advisory Board for supporting travel expenses for L. Rogers Bennett. Abalone and Urchin Interactions 753 LITER.ATURE CITED Andrew, N. L., D. G. Worthington. P. A. Bretl, N. Bentley. R. C. Chick & C. Blount. 1998. Interactions between the abalone fishery and sea urchins in New South Wales. N.S.W. Fish. Rep. Set: 12:1-108. Ault, J. S. & J. D. DeMartini. 1987. Movement and dispersion of red abalone. Hahoris nifescens. in northern California. Calif. Fish Game. 73:196-213. Cox. K. W. 1962. California abalone. family Haliotidae. Calif. Fish Game Bull. 118. 133pp. Daniels. R. & R. Floren. 1998. Poaching pressures on northern California's abalone fishery. / Shcl. Res. 17:859-862. Daylon. 1973. Two cases of resource partitioning in an intertidal commu- nity: making the right prediction for the wrong reason. ,\m. Naliiral. 107:662-670. Deacon, J. 1973. Habitat selection and competition among abalone and sea urchins at The Sea Ranch, California. M.Sc. Thesis. University of California. Davis. 83 pp. Ebeil, T. A. 1997. An analysis of the importance of AUee effects on management of the red sea urchin {StrongyloeenimiKs franciseaniis). In: Mooi & Telford, editors Proc. 9"^ Intl. Echinoderm Conf.. Rotter- dam: A. A. Balkema. pp. 619-627. Ebert. T. A. & M. P. Russell. 1988. Latitudinal variation in size structure of the west coast purple sea urchin: A correlation with headlands. Limnol. Oceanogr. 33:286-294. Ebert, T. A., S. C. Schroeter, J. D. Dixon & P. Kalvass. 1994. Settlement patterns of red and purple sea urchins (Strongyloeentrotusfranciscanus and S. purpiiratus). California. USA. Mar. Ecol. Prog. Ser. 1 1 1:41-52. Ebert. T. A.. J. D. Dixon. S. C. Schroeter. P. E. Kalvass, N. T. Richmond. W. A. Bradbury & D. A. Woodby. 1999. Growth and monality of red sea urchins tSlrongylocentrorus franciscaniis) across a latitudinal gra- dient. Mar. Eeol. Prog. Ser. 190:189-209. Haaker. P. L.. D. O. Parker. K. C. Barsky & C. S. Chun. 1998. Growth of red abalone. Haliotis nifescens (Swainson) at Johnsons Lee. Santa Rosa Island. Calif. / Shel. Res. 17:847-854. Harrold, C. & D. C. Reed. 1985. Food availability, sea urchin grazing, and kelp forest community structure. Ecology. 66:1 160-1 169. Harrold, C. & J. S. Pearse. 1987. The ecological role of echinoderms in kelp forests. In: M. Jangoux & J. M. Lawrence, editors. Echinoderm Studies 2. Rotterdam: A. A. Balkema. pp. 137-233. Hines. A. H. & J. S. Pearse. 1982. Abalones. shells and sea otters: dynam- ics of prey populations in Central California. Ecology. 63:1547-1560. Kalvass. P. E. & J. M. Hendrix. 1997. The California red sea urchin. Strongylocentrotus francisciamis. fishery: catch, effort, and manage- ment trends. Mar. Fish. Rev. 59:1-17. Karpov. K. A.. J. Geibel & P. Law. 1997. Relative abundance and size composition of subtidal abalone {Haliotis sp.). sea urchin (Strongylo- centrotus sp.) and abundance of sea stars off Fitzgerald Marine Re- serve. California. September 1993. Calif. Dept. Fish Game Mar. Res. Admin. Rep. No. 97-1. 16 pp. Karpov, K. A., P. Haaker. D. Albin. I. K. Taniguchi & D. Kushner. 1998 The red abalone. Haliotis nifescens. in California: Importance of depth refuge to abalone management. / Shel. Res. 17:863-870. Kalo. S. & S. C. Schroeter. 1985. Biology of the red sea urchin. Strongy- locentrotus franci.uianiis. and its fishery in California. Mar. Fish. Rev. 47:1-20. Lafferty, K. D. & A. M. Kuris. 1993. Mass mortality of abalone Haliotis cracherodii on the California Channel Islands: tests of epidemiological hypothesis. Mar Ecol. Prog. Ser 96:239-248. Lawrence. J. M. 1975. On the relationships between marine plants and sea urchins. Ocean. Mar. Biol. Ann. Rev. 13:213-286. Leighton. D. L. 1966. Studies of food preference in algivorous inverte- brates of southern California kelp beds. Pac. Sci. 20:104-1 13 Leighton. D. L. 1968. A comparative study of food selection and nutrition in the abalone, Haliotis rufescens (Swainson) and the sea urchin, Strongylocentrotus purpiiratus (Stimpson). Ph.D. Thesis, San Diego: University of California. 197 pp. Leighton, D. L. & M. Boolootian. 1963. Diet and growth in the black abalone. Haliotis cracherodii. Ecology. 44:227-238. Nicholson, N. L. 1970. Field studies on the giant kelp Nereocxstis. J. of Phyc. 6:177-182. North. W. J. & J. S. Pearse. 1970. Sea urchin population explosion in southern California waters. Science. 167:209. Parker. D. O.. P. L. Haaker & K. C. Henderson. 1988. Densities and size composition of red abalone, Haliotis rufescens. at five locations on the Mendocino and Sonoma county coasts, September 1986. Calif. Dept. Fish Game. Mar. Res. Div., Admin. Rep. No. 88-5. 65 pp. Pearse. J. S. & A. H. Hines. 1987. Long-term population dynamics of sea urchins in central California kelp forest: rare recruitment and rapid decline. Mar. Ecol. Prog. Ser 39:275-283. Rogers-Bennett. L. 1994. The spatial patterns in the life history character- istics of red sea urchins. Strongylocentrotus franci.wanus: implications for recruitment and the California fishery. University of California. Davis, California, USA. Rogers-Bennett, L., W. A. Bennett, H. C. Fastenau & C. M. DeWees. 1995. Spatial variation in red sea urchin reproduction and morphology: im- plications for harvest refugia. Ecol. App. 5:1 171-1 180. Rogers-Bennett. L. & J. S. Pearse. 2001. Indirect benefits of Marine Pro- tected Areas for juvenile abalone. Conserx: Biol. 15:642-647. Schroeter. S. C. 1978. Experimental studies of competition as a factor affecting the distribution and abundance of purple sea urchins. Strongy- locentrotus purpuratus (Simpson). Ph.D. thesis. Santa Barbara: Uni- versity of California. 1 84 pp. Shepherd, S. A. 1973. Competition between sea urchins and abalone. .\ust. Fish. 32:4-7. Tegner. M. J. 1980. Multispecies considerations of resource management in southern California kelp beds. Can. Tech. Rep. Fish. Acj. Sci. 954: 125-143. Tegner. M. J. & P. K. Dayton. 1977, Sea urchin recruitment patterns and implications of commercial fishing. Science. 196:324-326 Tegner. M. J. & L. A. Levin. 1982. Do sea urchins and abalones compete in California kelp forest communities'? In: J. W. Lawrence, editor. Intl. Echinoderms Conf.. Tampa Bay. Rotterdam: A. A. Balkema. pp. 265- 271. Tegner. M. J. & P. K. Dayton. 2000. Ecosystem effects of fishing in kelp forest communities. ICES. J. Mar. Sci. 57:579-589. Tegner, M. J.. P. A. Breen &. C. E. Lennert. 1989. Population biology of red abalone. Haliotis nifescens. in southern California and management of the red and pink. H. corrugata. abalone fisheries. Fish. Bull. U.S. 87:313-339. Tegner. M. J. & P. K. Dayton. 1991. Sea urchins. El Nifios. and the longlemi stability of southern California kelp communities. Mar. Ecol. Prog. Ser. 77:49-63. Tegner, M. J., J. D. DeMartini & K. A. Karpov. 1992. The California red abalone fishery; a case study in complexity. In: Abalone of the World: Biology. Fisheries and Culture. In: S. A. Shepherd. M. J. Tegner. & S. Guzman del Proo. Oxford: Blackwell Scientific, pp. 370-383. Thrush. S. F.. R. D. Pridmore & J. E. Hewitt. 1994. Impacts on soft- sediment macrofauna: the effects of spatial variation on temporal trends. Ecol. Appl. 4:31-41 Jouruul i>f Shclllish Rcscanh. Vol 20. No. 2, 755-7fi3. 2001. CLIMATE VARIABILITY, KELP FORESTS, AND THE SOUTHERN CALIFORNIA RED ABALONE FISHERY MIA J. TEGNER,^ t PETE L. HAAKER,- KRISTIN L. RISER,' * AND L. IGNACIO VILCHIS' 'Scripps liisrirutidii of Occuinii-niphy. ihiivcrsity of Califomia. San Diego. California 92093-0227; 'Califoinia Department of Fish and Game. Los Alamitos. California 90720 ABSTRACT Declines in landings of Southern Califomia abalone fisheries and the eventual collapse of many stocks over the last two decades coincided with a period of greatly increased environmental variability. This included massive storms, an increase in the frequency of warm-water El Nifio events after 1976. and an interdecadal-scale increase in sea surface temperatures. Kelp populations may be decimated by severe storms or warm water. Because of the strong inverse relationship between nitrate availability and water temperature, temperature is a good indicator of nitrate availability or stress SINCE kelp growth ceases in warm nutrient-depleted water, tissue decays, and standing stocks may be greatly reduced. Abalones are affected by the availability of the drift kelp on which they feed. Anomalously warm temperatures may affect reproduction, and altered current patterns may affect larval dispersal. Because water temperature varies with location in Southern California and each of the five exploited abalone species has its own thermal preferences, we chose to evaluate the role of environmental variability on populations of red abalone {Haliotis nifescens) on three northern Channel Islands spanning a temperature gradient. We compared water temperature regimes and anomalies, monthly aerial surveys of canopies of giant kelp (Macrocystis pyrifera). and field evidence of poor abalone growth and reproduction during El Nifio events. The severity of El Nifio disturbances and long-term changes in kelp standing stocks both correlated with the temperature gradient. Declines of red abalone total landings and area-specific landings on the warmer Santa Cruz and Santa Rosa Island began a decade after the large 1957-1959 El Nino. The subsequent collapse of many populations appears related to warm anomalies after the 1976-1977 regime shift, kelp declines, and poor reproduction coupled with fishing-mduced declines in adult abalone density. Red abalone populations have persisted on cooler San Miguel Island where thermal anomalies had less effect and kelp canopy biomass has been more stable. Southern California abalones evolved in this disturbance regime, but the combination of extended periods of increased envudnmenlal variability with intense fishing pressure may have led to the loss of local populations, especially in warmer areas. KEY WORDS: El Nifios, Macrocystis. storms, ocean climate, reproduction, stock collapse INTRODUCTION Declines in the landings of Southern Califomia abalone fish- eiies and the eventual collapse of many abalone stocks over the last two decades (Karpov et al. 2000) coincided with a period of changing ocean climate (McGowan et al. 1998). Since 1976 to 1977, there has been an increase in the frequency, duration, and intensity of warm- water El Nifio events and a corresponding de- crease in cold-water La Nina events (Fig. 1 ). Frequently described as a regime shift, this has been accompanied by a significant interdecadal-scale increase in sea surface temperature (Roemmich & McGowan 1995), and a shifting in the mean location of sea surface isotherms to the north along the West Coast of North America (Fig. 2). Furthermore, large waves in Central and South- em Califomia are strongly associated with El Nifio events, and there has been a marked increase in the number of broad area wave events exhibiting very large waves (Seymour 1996). El Nifios and large wave events are important agents of distur- bance to kelp forest communities, with severities that vary with latitude and exposure (Tegner & Dayton 1987, Seymour et al. 1989). El Niiio storms, such as those that produced the extraordi- nary number of large wave events in winter 1983, may decimate exposed kelp populations along the entire coast of Califomia. Storms also kill abalones directly (Cox 1962). The effects of the warm, nutrient-depleted waters, however, are stronger in the south- em portion of the range of giant kelp, Macrocystis pxrifera. the major source of productivity and structure in this community (Teg- ner & Dayton 1987). Sea surface temperature is the best predictor of harvestable Macmcystis biomass (Tegner et al. 1996). Because *Corresponding author. E-mail address: kriser@ucsd.edu tDeceased. of the strong inverse relationship between nitrate availability and water temperature, temperature is a good indicator of nitrate avail- ability or stress, as kelp growth ceases in warm nutrient-depleted water, tissue decays, and standing stocks may be greatly reduced. Community effects of the 1982 to 1984 El Nino were apparent long after the anomalous oceanographic conditions, as manifested by increased abundance of understory kelps relative to that of Macrocystis (Tegner et al. 1997). The long-temi increase in tem- peratures since 1977 is associated with a 2/3 decrease in the me- dian size of giant kelp plants (measured as a decline in number of stipes) and thus a decrease in drift kelp availability (Tegner et al. 1996). Furthermore, because strong El Nino events are adding to the secular increase in ocean temperatures, the 1997 to 1998 event was warmer than the very strong 1982 to 1984 event. While the effects on kelps are relatively well understood, the impacts of these changes in ocean climate on consumers are poorly known. Temperature affects animals directly, and temperature and storms both affect the food supply of grazers. The west coast of North America supports eight species of abalones, and their geo- graphic and depth distributions are largely determined by tempera- ture (Leighlon 1974). Thus, effects may be apparent in distribution as well as physiology. Macrocystis is the dominant source of drift kelp in Southern California (Tegner & Dayton 1991), and the major food for abalones (Tufschulte 1976). With up to 60% of the biomass of a healthy Macrocystis forest in its canopy (North 1968), the loss of that canopy and varying degrees of plant mor- tality lead to a corresponding decrease in drift availability. Under- story kelps produce far less drift per unit area than Macrocystis. Drift availability is very low after major storin disturbance, juve- nile giant kelp populations produce little drift, and thus drift pro- duction may remain low for months to years after an El Nino event (Tegner & Dayton 1991). Furthermore, warm, nutrient-depleted 755 756 Tegner et al. 1 0 -1 -2i MULTIVARIATE ENSO INDEX NOM-CIRES CHmab Oiognoeb'ca Canter (CDC), UnrvsraiV o) Colorado a^ Bouldar 1950 1955 1960 1965 1970 1975 1980 19B5 1990 1995 2000 Figure 1. Multivariate El Nino-Southern Oscillation Index (courtesy of National Oceanograpliic and Atmospheric Administration's Climate Diagnostic Center). Downloaded in January 2(1(10 from http://\vH\v.cdc.noaa.gov/ENSO.meiJnde\.html). El Nino events are positive, Ea Nina events are negative. This is a tropical index; events strong enough to affect the California Current region exceed some threshold about the mean. conditions also lead to a decrease in the nitrogen content of kelps (Tegner & Dayton 1987), resulting in both the quantity and quality of abalone food being altered. Previous observations indicate that abalones are affected by environmental variability. Cox (1962) noted that "during 1957. 1958. and 1959. the main food supply of abalones off California 40 N 125°W 120 W 115 110" Figure 2. Changing isotherm locations along Baja California (Mexico) and California (IS.A). The Southern California Bight extends from Point Conception to just below the Mexican Border. The present Cali- fornia Cooperative Oceanic Fisheries Investigations iCalCOFI) grid is illustrated by station lines. The range of Macrocystis pyrifera is from just south of San F'rancisco to the region just south of Punta Eugenia. COADS (Comprehensive Ocean Atmosphere Data Set, htip:// wvvw.cdc.noaa.gov/coads/) data were used to determine the location of the annual mean sea surface temperature (SST) isotherms before and after the mid m7(ls regime shift. The dashed line represents the posi- tion 1956 to 1976 and the solid line represents 1981 to 1992. practically disappeared when an unseasonal influx of w ami water destroyed all of the kelp beds. During this starvation period, the gonads of the abalones did not increase in size during their regular spawning season and many probably were unable to spawn." He also observed that growth was minimal or nonexistent, and that body tissues appeared to shrink. The flesh of abalones that did enter the fishery was shrunken and watery. Haaker et al. (1998) later quantified strong interannual variability in growth rates, in- cluding a dramatic reduction in the growth rate of Haliotis nife- sceiis (red abalone) during the 1982 to 1984 El Nifio event. Vega et al. (1997) observed an inverse correlation between sea surface temperature anomalies and total catch of abalones from Baja Cali- fornia which they related to impacts on algal food abundance. On the other hand. Shepherd et al. (1998) found total abalone catch at La Natividad. Mexico, to be positively correlated with mean sea surface temperature anomalies with a lag of eight years, but also reported evidence for recruitment failure during severe El Nifios. Field studies on the Palos Verdes Peninsula (33°46' N, 1 18°25' W) near Los Angeles during and after the strong 1982 to 1984 El Nino event suggest that anomalies in both temperature and food availability affect abalone reproduction and recruitment. We fol- lowed gonadal development of green (H. fiilgeiis) abalones from February 1982 through September 1983. Normally these animals spawn in both the spring and the fall. El Nino storms in January 1983 removed all Macrocyslis: the abalones had little if any food until kelp recruitment in May. Gonadal development in spring 1983 began much later than in 1982. was very low in comparison with the previous year, and it is unlikely that there was significant spawning (Tegner & Dayton. 1987, Fig. 9). The green abalone is a wami-water species near the northern end of its range at Palos Verdes (Cox 1962) and winter-spring 1983 was before the major temperature anomalies associated with this ENSO; thus the ob- served depression in reproduction may have been primarily caused by lack of food. Si/e-frequency distributions of cool-temperate red and warm-temperate pink {H. corriigatu) abalones collected in 1986 at Palos Verdes suggested that El Nino conditions affected these congenors quite differently. There was a near absence of two and three year old red abalones. animals that would have been spawned in 1983 and 1984: virtually all the pink abalones were in these size categories (Tegner & Dayton. 1987. Fig. 10). These data suggest that temperature anomalies may have favored pink abalo- nes over the colder water reds during the ENSO event. We cannot exclude large-scale alterations in cun'ent patterns during El Niiio events (Tegner & Dayton 1987) also affecting abalone recniitment. but this seems unlikely given the short larval period (McShane 1942). On smaller scales, the devastation of kelp forests and con- Red Abalone Fishery and Climate Variability 757 comilant reductions of their effects on flow could increase local dispersal of larvae (Tegner & Dayton 1987). While it is likely that all abalone species are affected by vari- ability in ocean climate, the effects will vary with species and location. Because Southern California is a mi.xing ground between cool temperate waters of the California Current coming from the north and warm-temperate flow from the south (Fig. 3). compari- sons of how populations in thermally distinct habitats respond to interannual \ ariability may offer insights into the relative effects of different factors. Here we compare red abalone populations across the thermal gradient from colder San Miguel Island to Santa Rosa Island to wanner Santa Cruz Island. We use long-term records of temperature, kelp canopy biomass. Tield assessments of abalone populations, and fisheries statistics to explore hypotheses relating environmental \ ariability to the decline of the red abalone fishery and the collapse of mans local stocks. .\L\TERI.4LS AND METHODS Historically the red abalone is the most important species in the California abalone fishery bv virtue of its large size, wide distri- bution, and long harvest history. Its range extends over 16° of latitude from Coos Bay. Oregon to Bahia Tortugas. Mexico (Cox 1962). From Southern California south, however, it is limited to the northwestern Channel Islands which are hea\ ily influenced by the cold waters of the California Current (Fig. 3). and to mainland locations where strong upwelling moderates temperatures (Tegner Figure 3. This satellite image illustrates thermal variabilit\ along the Central and Southern California coastline on November 21. 1996. The coldest temperatures are white and the warmest are black. Note the cold Haters of the California Current flowing south past Point Con- ception and the warm waters of the Southern California Kddj flowing north along the mainland: mixing of these two water masses creates temporal and spatial \ariahilit\ within the Southern California Bight. The three islands immediately helow Point Conception are. from west to east. San Miguel. Santa Rosa, and Santa Cruz Island. NOA.\ A\ HRR (.Advanced \ er> High Resolution Radiometer) sea surface temperature data provided by the NO.\.A Coastwatch Program for and displayed in Windows Image Manager. et al. 1992). Leighton (1974) reports that he found fecund, spawnable red abalones every month of the year (1969-1971) at Estero Bay in Central California, whereas in Southern California he obtained laboratory spawnings in January. February. April, Sep- tember, November, and December. Before the fishery was closed in 1997. size limits were 178 mm and 197 mm for the recreational and commercial harvests, respectively. Channel Island sea surface temperature data (courtesy of Jack Engle. Tatman Foundation, University of California, Santa Bar- bara) were generated from regional satellite-derived isotherm maps of the California coast posted several times a week by the National Weather Service (http://www.nws.mbay.net/sstl.gif). Data are interpolated to 0.5"C, and averaged by month for each island. Mean monthly temperatures calculated for the period 1982- 1998 were used to determine temperature anomalies through time. Biologists from International Specialty Products (formerly Kelco) have been making appro.ximately monthly aerial estimates of Macrocystis biomass available for harvesting (California law allows removal of the upper 1.2 m of the canopy) for more than three decades. As these estimates are made by a very small number of observers and continually refined against the actual harvest tonnage, we believe that these estimates are quite accurate. There were no estimates made during 1992 and 1993. To protect propri- etary harvest information and to eliminate the effect of the large differences in size among the islands, data are presented as a proportion of the maximum observed in the time series for each island. Linear models for all three islands were fitted with the percent of maximum annual relative canopy biomass as the re- sponse variable and year as the predictor variable. The error dis- tributions of the relative annual canopy biomass data were deter- mined to be normal according to visual analysis of residual-fit spread plots and normal quantile plots of residuals. These assured that all of the three fitted models adequately explain the variation in the data (Cook & Weisberg 1982. McCullagh & Nelder 1989). Sex ratio was assessed visually (Ebert & Houk 1984) at San Miguel Island in May 1997. Red abalones of varying sizes were collected from several locations on San Miguel Island with a target goal of 30 per sample. Animals were brought to the surface and gonad color was scored by at least two observers; dark colored (green to black) for females and light colored (cream to light brown) for males. Size-frequency distributions of red abalones from a single lo- cation on Santa Rosa Island and from four locations on San Miguel Island were collected in June 1999. The data were summed for the San Miguel Island sites. RESULTS Tiinperature Records Mean monthly temperatures for the three islands for the period 1982 to 1998 are plotted in Figure 4. It is apparent that San Miguel Island is consistently colder than Santa Rosa Island, which is con- sistently colder than Santa Cruz Island. The differences are mini- mal in the winter and maximal in late summer. Monthly averages were compared using Friedman's two-way ANOVA. The results showed significant differences between sites (P < O.OOI ) and an a posteriori test (Nemenyi 1963) indicated that each island's tem- perature regime was significantly different from the others". Note that this data set begins after the 1977 regime shift. Time series of ambient temperature in comparison with the mean monthlv values and the anomalies for 1982 to 1998 are 758 Tegner et al. 19 1 14 13 12 Santa Cruz Is. Santa Rosa Is. San Miguel Is. Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 4. Mean monthly temperatures, 1982 to 1998. for San Miguel, Santa Rosa, and Santa Cruz Island. plotted in Figure 5. Note the 1982 to 1984. 1992 to 1993. and 1997 to 1998 El Niiio warm anomalies. La Nina events took place in 1988 to 1989 and 1999 (Fig. 1 ). The timing and magnitudes of the anomalies were remarkably similar among the three islands, but the summer values were higher at Santa Cruz Island and the winter values lower at San Miguel Island because of the difference in average temperatures. Macrocystis Canopy Biumass Time series of aerial estimates of giant kelp canopy biomass {Fig. 6) exhibit the strong interannual variability characteristic of Miunicysiis populations (e.g. Tegner et al. 1996). Many of the peaks and depressions are common to all three islands, for example the regrowth in 1 984 and 1985 after the winter 1983 storms and the 1983 to 1984 El Nino warm anomaly disturbances. Because some of the within-year decreases in canopy biomass are due to natural disturbance and some to harvesting, the annual maximum biomass estimates for each island are compared in Fig. 7. Some of the patterns are strikingly similar in timing and magnitude, including the decline in canopy at all three islands from 1978 to the low 20% of maximum biomass from 1986 through 1999. San Miguel Island, the coldest of the three, had a higher relative canopy biomass than intermediate temperature Santa Rosa Island from 1995 through the end of 1999. Regressions of maximum annual relative canopy biomass against time were used to determine whether (a) kelp populations on the different islands had declined, (b) whether the regression coelTicients changed at the time of the 1976 to 1977 regime shift, and (c) whether the islands were different from each other, (a) Analysis of variance indicated that two of the three slopes were significantly different from zero. Relative canopy biomass de- clined significantly at Santa Rosa Island (slope -0.01 1, r" = 0.19. P = 0.04), and at Santa Cruz Island (slope -0.014. r- = 0.28. P = 0.001) from 1968 to 1999. (b) T tests comparing regression coefficients before and after the 1976 to 1977 regime shift were not significant for any island, (c) Analysis of covariance. however, found significant differences in the elevations of the regressions {P = 0.02). Multiple comparisons (Tukey test) of the elevations of the regressions indicated that San Miguel Island canopy biomass was significantly different from that at Santa Cruz Island, but Santa Rosa Island canopy biomass was not significantly different from either San Miguel or Santa Cruz Island canopy biomass. Thus, kelp canopy biomass has persisted better at San Miguel Island relative to Santa Cruz Island. U IS o 14 San Miguel Is. 1982 1983 1984 1985 198S 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 %rtf"| ^n|||YfS|r^ Santa Rosa Is. O 16 o 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 c I < -2 . U I J. 11 I. I ffifl '1 "T A 22 20 O '6 O Santa Cruz Is. 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Year Figure 5. Time series of monthly ayerage temperatures (solid lines) in comparison ^vith mean monthly temperatures (dotted lines) for the period 1982 to 1998 (upper panels). Monthly anomalies (hars) from mean monthi) temperatures, 1982 to 1998 (loyyer panels). Sec text for details. San Miguel, Santa Rosa, and Santa Cruz Island. Ahalone Field Data In May 1997, red abalone sex ratios (assessed visually) on San Miguel Island were heavily skewed toward females. The sex ratio was 5.8:1 (female:male) on the west side of the island (/; = 34). 5.7:1 on the south side of the island (/i = 60). and 5.6:1 on the north side (« = 33). At this time near the onset of the 1997 to 1998 El Nino, the surface temperature on San Miguel Island was about \''C above normal (Fig. 5) and kelp canopy biomass was only 10% Red Abalone Fishery and Climate Variabilh y 739 l»72 iyJ4 ^ ioo<»-- d 75%' U 50% ■ lu Santa Cruz Island ifU* ,*H [If ipjl|i,i ^. ,>,itv^.»i^^ 'r^H I* f *, i4iiffc ii*. C 25% - rt) 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2O00 Survey date Figure 6. Monthly aerial estimates of Macrocystis pyrifera canop> biomass from San Miguel, Santa Rosa, and Santa Cruz Island An asterisk (*) indicates no survey. of niuxiinum (Fig. 6). The sexes of 13 animals could not be de- ternimed. Observers noted that sex determination was more diffi- cult than in February 1997. when temperatures were slightly below the long-term mean. The red abalone size-frequency distribution on San Miguel Is- land in June 1999 was strongly bimodal (Fig. S). The near absence of sizes representing two and three year old indi\iduals suggests poor reproduction and recruitment in 1997 and 1998. The strong mode at 20 mm suggests that reproduction quickly followed the onset of the La Nifia and cooler temperatures in November 1998; the kelp canopy recovered to 479r of maximum by January 1999 (Fig. 6). Limited size-frequency data collected at San Miguel Is- land in 1993 revealed very few recruits; a collection from Castle Rock on western San Miguel Island in 1994 after the 1992 to 1993 El Nifio found a burst of small animals (Haaker unpub. data). The small 1999 sample from Santa Rosa Island, where red abalone densities are very low (Fig. 8), suggests that reproduction and recruitment have not been successful on this island for the last four years. DISCUSSION Snapshots of sea surface temperature (Fig. 3). mean monthly temperatures (Fig. 4). and biogeographic differences in species composition (e. g., Murray et al. 1980) all indicate that there is a significant thermal gradient across the northern Channel Islands troni coldest al San Miguel Island in the west to warmest at Santa Cruz Island in the east. Temperatures are highly variable in time. but the data in Figure 3 suggest that approximately constant aver- age temperature differentials among the three islands are main- tained through El Niiio and La Nifia events, as well as interdecadal scale changes. Temperature is directly important to abalone physi- ological processes and larval survival (Leighton 1974). to harvest- able Macrocystis biomass (Tegner et al. 1996). and to the quantity and quality of drift Macrocystis (Tegner & Dayton 1987). the major food source of abalones (Tutshulte 1976). Because of the strong inverse relationship betw een temperature and nitrogen availability. Macrocystis populations are very sensi- tive to extended periods of warm water. Warm water per se does not harm giant kelp which, if provided with abundant nutrients, grows well even at 25°C (reviewed by North & Zimmerman 1984). There are negligible amounts of nitrate above 15°C (Jackson 1977, Gerard 1982. Zimmerman & Kremer 1984) but giant kelp thrives at 14"C. so very small differences in temperature can have impor- tant community effects. Macrocystis can build internal N reserves when external concentrations are high, and then use these reserves to maintain relatively rapid growth for up to a month in the ab- sence of significant external nutrients (Gerard 1982, Zimmerman & Kremer 1986). When kelp tissue N drops below K/r (dry weight), internal reserves are depleted, and N starvation will lead to rapid deterioration. During most summers when surface tem- peratures exceed 16°C, upwelling and internal wave-induced ther- mocline motions provide nutrient input to maintain giant kelp populations (Zimmemian & Kremer 1984). During strong El Nino events, the thermocline is often depressed to levels too deep for 760 Tegner et al. E 100%- © c « 9 S S 05 50% 01 X 25% San Miguel Island Santa Rosa Island Santa Cruz Island o%i — 1 — \ — I — I — I — I — I — I — I — I — I — I — 1 — I — 1 — I — 1 — I — 1 — I — 1 — I — 1 — I — 1 — I — 1 — I — I — [ — 1 — I 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Year Figure 7. Maximum annual estimates of Macrocystis canopy biomass from San Miguel, Santa Rosa, and Santa Cruz Island. these processes to be effective (Zimmerman & Robertson 1983). Macrocystis canopies were completely lost on the southern-most Santa Calalina and San Clemente Islands and along the southern- most counties. Orange and San Diego, in California during the 1982 to 1984 El Nino (Tegner & Dayton 1987). Further to the north, given temperature decrease with latitude, the warm anoma- lies added to lower temperature bases and there was .some degree of giant kelp canopy formation. Conversely. Macrocystis thrives during the cold water and enhanced upwelling of La Nina events (Tegner et al. 1997). Santa Cruz. Santa Rosa, and San Miguel Island experienced temperature anomalies during recent El Nino events that were remarkably similar in timing and magnitude; the difference is that the anomalies added to a different temperature base at each island. Thus, the anomalies were more likely to make the environment unsuitable for Macrocystis at Santa Cruz Island, and that may explain low standing canopy biomass esti-nates at that island from the mid 1980s through 1999 (Figs. 6. 7). There were significant declines in relative canopy biomass at Santa Cruz and Santa Rosa Island, but not at cooler San Miguel Island. We stress, however, that aerial surveys cannot detect sea urchin banens. another po- tential explanation for poor performance of giant kelp. Because up to 60% of the biomass of a healthy Macrocystis population is in the canopy (North 1968). we can safely infer that poor canopy con- dition is reflected in low drift availability (Tegner & Dayton 1991 ) for grazers. Several lines of evidence suggest that red abalone reproduction and recruitment are depressed by warm temperatures and lack of food. Cox ( 1962) reported the lack of gonad development and the low probability of spawning during 1957 to 19.59. a warm anomaly that only later would be recognized as a strong El Nifio event. There was a near absence of animals of the size classes that would have been spawned in 1983 and 1984 in a 1986 size-frequency distribution from the Palos Verdes Peninsula (Tegner & Dayton 1987). Visually assessed sex ratios from San Miguel Island were highly skewed toward females in 1997. near the onset of the 1997 to 1998 El Nino. A 1:1 sex ratio is expected in abalones. but when gonads are poorly developed, it is impossible to distinguish the grayish-brown digestive gland from an enveloping dark-green ovary (Ebert & Houk 1984). Later histological examination of 97 of these animals revealed a 1 : 1 sex ratio (C. S. Friedman unpub. data), indicating that many adult red abalones stressed by environ- mental conditions at San Miguel Island had undeveloped gonads. Finally, size-frequency data from San Miguel Island in 1999 (Fig. 8) suggests little to no recruitment during the 1997 to 1998 El Nifio. Warm temperature anomalies of the magnitude observed in 198.3 to 1984 and 1997 to 1998 were also observed in 1992. and that anomaly persisted into early 1994 (Fig. 5). Furthermore, McGowan et al. ( 1998) illustrate regional sea surface temperature anomalies of about 2°C in 1977 and 1978, and about 1°C in 1980. a period of the decline in Macrocystis canopy biomass at all three islands (Fig. 7). Thus, there have been a large number of warm events since 1977 that may have affected red abalone reproduction and recruitment. Haaker et al. ( 1998) assessed the annual growth of tagged red abalones on Santa Rosa Island from 1978 to 1982. and revisited the site in 1984. a two year hiatus that included the strong 1982 to 1984 El Niiio event. Growth parameters changed significantly throughout the study period, with L^. generally declining during each successive period. Warm anomalies in 1977. 1978. and 1980 (McGowan et al. 1998) followed by the 1983 El Nino winter storms (Tegner & Dayton 1987) were associated with a substantial decline in Macrocystis canopy biomass from 1978 to 1983 at all three islands (Fig. 7). Kelp recruited in 1983. but with the thermal stress of the El Nifio. canopy formation was not significant until 1984. Thus, the very slow abalone growth observed during 1983 to 1984 can be attributed to thermal stress (Fig. 5), low kelp biomass (Fig. 6, Fig. 7), and the low drift production of a young Macro- cystis stand (Tegner & Dayton 1991 ). Statewide commercial red abalone landings began a significant Red Abalone Fishery and Climate Variability 761 San Miguel Is. 0) 3 n = 798 75 100 125 150 175 200 225 250 Size (5 mm groups) o E 3 90 80 70 60 50 40 30 20 10 0 Santa Rosa Is. n = 43 ,„.i,ii. .iii.X 0 25 50 75 100 125 150 175 200 225 250 Size (5 mm groups) Figure 8. Red abalone size-frequency distributions from San Miguel and Santa Rosa Island. June 1999. decline iii the late 1960s (Fig. 9). When coinmereial fishing was closed in 1997. total red abalone landings were at 87 metric t, almost a 90% decline from the historic peak in the early 1960s. Karpov et al. (2000). analyzing area-specific landings, showed a spatial trend with higher catches coming from the mainland or nearshore islands, shifting over tiine to more remote areas. Catch from Santa Cruz Island, the closest to shore, peaked before Santa Rosa Island: at the time of the closure, catch on the two islands was I and 3% of their respective peaks. In contrast. San Miguel Island was the source of 82% of the red abalones landed in 1996. and these represented 23% of the peak catch on this island. Fishery- independent abundance estimates collected from 1983 to 1998 confirined that populations declined in conjunction with the catch (Fig. 9). Red abalone population estimates in Santa Cruz Island study sites (see Davis 1989 for site locations and survey tech- niques) went to zero in the late 1980s; Santa Rosa Island had higher initial abundance and density estimates that also declined to low levels by the end of the 1990s. Abundance estimates at San Miguel Island niiiTored the continuing catch until the fishery was closed in 1997 (Karpov et al. 2000). The temperature, kelp, and abalone reproduction and recruit- nient data suggest that environmental variability contributed to the decline of the fishery and the subsequent collapse of some stocks. II. as Cox ( 1962) reported, there was little to no recruitment of red abalones during 1937 to 1959. there would ha\e been a corre- sponding decline of animals entering the fishery 10 to 15 years later. The decline in statewide landings and in the Santa Cruz and Santa Rosa Island catches in fact began about 1968 (Fig. 9). Under intense fishing pressure, few abalones exceed cominercial size limits in natural populations (Tegner et al. 1989) and aggregations are fewer and smaller (Shepherd & Partington 1995). As the re- duced or missing age classes passed through the size-frequency distribution to fishable size, stocks would experience a correspond- ing decrease in density and egg production. While stock- recruitment relations are poorly known for abalones, it appears that major environmental disturbances such as the 1957 to 1959 El Nino coupled with continued intense fishing pressure establish the conditions for recruitment overfishing. Shepherd and Partington (1995) report that below densities of 0.15 to 0.2 m~', abalone populations are increasingly vulnerable to recruitment failure. Fishery independent data (Fig. 9) indicate that the Santa Cruz Island red abalone population was below this threshold in 1983, at the beginning of a strong El Niiio event, and these animals disap- peared from the study sites within four years. At Santa Rosa Is- land, the 1983 abundance estimate was 0.15 nr~. but this study population did not survive the environmental disturbances of fol- lowing years. Fig. 8 indicates that there are still red abalones on Santa Rosa Island, but the lack of recruitment during conditions in late 1998 through 1999 which produced strong recruitment at San Miguel Island suggests that Santa Rosa Island densities are below the threshold for successful fertilization. Once all three of these northern Channel Islands were highly productive red abalone habitat (Karpov et al. 2000). San Miguel Island populations have persisted not only because it is the furthest of the three islands from ports, but because its ocean climate has apparently buffered recent environmental variability relative to the warmer islands. Indeed, fishermen find it easier to get to San Miguel Island than Santa Rosa Island by following the mainland coast to Point Conception before crossing the Santa Barbara Chan- nel (J. Colgate pers. comm.). The role of climate in exacerbating overfishing is supported by catch statistics from other locations. Red abalone catch from the San Diego region, apparently heavily impacted by the 1982 to 1984 El Nino (Tegner & Dayton 1987). dwindled to very low levels in the late 1980s, whereas Point Con- ception stocks were still productive when the fishery was closed in 1997 (California Department of Fish and Game 1997). One of the most disturbing implications of El Niiio events and interdecadal scale increase in temperatures for abalones involves Withering Syndrome (WS). a disease first identified in black aba- lones {H. crachciodii) and now observed in other California spe- cies as well. Elevated temperatures accelerate mortality rates (Friedman et al. 1997). and Tissot { 1995) suggests that temperature is the single most important factor influencing population recov- ery. WS is now progressing through Central California (Altstatt et al. 1996). During the elexated temperatures of the 1997 to 1998 El Niiio, several abalone farms in California experienced a dramatic increase in the proportion of red abalones showing clinical signs of WS (Moore et al. 2000). Management of California abalone fisheries was heavily based on the high fecundity of these gastropods; a size limit large enough that the animals would reproduce several times before entering the fishery was considered adequate to protect the stocks. Later egg- per-recruit analysis indicated that the size limits allowed for healthy egg production, even at high fishing mortality rates (Teg- ner et al. 1989). The implicit assumptions of the size-limit ap- proach, however, didn't consider extended periods of environmen- 762 300t c 200 100 300 T 200 100 300 T c 200 100 1400 - » 1200; g 1000 * *- 600* ^ 600- I 400 1 200 t I 04»t»- Tegner et al. San Miguel Island 800 ■600 II i.liiiiiiiiiiiiii...iii Aiitb X i 400 200 I Santa Rosa Island l..l.«,.-llllllllllilll,illH... Santa Cruz Island ■Jhlllllllllilliilln... Statewide Red Abalone Landings l1l,l|ll|lll|l|lll|il,l,l|IHa 20O0 S + 1500 I X 1000 & 500 § z tKIO 600 I 400 I I 200 I 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 Ymt Figure 9. Red abalone harvest (melric tonsi from San Miguel, Santa Rosa, and Santa Cruz Island. 1952 to 1996. and non-invasive population counts on 60 nr transects from the Channel Islands National Park annual monitoring program. 1983 to 1998. Statewide landings. 1942 to 1996. Note the change in scale on the vertical axis. Adapted from Karpov et al. (2000). tal \uriability leading to protracted recruitment tailnre or Allee effects on fertilization (Shepherd & Partington 1995). The organ- isms of west coast kelp communities are adapted to environmental variability on scales including episodic El Nifio and La Nina events, and interdecadal regime shifts. The Scripps pier tempera- ture record reveals three regimes since the beginning of the data set in 1916: a warm period until about 1940, a cool period lasting until 1976. followed by the present warm period (McCall 1996). Red abalone reproduction and growth may be depressed during periods of warm water, but most animals survive the adverse environmen- tal conditions; the addition of intense fishing pressure onto climate disturbances, however, makes population decline almost inevi- table. To be successful, future management must take both envi- ronmental variability and Allee effects into account; no-take re- serves to protect aggregations of brood stock are probably the only way to reduce Allee effects. Our data suggest that remote moni- toring of temperature and kelp canopy biomass could be calibrated to offer an inexpensive and practical approach to incorporate en- vironmental variability into management of abalones and other grazers. ACKNOWLEDGMENTS We thank Mali Kahru for the satellite image. Jack Engle for the island temperature data. Dale Glantz and Keith Ullrich of Interna- tional Specialty Products for the kelp data. Kon Kaipo\' for the red abalone landings data. Alistair Hobday and two anonymous re- viewers for their comments on the manuscript, and the Pew Fel- lows Program in Marine Con.servation for financial suppoil. Red Abalone Fishery and Climate Variability 763 LITERATURE CITED Altstall. J. M. R. I-. Ambrose. J. M. Engle, P. L. Haaker. K. D. Lafferty & P. T. Raimondi. 1996. Recent declines of black abalone Halious cra- cherodii on the mainland coa.st of central California. Mar. Ecol. Prog. Ser. 142:185-192. California Department of Fish and Game. 1997. The ecology and manage- ment of California abalone resources. Abalone Informational Docu- ment. Sacramento: CDFG. Cook. R. D. & S. Weisberg. 19S2. Residuals and Inlluence in Regression. Monographs on Statistics and Applied Probability. New York: Chap- man and Hall. 230 pp. Cox, K. A. 1962. California abaloiies, family Haliotidac, Culit. Fi.\h Gtime 118:133. Davis, G. E. 1989. Design of a long-term ecological monitoring program for Channel Islands National Park. California. Nauiral Areas J. 9(2): 80-89. Ebert, E. E. & J. L. Houk. 1984. Elements and innovations in the cultiva- tion of red abalone Haliotis riifescens. Aquaculliire. 39:375-392. Friedman. C. S.. M. Thomson. C, Chun. P. L. Haaker & R. P. Hednck. 1997. Withenng Syndrome of the black abalone. Haliotis cracherodii (Leach): water temperature, food availability, and parasites as possible causes. / Shellfisli Res. 16:403-111. Gerard, V. A. 1982. Growth and utilization of internal nitrogen reserves by the giant kelp Macrocysris pyrifera in a low-nitrogen environment. Mar. Biol. 66:27-35. Haaker, P. L., D. O. Parker, K. C. Barsky & C. S. Y. Chun. 1998. Growth of red abalone, Haliotis nifescens (Swainson), at Johnsons Lee, Santa Rosa Island, California. / Shellfish Res. 17:747-753. Jackson, G. A. 1977. Nutrients and production of giant kelp, Macrocxstis pyrifera. off Southern California. Limnol. Oceaiwgr. ll:'^19-995. Karpov, K. A., P. L. Haaker, I. Taniguchi & L. Rogers-Bennett. 2000. Serial depletion and the collapse of the California abalone fishery. Can. Spec. Piibl. Fish. Aqitat. Sci. 130:11-24. Leighton, D. L. 1974. The influence of temperature on larval and juvenile growth in three species of Southern California abalones. Fish. Bull. 72:1137-1145. McCall. A. D. 1996. Patterns of low-frequency variability in fish popula- tions of the California Current. Calif. Coop. Oceanic Fish. Invest. Rep. 37:100-110. McCullagh. P. & J. A. Nelder. 1989. Generalized Linear Models, Mono- graphs of Statistics and Applied Variability 37. London: Chapni.in and Hall. 511 pp. McGowan. J. A., D. R. Cayan & L. M. Dorman. 1998. Climate-ocean variability and ecosystem response in the northeast Pacific. Science 281:210-217. McShane. P. E. 1992. Early life history of abalone: a review. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the Worid, Biology, Fisheries, and Culture. O.xford: Fishing News Books, pp. 120-138. Moore. J. D., T. T. Robbins & C. S. Friedman. 2000. Withering Syndrome in fanned red abalone, Haliotis nifescens: thermal induction and asso- ciation with a gastrointestinal nckettsiales-like prokaryote. / .\quai. Aniin. Health 12:26-34. Murray, S. N.. M. M. Littler & 1. A. Abbott. 19S0. Biogeography of the California marine algae with emphasis on the southern California is- lands. In: D. M. Power, editor. The California Islands: Proceedings of a Multidisciplinary Symposium. Santa Barbara: Santa Barbara Mu- seum of Natural History, pp. 325-339. Nemenyi. P. 1963. Distribution-free multiple comparisons. Ph.D. dissena- tion. Princeton: Princeton University. North, W. J. 1968. Effects of canopy cutting on kelp growth: comparison of e.xpenmentation with theory. Calif Dep. Fish Game. Fish Bull. 139:223-254. North, W. J. & R. C. Zimmerman. 1984. Influences of macronutrients and water temperatures on the summertime survival of Macrocystis cano- pies. Hydrohiologia 1 16(1 17):4 19^24. Roemmich. D. & J. A. McGowan. 1995. Climatic warming and the decline of zooplankton in the California Current. Science 267:1324-1326. Seymour. R. J. 1996. Wave climate variability in Southern California. J. Walenvay. Port. Coast. Ocean Eng. 122(4):I82-I86. Seymour, R. J., M. J. Tegner, P. K. Dayton & P. E. Pamell. 1989. Storm wave induced monality of giant kelp, Macrocystis pyrifera. in Southern California. Esliiar.. Coast. Shelf Sci. 29:151-160. Shepherd. S. A. & D. Partington. 1995. Studies on Southern Australian abalone (Genus Haliotis). XVI. Recruitment, habitat and stock rela- tions. Mar. Freshwater Res. 46:669-680. Shepherd, S. A., J. R. Turrubiatcs-Morales & K. Hall. 1998. Decline of the abalone fishery at La Natividad, Mexico: overfishing or climate change 'ly. Shellfish Res. 17:839-846. Tegner, M. J. & P. K. Dayton. 1987. El Niiio effects on Southern California kelp forest communities. Adv. Ecol. Res. 17:243-279. Tegner, M. J. & P. K. Dayton. 1991. Sea urchins. El Niiios, and the long term stability of Southern California kelp forest communities. Mar. Ecol. Prog. Ser. 77:49-63. Tegner, M. J., P. A. Breen & C. L. Lennert, 1989. Population biology of red abalones, Haliotis nifescens. in Southern California and manage- ment of the red and pink, H. corrugata. abalone fisheries. Fish. Bull. 87:313-339. Tegner, M. J., J. D. DeMartini & K. A. Karpov. 1992. The California red abalone fishery: a case study in complexity. In: S.A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the World, Biology, Fisheries, and Culture. Oxford: Fishing News Books, pp. 370-383. Tegner, M. J., P. K. Dayton, P. B. Edwards & K. L. Riser. 1996. Is there evidence of long-term chmatic change in Southern California kelp forest communities? Calif Coop. Oceanic Fish. Invest. Rep. 37:111- 126. Tegner. M. J., P. K. Dayton, P. B. Edwards & K. L. Riser. 1997. Large scale, low frequency oceanographic effects on kelp forest succession: a tale of two cohorts. Mar. Ecol. Prog. Ser. 146:1 17-134. Tissot, B. N. 1995. Recruitment, growth, and survivorship of black abalone on Santa Cruz Island following mass mortality. Bull. Soiilhern Cali- fin-nia ActuL Sci. 94(3):I79-I89. Tutschulte, T. C. 1976. The comparative ecology of three species of aba- lones. Ph.D. thesis, San Diego: University of California. Vega. A.. D. Lluch-Belda, M. Mucirio, G. Leon. S. Hernandez. D. Lluch- Cota. M. Ramade & G. Espinoza. 1997. Development, perspectives and management of lobster and abalone fisheries off Nonhwest Mexico under a limited access system. In: D. A. Hancock, D. C. Smith, A. Grant & J. Beumer, editors. Developing and Sustaining World Fishery Resources. Victoria. Australia: CSIRO, Collingwood. pp. 136-142. Zimmerman, R. C. & J. N. Kramer. 1984. Episodic nutrient supply to a kelp forest ecosystem in Southern California. J. Mar. Res. 42:591-604. Zimmerman. R. C. & J. N. Kremer. 1986. In situ growth and chemical composition of the giant kelp. Macrocystis pyrifera: response to tem- poral change in ambient nutrient availability. Mar. Ecol. Prog. Ser. 27:277-285. Zimmerman, R. C. & D. L. Robertson. 1985. Effects of the 1983 El Niiio on growth of the giant kelp. Macrocystis pyrifera. at Santa Catalina Island. Limnol. Oceanogr. 30:1298-1302. JtHdiial of Shellfish Research. Vol. 21), No. 2. 765-770. 2001. ''I POPULATION GENETICS OF THE YELLOW ABALONE. HALIOTIS CORRUGATA, IN CEDROS AND SAN BENITO ISLANDS: A PRELIMINARY SURVEY MIGUEL A. DEL RIO-PORTILLA' * AND JOSE G. GONZALEZ-AVILES' DepurtameiUo dc Aciilculnim. Centra de Investigaciofi Cientificu v de Ediicacidn Superior de Ensenuda. Km 107 Ciirr. Tijiiana-Eiisenuda. Ensenada. B.C. Mexico. A.P. 2732. Ensenada, B.C. Mexico 22800: -S.C.P.P. Pescadores Nacionales de Abidon. S.C. de R.L.. Rxersoii 117 Col. Centra. Ensenada. B.C. Mexico 22800 ABSTRACT The yellow ;ihalone, tialunis comii^aui. is one of the principal specie.s caught in Central Baja California. Mexico. Around the Cedros and San Benito Islands, the yellow abalone is mainly distributed in three zones: the North (Punta Norte) and South (San Agustin) of Cedros Island and around the small islands of San Benito. The main goal of this work was to characterize genetically the populations of the yellow abalone in these three zones as a baseline to help the fishery. Allozyme electrophoresis at eight loci was carried out with six samples from two years in the three localities. The average number of alleles per locus was 2.3 with a 67% polymorphism. Mean unbiased heterozygosity ranged from 0.15 to 0.25. which is similar to that of other abalone species. Only three out of 34 cases did not agree with the Hardy-Wemberg equilibrium and there was no tendency for either heterozygote excess or deficit. It was concluded that H. corni.qatu shows differentiation between localities and these may be considered as independent populations for fishery management. KEY WORDS: abalone. Halioiis cornigata. population genetics, allozymes. Cedros Island. San Benito Islands I.NTRODUCTION In Mexico. Hiiliutis corni^ahi is known as the vellow abalone because of the color of its foot. It is distributed from the Coronado Islands near the border with the United States to Margarita Island, and in depths from 10 to 20 m (Guzman del Proo 1992). The yellow abalone together with the blue abalone. Haliotis fidiiens. are the principal products from the Mexican fishery with annual production about 709 t (meat weight) in 1998 (SEMARNAP 1999). Along the Peninsula of Baja California. H. cornigata represents about 13% of the fisheiy's production (Leon & Ramirez 1992). The Cooperative "Pescadores Nacionales de Abulon" (PNA) capture abalone from the Cedros and San Benito Islands (Fig. 1) and the annual production in the past 1 1 y is shown in Table 1. In these islands, abalone is mainly distributed in three zones: (a) the North and (b) South of Cedros Island, and (c) around the small islands of San Benito. There are few studies on abalone population genetics on the Pacific Coast of North Amenca (Gaffney et al. 1996. Kirby et al. 1998. Gutien-ez-Gonzalez 2000. Hamm & Burton 2000. Zufiiga et al. 2000). but to our knowledge there are no studies on population genetics of the yellow abalone. Therefore, the aim of this study was to carry out a preliminary genetic characterization of the yel- low abalone. Haliotis cornigata. population at the Cedros and San Benito Islands using allozyme electrophoresis, to determine the population genetic structure and to obtain initial genetic data for fishery management and for population enhancement. MATERIALS AND METHODS Adults of H. cornigata were sampled from the commercial catch and from a population evaluation program. These abalone were caught from six localities: "El Faro." "La Bandera." "An- egados," and "El Chual" during 1997. and "Punta Norte" and "San Agustin" during 1998 (Fig. 1). Their shell length was measured ♦Corresponding author: P.O. Box 434844. San Diego. CA. 92143-4844. USA. E-mail address: midelrio@cice.se. mx and locality mean shell lengths were compared with a one-way ANOVA. followed by an a posteriori Tukey test. Gill and diges- tive gland samples were individually frozen in 1.5-ml tubes and transpoiled to the Laboratory of Genetics at Centro de Investi- gacion Cienti'tlca y de Educacion Superior de Ensenada. B.C. (CICESE) and preserved at -70 °C until electrophoresis analysis. Abalone tissues were homogenized individually with 0.5 ml of TME buffer (0.1 M Tris. 0.1 M maleic acid. 0.01 EDTA. and 0.01 MgCL. pH 7.8). and allozyme electrophoresis on 12% starch gels was carried out for 1 1 enzymatic systems: aspartate aminotrans- ferase ( AAT. 2.6. 1.1). glucose phosphate isomerase (GPL 5.3. 1 .9). glutathione reductase (GSR. 1.6.4.2). hexokinase (HK 2.7.1.1). isocitrate dehydrogenase (IDH. 1.1.1.42). leucine amino peptidase (LAP. 3.4.11. -). malic enzyme (ME 1.1.1.40), malate dehydro- genase (MDH 1.1.1.37). mannose-phosphate isomerase (MPI. 5.3.1.8). 6-phospho-gluconic dehydrogenase (PGD. 1.1.1.44). and phospho-glucomutase (PGM. 2.7.5.1). Gels were run in the TME buffer (pH 7.8) and staining was performed using standard proce- dures (Harris & Hopkinson 1976), except for LAP which was stained according to procedures described by Beaumont et al. (1983). Meldola blue (0.8%) was used instead of PMS (Turner & Hopkinson 1979). When more than one locus was observed, the most cathodical locus was named as "1.' The most common allele was designated as "100" and the rest were designated according to their relative mobility to the 100 allele. Allele frequencies, polymorphism, unbiased and direct count heterozygosity, heterozygote deficiency (D), and Nei's ( 1978) ge- netic distances were evaluated using the Biosys-1 (Swofford & Selander 1981), and an un weighed pair group method with arith- metic means (UPGMAl phenogram was built. Deviations from the agreement with the Hardy-Weinberg (H-W) equilibrium were tested with the Genepop"s exact test (Raymond & Rousset 1994). Wright's (1951. 1965) F (F,s, F^-j. and F,t.) statistics were evalu- ated using Weir and Cockerham's (1984) estimators by jacknifing and their significances were tested by bootstrapping with the Fstat program (Goudet 1994). Pairwise comparisons among all popula- tions were done usina exact tests (Miller 1997). Whenever there 765 766 DEL RiO-PORTILLA AND GONZALEZ-AVILES 116°'30'W 115''l5'W Punta * Norte, El Faro La Bandera i 1 San Benito Islands Anegados ^ 28- ism. Cedros Island Jl tT El Chual San ' V Agustin "^— ''" M 115*15'W 28'00'N. Figure 1. Haliotis corrugata sampling localities in Cedros and San Benito Islands. was a high number of tests, the level of significance (a) was Bonferroni adjusted using the Dunn-Sidak method [a" = (1 - (1 - 0.05)"^); where k = number of testl (Sokal & Rohlf 1995). RESULTS Yellow abalone mean sizes were between 128 and 157 mm; there were significant differences between localities, but four out of six localities had similar sizes (Table 2 1. From the 1 1 allozyme systems tested, only eight loci gave reliable results (Table 2). PGD, GSR, and MPI did not give posi- tive staining results. Two loci were found at LAP and MDH. but the first locus of each one was not reliable and thus, excluded from further analysis. The mean sample size was between 13.4 to 31.8 and the num- ber of alleles per locus was from 1.9 to 2.8. The percentage of polymorphism without criterion ranged from 62.5% to 87.5%. while direct count frequency of heterozygosity was from 0.094 to 0.201 and unbiased frequency of heterozygosity was from 0. 141 to 0.258. In all cases El Chual had the lowest values (Table 2). Frequencies of the 100 allele were similar between localities ex- cept at the Hk in Punta Norte and at Mdh-2 in La Bandera, which both showed substantially reduced frequencies of the most com- mon allele (Table 3). Agreement with H-W was present in all cases, except in 3 loci from Punta Norte (after Bonferroni adjustment), all of which had a heterozygote deficiency (Table 3). Furthermore, all positive and negative cases of D were counted, the total was D+ = 21 and D- = 13, which was not significant from a 1:1 proportion (sign test, Sokal & Rohlf 1995). The UPGMA phenogram using Nei"s genetic distance grouped localities; first La Bandera with El Chual. then El Faro with An- egados, and then these two groups clustered together, then San Agustin joined and finally Punta Norte (Fig. 2). Wright's F statistics showed positive and negative values for the loci (Table 4). F,^ had three positive values significantly dif- ferent from zero at the Lap2. Mdh2, and Hk loci (Table 4). Four cases of Fs^ were significantly different from zero at the Lap2. Me. Mdh2, and Hk loci. While the F,s had two negative and significant values at Me and Aat and only one negative and non-significant value at the Pgm locus. All mean F values were significantly TABLE 1. Haliotis curriigata production from the Cooperative "Pescadores Nacionales de Abulon" at Cedros and San Benito Islands. Year Meat weight (kg I Percentage from total' 1989 1990 1991 1992 199.1 1994 199.'i 1996 1997 1998 1999 2(100 8,241 12,930 10,047 6,542 a 2,384 23,807 23,912 23,174 24.698 16.612 8.31% 7.87% 5.17% 3.46% 1.26% 12.67% 12.05% 12.84% 18.57% 24.13% Source Alejo Ojeda Ibarra, bookkeeper S.C.P.P. "Pescadores Nacionales de Abulon" S.C. de R.L. ■' There was no capture of yellow abalone in 1993. Population Genetics. Hauotis corrucata 767 TABLE 2. Mean shell length, mean sample size per locus (n), mean number of alleles per locus (al), percentajje of loci polymorphic t'l , «i(hout criterion) and heterozygosity (dc = direct count: and Nei's unbiassed H) of Haliotis corrugata in Cedros and San Benito Islands (SE). Population 1997 1998 Locus El Faro La Bandera .\negados El Chual San .\gustin Punta Norte Shell length* IS^V?.^' I28.7.V 138.11" 156.70'' 154.07" 1 56.00- (1.80) (2.15) (1.84) (2.49) (1.80) (2,52) n 22.6 15.5 24.4 15.4 31,8 22.9 (2.2) (1.3) (2.2) (1.1) (1.4) (2.4) al 2.3 2.1 2.8 1.9 2.3 2.3 (0..^) (0.4) (0.5) (0.3) (0.3) (0.4) Polvmorphisni** 87.5 62.5 75.0 62.5 87.5 75.0 Hdx. 0.201 0.196 0.139 0.094 0.193 0.118 (0.076) (0.069) (0.039) (0.033) (0.040) (0.031) H 0.174 0.200 0.166 0.141 0.203 0.258 (0.0.59) (0.071) (0.049) (0.0631 (0.043) (0.075) * Superscripts show statistically equal mean shell lengths. ** A locus was considered polymorphic if more than one allele was detected. different from zero, indicating heterozygote deficiency and popu- lation dit'ferentiation. A pairwise comparison among populations was carried out and it did not show differentiation between La Bandera. El Chual and Anegados after Bonferroni adjustment (Table 5). These differences were mainly produced by Mdh2 and Hk loci (Table 5). DISCUSSION This is a preliminary survey where the analysis of 8 allozyme loci shows genetic differences between six localities of the yellow abalone Halioiis cornigara in Cedros and San Benito Islands. B.C. Mexico. The number of loci sampled is approximately the mean of the numbers of other tested loci in different abalone species (Table 6). Hamm and Burton (2000) found genetic differentiation be- tween localities of the black abalone. Haliotis cracherodii. using three allozyme loci (Aar-I. Pt;i [ = Gpi], and Pi;m). although com- parisons with partial DNA sequences of the mtDNA encoded cy- tochrome oxidase subunit I gene failed to differentiate between populations. These loci did not differentiate between three popu- lations of the red abalone. Haliotis nifescens (Burton & Tegner 2000). Similarly, 7 loci tested with the blue abalone. Haliotis fiilgens, failed to separate six localities in the middle of the Pen- insula of Baja California (Ziifiiga et al. 2000). and 16 loci did not separate four populations between Isia de Cedros and Isia Magdalena (Gutierrez-Gonzalez 2000), Size difference among samples is not considered to have an influence in population differentiation because localities with simi- lar abalone size did not join together (Fig. 2. Table 2). Sample sizes are similar to those from the blue abalone (Zdniga et al. 2000) and may be considered as small satiiple size, but the statistical methods used here are considered to be robust enough to avoid the bias produced by small sample size. The polymorphism, mean allele number per locus, and the ob- served heterozygosity found in H. corrugata are similar to other abalone species, marine invertebrates and invertebrates in general (Table 6). Only Punta Norte population was not in H-W equilibrium and was genetically more separated from the other populations. which may be due to its location on the East coast of Cedros Island. In this area, an anticyclonic eddy is formed during August and September due to an upwelling event on the coast of the Peninsula of Baja California (Mancilla-Peraza et al. 1993, Palacios-Hemandez et al. 1993, Amador-Buenrostro et al. 1995). The main stream east of the Cedros Island goes north around the island and then west towards San Benito Islands, and from these to the San Agustin area (Palacios-Hemandez et al, 1993). The spawn- ing peaks in these areas are in August and in Novem- ber, then, if spawning occurs when currents go north-south to the west of Cedros Island, cuiTents could transport some larvae from one place to another: these larvae could be carried from Anegados and San Benito towards the south of Cedros Island. There are no studies regarding abalone larval dispersal and oceanographic conditions around Cedros and San Benito islands during spawning time. However, it is considered that migration is small among abalone populations (Guzman del Proo et al. 2000), Although, Guzman del Proo et al. (2000) consider that abalone larvae usually settle around parental beds, in some cases, larvae may be transported about 3-5 km; because this distance is smaller than the actual distance among islands, it is considered that geo- graphical distance may be one of the most important factors in population differentiation and in H-W equilibrium of H. corrugata. Other factos may infiuence the agreement with H-W equilib- rium (i.e. selection, migration, and inbreeding). Fishing may be considered as a selection pressure, because abalone larger than 140 mm are commercially caught. Environmental factors may have an effect on the population structure. It has been observed that sea surface temperature anomalies have a positive effect on H. corru- gata population (Shepherd et al. 1998). This may reduce stock density and the possibility of genetic differentiation due to genetic drift increases. Non-random mating may be considered to be important in population differentiation because of the spawning behavior of abalone and because larvae tend to settle in adult beds (Guzman del Proo et al, 2000). Another source of non-random mating is the introduction of cultured organisms and may have an influence on 768 DEL RiO-PORTILLA AND GONZALEZ-AVILES TABLE 3. Haliotis corrugata. Allele frequencies and heterozygote deficiencies (D) and their significance after a adjustment. Population 1997 1998 Locus El Faro La Bandera Anegados El Chual San .\gustin Punta Norte Aat 92 0.019 0.000 0.000 0.000 0.000 0.000 100 0.926 0.900 1 .000 1.000 0.939 1.000 140 0.056 0.100 0.000 0.000 0.061 0.000 D 0.0+4 0.074 0.048 Cpi 85 0.089 0.075 11.145 0.026 0.135 0.100 90 0.018 0.000 0.000 0.000 0.000 0.000 100 0.893 0.900 0.855 0.974 0.851 0.900 110 0.000 0.025 0.000 0.000 0.014 0.000 D 0.082 0.061 -0.361 0.000 -0.065 0.093 Hk 91 0.000 0.000 0.000 0.000 0.000 0.429 100 0.938 1.000 1.000 1.000 0.870 0.571 106 0.063 0.000 0.000 0.000 O.I.W 0.000 D 0.044 0.128 -0.620* Itih 87 0.000 0.000 0.067 0.000 0.000 0.000 91 0.000 0.056 0.050 0.000 0.000 0.000 100 0.957 0.861 0.833 0.941 0.939 0.904 104 0.000 0.028 0.017 0.000 0.000 0.000 107 0.000 0.000 0.000 0.000 0.000 0.038 112 0.043 0.056 0.033 0.059 0.061 0.058 D 0.023 0.074 -0.228 0.031 0.048 0.058 Lap-2 90 0.000 0.000 0.000 0.036 0.000 0.000 100 1.000 1.000 0.925 0.929 0.780 0.682 104 0.000 0.000 0.025 0.000 0.220 0.318 110 0.000 0.000 0.025 0.000 0.000 0.000 120 0.000 0.000 0.025 0.036 0.000 0.000 D 0.026 0.019 -0.200 -0.800* Mdh2 100 0.630 0.500 0.780 0.647 0.891 0.600 105 0.217 0.000 O.OOO 0.000 0.078 0.033 110 0.000 0.000 0.020 0.000 0.000 0.000 120 0.152 0.500 0.200 0.353 0.031 0.367 D 0.279 -0.352 -0.219 -0.500 -0.2.W -0.872* Me 80 0.132 0.000 0.000 0.000 0.000 0.000 92 0.000 0.000 0.024 0.000 0.000 0.000 100 0.868 1 .000 0.976 1 .000 1.000 1 .000 D 0.121 0.000 Pf;m 89 0.000 0.056 0.017 0.150 0.000 0.033 100 0.964 0.722 0.879 0.800 0.792 0.883 109 0.036 0.222 0.052 0.050 0.153 0.050 115 0.000 0.000 0.052 0.000 0.056 0.033 D 0.019 0.232 0.073 -0.148 0.027 -0.085 ' Significant deviation from H-W after a adjustment. the genetic structufe of a population, by reducing the genetic pool (Gaffney et al. 1996). With the mean value of F^j 0.093, it can be considered that the localities of H. cormgata conform to discrete populations with moderate diversification, since this value is between 0.05 and 0. 15 (Wright 1951, Wright 1965). This contrasts with the low diversi- fication of H. fulgcns in Mexico which shows low values of Fsy, 0.022 to 0.036 (Table 6). and it has been considered that H. fiilgens forms a single population from Cedros Island to Isia Magda- lena (Gutierrez-Gonzalez 2000). H. corrugata also shows a Population Genetics. Haliotis corrugata 769 Distance 0 025 I — La Bandera ■—El Chual - Anegados -San Agustin - Punta Norte Figure 2. Nei's (1978) unbiased genetic distance for tlie yellow aba- lone, Haliotis corrugata, at six localities on Cedros and San Benito islands. Cophenetic correlation = (1.904. Locus Aat Gpi Hk El Faro /^„, Lc,p2 Mdh2 Me P^}ii Mean TABLE 4. Haliotis corrugata F statistics. Fsr 0.044 0.008 -0.052 0.074 -0.004 0.078 0715* 0.360* 0.411 0.072 0.007 0.065 0.436* 0.129* 0.333 0.336* 0.110* 0.225 0.004 0.154* -0. 1 7 1 0.037 0.038 -0.001 0.240* 0.093* 0.160 * Significanlly different from 0 after a adjustment. T.ABLE 5. Probability values (below the diagonal) obtained after pairwise comparisons among all populations using exact tests.' Loci where there was a difference between populations (above the diagonal) after a adjustment." Population 1 2 3 4 5 6 1 El Faro *** Mdh2 Mdh2 Pk"i M 2 La Bandera 0.0007 *** Mdh2 Hk 3 Anegados 0.0034 0.1133 *** Lap2. Hk 4 El Chual 0.0014 0.6875 0,2752 *** Mdh2. Pgm Lap2. Hk 5 San Agustin 0.0006 <0.0001 0.0001 <0.0001 *** Mdb2. Hk 6 Punta Norte <().000l <0.0001 <0.0001 <0.()001 <0,0()0l *** 10,000 dememorization steps. 30 batches and 10.000 permutations per batch, -a" = 1 -(1 -0.05)'"-' = 0.00341. TABLE 6. Genetic parameter comparison between different abalone species. Mean allele No, No, Polymorphism number loci sites It Ho {'yc) per locus Frr FsT F,s Reference H. cracherodii 3 7 -427 0.039 0. 1 25 Hamm & Burton 2000 H. fulgens 0.067 Brown 1993 H. fidgens 7 5 102 0.054-0.195 14.3-100.0 1.7-2.0 0.335 0.036 0.318 Ziiiiiga et al. 2000 H. fulgens 12 4 377 0.061-0.075 38-44 1.96-2.06 0.637* 0.022 0.629* Gutierrez Gonzalez 2000 H. Icievigalti 15 8 0.195 73.3 2.67 0.062 0.014 0.076 Brown & Murray 1992 H. rubra 15 15 0.140 63.1 2.58 0.056 0.016* 0.071 Brown & Mun-ay 1992 H. rubra 0.022 Brown 1991 H. rufescens 3 3 131 0.012 Burton & Tegner 2000 H nifescens' 13 7- 714 0.056-0.291 38.46-75 1.5-2.1 Licona-Chavez 1999 H. luberculata 0.284 2.3-3.0 Mgayaet al. 1995 H. virginea 0.125 Brown 1993 H. corrugata 8 6 183 0.094-0.201 62.5-87.5 1 .9-2.8 0.240* 0.093* 0.160* This work M arine invertebrates 0.061-0.216 Fujinoet al. 1983 In i.ertebrates O.IOI Nevo 1978 * Significantly different from zero. ' Cultivated. " Number of batches. 770 DEL RlO-PORTILLA AND GONZALEZ-AVILES slightly higher Fsx value in comparison with other abalone spe- cies, showing more diversification (Table 6). Yellow abalone localities at Cedros and San Benito Islands form different populations, which should be considered in fishery management. More studies on other populations of the yellow abalone need to be performed to test whether population differen- tiation is also present along the Peninsula of Baja California. Also, studies on larval ecology should be carried out to determine mi- gration amona islands. ACKNOWLEDGEMENTS We thank all people from the Cooperative "'Pescadores Na- cionales de Abulon" and Isla de Cedros. including Jose Mario Espinoza and Rogelio Amaro. We thank Diana Rodriguez for tech- nical assistance. This study was supported with funds of CICESE project No. 6153. LITERATURE CITED Amador-Buenrostro, A.. M. L. Argote-Espinoza, M. Mancilla-Peraza, & M. Figueroa-Rodrigueez. 1995. Variaciones de perdiodo corto de la circulacion anticiclonica en Bahi'a Sebastian Vizcaino, B.C. Cieiiciai Marinas 21:201-223. Beaumont, A. R., CM. Beveridge. & M. D. Budd. 19S_V Selection and heterozygosity within single families of the mussel Mytiltix ediilis (L.). Mar. Biol. U'tt. 4:151-161. Brown, L. D. 1991. Genetic variation and population structure in the black- lip ahalone, Haliittis rubra. Ausi. J. Mar. Freshwater Res. 42:77-90. Brown, L. D. 199.^. Biochemical genetics and species relationships within the genus Haliotis (Gastropoda: Haliotidae). / Moll. Stud. 59:429—143. Brown, L. D. & N. D. Murray, 1992, Population genetics, gene flow and stock structure in Haliotis rubra and Haliotis laevigata. In: Shepherd, S. A., S. A. Guzman del Proo, & M. J. Tegner, editors, Abalone of the world. (Biology. Fisheries and Culture). Proceedings of the 1st Inter- national Symposium on Abalone. Fishing News Books, pp. 24-33. Burton. R. & M. J. Tegner. 2000. Enhancement of red abalone Haliotis rufescens stock at San Miguel Islands: reassessing a success story. Mar. Ecol. Prog. Sen 202:303-308. Fujino, Y„ R, Yamanaka & P, J, Smith. 1983. Genetic variation in marine molluscs. Bull. Jap. Soc. Sci. Fish. 49:1809-1817. Gaffney, P. M.. R. V. Powell, D. Hedgecok. D. A. Powers, G. Moms & L. Hereford. 1996, Genetic effects of artificial propagation: signals from wild hatchery populations of red abalone in California. Acjiiaculture 143:257-266. Goudet, J. 1994. FSTAT, a program for IBM PC compatibles to calculate Weir and Cockerham's ( 1984) estimators of F-statistics. (version 1.2). Institut de Zoologie et d'Ecologie Animale Universite de Lausanne Switzerland, Gutierrez-Gonzdlez, J. L. 2000. Vanabilidad genetica de cuatro pohla- ciones de abulon azul Haliotis fulgens. en la costa Occidental de Baia California. M. Sc. Thesis. Universidad Autiinoma de Baja California Sur. B.C.S. Mexico. 49 pp. Guzman del Proo, S. A. 1992. A review of the biology of abalone and its fishery in Mexico. In: Shepherd, S. A., S. A. Guzman del Proo & M. J. Tegner, editors, Abalone of the World. (Biology, Fisheries and Cul- ture), Proceedings of the 1st International Symposium on Abalone. Fishing News Books, pp. 341-360. Guzman del Proo, S. A., F. Salinas, O. Zaytsev, J. Belmar-Perez, & J. Carrillo-Laguna. 2000. Potential dispersion of reproductive products and larval stages of abalone (Haliotis spp.) as a function of the hydro- dynamics of Bahia Tortugas Mexico. J. Shellfish Res. 19:869-881. Hamm, D. E. & R. S. Burton. 2000. Population genetics of black abalone, Haliotis cracherodii. along the central California coast. J. Exp. Mar. Biol. Ecol. 254:235-247. Harris, H, & D. A. Hopkinson. 1976. Hand book of enzyme electrophore.sis in human genetics, Amsterdam: North Holland Publishing Co. 259 pp. Kirby, V. L, R. Villa & D. A, Powers. 1998. Identification of microsatel- lites in the California red abalon, Haliotis rufescens. J. Shellfish Res. 17:801-8(J4. Leon-C, G. & M. Ramire/-R. 1992. The abundance of abalone in 23 fishing areas of Baja California Sur, Mexico. In: S. A, Shepherd, S. A. Guzman del Proo, M. J. Tegner, editors. Abalone of the world. Biol- ogy, fisheries and culture. Supplementary papers. Proceedings of the 1st International Symposium, Adelaide, South Australia. Australia Dept of Fisheries 24:5-9, Licona-Chavez, A, L. 1999, Caracterizacion genetica del abulon rojo cul- tivado por la empresa Abulones Cultivados S.A. de C.V." M. Sc. The- sis. Centro de Investigacion Cientffica y de Eduacidn Superior de Ensenada. Mexico. 108 pp. Mancilla-Peraza, M„ E. Palacios-Hernande/. G. Lopez-Castillo. 1993. Vanabilidad hidrografica de Bahia Vizcaino, B.C. Ciencias Marinas 19:26.5-284. Mgaya, Y. D., E. M. Gosling, J. P. Mercer & J. Donlon. 1995. Genetic variation at three polymorphic loci in wild and hatchery stocks of the abalone Haliotis tubercuhitu Linnaeus, Aquaculture 136:71-80. Miller. M. P. 1997. Tools for population genetic analyses (TGPGA). 1.3: A windows program for the analysis of allo/yme and molecular popula- tion genetic data. Computer software distributed by author. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590. Nevo, E. 1978. Genetic variation in natural populations: patterns and theory. Theor Pop. Biol. 13:121-177. Palacios-Henandez, E., M. L. Argote-Espmoza, A. Amador-Buenrostro & M. Mancilla-Peraza, 1993. Circulacion barotropica de Bahia Sebastian Vizcaino B.C. simulada por un modelo numenco. Secretaria de Marina. Direccion General de Oceanografia Naval. Mexico. 61 pp. Raymond, M. & F. Rousset. 1994. GENEPOP Ver 1.0. Population genetics software for exact tests and ecumenicism. / Heredity 86:248-249. Salas-Garza, A. E„ & R. Searcy-Bemal, 1992, Development and present status of abalone culture in Mexico, In: S, A. Shepherd, S. A. Guzman del Proo & M.J. Tegner, editors, Abalone of the World. (Biology. Fisheries and Culture). Proceedings of the 1st International Symposium on Abalone. Fishing News Books, pp. 538-546. SEMARNAP 1999. Anverio Estadi'stico de Pesca. 1998. Mexico. Shepherd, S. A., J. R. Turrubiates-Morales & K. Hall. 1998. Decline of the abalone fishery at La Natividad, Mexico: Overfishing or climate change'/ Shellfish Ret. 17:839-846. Sokal, R. R. & F. J. Rohlf 1995. Biometry. 3rd edition. New York: W. H. Freeman and Co, 887 pp, Swofford, D, L, & R, B, Selander. 1981. BlOSYS-1. A program for the comprehensive analysis of electrophoretic data in population genetics and systematics. J. Heredity 72:281-283. Turner, V. S, & D, A. Hopkinson. 1979. The use of meldola blue in isozyme stains after electrophoresis. F.E.B.S. Letters 105:376-378. Weir, B. S. & C. C. Cockerham. 1984. Estimating F-statislics for the analy- sis of population structure. Evolution 38:1358-1370. Wright, S. 1951. The genetical structure of populations. /lnH(;/.v o/'f/f^^fH/cs 15:323-3.54. Wright. S. 1965. The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution 19:395—420. Zuniga, G., S. A. Guzinan-del Proo, R. Cisneros & G. Rodriguez. 2000. Population genetic analysis of the abalone Haliotis fulgens (Mollusca: Gastropoda) in Baja California, Mexico. J. Shellfish Res. 19:853-859. Jniirihil ,>i Slwlllish Research. Vol. 20. No. 2, 11\-11^). 201)1. MOVEMENT AND RE-AGGREGATION OE THE BLACKLIP ABALONE, HAUOTIS RUBRA LEACH, AETER FISHING Rlt KARD A. OFFICER,' * CAMERON D. DIXON," AND HARRY K. GORFINE" Tiisiiuiiiicin Ac/ticiculturc and Fisheries Institute. University of Tasmania. Niibeena Crescent. Taroona. Tasniaina 7U5J. Australia: -Mariiu- ami Freshwater Resources Institute. PO Box 114. Queenscliff. Victoria 3225. Australia ABSTRACT Movement and re-aggregation after fishing of a stock may have a major impact on estimates of abundance and stock assessment parameters such as natural mortality. The propensity of blacklip ahalone {Huliotis rulini Leach) to re-aggregate after fishmo was studied using a combination of /« siiu tagging with fine-scale mapping of two distinctly different abalone populations. Controlled fishing of each population was conducted to test the hypothesis that removal of abalone by fishing stimulates movements that result in re-aggregation. Declines in abundance due to fishing were evident but less than those expected had no recovery occurred. Changes in the fine-scale spatial distribution of abalone suggested that re-aggregation occurred through a series of contiguous displacements. Smaller emerging abalone did not appear to be the prime source for this re-aggregation. These results undermine the utility of change-in-ratio and catch-per-unit-effort methods of abundance estimation. They also highlight the need to develop methods of abundance estimation that accommodate the impact of aggregating behavior on blacklip abalone populations. KEY WORDS: Hutioiis rubra, abalone. nio\cmem. aagregation INTRODUCTION Blacklip abalone. Haliotis rubra, form the basis of a large and extremely important fishery in southern Australian waters. The year 2000 total allowable catch of 2720 tons was worth about $AUD 80 million at the point of first sale. Despite the economic importance of abalone fisheries, stock assessment of ahalone has remained problematic (Breen 1992). These difficulties are partly a product of the biology of abalone and the way they are fished. The tendency of abalone to live in aggregations renders them more vulnerable to overfishing by increasing their catch rates (McShane 1995), Divers are able to maintain high catch rates by moving from aggregation to aggregation and by relocating to a different area once catch rates drop below an acceptable level (Prince 1989. Breen 1992. McShane 1995), If the abalone remaining after fishing then re-aggregate, a diver returning to the area may well fish the area at a similar catch rate to the first excursion. These processes may all contribute to hyper-stability in catch rates and conceal real fluctuations in population size (Breen 1992. Prince & Shepherd 1992. McShane 1994). Understanding how abalone reform aggregations is therefore central to understanding the resilience of abalone populations to fishing. Furthermore, because stock assessment techniques are based on assumptions about the spatial and temporal distribution of animals, understanding the nature of re-aggregation is crucial to the choice of tools used for stock assessment. This paper first describes the response of two blacklip abalone populations to experimental fishing. It then focuses on determining where '"replacement" abalone come from to reform aggregations and how these processes change with time. Several hypotheses that could explain re-aggregation are examined. These include emer- gence (from cryptic habitat within the fished area), immigration (from outside the fished area), relocation (from sparsely populated regions within the fished area), and a combination of these pro- cesses. A fifth hypothesis of growth (of pre-recruits to recruits) is *Corresponding author. Current address; Marine Institute, Marine Fisher- ies Services Division. SO Harcourt St. Dublin 2. Ireland. E-mail address: rick.ofncer(n' marine. ie. not examined because growth was negligible during the period of the study. MATERIALS AND METHODS Experimental sites were selected off the coast of Victoria, Aus- tralia, at Point Cook and at Flinders (Fig, 1). Both sites were subject to little illegal or recreational abalone fishing, and were closed to commercial fishing for six months prior to and through- out the experiment. The sites were chosen to lepresent extremes of habitat type and abalone distribution. The Point Cook site (37°55.893' S. 144 47.104' E) was 3-4 ni in depth and was not subject to strong swell or current. The reef consisted of basalt boulders, rarely over I m in height, on a sandy substrate. Conditions appeared anaerobic beneath the boulders and there was very little cryptic habitat that could not be thoroughly searched by a diver. The dominant macro-algae included Ecklonia radiata. Caulerpa spp,. Enteromorpha linza. Cystoseria spp,. and. during summer months, a thick mat of drift red algae (mostly Jeannerettia pedicellata). Abalone were abundant at Point Cook and were relatively evenly distributed. At Flinders (38°29.397' S. 145-01.274' E) the depth was 7 to 9 m and the site was prone to strong swells and currents. The reef consisted of fractured and stepped basalt platforms, often over 1 m in height, and interspersed by sand gutters. The fractured nature of the reef provided an abun- dance of unsearchable cryptic habitat. The dominant algae in- cluded Phyllospora spp,. Ecklonia radiata. and coralline algae. Abalone were less abundant at Flinders than at Point Cook and were more patchily distributed. Four square plots were set up in fixed positions at each site. The boundaries of each plot measured 24 in and each plot was sepa- rated from the other plots by 10-m conidors (Fig, 2). At each site two plots were fished to assess the impact of fishing and two plots were left unfished to compare the fishing impact with unfished areas. Fished and control plots were chosen randomly, but were constrained to being diagonally opposed. Surveys were conducted prior to fishing. 3—4 wk after fishing (to assess the impact of fishing), and. 10 wk later to assess the recovery from fishing. This design allowed for temporal and spatial comparisons between plots, both fished and control, and between sites. 771 772 Officer et al. Location of Study Area Melbourne Flirxlers''^--"^c Figure I. Location of the experimental sites at Point Cook and Flinders. Changes in abundance were assessed by counting all emergent abalone within the plots using an exhaustive 1 x 1 m grid system. These grid counts were also used to calculate Morisita's index of dispersion (Krebs 1989) (Eq. 1): I, = n where /,, is Morisita's index of dispersion, n is the sample size. X-v is the sum of grid counts, and S-v" is the sum of grid counts squared. This index is a useful descriptor of the degree of aggregation in observed spatial patterns. The significance of observed spatial pat- terns was tested using the standardized Morisita index of disper- sion (described in Krebs 1989). This index allows comparisons between plots and sites because it is independent of both popula- tion density and sample size. Differences in the spatial distribution of abalone were also studied by measuring the distance between abalone and their first, second, and third nearest neighbors. The source individuals for these measurements were selected randomly from cells with the greatest and least abundance of abalone within each plot. A chi- square goodness-of-fit test was also used to compare the observed and expected distributions of nearest-neighbor distances (Camp- bell & Clarke 1971). These tests indicated the significance of any departure from random spacing. The shape of the observed distri- butions indicated any departure from a random pattern. The dis- tribution of nearest-neighbor distances at each survey was com- pared to assess temporal changes in the degree of aggregation. During each survey the abundance of abalone outside each plot was estimated by making counts along fixed transects (Fig. 2). These transects measured 24 ni in length by 1 m in width. These J 1 m by 1 m gnd vvittiin each plot. Fishe(d h 10 m 24 m Figure 2. Diagram showing the set-up of experimental areas at each site. Shaded areas outside the plots indicate areas in which abalone were mass marked. Dotted lines indicate the position of fixed transects swum outside the plots. counts were made to assess the extent of immigration into the plots. Lengths were recorded to assess fishing-induced changes in the size composition of the population. The extent of weed growth on shells was also noted. These data were collected under the assump- tion that abalone emerging from cryptic habitat were smaller and carried less algal cover than the conspicuous population. To examine immigration and relocation abalone were individu- ally tagged. This was done with a rivet tag through a respiratory pore using a method modified from that of Prince (1991). All abalone were tagged and measured in siiu to minimize disturbance and displacement. To spread the tagging within the plots across the range of conspicuous abalone sizes and across all habitats, divers attempted to tag every fifth abalone sighted. The positions of tagged abalone were then recorded within the grid to an accuracy of ±0.1 m. The 1322 and 819 abalone tagged within the plots at Point Cook and Flinders represented 19 and 18 percent of the unconcealed populations, respectively. To assess the extent of immigration into the plots after fishing, abalone on the outer perimeters of the plots were also marked (Fig. 2 1. Starting closest to each plot boundary and working outwards, all abalone encountered were marked until at least 170 had been either tagged or painted with a uniquely colored paint-stick (509r for each marking method). Abalone between the inner borders of the plots were not marked so that interaction effects along the internal borders between plots were avoided. Approximately 1400 abalone were marked outside the plots at each site. The relocation of tagged abalone within the plots was examined by using grid-cell abundance as an index of habitat quality. The assumption was that the greater the number of abalone initially resident within a grid cell, the better the habitat quality of thai cell. The initial abundance in cells that tagged abalone occupied prior to moving was compared with that of the cells in which they settled. Abalone were determined to have settled when subsequent surveys revealed that they had moved less than one meter from their pre- vious position. Movement and Re-aggregation oi- Blacklip Abalone 773 Experimental fishing was done above minimum size limits. The size limits used were less than those normally in use in the Vic- torian abalone fishery (Point Cook: normally 100 mm. used 93 mm; Flinders: normally 1 10 mm. used 100 mm). These size limits made a greater proportion of the population axailable to fishing and hence, allowed a greater fishing impact. About 50% of the population at each site were over the respective size limits. Of those, 209^ were tagged and therefore unavailable to fishing. The remaining 40% of the population oversize were fished until a reduction of 35% of the abundance prior to fishing was achieved. To allow comparison between sites the same relative reduction in abundance was applied at both sites. RESULTS Changes in Ahnnclance Changes in the total abundance of abalone counted within the plots were compared as a percentage of the initial abundance counted prior to fishing (Fig. 3). At each site changes in abundance were consistent within each type of plot and were therefore com- bined for analysis. In all cases the observed change in abundance was different to that expected. If no movement had occurred after fishing, a 35% reduction was expected in fished plots, and no change was expected in control plots. At both sites a decrease in abundance of less than 20%' was observed in fished plots. An increase in abundance was observed in all control plots. At Point Cook the apparent recovery of fished areas was sustained but at Flinders the recovery slowed or possibly waned. Changes in Distribution At both sites and during all surveys the distribution of abalone was found to be significantly aggregated (standardized Morisita (a) Point Cook 100 80 60 40 20 &t' . rt ■ ( ■ — ■ II II c c I - 1 < 0) (D) Flinders Q Fished □ Control 100 80 60 40 20 ^ ' p 53 og II II n Prefishing Expected 1st 2nd Abundance Reduction Survey Survey Figure 3. Changes in abundance iif abalone within fished and control plots at (a) Point Cook and lb) Flinders over the course of the experi- index of dispersion >0.5. therefore >95% confident of aggregated pattern). Changes in the non-standardized Morisita index of dis- persion indicated site-specific differences in distribution (Fig. 4). The Flinders plots showed a consistently higher degree of aggre- gation. This was indicative of there being many areas with no abalone and a few areas with many. The lower degree of aggre- gation at Point Cook was indicative of a more homogenous spatial distribution. These initial distributions influenced the change in distribution observed after fishing. At Flinders the removal of abalone from aggregations made the size of aggregations less ex- treme and lowered the degree of aggregation. At Point Cook the even removal of abalone across the entire area created grid cells with no abalone. thereby causing an increase in the degree of aggregation. The longer-term recovery after fishing generally showed a return toward the initial distribution. The complete enumeration of all abalone within the plots meant that a Chi-square test of goodness-of-fit (Campbell & Clarke 1971) could be used to compare the observed and expected distributions of first, second, and third nearest-neighbor distances. These tests indicated that significantly aggregated patterns existed in all plots at both sites during each survey (P < 0.01, df = 5). Prior to fishing the distribution of first, second, and third nearest-neighbor dis- tances was similar in fished and control plots at each site (Fig. 5 and Fig. 6). At both sites approximately 80% of abalone were located within 30 cm of their nearest neighbor. This aggregated pattern was more extreme at Flinders where about 50% of abalone were located within 5 cm of their nearest neighbor (compared with about 40% at Point Cook). Distances to second and third nearest neighbors were also less at Flinders than at Point Cook. About 25% of second and 10% of third nearest neighbors were located within 5 cm of the source abalone at Flinders compared with 10% and <5% respectively, at Point Cook. This pattern prior to fishing is indicative of the more aggregated pattern at Flinders and the more homogenous distribution at Point Cook. The apparent impact of fishing was most noticeable at Flinders (Fig. 6). In both control and fished plots the proportion of first, second, and third nearest neighbors found close to the source aba- lone decreased during the first post fishing survey and the propor- tion located some distance from the source abalone had increased. This change was most noticeable in the fished plots. At Point Cook these changes were only noticeable in the fished plots (Fig. 5) and to a lesser extent than at Flinders. Chances in the distribution of Flinders Fished Flinders Control Point Cook Fished Point Cook Control Prefishing 1st Survey 2nd Survey Figure 4. Changes in the non-standardized Morisita Index of Disper- sion for abalone within fished and control plots at Point Cook and Flinders over the course of the experiment. 774 Ofhcer et al. 30 20 I I Prefishing Survey r~l 1st Survey rTl Control Plots 1 n-B Ht. . -■ 2nd Survey \}M Fished Plots n= rrrei nsT\ |m HM. L I^^^^Tb ^i^l^r.,,^ >25-J5 >35-t5 >4555 >45-55 >55 < 5 >5-15 >15-25 >2 Distance to Neigtibour (cm) Figure 5. Relative frequency distriljutions of distances from abalone to tlieir nearest, second nearest, and third nearest neighbors at Point Cook prior to fishing, and at each post-fishing survey. nearest-neighbor distances in the control plots at Point Cooi< be- tween the pre-tishing and first post-fishing survey were erratic. Between the first and second post-fishing surveys changes in these distributions in control plots at Point Cook remained erratic. of first, second, and third nearest neighbors at Flinders and in the fished plots at Point Cook had shown a general recovery toward the distribution found prior to fishing. .At Flinders this recovery was most complete in the control plots. In fished plots the propor- Conversely. by the second post-t"ishing survey the distribution tion of first, second, and third nearest neighboring abalone found 1 1 Prefishing Survey P] IslSurvey 60 50 . Control Plots fn. 40 . 30 . ■ 1 iz! 1 risri mi 1 20 . 10 . n U .rr^.rTm.n-m.-^r^ . 1 . 1^ 2nd Survey Fished Plots n= rrari n34] ufj JLi -T-a .^j-m .-j^ In rri.^i.i-ri 1 1 ^ rfl.r^ ■ . ' m . ,rrl >5-15 >15-25 >25-35 >35^5 >45.55 >55 < 5 >5-15 >15-25 >25.35 >3545 >45-55 >55 Distance to Neighbour (cm) Figure 6. Relative frequency distributions of distances from abalone to their nearest, second nearest, and third nearest neighbors at Flinders prior to fishing, and at each post-Hshing survey. Movement and Re-aggregation of Blacklip Abalone 775 within 5 cm of the source abalone was similar on the second post-fishing survey to that found prior to fishing. In contrast, the increase noted on the first post-fishing survey in the proportion of abalone located some distance from the source abalone decreased only marginally on the second survey. Emergence Categorizing the shell cover of abalone proved to be difficult and imprecise. Therefore, these data were not used for analysis. Figure 7 shows the relative length frequency distributions of the abalone in fished areas at each site. Length frequency distributions before and after fishing did not show great change in the propor- tions of animals above and below the size limit. Immigralion Figure 8 shows changes in the number of abalone counted on fixed transects around fished and control areas at each of the sites. At Point Cook there was a general and sustained decline in the number of abalone around fished areas after fishing. This result was not as apparent around control plots. At Flinders there was a decline in abundance around both fished and control areas but this decline was not as marked as at Point Cook. The immigration of abalone tagged outside the plots is illus- trated for the Point Cook site in Figure 9. The movement of aba- lone into control plots was just as significant as the movement into fished areas. While the number of tag movements at Flinders was not as sreat as at Point Cook, the results were similar. (a) Point Cook o c Oi 0) > ra 25 CC 20 15 10 5 0 5 10 15 20 25 I I Prefishing I 1 1 St Survey (b) Flinders n I □ I □ I D I 49 54 59 64 69 7A JS 8A □ 8£ 64 3£ -04 10!) ¥ 119 Shell length (mm) Figure 7. Relative length frequency distributions of abalone within fished plots at (a) Point Cook and (b) Flinders. The distribution prior to fishing is shown above the \-axis and the distribution after fishing is shown below the x-axis. Vertical lines indicate the size limits used at each site. 776 Officer et al. (a) Point Cook DISCUSSION 100 80 60 40 20 943 s en "'*■'' T*'? ■■' ■ 'P II M j ■: --.:' c c 1 1 t*>-y.v i (b) Flinders I I Fished Q Control 100 80 60 40 20 t 1 I'-'-l 1 r^ - CO 547 II II 1 . . '. '-^■■:i c C 1 -: l-;; i:;..-i ■ " 1 Prefishing Abundance 1st Survey 2nd Survey Figure 8. Changes in llie abundance of abalone counted on fixed transects outside fished and control plots at (al Point Cook and (b) Flinders over the course of the experiment. Relocation Figure 10 shows the number of abalone that moved and sub- sequently settled at each site plotted against the initial abundance of the cell in which they were tagged. At both sites abalone settled in cells of lower initial abundance than that of the cells in which they were tagged. This movement occuired in both control and fished plots but was more apparent in fished plots. ^ ^kX 1- i y Fished ^ -r Control Fished / / ^ ^ 1 < % \ ' ) , ^ tftt tl Figure 9. Schematic diagram showing the minimum distance moved of peripherally tagged abalone that moved into the plots at Point Cook. Many traditional methods of fisheries stock assessment assume that movement-driven changes in spatial distribution have little impact on estimates of abundance or important parameters such as catchability and natural mortality. Such methods have been applied in abalone population surveys despite many anecdotes describing the propensity for abalone to re-aggregate after fishing and during spawning (Shepherd 1986b. McShane 1996. Hart & Gorfine 1997). Although there have been previous studies of abalone movement (Newman 1966. Poore 1972. Shepherd 1986a. Ault & DeMartini 1987. Prince 1992) and the frequency of aggregations (Shepherd 1986b. McShane 1996). a synthesis of these results has remained problematic. This study attempted to redress this defi- ciency by employing a combination of ('/; situ tagging with fine- scale mapping of two distinctly different abalone populations. Controlled fishing of each population was conducted to test the hypothesis that removal of abalone by fishing stimulates move- ments that result in re-aggregation. Experimental fishing clearly had an impact on the populations, evident as declines in the abundance of abalone counted in all fished areas. However, these reductions were less than those ex- pected. Without fishing induced movement a 35% reduction in fished areas was expected and no change was expected in control areas. The increase in abundance in control areas, while not as great, may suggest a response to fishing at some distance. The fact that these changes in abundance within the plots were the opposite of those shown outside the plots suggested that recovery of the populations after fishing was occurring. Substantial movement was observed and both populations showed a trend to return to their original spatial distributions. However, understanding how this redistribution was achieved proved difficult because contradictory results were obtained from two parts of the study. Analysis of the movements of tagged aba- lone showed an apparent dis-aggregation during the recovery, whereas the analysis of changes in the spatial distribution of aba- lone during the recovery suggested a re-aggregation. It was antic- ipated that the movement of tagged abalone would provide data that would best describe the nature of recovery of the abalone populations to fishing. Unfortunately these data were apparently the least reliable as the number of tagged abalone available for this analysis was only a small proportion of the lagged population. Furthermore, the repeated disturbance during the tagging, fishing, and re-surveying could well have caused abalone to disperse, re- ducing the likelihood of recapture, and resulting in dis- aggregation. Increased movement activity has been noted in other abalone species subject to disturbance (Shepherd & Godoy 1989. Werner et al. 1995). One of the important, but unforeseen, out- comes from this study was the impact that tagging-induced dis- persal may have on estimates of natural mortality (Dixon et al. 1998). Changes in the distribution of nearest-neighbor distances ap- peared to better describe the recovery of the populations after fishing. These data represented the entire population under study and were therefore less likely to be affected by tagging induced disturbances. These data suggested that restocking of aggregations occurred at both sites but that it was less noticeable at Point Cook where the abalone displayed a more homogenous distribution. The attenuated recovery at Flinders suggests that reformed aggrega- tions contained fewer individuals than they did prior to fishing and that sparsely populated areas surrounding aggregations were not Movement and Re-aggregation oh Blacklip Abalone 777 Point Cook 14 Control 6 12 1 1 5 10 ^~ 4 8 6 4 1 3 2 2 " 1 1 " 0 -1 0 6 5 ^ 14 IL. 12 Fished 10 8 6 4 1 4 - r" ; 1 — ■ 3 2 2 - 1 1 0 I ' 1 ^ 0 Flinders Control Fished 0-1 2-4 5-9 >10 0-1 2-4 5-9 >10 Initial abundance of cell (No. per m^) I I Moving n Settled Figure 10. Position of moving and settled abalone relative to the Initial abundance of the cell in which they were sighted. quickly restocked following fishing. The recovery of u population through re-aggregation without complete restoration of pre-fishing population densities has important implications for area-hased population surveys because such a re-distribution of abalone will mask depletion caused by fishing. Moreover, the tendency for recruited abalone to re-aggregate ensures that catchability will be maintained despite stock depletion. This means that abalone popula- tions are more vulnerable to over-exploitation than populations of species whose catchability declines as stock numbers decrease. This is not surprising given the global history of abalone stock collapses. Smaller abalone were not the prime source of re-aggregation particularly where emergence from crevices was a factor. This result is contrary to the conclusions of other studies (McShane & Smith 1989. Hart & Gorfine 1997. Hart et al. 1997). It is. however, consistent with the notion that smaller abalone seek the shelter afforded to them by cryptic habitat until they are large enough that the risks of occupying more exposed locations are outweighed by the benefits of improved food availability and the need to repro- duce. Most of the observed movements involved relatively small displacements that, when considered together with post-fishing changes in abundance, suggest re-aggregation resulted from a se- ries of contiguous displacements similar to a "domino effect." This implies that post-fishing spatial dynamics involves size-related competition for preferred home-sites. Our results cast serious doubt on the validity of change-in-ratio methods that have sometimes been used for estimating abalone abundance (Nash et al. 1994). Our results indicate that violation of the critical assumption of equal catchability of two animal types (pre-recruits and recruits) is inevitable. Fine-scale hyper-stability in catch per unit effort (CPUE) is another consequence of re- aggregation and renders CPUE useless as an index of abundance except under conditions of severe depletion. This is not to suggest that instantaneous catch rate may not be sensitive to localized stock depletion. Iinmigration of abalone into fished areas, such as that suggested in this study, would further undermine the assump- tions of change-in-ratio analyses. This will be even more apparent if immigration is most noticeable in recruits only. Blacklip abalone populations can be regarded as comprised of 778 Officer et al. two parts, the aggregations that collectively constitute the stock commercial abalone divers are adept at targeting, and the more sparsely distributed abalone that serve to re-supply and thus main- tain the aggregations after fishing. Although estimates of the num- bers and sizes of aggregations provides a picture of the size of the stock, it is likely that these estimates tend to be hyper-stable whereas the density of sparsely distributed abalone may be a more sensitive indicator of the impact of fishing. Future research should be directed toward developing methods of abundance estimation that accommodate re-aggregation. Further analysis of our data us- ing techniques developed for terrestrial ecosystems that estimate density using distances between nearest neighbors and distances from randomly selected sampling points offer promise in this re- gard (Byth & Ripley 1980. Byth 1982, Officer et al. 2000). Here. the distances are considered as radii that define areas occupied by both sparse and clumped individuals. As the degree of aggregation increases, the distances between nearest neighbors decreases while the distances from randomly selected points tends to increase. The geometric mean between the overall density of nearest neighbors and the overall density relative to random points gives an estimate of density with less bias than more conventional estimates from quadrats and transects. Regardless of the survey methodology em- ployed, the results of this study indicate that aggregating behavior must be considered when assessing the impact of harvesting on blacklip abalone populations. ACKNOWLEDGMENTS Without the support of the Victorian abalone industry this proj- ect would not have been possible. Commercial abalone divers refrained from diving at the two study locations for 6 mo prior to field sampling, during the study, and for the following 6 mo. Miriana Sporcic provided extensive assistance with data analysis. Scientific advice was generously provided by Warwick Nash (Australian Maritime College) and Neil Andrew (NIWA. Welling- ton). TeiTy Walker (MAFRI) re\'iewed initial drafts of this manu- script and provided many valuable suggestions for its improve- ment. Acknowledgment must be made of the substantial effort made by the project's field staff. Assistance with vessel operations and diving support in the initial establishment of experimental sites was pro\ ided from within MAFRI by Mark Ferrier. David Beyer. Dale Thomson, and Ian Duckworth. Andrew Tulloh provided in- valuable expertise for the design and fabrication of tagging equip- ment. External charter vessels were provided at Flinders by Geoff Rodda (Peninsula Diving Instruction), at Point Cook by Graham Leckie (Ultimate Fishing Charters) and Alistair MacDonald (Port Campbell SCUBA and Marine). Diving contractors employed on the project included David Forbes and Bnice Waters of Aquatic Research Support Services. Rod Knights from Professional Diving Services. Geoff Rodda from Peninsula Diving Instruction, and Michael Callan. The research divers collectively accumulated 1000 hours underwater during the surveys. We are extremely grateful to Joanne Merry who generously volunteered her time and expertise to support the field team with the logging of dives and the management of data collection while at sea. and who subse- quently entered and processed the data on computer. The project was funded by the Fisheries Research and Development Corpora- lion (Project 95/165) and the Marine and Freshwater Resources Institute. Auk. J. S. & J. D. DeMartini. 1987. Movement and dispersion of red abalone. Hatknis rufesceiis. in Northern California. California Fish and Game 73:196-21.^. Breen. P. A. 1992. A review of models used lor stock assessment in abalone fisheries. In: S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo. editors. Ahalone of the world: biology, fisheries and culture. Oxford: Blackwell. pp. 253-275. Byth. K. 1982. On robust distance-based intensity estimators. Biomeirics 38:127-135. Byth, K. & B. D. Ripley. 1980. On sampling spatial patterns hy distance methods. Biometrics 36:279-284. Campbell. D. J. & D. J. Clarke. 1971. Nearest neighbour tests of signifi- cance for nonrandomness in the spatial distribution of singing crickets (Teleogr\lliis commodus |Walker|). Animal Behaviour 19:750-756. Dixon, C. D.. H. K. Gorfine. R. A. Officer & M. Sporcic. 1998. Dispersal of tagged blacklip abalone. Haliolis rubra: Implications for stock as- sessment. J. Shellfish Res. 17:881-887. Hart, A. M. & H. K. Gorfine. 1997. Abundance estimation of blacklip abalone {Haliolis rubra) II. A comparative evaluation of catch-effort, change-in-ratio. mark-recapture and diver-survey melhods. Fisheries Research 29:171-183. Hart. A. M.. H. K. Gorfine & M, P. Callan. 1997. Abundance estimation of blacklip ahalone {Haliolis rubra) 1. An analysis of diver-survey meth- ods for large-scale monitoring. Fisheries Research 29:159-183. Krebs. C. J. 1989. Ecological methodology. New York: Harper Collins, p. 654. McShane, P. E. 1994. Estimating the abundance of abalone {Haliolis spp.) stocks — examples from Victoria and southern New Zealand. Fisheries Research 19:379-394. LITERATURE CITED McShane. P. E. 1995. Estimating the abundance of abalone: the importance of patch si/e. Marine and Freshwater Research 46:657-662. McShane. P. E. 1 996. Patch dynamics and effects of exploitation on aba- lone {Haliolis iris) populations. Fisheries Research 25:191-199. McShane. P. E. & M. G. Smith. 1989. Direct measurement of fishing mortality in abalone {Haliolis rubra Leach) off Southeastern Australia. Fisheries Research 8:9.3-102. Nash, W. J., T. L. Sellers. S. R. Talbot. A. J. Cawthorn & W. B. Ford. 1994. The population biology of abalone {Haliolis species) in Tasma- nia. I . Blacklip abalone {H. rubra) from the North Coast and the islands of Bass Strait. Technical Report. Department of Primary Industry and Fisheries. Tasmania: Hobart. Newman. G. G. 1966. Movements of the South African abalone, Haliolis inidae. Invest. Rep and Div. Sea Fish and Rep. South Africa 56:20. Officer. R. A.. M. Haddon & H. K. Gorfine. 2001. Distance-based abun- dance estimation for abalone. / Shellfish Res. 20:781-786. Poore. G. C. B. 1972. Ecology of New Zealand abalones. Haliotis species (Mollusca. Gastropoda). 2: Seasonal and diurnal movement. New Zealand Journal of Marine and Freshwater Research 6:246-258. Prince. J. D. 1989. The fisheries biology of Tasmanian stocks of Haliotis rubra. PhD Thesis: Department of Zoology, University of Tasmania: Hobart. Prince. J. D. 1991. A new technique for tagging abalone. Australian Jour- nal of Marine and Freshwater Research 42:101-106. Prince. J. D. 1992. Using a spatial model (o explore the dynamics of an exploited stock of the abalone Haliotis rubra. In: S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo. editors. Abalone of the world: biology, fisheries and culture. Oxford: Blackwell. pp. 305- 317. Movement and Re-acgrhgation of Blacklip Abalone 779 Prince. J. D. & S. A. Shepherd. 1992. Australian abalone fisheries and their management. In: .S. A. Shepherd, M. J. Tegner & S. A. Guzman del Prcio. editors. Ahalone of the world: biology, fisheries and culture. Oxford: Blackwell, Shepherd. S. A. |yX6a. Mo\ement ol the southern Australian abalone Haliotis laevigata in relation to crevice abundance. Australian Jaimial of Ecology ll:29.'i--^02. Shepherd. S. A. 1986b. Studies on southern Australian abalone (genus Haliotis) VII. Aggregative behaviour of H. laevigata in relation to spawning. Marine Biology 90:231-236. Shepherd, S. A. & C. Godoy. 1989. Studies on southern Australian abalone (genus Haliotis) XI. Movement and natural mortality of juveniles. Joiiriial of the Malacological Society of Australia l():S7-9.'i. Werner. I.. S. Flothmann & G. Burnell. 199.S. Behaviour studies on the mobility of two species of abalone {Haliotis iiihercalata and H. discus haiiiiai) on sand: implications for reseeding programmes. Marine mid Freshwater Research 46:681-688. Jotinuil ,ij Slullfish Research. Vol. 20. No. 2. 78I-7S6. 201)1. DISTANCE-BASED ABUNDANCE ESTIMATION FOR ABALONE RICKARD A. OFFICER,' * MALCOLM HADDON,' AND HARRY K. GORFINE" 'Tii.siiHiiiiiiii Aqiiaciihurc and Fisheries Institute. University of Tasmania. Taroona. Tasmania 7053, Australia: 'Mtaiiw and Freshwater Resaurces Institute. Qiieenscliff. Vietoria 3225. Australia ABSTRACT Indices, of abundance are an iniporlani component of stock assessment models. Unfortunately, conventional attempts to estimate the abundance of abalone are hampered by the patchy spatial distribution characteristic of abalone. An alternative distance- based abundance estimator was evaluated by the simulated sampling of a large natural population of abalone whose relative positions had been accurately mapped. Distance-based abundance estimators can better accommodate the aggregated distribution of abalone. The simulated sampling of the real population of blacklip abalone {Haliotis nilyni Leach) was u.sed to examine the accuracy, bias, and sensitivity to changes m abundance of the distance-based approach. In each sample the distances from a randomly selected point and from a randomly selected abalone to the nearest abalone were used as the basis for a compound estimate of abundance. A Monte Carlo procedure was used to obtain percentile confidence limits about abundance estimates. The distance-based method was found to approximate the true abundance of the population and therefore may be useful as an indicator of absolute abundance. The method was al.so sensitive to changes in abundance. This sensitivity was examined by simulating the effects of fishing at varying rates of exploitation. Simulated reductions at moderate levels of exploitation {20-25% overall reduction) showed that the method was able to detect changes in abundance with reasonable confidence (90%). The resolution of some technical difficulties may enhance the ability of the distance-based method to detect changes in abundance in a real fishery. KEY WORDS: Hdlialis rubra, abalone. blacklip. stock assessment, abundance estimation, nearest neighbor distance, distance INTRODUCTION Blacklip abalone. Hiiliotis nihra. form the basis of u large and valuable fishery in southern Australian waters. Despite the eco- nomic importance of the fishery, stock assessment of blacklip abalone has remained problematic. Proper estimation of the sus- tainable yield from the fishery requires assessment of how the size of the stock changes over time. For assessment methods to be useful in routine monitoring they must: provide a reliable relative or absolute index of abundance for the stock, be sensitive to changes in abundance in the population, and be easily applied across a representative proportion of the fishery. Fishery-dependent methods of stock assessment often fail for abalone fisheries because of the inherent hyper-stability in catch- rate indices (Breen 1992. McShane 1994). The tendency of black- lip abalone to live in aggregations renders them more vulnerable to overfishing by increasing their catchability. Divers are able to maintain high catch rates by moving from aggregation to aggre- gation and by relocating to a different area once catch rates drop below an acceptable level. If the abalone remaining after fishing then re-aggregate, a diver returning to the area may well fish the area at a similar catch rate to the first excursion. These processes may all contribute to hyper-stability in catch rates and conceal real fluctuations in population size due to local depletion. Unfortunately, fishery-independent methods of assessment are also difficult to apply to abalone populations. Area-based abun- dance estimators can provide accurate measurement of abundance and an ability to detect change (Hart & Gorfine 1997, Hart et al. 1997). However, the patchy distribution of abalone unfortunately introduces high variance into such estimates and sample sizes must therefore be extremely high to achieve acceptable levels of preci- sion (Nash 1995). Methods such as mark-recapture and change- in-ratio both may show reasonable precision and sensitivity (Hart and Gorfine 1997). Unfortunately, the sampling intensity needed *Corresponding author. Current address: Marine Institute. Marine Fisher- ies Services Division, 80 Harcourt St., Dublin 2, Ireland. E-mail address: rick.officer@marine.ie to achieve this makes these measures useful only for the assess- ment of small populations. The time required to apply such mea- sures across a large fishery would make them prohibitively expen- sive as a routine monitoring tool. Fishery-independent assessment methods that are less time consuming offer scope for the monitoring of large areas but may forsake precision. Estimates provided by techniques such as timed collections suffer from variation in collection efficiency between divers (Shepherd 1985, McShane 1994) and in sea conditions at the time of sampling. Such methods are also prone to hyper- stability due to the tendency of abalone to aggregate and the un- intentional attraction of research divers to these aggregations (Mc- Shane 1994). Where the research diver spends most time within an aggregation handling abalone, rather than searching, the estimated density tends unrealisticaily toward infinity (Hart et al, 1997). The development of assessment methods that better handle the aggregated distribution of abalone would clearly be an advantage. Distance-based methods of abundance estimation offer scope for species such as abalone that are largely sedentary and aggregated. Techniques developed for terrestrial ecosystems that estimate den- sity using distances between nearest neighbors and distances be- tween randomly selected sampling points and the nearest indi- vidual of the species of interest offer potentially workable methods (Byth & Ripley 1980, Byth 1982). As the degree of aggregation increases, the distances between nearest neighbors tend to decrease while the distances from randomly selected points to individuals tend to increase. Compound estimates of the overall density using these two measures may therefore accommodate a wider range of spatial distribution types than more conventional abundance esti- mates (Diggle 1975). While such methods have been in existence for some time (Clark & Evans 1954) they have not been applied to abalone. This is perhaps because the data requirements necessary to test the efficacy of the methods are particularly difficult to achieve — samples are required for a population of known abun- dance and distribution. The objective of this study was to evaluate the efficacy of a relatively simple distance-based abundance estimator. This was done using simulated measurements and removals of abalone from a blacklip abalone population of known abundance and spatial 781 782 Officer et al. distribution. Tlie evaluation examined tlie accuracy, precision, and sensitivity to changes in abundance of the estimator. MATERIALS AND METHODS Sliiilv Site The mapping ot abalone distributions was undertaken at Point Cook. Victoria. Australia (37 5.'i.893' South. 144' 47. 104' East). The site was 3 to 4 m in depth and was not subject to strong swell or current. The reef consisted of basalt boulders, rarely over 1 m in height, on a sandy substrate. There was very little cryptic habitat that could not be thoroughly searched for abalone. The study site covered an area of 38 x 58 m. subdivided into four plots of 24 x 24 m separated by conidors 10 m wide. Each plot was further divided into 1 m" cells using a grid system. The grid was used to record the position of all abalone observed within the plot within an accuracy of ±10 cm. To simplify the application of the computer simulation procedure the 10-m corridors between plots were removed, creating one plot measuring 48 x 48 m. Simulalion Procedure The position of all abalone observed by divers in the plot was recorded in two dimensions as an X-Y coordinate and saved in a Microsoft Excel spreadsheet (Fig. 1 ). A total of 8009 abalone was observed by divers in the plot. The length composition of the population was determined for a sample of 1809 abalone measured in the study area. To allow for comparisons between the size structure of abalone occupying densely populated habitat and these in sparsely populated areas, lengths had been recorded separately for abalone in each type of habitat. Each of the 8009 abalone was assigned a length from the known length composition data. An interactive macro program was written in Visual Basic to calculate a series of distances between source positions (either randomly chosen points or randomly chosen abalone) and the near- est abalone. The method assumes that distances can be measured from source positions within the plot to abalone located outside the plot. Because there was no data for abalone outside the plot, source positions were restricted to lie inside a 2-m border region around the plot (Fig. 1 ). It was possible to measure from a source position located within the inner plot to an abalone located within the border region but no measurements originated within the border. In all simulations reported here the border region used was 2 m. This reduced the size of the area from which source positions could be chosen to 44 x 44 m. 6596 abalone were observed within the 44 x 44 m inner plot. Calculation of Randomly Chosen Point to Nearest Abalone Distances A random number generator was used to generate the reqtnred number of randomly chosen points within the inner plot. For every randomly chosen point the distance to the nearest abalone was recorded. The nearest abalone could be located in the 2-m border region outside the inner plot. Calculation of Randomly Chosen Abalone to Nearest Abalone Distances Because the number of abalone within the plot was known it was possible to assign each abalone a random number. A table of random numbers was then used to choo.se abalone located within the inner plot from which distances to the nearest abalone would be calculated. The nearest abalone could be located in the 2-m border region outside the inner plot. The number of randomly fe* _^ ^ _ _ ^ Fijjure 1. Diagram sho«in}> the position of all SdO"* abalone within the study area. Abalone are plotted as untllied circles se\eral limes their actual size. Where abalone are densely aggregated the plotted circles overlap creating dark patches. Source positions for either randomly chosen points or randomly chosen abalone were prevented from origi- nating within a 2-m border region (shaded area! to prevent the pos- sihllit\ that the nearest ahalone to a source position was outside the mapped region. However, it was possible to measure from abalone located within the inner plot to abalone located within the border region. chosen abalone from which to calculate distances could be speci- fied. The macro prevented abalone from being chosen more than once in any sample taken from the plot. This avoided the re- measurement of nearest neighbor distances from the same abalone. For every randomly chosen abalone the distance to the nearest abalone was calculated. The lengths assigned to each abalone were used to compare the calculated nearest neighbor distance with the minimum possible distance between the centers of each abalone. The greater of these two distances was recorded. This was done to avoid assigning nearest neighbor distances smaller than if nearest neighbors were touching. Calculation of Distance-based Density Estimates The formula suggested by Byth and Ripley ( 1980) was used to generate estimates of population density (Eq. 1 ); A': ■^2j^n; Where A' is an estimate of population density. II is the sample size (number of distances measured), and r, is the distance from the /th random source position to the nearest abalone. Density estimates were calculated separately from the series of randomly chosen point-abalone distances and from the series of randomly chosen abalone-abalone distances. Unfortunately the density estimates derived from each type of calculation are sensi- tive to the spatial pattern of the organisms under study (Diggle 1975. Krebs 1989). Diggle (1975) found that as the degree of Distance-based Abundance Estimation for Abalone 783 population aggregation increased, point-organism distances also tended to increase resulting in underestimates of population den- sity. Conversely, organism-organism distances tended to decrease and result in overestimation of population density. He showed that an unbiased estimate of the population density could be obtained by calculating the geometric mean of the point-organism and or- ganism-organism density estimates. In this study Diggle's protocol was followed and a compound and theoretically unbiased estimate of density was calculated as (Eq. 2): N-- V^, N, Where N is a compound estimate of population density, and ^/tcfi ''"'J ^KCAh '"''5 respectively the randomly chosen point- abalone and randomly chosen abalone-abalone estimates of population density (derived from Equation I). Cakiilaliiiii of Expected Oensilies The expected density of abalone in the plot was calculated as the total number of abalone within the inner plot divided by the total area of the entire plot (48 x 48 m = 2.^04 m"). To more closely redect the approach taken by the distance-based density estimation, the number of abalone within the inner plot was di- vided by the total area, rather than the area of the inner plot. The distance-based method estimates density for a larger area than that of the inner plot because it was possible to measure from a source position located within the inner plot to an abalone located within the 2-m border region. Expected densities were calculated for the initial population and for three levels of population reduction. The levels of reduc- tion simulated were 25%, 50%, and 75% less than the initial popu- lation size. Little difference was found between the size compo- sition of abalone located in densely populated habitat with that of abalone located in sparsely populated areas. The simulated re- moval of abalone from the plot was therefore done randomly to effect each level of population reduction. Estimation of Optimal Sample Size By successively increasing the number of distance measure- ments used in the calculation of the compound distance-based estimate, the simulator was used to determine the sampling inten- sity required for consistent density estimation. The number of distance measurements of each type was increased from 50 to 800 in steps of 50. Five iterations were run at each step to allow variability to be examined for each number of measurements. Estimation of the Miiiinitim Detectable Difference Subsequent distance-based estimates of the initial population density were calculated from 250 distance measurements of each type of density estimate. This number was chosen for two reasons: Further increases in the number of distance measurements did not appear to greatly increase the precision of density estimates, and. 250 was thought to be a practical number of measurements to take if conducting the sampling in the field. 250 randomly chosen abalone represented 3.8% of the initial population of 6596 abalone within the inner plot of the study area. 100 iterations of the simulation procedure were run to generate a distribution of estimated densities for the population within the entire plot. 807r, 90%, and 95% percentile confidence limits were calculated from this distribution. To estimate the minimum detectable difference the population was reduced in successive steps and the simulation procedure re- run 10(1 times at each level of reduction. The number of distance measurements of each type was kept at the same percentage (3.8%) of the residual population within the inner plot to maintain the same level of sampling intensity as that used in the estimation of density of the initial population. 80%, 90%, and 95% percentile confidence limits were calculated for each level of reduction. A comparison was made between the 100 estimated densities calculated at each level of reduction and the corresponding ex- pected densities. The relationship between the two was described using linear regression. Regression lines were also drawn between the various upper and lower percentile confidence limits. These relationships were used to estimate the ability of the method to detect changes in abundance at various levels of confidence and population density. RESULTS Estimation of Optimal Sample Size Successive increases in the number of distance measurements used in each simulation improved the consistency of density esti- mates calculated for the initial population (Fig. 2). Variability in repeated simulations reduced markedly when 300 or more dis- tances of each type were measured. Density estimates of the initial population approached a central limit of about 2.7 abalone/m^. Estimation of the Minimum Detectable Difference Negligible differences in si/.e structure were found between abalone in densely populated habitat and those in sparsely popu- lated areas (Fig. 3). This justified the random removal of abalone from the plot to affect each level of population reduction. Estimates of density were less variable at reduced population sizes (Table 1, Fig. 4 1 despite there being no increase in sampling intensity in simulations run at lower population sizes. Density 3.3 3.1 - 2.9 2.7 1 .2- 0) TO E CD ^ 2.5 l- c 0) ° 2.3 2.1 {{{ j}*'Mn{nn 200 400 600 Number of distances measured 800 Figure 2. ,\vera(;e density estimates (+SK) eiileulated In the simulator at various sampling intensities. Five iterations were run at each sam- pling intensity. The sampling intensity is indicated by the number of distances measured of each type (i.e., 2(10 indicates (hat 200 randomly chosen point-ahalone and 200 randomly chosen abalone-abalone dis- tances were measured). 784 Officer et al. 25% 20% >. o c 0) =1 15% cr ■^ 10% TO 0) 5% 0% _n n ■=jctl B Dense, n = 849 □ Sparse, n = 957 1].=. 64 69 74 79 84 89 94 99 104 109 114 119 124 Shell length (mm) Figure i. Comparison of the size structure of abalone collected from densely populated habitat and sparsely populated habitat at the Point Cook site. There is little difference in the size structure of abalone collected from each habitat type and above the minimum legal length of 100 mm these differences are negligible. estimates at each population si/e were normally distributed and reflected relative reductions in population si/e (Fig. 4). The relationships between estimated and expected densities, and between the various upper and lower percentile confidence limits were linear (Fig. 5). Estimated densities closely estimated the true density of the population, and estimated reductions were close to the true levels of reduction (Table I). Estimated densities appeared to be a good indicator of absolute abundance. This was demonstrated by the strong proportional relationship found be- tween estimated and absolute abundance (Estimated density = 0.92 X Expected density + 0.10. R" = 0.96). The ability to detect proportional differences in abundance was dependent on the size of the population and declined in proportional terms with reducing population size. The linear relationship between estimated and expected densi- ties (and the confidence limits about this relationship) allowed an evaluation of the power of the method to detect changes in abun- dance at a range of initial population densities. At greater initial population densities (i.e.. >3 abalone/m") the method was able to detect declines in abundance of as little as 22% with 90% confi- dence (Fig. 6A). Power to detect declines with greater levels of confidence was high at greater initial population densities (e.g.. at an initial density estimate of 3 abalone/nr. declines in abundance of 24% could be detected with 93% confidence) (Fig. 6A). Power to detect proportional declines in abundance reduced markedly with declining initial density estimates less than 1 abalone/m" (Fig. 6A). Power to detect increases in abundance followed a similar pattern to that for detection of declines (Fig. 6B). At an initial population density of 3 abalone/m" the method was able to detect increases in abundance of about 26% with 90% confidence (Fig. 6B). DISCUSSION The distance-based abundance estimator described here was able to detect measurable changes in the absolute abundance of an abalone population. The minimum differences detectable were relati\ely small even at reasonably high levels of confidence. The method appears to satisfy two primary requirements of abundance estimation methods: it is a good indicator of abundance, and it is able to detect changes in abundance. A third requirement, that the method be easily applied across the fishery, was not examined completely in this study. The technical requirements of the method appear to be minor. The time consuming repeated surveys char- acteristic of change-in-ratio and mark-recapture methods are not required. The method therefore minimizes reliance on prolonged periods of good diving conditions. The measurements required are relatively simple to achieve and the skill required to take the mea- surements would probably be acquired quickly by an observer with limited experience. This satisfies the need of long-term monitoring programs to accommodate changes in the survey team and avoids the requirement for a standard dive team when using timed col- lection methods. The simulated sampling using real field data presented here suggests that the method has the accuracy of area-based measures but the ease of application of timed collections. The levels of sampling demonstrated here do not represent an insurmountable amount of work. At lower densities the number of measurements that need to be taken could be increased. The method accommo- dates an uneven amount of randomly chosen point-abalone and randomly chosen abalone-abalone measurements, making it adapt- able to field situations. Furthermore, by considering the randomly chosen point-abalone and randomly chosen abalone-abalone den- sity estimates separately, the method provides an opportunity to describe the patchiness and patch density in the population and TABLE 1. Results of simulations performed at various population sizes. Population Reduction Density (abs/ni") Percentiles Population Size Measured Actual Estimated Actual Estimated 10th 90th 6596 2.5(1 ~ ~ 2.86 2.73 a,. 2.40 3.06 494} 187 25% 25% 2.15 2.05 „,, 6 1.81 2.32 .^.^1)0 124 50% 48% 1.43 l-42o„ 1.16 1.66 lh4S 62 75% 72% 0.72 0.75 „,„ 0.63 0.90 Population reductions were effected over the entire plot. Therefore the population size of reduced populations represents the average number of abalone remaining within the inner plot. To maintain the same relative sampling intensity the number of distances measured was less for simulations performed on reduced populations. The number of distances measured indicates the number measured of each type (i.e.. 62 indicates that 62 randomly chosen point-abalone and 62 randomly chosen abalone-abalone distances were measured). Estimated reductions were calcu- lated by comparing the estimated density of reduced population sizes with that of the initial population, .\ctual densities were calculated as the population size divided b\ the area of the entire plot (2.V)4 m'). The estimated density is the mean of 100 density estimates calculated at each population size. Sub-scripted numbers indicate the standard deviation about the mean. Distance-based Abundance Estimation for Abalone 785 50 r 30 15 0.0 0.5 1.0 1.5 2.0 2.5 Density estimate (abs/m^) 3,0 Figurt' 4. Distributions of density estimates generated from KM) simu- lations calculated at various population sizes: initial population (white bars), 25*^^ reduction (light-shaded bars). 50'^f reduction (dark- shaded barsi and TS'^r reduction (black bars). Overlaid lines indicate normal distributions lltted to the estimated data for each population size. Note that the variance reduces as the mean density estimate de- creases. changes in these measures over time. Information on the patch structure of abalone populations may be an important indicator of productivity (McShane 1995. Taylor et al. 2000). but this infor- mation is not collected by most inonitoring methods. If the field requirements of the method could be satisfied then the method could be particularly useful as a routine estimator of abundance. However, the simulations reveal several limitations of the method and indicate some further areas for investigation. Per- haps the most important limitation of the analysis presented here is that the fishing patterns simulated were not real. The removal of abalone was effected by random selection of abalone from the entire plot. True fishing patterns are focused on aggregations of abalone (Hart et al. 1997). Random fishinc assumes that all fish in the plot are of legal size and equally likely to be removed. In most abalone fisheries adherence to minimum size limits makes only a proportion of the population available to fishing. If the size struc- ture of abalone occupying sparsely populated habitat differs from that in aggregations, fishing will dramatically affect the spatial structure of the population. In this study no evidence for differing size structure between densely and sparsely populated habitats was found in the Point Cook population. Similar results have been found at another site (Officer et al. 2000). Even if no differences exist between size structures in densely and sparsely populated habitat, fishing patterns that focus on ag- gregations are likely to have both immediate and prolonged effects on the population. In the short term, the removal of abalone from aggregations would create a more dispersed and homogenous dis- tribution pattern after fishing. However, there is evidence that abalone are capable of moving considerable distances if disturbed and that this enables re-aggregation into preferred home-sites to occur rapidly (Gorfine et al. 1998, Officer et al. 2000). At the Point Cook site this behavior returned the distribution of abalone to the initial distribution pattern (albeit at a lower density) within weeks of intensive fishing (Gorfine et al. 199S. Officer et al. 2000). The random removal of abalone simulated here is therefore likely to 80% Q) 70% CD Q 60% - 0) ■D CD 50% - £! 40% - ■D 30% - E F 20% - c is 10% - 0% (A) 1 03 3.5 3.0 2.5 ^ 2.0 c O) "D "D B ro E to UJ 1.5 1.0 0.5 0.0 0.0 0.5 1.0 1.5 2.0 Expecte(j density (abs/m^) 2.5 3.0 Figure 5. Relationship between estimated and expected density (F^sti- mated density = 11.92 x Expected density -i- tl.ll), R- = l).96l. Dotted lines indicate 90% confidence intervals. 80% CD 70% £ 60% _Q) 50% CO o 40% ■^ 30% .i 20% - c ^ 10% 0% (B) 0 12 3 4 Initial (Jensity estimate (abs/m'^) Figure 6. Povier of the distance-based method to detect change. Mini- mum detectable changes are expressed as a percentage of the initial density estimate for (A) the minimum detectable proportion decrease, and (Bl the minimum detectable proportion increase. 786 Officer et al. approximate the distribution of abalone found in the real Point Cook population several weeks after fishing. Whether or not this relationship would hold generally is uncer- tain. In a study of the movement and re-aggregation of abalone. Officer et al. (2000) subjected the Point Cook site and a second abalone population (Flinders) to similar levels of experimental fishing. At Flinders the reef had much greater relief and cryptic habitat than at Point Cook and the abalone were less abundant but more patchily distributed. Recovery of the initial distribution pat- tern took much longer at Flinders (Officer et al. 2000). The Point Cook site was unusual in that it contained little cryptic habitat and low relief This apparent success of the distance-based method at Point Cook may result from the two-dimensional area assumed by the estimator correlating well with the actual area of the site. Testing the efficacy of the distance-based method in areas of more complex topography and cryptic habitat will require further ex- perimental work. Unfortunately the role of the cryptic component of abalone stocks remains poorly understood despite its impor- tance to stock assessment (Haddon 2000). Another shortcoming of the method is that it is only sensitive to changes in abundance when the population density is relatively high. In this respect the method differs from other assessment techniques that may become hyper-stable at higher population den- sities (e.g.. the timed collection (Hart & Gorfine 1997. Hart et al. I997)|. Fortunately, at lower population densities methods such as the timed collection come into their own as measures of relative abundance. A benefit of applying the method at higher population densities is that the area that has to be searched between nearest neighbors by research divers will tend to decrease. This will in- crease the efficiency of the method in the field. The method may therefore be particularly useful for H. rubra populations in Tas- mania where the abundance of abalone appears to be relatively high and the amount of coast that needs to be monitored is ex- tremely large. There are also some technical limitations to the simulations applied here. Designing a field protocol that selects source posi- tions and abalone randomly will not be a trivial task (Ripley 1981 ). Selecting a randomly chosen point underwater is relatively easy. The main problem may occur when the point falls on inappropriate habitat. This suggests a need to stratify sampling by habitat when applying the distance-based method. Selection of a randomly cho- sen abalone is far more complicated and may be a source of po- tential bias in the method (Ripley 1981 ). For the method to remain effective it will be important to avoid one of the shortcomings of the timed collection procedure: attraction of research divers to aggregations. The apparent utility of the method as a sensitive abundance estimator suggests that attempts to overcome these dif- ficulties warrant serious attention. ACKNOWLEDGMENTS The data used in these simulations could not have been col- lected without the support of the Victorian abalone industry. Com- mercial abalone divers refrained from diving at the study location during the course of the study. Mapping the position of over 8000 abalone required substantial effort from the project's field staff Assistance with vessel operations and diving support was provided within MAFRI by Mark Feirier. David Beyer. Dale Thomson, and Ian Duckworth. Charter vessels were provided by Graham Leckie (Ultimate Fishing Charters) and Alistair MacDonald (Port Camp- bell SCUBA and Marine). Diving contractors employed on the project included David Forbes and Bruce Waters of Aquatic Re- search Support Services. Rod Knights from Professional Diving Services. Geoff Rodda from Peninsula Diving Instmction. and Michael Callan. The research divers collectively accumulated 1000 hours underwater during the surveys. We are extremely grateful to Joanne Merry who generously volunteered her time and expertise to support the field team with the logging of dives and the management of data collection while at sea, and who subse- quently entered and processed the data on computer. The initial project was funded by the Fisheries Research and Development Corporation (Project 95/165) and the Marine and Freshwater Re- sources Institute. LITERATURE CITED Breen. P. A. 1992. A review of models used for stock assessment in abalone fisheries. In: S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo. editors. Abalone of the world: biology, fisheries and culture, Oxford: Blackwell. pp. 253-275. Byth. K. 1982. On robust distance-based intensity estimators. Biometrics 38:127-135. Byth. K. & B. D. Ripley. 1980. On sampling spatial patterns by distance methods. Biometrics 36:279-284. Clark. P. J. & P. C. Evans. 1954. Distance to nearest neighbour as a measure of spatial relationships in populations. Ecology 35:445—453. Diggle. P. J. 1975. Robust density estinialion using distance methods. Biometrika 62:39-48. Gorfine. H. K., R. A. Officer & C. D. Dixon. 1998. Movement and re- aggregation of the blacklip abalone, Huliotis rubra Leach, following intensive fishing. Final Report to the Fisheries Research and Develop- ment Corporation. Melbourne: Marine and Freshwater Resources In- stitute. Haddon, M. 2000. Size-structured models of abalone populations with a cryptic component to the stock. J. Shellfish Res. 20. Hart, A. M. & H. K. Gortlne. 1997. Abundance estimation of blacklip abalone (Haliotis rubra) II. A comparative evaluation of catch-effort, change-in-ratio. mark-recapture and diver-survey methods. Fisheries Res. 29:171-183. Hart, A. M., H. K. Gorfine & M. P. Callan. 1497. Abundance estimation of blacklip abalone (Haliotis rubra) I. An analysis of diver-survey meth- ods for large-scale monitoring. Fisheries Res. 29:159-183. Krebs, C. J. 1989. Ecological methodology. .New York: Harper Collins, pp. 654. McShane, P. E. 1994. Estimating the abundance of abalone (Haliotis spp.) stocks - examples from Victoria and southern New Zealand. Fisheries Res. 19:379-394. McShane. P. E. 1995. Estimating the abundance of abalone: the importance of patch size. Marine anil Freshwater Res. 46:657-662. Nash. W. J. 1995. The development of new techniques for assessing and managing the Australian abalone fisheries. Final Report to the Fisheries Research and Development Corporation. Department of Primary In- dustry and Fisheries. Tasmania: Hobart. Officer. R. A.. H. K. Gorfine & C. D. Dixon. 2001. Movement and re- aggregation of the blacklip abalone. Haliotis rubra, after tlshing. J. Shellfish Res. 20:11 \-119. Ripley. B. D. 1981. Spatial statistics. In: Wiley series in probability and mathematical statistics. Shepherd. S. A. 1985. Power and efficiency of a research diver, with a description of a rapid underwater measuring gauge: their use in mea- sunng recruitment and density of an abalone population. Diving fi)r Science 85:263-272. Taylor. N. A., R. McGarvey & S. J. Hall. 2000. A parameter estimation model for greenlip abalone [Haliotis laevigata) population dynamics. J. Shellfish Res. 20. Joimuil cf Shellfish fic.scanh. Vol. :(). No. 2. 787-794. 2001. DIVER BEHAVIOUR AND ITS INFLUENCE ON ASSESSMENTS OF A QUOTA-MANAGED ABALONE FISHERY HARRY K. GORFINE* AND CAMERON D. DIXON Marine & Freshwater Resources Institute. P.O. Bo.x 114. Queenscliff. Victoria. Australia ABSTHiCT During I WS. we inituiled an on-bodrd obhener program to gain a better tinderstandiiig of spatial and temporal patterns of catch and effort in the blacklip abalone fishery of Victoria, Australia. Although average catch per unit effort (CPUE) for this fishery has been increasing, there has also been significant spatial contraction of the fishing grounds away from reefs of low productivity as a consequence of quota introduction during 1988 to 1989. It is this shift away from reefs of low productivity rather than an increase in abalone abundance that is responsible for the trend in CPUE. From our on-board observations, divers do not operate in an area if they believe that they will not meet their daily catch expectations; they have a relatively high catch rate threshold for deciding when to shift to another reef Catch rates per bag of abalone are several times higher than the daily reported CPUE rates, but v;iry substantially. We conclude that the incentive to maintain high catch rate thresholds for cessation of fishing has led to patterns of rotation of effort at the reef scale that tend to mitigate against high rates of exploitation. This has provided an effective tier of fme-scale self-governance below that controlled by management regulations. However, contemporary changes in the fishery, such as reductions in the number of divers who own access entitlements and their levels of experience, may lead to reduced catch rate thresholds and unfavorable patterns of fishing behavior. There is also a tendency for divers who have recently entered the fishery to concentrate their effort on shallower reefs for reasons of health and safety. Although there is no evidence to date that the over-all status of Victorian abalone stocks has been compromised by these changes, managers need to be aware that regulations affect diver behavior and that it is the divers not the resource that is managed. Our studies reinforce the importance of identifying and promoting behaviors among divers that are desirable in terms of sustainable production within the conteKt of contemporary management strategies. KEY WORDS: quotas, management, fishing behavior, abalone INTRODUCTION Although quota iiianagenieiit is ubiquitous among productive abalone fisheries in the Southern Hemisphere, management agen- cies in the Northern Hemisphere are still considering the merits of quotas as output controls for abalone (CDFG 1997). In an assess- ment of the efficacy of quota management, it would be imprudent to assume that total allowable catches (TACs) per se have been the reason for the relatively prosperous state of abalone fisheries in the Southern Hemisphere. Quota management regimes are usually implemented over broad scales that provide considerable latitude for abalone divers to choose where and when to expend their fishing effort. For example, the abalone fishery in Victoria. Aus- tralia, which provides about 14% of global wild-stock production and spans about 1,300 km of coastline, is divided into three rela- tively large quota management zones that are not subject to sea- sonal closure (Fig. 1 ). Such expansive zones do not facilitate con- trol of catch outputs at the metapopulation scale that delimits population recruitment (Shepherd & Baker 1998). Not only may quota management regimes not be as effective in controlling outputs as intended, they may also have inadvertent effects on diver behavior. Prince and Shepherd (1992) identified potential for spatial maldistribution of effort arising from the implementation of quotas, with overfishing of the more productive areas of the fishing grounds proximal to port and underutilization of the more remote and less productive reefs. Since the introduc- tion of TACs in the Victorian fishery, the fishing grounds have contracted substantially. There has been a progressive shift in ef- fort away from reefs with lower catch rates toward more highly productive reefs (Gortlne et al. this publication). Although improb- able, there are no regulatory constraints to prevent divers within a management zone from intensively harvesting the TAC from only one region in a relatively short time period (currently an aggregate 'Corresponding author. E-mail address: Harry. Gorfme@nre.vic.go of 50-55 days effort per diver is required to harvest annual quota allocations in Victoria). Quotas have also brought about structural changes in the Victorian abalone fishery. Mean experience levels among those participating in the catching sector has decreased markedly in all zones of the Victorian fishery since quotas were introduced. At the same time, the number of fishery access license owners choosing to engage the services of contract divers to har- vest their quota allocations has increased commensurately (Fig. 2). It is reasonable to speculate that some impetus for this transition from owner-operator to contract diver came from increased prof- itability associated with an escalation in market demand for Aus- tralian abalone products that saw the beach price more than double between 1991 and 1992 and 1993 to 1994 (Fisheries Victoria 1998). Despite the potential for intense fishing pressure to cause serial depletion of the more productive reefs, this has not been manifest in the Victorian abalone resource (McShane 1992). although recent modeling predicts a slow decline in relative biomass (Gorfine & Dixon 2000), and there are some recent instances of localized depletion. Possible explanations for the apparent sustainability of Victorian abalone stocks include: 1 . Reductions in real and latent effort through a combination of limited entry, nontransferability, and diver attrition during the two decades preceding quota introduction in 1988 (Sanders & Beinssen 1972. McShane 1992). Further effort reductions when transferability on a "two divers out for each diver in" basis was introduced during 1984. 2. Divers varying their effort at a given reef considerably from year to year. In many instances, effort is rotated among reefs to maintain catch rates and allow substocks on recently fished reefs time to recover (McShane 1992). This is akin to the rotational harvest strategies discussed by Perry et al. (1999). However, although many reefs may recover periodi- cally over short time scales, the extent to which this is sustainable for each reef complex is presently unclear. 787 788 GORFINE ET AL. Central Zone (a) Weslern Zone Warrnambooi (b) Central Zone Figure 1. Fishing grounds (stippled) and key reef areas (hatched) in the (a) Western Zone, (h) Central Zone and (c) Eastern Zone, that were included in on-board observations of the \ ictorian abalone fishery. 3. Self-imposed daily quotas. For example, in the Eastern Zone of the Victorian fishery, the Abalone Fishermen's Coopera- tive, which includes about 21 out of the 23 fishery access licenses, has a self-imposed daily limit of twelve bins (ap- proximately 480 kg live weight) per diver. Similar forms of self-governance have been evident among Victorian divers generally during the formative first decade of the fishery (Sanders & Beinssen 1972). Eastern Zone Western Zone 1969 1972 1975 1978 1981 1984 1987 19' Year 1993 1996 1999 Figure 2. Changes in diver experience (mean ± SE; broken line) and the proportion of fishery access licences where the diver was not the license owner (solid line), in each of the three management zones dur- ing 1969-20(10. The latter two points describe an effective second tier of finer- scale self-management below the controls imposed by fisheries regulations. It is our belief that maintenance of this voluntary regime is both critical to the future sustainability of the resource and vulnerable to changes in industry structure that have a pro- pensity to diminish stewardship. With quotas comes the need to establish data collection pro- grams that will support modeling and assessment of stocks for future TAC adjustment (Ramade-Villanueva et al. 1998). Diver- repoiled catch and effort statistics will generally be perceived as inadequate for this purpose (McShane 1992, although see Andrew et al. 1997). Like the limitations of quota management, this is partly a consequence of geographic scale that renders catch per unit effort data hyperstable to serial stock depletion (Breen 1992), although the abundance of dense aggregations of abalone also plays a role (Hart et al. 1997. Prince 1992). Inevitably, fishery- independent survey methods will be considered to provide the necessary data (McShane 1994). However, government-led inde- pendent surveys can be an expensive approach beset with the problems identified by Prince et al. (1998) about shrinking public sector resources in the face of the need to cover extensive fishing grounds. In contrast with fishery-independent survey data, opportunities Diver Behaviolir and Assessment of Abalone Fishery 789 to sample dependent data are more numerous, and their acquisition is cheaper (Kesteven 1997). This led us to speculate as to whether some of the impediments to using catch-effort data could be over- come if we considered divers as an integral part of the fishing gear. In trawl fisheries, for example, much data is collected about ves- .sels. their fishing power, and tleet dynamics. Fisheries observer programs in southeastern Australia are well established as methods for collecting detailed information about trawl fishing fleets and their catches. Although the fishing power of divers has been stud- ied in Australia (Beinssen 1979. Hart & Gorfine 1997). this has not been extended to monitoring programs involving fisheries observ- ers. During late 1998. we initiated an on-board observer program directed at gaining a detailed understanding of the fleet dynamics, fishing effort, and contemporary fishing grounds of the Victorian abalone fishery (Gorfine et al. this publication). Although the quantity of data acquired to date is relatively small, these data provide information unavailable from catch and effort reporting. In this paper, we examine some descriptive statistics from our ob- server program and the insights they reveal about diver behavior. We draw comparisons with reported catch and effort statistics and then consider the implications of these behaviors for fishery man- agement and stock assessment. MATERIALS AND METHODS Casual fisheries observers were employed at several major aba- lone-fishing ports along the Victorian coast. Each observer had contact information for licensed abalone divers in their vicinity who had agreed to participate in the program. The observers were instructed to spread their observations among as many divers and fishing locations as possible, because random sampling of the abalone fleet was impractical. Where possible, observers followed the catch through to processing to enable each day's obsei^ations to be linked to catch and effort information reported by divers and processors. A more extensive commercial catch-sampling program also operated in parallel with the observer program. Observations made during each day aboard a particular diver's vessel comprised three levels of infonnation: 1. An overview of the day's fishing, which included intended diving location, reason for its selection, catch expectation, type of abalone sought and reason why, details about the size, power, and hull style of the vessel, and weather and sea conditions; 2. GPS coordinates and divers" observations of the underwater habitat and environmental conditions, qualitative abundance of competitors and predators, and condition of abalone population as compared with previous visits at each dive location; and 3. Quantitative estimates of catch weight, number in the catch, and effort for each catch bag of abalone, qualitative esti- mates of the amount of shell epibiota and proportions of "tiger" variants and greenlip abalone, and the dive profile of the diver (depth recorded every 3 min by dive computer). Where possible, length frequency samples were taken from the day's catch after landing. However, to date length- frequency data have been collected too sporadically to ana- lyze. The three levels of information were entered into a relational database (M.S Access"), and summary statistics were produced. RESULTS Summary Six observers collected data from eighty-five dives over sixty- seven diver days. Observed catches totaled 32.3 tons live weight har\ested by expending 215.6 h diver effort. Vessels Vessel length ranged from 5.6 to 8.5 m and engine power from 90—450 horsepower. Catamaran hulls were the most popular in the Western and Eastern Zones at 63 and 87%, respectively; however. Central Zone divers favored monohulls slightly, with only 40% catamarans. Diving Practices The maximum depth recorded out of 3, 1 1 2 observations was 27 m, and the minimum was 1 m, with an average of 9 m and 95% confidence interval of 8.8-9.2 m. Percentiles of the distribution of depths show that 80% of all harvesting occun'ed shallower than 1 3 m or 43 ft (Table 1 ), Diving practices ranged from lower to higher risk as demon- strated by recorded depth profiles (Fig. 3a). High-risk profiles involving progression to deeper depths later during the day were not necessarily associated with improved catch rates (Fig. 3b). Multiple ascents were a feature of most dive profiles recorded. Site Selection Of those occasions where divers gave specific reasons for site selection 50% related to weather. 38% to high catch expectation, 15% to diving safety, 12% to market requirements, and 4% to exploration of unfamiliar territory. Note that there were no con- straints on the variety of reasons that could be recorded and that totals exceed 100% because of more than one reason given on some occasions. Only one or two reefs were visited daily on 91% of the days observed during this study (Table 2). Although movement between reefs during a day's fishing was infrequent, divers often relocated within a reef complex, depending on their personal diving prac- tices. TABLE L Quantiles of the distributions of diving depths and catch rates observed aboard commercial divers' vessels in the Victorian abalone fishery. Percentile 5 10 20 30 40 50 60 70 80 90 100 Deptli (m) Catch rate (kgh ') 65 ?, 19 4 106 5 120 7 8 I.SO 10 165 11 183 13 204 16 240 27 386 790 GORFINE ET AL. (a) Dissimilar mean catch rates a 0) Q Time (minutes) 0 30 60 90 120 150 180 210 240 270 12 '^-Mean daily CPUE = 101 kg/h Mean daily CPUE = 226 kg/h 16 274 300 225 220 182 200 (b) Similar mean catch rates 30 60 Time (minutes) 90 120 150 180 210 240 270 300 Mean daily CPUE = 173 kg/h Mean daily CPUE = 181 kg/h ^■^ — ^« — > -^ — ^-e- 24 168 135 107 86 184 210 219 227 288 280 Figure 3. al Different dive profiles yielding dissimilur mean dally catch rates (CPUE) for two divers in the same management zone of the Victorian abalone fishery, b) Different dive profiles yielding similar daily catch rates (CPUE) for two divers in the same management Z(me of the Victorian ahulone fishery. (Broken line depicts low-risk dive profile and solid line depicts high-risk profile. Catch rates (kg/h) for each interval between surfacing are shown for each profile). Harvesting Objectives Divers reported that their principle objective tor the day's fish- ing was large abalone on 28%, small abalone on 4%, and live abalone on 6% of occasions. Tigers; that is, a blacklip variant exhibiting a striped epidodiiim, were the target on 19% of occa- sions, usually in conjunction with small- or medium-sized abalone with the usual black epipodium. Catch Expectation The 95% confidence interval for daily catch expectation was 430-490 kg. The highest catch expectation was about 900 kg. and the lowest was 60 ks. TABLE 2. Frequency of number of reefs visited daily during on-board observations of the Victorian-abalone fishery. Number of Reefs Visited 1 2 3 4 5 Propdrtum of days (%) 55 36 .5 2 ^ DivHR Bhhaviour and Assessment of Abalone Fishery 791 900 800 ^ 700 600 500 200 100 ■ 200 300 400 500 Expected catch (kg) 600 800 Fissure 4. Relationship between expected and observed daily catches from observations on-board commercial divers' vessels in the Victorian abalone llsherv. Regression of observed catches on daily catch expectations showed that catches increased linearly with expectation (Fig. 4) and tended to exceed expected weights by 24% (df = 1 , 4 1 . r" = 0.69, t = 40.29, P < 0.0001). Comparison of expected, observed, and reported catches for a sample of the observations shows that reported catches were 8 ± 4% (mean ± SE) less than observed and expected catches were 15 ± 9% less than observed (Table 3). Only a small satiiple (/( = 7) was available because of difficulties in matching obser\ations with reported catches. Catch Rates Reef-specific catch rates ranged from 0-300 kgh"' with 95% of values between 134-156 kgh"'. Net bag catch rates ranged from 0-386 kgh*' with 95% of values between 1.50-161 kgh"' (Table 4). Reported effort values used to calculate catch per unit effort were 1 7 ± 9% higher than those we observed for the same catches (Table 3). Calculation of catch per unit effort (CPUE) using the mostly lower reported catch and lower reported effort gave values 12-29% (95% CL) less than the actual rates. Althouah it is difficult to identifv a threshold catch rate for the fishery explicitly, it is informative to consider that 80% of catch rates were greater than 106 kgh"' (Table 1). The relatively small number of values below 50 kgh"' (five observations) was for efforts of less than 20 min. Generally, there was no relationship between catch rates and duration of effort. However, unfavorable weather and low abundance accounted for the relatively few in- stances when catch rates were low. The former involved an ex- New South Wales diver new to the fishery, and the latter was associated with efforts of short duration. Maximum catch rates were associated with efforts of about 0.5 to 3.5 h, and longer efforts tended to produce slightly lower catch rates. DISCUSSION This modest study reinforces perceptions that commercial divers in the Victorian abalone fishery have developed fishing behaviors that maintain high catch rates enabling them to predict the quantity of abalone they expect to harvest from a specific area on a particular day reliably. To some extent, these patterns in behavior provide a degree of resource protection abo\e that af- forded by management regulations, because they tend to ensure that fishing mortality is generally low (McShane & Smith 1989). TABLE 3. Sample of divers" expected catches, estimated v>eights on landing and observed daily catch, effort, and CPUE, matched with reported catch, effort, and CPUE for the same observations. Catch (kg) Effort ( min) CPUE (kg/h"') Expected Divers" Estimate Observed Reported Observed Reported Observed Reported — 550 567 539 293 480 116 67 400 350 474 344 121 120 235 172 300 230 273 232 219 240 75 58 350 400 470 414 160 150 176 166 350 300 378 36S 179 210 127 105 600 400 517 476 166 180 187 159 350 450 527 5S5 170 220 186 160 792 GORFINE ET AL. TABLE 4. Comparison of catch rate statistics at different scales of sampling during on-board observations of the Victorian abalone fishery. Scale of Sampling By Catch Bag Statistic Dail V (kgh-') By Reef 1 kgh') (kgh-'i Mean 144 156 156 Maximum 234 300 386 Minimum 7 0 0 Upper 93% CL 138 156 161 Lower 9'i"r CL 131 134 1 30 Not surprisingly, the main harvesting objective of divers during this study was to maximi/,e catch rates. However, on some occa- sions, abalone with striped epidodia (tigers) and smaller abalone were sought as the prime objective because of market pi'eference. Smaller abalone. particularly fi-om Port Phillip Bay. where the legal minijiium length (LML) is 100 mm. are preferred for can- ning, because cans containing four or five whole pieces attract the best prices. The targeting of these specific morphotypes did not compromise divers' abilities to achieve average daily catches. Al- though daily catch expectations were mostly high, low catch ex- pectations occurred when divers only required small quantities to obtain the balance of their quota allocations. Divers' catches ex- ceeded their expectations in most instances. This is likely to have occurred because the divers base their catch expectations on weights used to deoement quota allocations. Consequently, we suspect that the divers have made allowances for losses attribut- able to the re)T)oval of epibiota and drainage during postharvest transport and storage (Gorfine this publication). The faster catch rates observed at a within-reef spatial resolu- tion highlight the loss of accuracy in effort associated v\ ith aggre- gating values at the reef scale. These inaccuracies are further a]n- plified when catch rates are reported as total effort for the entiie time spent at sea. When combined with post-harvest weight losses caused by retnoval of epibiota and d)ainage, it becomes evident why CPUE values extracted from Fisheries Victoria's catch and effort reporting system are often substantially lower than those measured directly during the course of this study. Inspection of fisheries observers' comments noted on datasheets shows that the highest catch rates were for locations with very large abalone, each about 600-650 g live weight, that had large amounts of such shell epibiota as barnacles and sponges accounting for up to 30% of their weight. Allowing for retiioval of epibiota. the highest catch rate of 386 kgh"' may have reduced to 270 kgh~'. The amount and composition of epibiota getierally reflects the particular habitat occupied by the abalone. with cun- jevoi adding the most weight, and divers inform us that they be- lieve abalone with clean shells are usually those that have recently emerged fiom cryptic habitat. Previous studies we have undertaken at a limited nurnber of sites have not supported this hypothesis (Officer et al. this publication). Weather played a dotiiinant role in divers' decisions about where to fish. Generally, most Victorian divers do not head to sea if weather conditions ai'e likely to prevent them froin satisfying their catch expectations. This is reflected in some of our fisheries observers' experiences of aborted trips despite pi'eplanning and weather forecasting. Examination of the relationship between ef- fort and catch rate showed that a small nu)nber of low catch rates for huge effort expenditure related to weather, lack of experience, or a cotnbination of both these factors. Indeed, one of the divers had only recently commenced in the Victorian fishery after having fished in NSW where tnaxi)num catch rates are tvpically about 25 kgh"' (Worthington et al. 1999). Although an experienced and competent diver, he may have been prepared to fish at lower catch rates than most Victorian divers would tolerate, because the 70 kgh"' he achieved was more than double the late he would have expei'ienced in the NSW fisherv. The vessels used by Victorian divers reflect a co)nbination of geography and economics. In some instances, individual divers may own different vessels to access different pajis of the coast. Monohulls are preferred where on-road towing distances are long and at-sea distances short and where beach launching across sand is required. This applies more to Central Zone divers than to those from the other /ones of the fishery. The lower operating costs of )nonohulls as compared to catatnaran hulls also jiiake thon attrac- tive to contract divers. Twin hulls are preferred where bar cross- ings are made and where long distances on the water are necessary in sometimes fickle weather conditions, such as in the Eastern and Western Zones. As more contract divers participate in the fishery, we may see an increase in the number of monohulls. To some extent, this may limit the intensity of fishing at remote locations and concentiate effort in relatively safer areas closer to ports of access. Despite the consistency in high catch rates ajiiong divers, there is considerable variability in diving practices. A sjiiall number of divers spend a large proportion of their time continuously under- water, preferring to send their catch bags of abalone to the surface using inflated parachute-style lift bags. These divers ascend infre- quoitly; whereas, others make frequent multiple ascents, some- times from relatively deep depths. Sotne divers use large catch bags with a capacity of about 150 kg; whereas, others use small bags that hold about 50 kg. Time spent transporting the catch across the substrate and to the surface is counted as effort, and the size of the catch bag could reasonably be expected to influence catch rates. Two divers reported that they only dived to shallow depths. One diver never exceeds 10 m; whereas, the other, who suffered two episodes of decompression illness (DCI) during 1998, does not exceed 12 m, in accordance with medical advice. Although these divers believe their practices are conservative, examination of their profiles shows an ignorance of the effects of multiple ascents in precipitating DCI (Fig. 3a). The trade-off for these "con.servative" profiles was reduced catch rates in some instances but not in oth- ers, so it does not necessarily follow that diving to deeper depths will yield better catches. In contrast, another diver showed a will- ingness to defy accepted wisdom by diving deeper on successive dives in an attempt to maximize his catch rates (Fig. 3b). Ironi- cally, it was during a shallow '■deco)npi"ession" dive toward the end of the day that he achieved his best results. Although diving depth may be unrelated to the ability to satisfy catch expectations, at present it may becotne a factor if stocks in shallower locations decline because of increased intensity of effort. The relatively shallow depths of most of the diving observed dur- ing this study reflect a tendency for younger contract divers to adopt safer practices and concentrate their effort on shallower reefs. In recent instances where localized stock depletion has oc- Diver Behaviol'r and Assessment of Abalone Fishery 793 curred (Gorfine et al. this publication), divers have reloeated effort previously expended on the depleted reefs to other regularly fished reefs in their zone. There is no evidence that divers have shifted their effort offshore to those deeper or more remote reefs that have been infrequently fished since quotas were introduced. This study has illustrated some aspects of diver behavior that may have substantial influence on the state of Victorian abalone resources and the long-term sustainability of the fishery. One of the key observations was the high, but variable, catch rate thresh- olds that prompted divers to shift their effort within a reef com- plex. Net reductions in these thresholds o\er time will signal over- fishing of the most productive reefs. It is likely that such changes will lead to an over-all decrease in the performance of the fishery, because underutilized reefs where effort can be redeployed are mostly those with historically low catch rates. Departures from established such practices as self-imposed daily quotas and rotational harvesting will threaten abalone stocks regardless of current zonal quotas and regional size limits. This is not merely speculative, given that the fishery is progressively shift- ing to a new generation of contract divers (Fig. 2) motivated by different values, past experiences, and expectations that will be reflected in their fishing behaviors. Although many contract divers seem to be genuinely interested in the on-going viability of the fishery, they do not necessarily have the same incentive as quota- owning divers to ensure that their fishing practices are compatible with long-term sustainability (Prince & Shepherd 1992, Parliament of Victoria 2000J. Victorian abalone divers are understandably proud of their history of resource stewardship and the altruism they displayed in reducing their catches in the shift to quota-based management twelve years ago. In a brief discussion on the evolu- tion of altruism, noted ecologist Simon Levin (1999) points out that altruistic behavior is the manifestation of "enlightened self- interest" and is strongest when there are expectations of rapid payoffs to individuals or coalitions sharing common aspirations. Obviously, the self-interests of divers who are either quota owners or close relatives likely to inherit some proportion of quota en- titlement will differ from those without equity in quota. The pro- portion of the beach price (usually less than lOVr) received by contract divers from the owners of fishery access licenses is suf- ficiently small to ensure that the contract divers are likely to be more concerned with reducing the costs of fishing effort to maxi- mize their incomes. In an assessment of the Tasmanian abalone fishery, where there is clear separation between quota ownership and the licensing of divers. Officer ( 1999) expressed similar con- cerns that contract divers reluctant to incur the costs of travel to remote locations may be prepared to endure lower catch rates than owner-divers on more accessible reefs. Currently, those charged with developing a management plan for the Victorian fishery are proposing changes toward a system similar to the one adopted in Tasmania. These changes will facili- tate further separation between quota ownership and fishery access licenses that confer the rieht to dive or enaace a contractor to dive for abalone. Whether additional management arrangements, such as small-scale subzonal quotas, are required to modify potential changes in diver behavior to be within acceptable limits is unclear. At this stage, there is no evidence that the over-all status of Victorian abalone stocks has been compromised by changes in the licensed catching sector. However, there is clear evidence that the fishing grounds have contracted substantially since quota introduc- tion and that the fishery is now dependent on a limited number of relatively shallow and highly productive reefs (Gorfine et al. this publication). There is no doubt that this makes the Victorian aba- lone fishery more vulnerable to the impact of serial depletion and threatens its sustainability. We believe it would be preferable for industry to remain self-regulating at finer spatial scales, but this view is contingent on the Victorian abalone industry's ability to maintain the resource stewardship characteristic of the retiring generation of owner-divers. The concept of a Territorial Users" Rights Fishery (TURF) espoused by Prince et al. (1998), with potential benefits of fine-scale monitoring and industry self- governance, could be an effective framework for such stewardship. However, the public acceptance of this kind of scheme would be improbable for a resource that is the common property of the whole community of Victoria. As fisheries management agencies in the Northern Hemisphere look toward quota management as a means of imprt)\ ing the sus- tainability of their abalone fisheries they should be cognizant of the limitations of this approach, particularly for a fishery in need of restoration. Although quotas may prove to be the best option, how they are applied as part of a suite of management arrange- ments will be critical for their effectiveness. Abalone fisheries managers must be mindful that regulations affect diver behavior and that principally it is the divers, not the resource, that is man- aged. We believe that this study highlights the importance of iden- tifying and promoting diver behaviors that are compatible with resource sustainability within the context of contemporary man- agement strategies. Such fisheries observer programs as ours can provide valuable infonnation to managers for achieving this ob- jective. ACKNOWLEDGMENTS This study would not have been possible without the willing- ness of commercial abalone divers to have fisheries observers aboard their vessels. This level of cooperation between abalone divers and fisheries biologists has been a hallmark of the Victorian fishery for many years, for which we owe a debt of gratitude to both industry and our predecessors. We thank our fisheries ob- servers who are required to go to sea at short notice, sometimes in adverse weather that inakes conditions aboard small vessels un- pleasant for most people. Sonia Talman is thanked for efficiently entering the data into the relational database expertly erected by Masaaki Machida. The manuscript was greatly improved as a re- sult of the constructive criticisms of an erudite anonymous re- viewer. LITERATURE CITED Andrew. N. L.. D. G. Worthington & P. A. Brett. 1997. Size-structure and growtli of individuals suggest high exploitation rates in the fishery for blacklip abalone. Hatiotis nibni. in New South Wales. Australia. Moll. Rei. 18:275-287. Beinssen, K. 1979. Fishing power of divers in the abalone fishery of Victoria. Australia. Rapp. P.-v. Reiin. Com. Int. Explm: Mer. 175:20-22. Breen, P. A. 1992. A review of models used for stock assessment in abalone fisheries. In: S. A. Shepherd, M. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: biology, fisheries, and culture. London: Blackwell. pp. 253-275. CDFG. 1997. Draft fisheries recovery/managemeni plan for California aba- lones. draft envir. doc. app. I. 794 GORFINE ET AL. Fisheries Victoria. 1998. Catch and effort information bulletin 1998. 5th ed. Queenscliff, Victoria: Marine and Freshwater Resources Institute. 31 pp. Gorfme, H. K. 2001. Post-harvest weight loss has important implications for abalone quota management. / Shellfish Res. 20:795-802. Gorfine. H. K. & C. D. Dixon, editors. 2000. Abalone— 1999. Compiled by the Abalone Fishery Assessment Group. Fisheries Victoria Assessment Rept. 27. Queenscliff. Victoria: Marine and Freshwater Resources In- stitute. Gorfine. H. K.. B. L. Taylor & T. I. Walker. Triggers and targets: what are we aiming for with abalone fisheries inodels? J. Shellfish Res. 20:80.V 811. Han. A. M.. H. K. Gorfine & M. C. Callan. 1997. Abundance estimation of blacklip abalone [Haliolis rubra) I. An analysis of diver-survey methods for large-scale monitoring. Fish. Res. 29:159-169. Hart. A. M. & H. K. Gorfine. 1997. Abundance estimation of blacklip abalone (Hiilknis rubra) 11. A comparative evaluation of catch-effort, change-in-ratio, mark-recapture, and diver survey methods. Fish. Res. 29:159-183. Kesteven. G. L. 1997. Fishery-independent surveys and data. Fish. Res. 32: 271-272. Levin, S. A. 1999. Fragile dominion — complexity and the commons. Read- ing, MA: Perseus. 250 pp. McShane. P. E. 1992. Exploitation models and catch statistics of the Vic- torian fishery for abalone Haliotis rubra. Fish. Bull 90:139-146. McShane. P. E. 1994. Estimating the abundance of abalone (Haliuiis spp.) stocks: examples from Victoria and southern New Zealand. Fish. Res 19:379-394. McShane. P. E. & M.G. Smith. 1989. Direct measurement of fishing mor- tality in abalone {Haliuiis rubra Leach I off southeastern Australia. Fish. Res. 8:93-102. Officer, R. A. 1999. Fishery Assessment Report: Tasmanian Abalone Fish- ery 1997/98. TAFl Marine Research Laboratories, University of Tas- mania, 68 pp. Officer. R. A., C. D. Di,\on & H. K. Gorfine. Movement and re-aggregation of the blacklip abalone Haliotis rubra Leach, after fishing. J. Shellfish Res. 20:771-779. Parliament of Victoria. Environment and Natural Resources Committee. 2000. Inquiry into Fisheries Management. Discussion Paper. Peiry, R. I., C. J. Walters, & J. A. Boutillier. 1999. A framework for providing scientific advice for the management of new and developing invertebrate fisheries. Rev. Fish Biol. & Fisher. 9:125-150. Prince, J. D. 1992. Using a spatial model to explore the dynamics of an exploited stock of the abalone Haliotis rubra. In: S. A. Shepherd, M. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: biology, fisheries, and culture. London: Blackwell. pp. 305-317. Prince, J. D. & S. A. Shepherd. 1992. Australian abalone fisheries and their management. In: S. A. Shepherd, M.Tegner & S. A. Gu/.man del Proo, editors. Abalone of the world: biology, fisheries, and culture. London: Blackwell. pp. 407-426. Prince. J, D., C. Walters, R. Ruiz-Avila & P. Sluczanowski. 1998. Terri- torial user's rights and the Australian abalone {Haliotis sp.) fishery. In: G. S. Jamieson & A. Campbell, editors. Proceedings of the North Pacific symposium on invertebrate stock assessment and management. Can. Spec. Piibl. Fish. Aqiiat. Sci. 125:367-375. Ramade-Villanueva, M. R., D. B. Lluch-Cota, S. E. Lluch-Cota, S. Her- nandez-Vazquez, A. Espinoza-Montes & A. Vega- Velazquez. 1998. An evaluation of the annual quota mechanism as a management tool in the Mexican abalone fishery. ./. Shellfish Res. 17:847-851. Sanders. M. J. & K. H. Beinssen. 1972. Licence limitation in the Victorian abalone industry. Austral. Fisheries. 31(3):25. Shepherd. S. A. & J. L. Baker. 1998. Biological reference points in an abalone fishery. In: G. S, Jamieson & A. Campbell, editors. Proceed- ings of the North Pacific symposium on invertebrate stock assessment and management. Can. Spec. Publ. Fish. Aquat. Sci. 125:235-246. Worthington, D. G., C. Blount, R. C. Chick & P. T. Gibson. 1999. An assessment of the NSW fishery in 1998. NSW Fisheries Research Institute. Fishery Resource Assessment Series No. 6:55. Jimnuil itf Shellfish Research. Vol. 20, No. 2, 795-802, 2001. POST-HARVEST WEIGHT LOSS HAS IMPORTANT IMPLICATIONS FOR ABALONE QUOTA MANAGEMENT HARRY K. GORFINE* Marine and Frcshwalcr Resources Insriluie. Qiieensvliff. Victoria. Australia ABSTRACT Regulations lorminj; part of the abalonc quota management system in Victoria. Australia require that abalone processors notify fisheries authorities of quantities of abalone consigned from commercial divers within 25 h of landing. The regulations also require that the abalone are lo be landed whole in the shell, and transported and stored in sealed bins until confirmatron of official notification. Although the bins cannot be opened before notification, the 25-h limit for notification provides a window of opportunity to make potential gains through weight loss in those abalone destined for canning. This arises because notified weights, rather than weights estimated upon landing, are used to decrement the quota allocations of individual divers, I investigated the potential for post-harvest weight loss by subjecting abalone to three experimental treatinents selected to simulate a range of possible transport and storage regimes. My results demonstrate that substantial weight loss can occur in whole abalone during both ambient and refrigerated storage. Loss of weight in abalone during storage results from the release of water and body fluid associated with physiological responses to hypoxic stress caused by exposure to air. In Victoria, about 60% of the landed catch is used to produce canned product. Because of the observed weight losses during storage, divers supplying abalone for canning have to harvest more abalone to achieve their quotas than those supplying abalone for live export. Losses in weight of 10-20% observed during this study equate to a harvest of 6-12% (350,000-700.000 abalone) more than the total allowable catches. Harvesting these additional quantities of abalone may exceed desired fishing mortalities for long-term population sustainability. If beach weights were deducted from quotas, then this problem could be resolved. KEY WORDS: quotas, management, weight loss, air exposure, abalone INTRODUCTION The Victorian fishery for blacklip abalone (Huliotis rubra Leach) has operated under quota management since 1988, Total allowable catches (TACs) were initially based on average annual catches during the preceding 5 y, A quota docketing system was implemented concurrently to track divers' progress in attaining their annual TACs, It became apparent during the early 1990s that some processors were taking advantage of regulations that permit- ted a transportation period of up to I day post-harvest before the consignment of catches and completion of quota dockets. Potential fluid loss during this period enabled some processors to pay less for the same nutnber of abalone. Correspondence from this period shows that some of the larger processors and diver associatiiins sought to have this loophole in the regulations removed. During 1994 additional regulations were proposed to ensure that, immediately upon landing, abalone would be placed in bins sealed with nonreusable identification tags and weighed at the beach. The principal objective was to eliminate the potential for overquota harvesting through understatement of estimated catches on quota dockets. Sealed bins prevent the removal of abalone between the point of landing and delivery to registered abalone processing or storage preinises. Beach weighing ensures that pro- cessor weights can be reconciled with landed weights. New legu- lations requiring the sealing of bins and establishing a comprehen- sive audit trail were introduced during 1996, However, attempts to introduce beach weighing for decrementing quota allocations were unsuccessful. This was largely because processors close to points ot landing argued thai it was an unnecessary imposition when their factories were only about 10 min away. Ironically, processors dis- tant from points of landing, and the divers supplying these pro- cessors, lobbied for sufficient time before notification lo allow abalone harvested at one end of the state to be transported to a processor at the other end (a distance of almost 900 km). Such *E-mail address: Harry, GorfinetS'nre. vie. gov.au arguments by the abalone industry tend to be successful because of the Australian federal government's National Competition Policy legislation, which, in seeking to remove bamers to business and trade, requires any commercially restrictive regulations to be jus- tified on resource protection grounds. The 1996 regulations, forming part of the Abalone Quota Man- agement System, require abalone processors to notify Fisheries Victoria of quantities of abalone (as weights to the nearest 0, 1 kg) consigned from commercial divers within 25 h of landing. This is done via electronic facsimile to an automated notification service that then provides confirmation by return transmission within I.S min of receipt. Processors must not break the seals on the bins until this confirmation is received. The particular diver's quota is then decremented accordingly. The cuiTcnt practice among many aba- lone processors is to take possession of abalone during the after- noon or evening on the day of landing, store them in a cool room overnight, and not effect consignment and notification until early the following morning before processing. This means the abalone may be emersed for 1 2-25 h before weighing. Despite the paucity of literature on the effects of emersion on haliotids. there is a substantial body of work on intertidal gastro- pods, including prosobranchs. Many intertidal prosobranchs have adapted to survive extensive periods of hypoxic stress and desic- cation associated with exposure to air. as well as temperatures beyond the range experienced during immersion. Physical adaptive strategies include a relatively large water storage capacity within the mantle cavity (Hyinan 1967). a shell morphology that provides a large volume to surface area ratio (Newell 1979). capacity to key the irregularities of the shell margin to fit the perimeter of the homesite occupied closely (Hyman 1967). closure of the shell aperture with opercula (Newell 1979). and vertical orientation (Hyman 1967), Physiological strategies include enhanced osmo- regulatory ability (Gardiner 1972). excretion of uric acid rather than ammonia compounds to conserve water (Hickman 1973), uptake of seawater into the gland cavities of the foot, remaining motionless while exposed, and acclimation by lowering the meta- 795 796 GORFINE bolic rate to reduce oxygen consumption (Hyman 1967). Among all these strategies, water-holding capacity is the most important, because it prevents desiccation of the gill tissues, provides limited oxygen for metabolic activity, reduces thermal stress through evaporative heat loss, and dilutes otherwise lethal concentrations of solutes in body fluids (Newell 1979). Although they are mostly subtidal. air-exposed blacklip aba- lone have been observed under vertical overhangs inmiediately above the low-water mark and may possess some of the adapta- tions of predominantly intertidal species from the same order. Other species of abalone are more frequently exposed to air in the lower intertidal. For example, the tropical species Haliotis asinina (Linne) forages atop coral heads exposed at low tide (Donovan et al. 1999), and some populations of Haliotis nifescens (Swainson) may be exposed to air for up to 1 h during normal tidal cycles (Tjeerdema et al. 1991). This may explain why blacklip abalone are able to retain water after removal from the sea; furthermore, it may explain the pattern of progressive water loss while these aba- lone attempt to survive extended periods of air exposure. Survival time is likely to affect the rate of water and body fluid loss, and abalone biochemistry shows some adaptation in glycolytic path- ways that promotes survival of hypoxia during air exposure (Donovan et al. 1999. Wells & Baldwin 1995). The present study aims to measure the weight loss associated with up to 25 h emersion by simulating conditions typically en- countered at abalone processing facilities. I was originally re- quested to investigate this potential for weight loss during storage in 1994, because some sectors of industry had indicated they would seek compensation if beach weighing was introduced. My initial research demonstrated that industry claims regarding po.st- harvest weight loss could be substantiated. However, recent de- velopment of a fishery model, aimed at establishing the risks to sustainability of different TACs, created the need for additional research to confirm the initial results and enable more accurate catches to be input to the model. MATERIALS AND METHODS Two experiments were conducted to investigate post-harvest weight loss in air-exposed blacklip abalone (H. iiihra) during stor- age. The first experiment measured losses through drainage from commercially used bins of abalone. and the second experiment measured fiuid loss from individually contained abalone. Experiment I: Weight Loss from Commercial Bins of Abalone In warm to hot weather between 0930 h and 1.^30 h during December 1994. research divers harvested 300 kg whole weight of abalone from Bushranger's Bay near Cape Schanck. Victoria (Fig. 1 ). The majority of these abalone were marginally larger than the legal minimum length of 110 mm and did not seem to have re- cently spawned. The abalone were packed into prelabeled Nallys* No. 12 fish bins in a primary layer consisting of three longitudinal rows of abalone oriented vertically and a secondary layer placed in a horizontal plane. The total weight for each bin was about 20 kg. The bases of the bins were perforated to allow for the drainage of water and abalone body fluids. Abalone harvested at different times were equally distributed among experimental treatments, and all abalone were landed at 1445 h when initial weights were mea- sured to the nearest 0.05 kg. Each bin was re weighed every half- hour for a period of 5 h. and final weights were determined after an additional 12 h. 150° E 38° S N • Cape Schanck Bass Strait Figure 1. Locations of blucklip abalone {H. rubra Leach) collection sites along the coast of Victoria, Australia. Three experimental treatments were applied to sets of five rep- licate bins of abalone: 1. coi'l — involved refrigeration at 6°C immediately post land- ing: 2. ambient — required the abalone to remain uncovered at am- bient temperature (24.5°C @ 1445 h to 22°C @ 2000 h) for 5 h before refrigeration at 6°C for a further 12 h; and 3. moist — the abalone were covered with moist cloth towels at ambient temperature (24.5X @ 1445 h to 22°C @ 2000 h) for 5 h before refrigeration at 6°C for a further 1 2 h. Treatments 2 and 3 were conducted under covered walkways. so the bins of abalone were not exposed to direct sunlight. The air temperature at 1200 h in direct sunlight was close to 30°C. Rela- tive humidity was not measured. Experiment 2: Weight Loss from Individual Abalone This experiment was conducted on a warm day (26°C at 1200 h in direct sunlight) during November 1999. Two storage environ- ments were used, an air-conditioned laboratory in which tempera- ture was maintained in the low 20s (°C) and a cool room with the thermostat set at 3°C. Abalone ranging in shell length from 83-120 mm were collected from Point Cook in Port Phillip Bay (Fig. 1 ) and held in one ton capacity, flow-through aquaria for several The Implications of Post-Harvest Weight Loss 797 months before this experiment. Abalone were removed from the aquaria, blotted on absorbent towels to remove excess external seavvater. individually placed in preweighed 1.0-L plastic contain- ers, and weighed to the nearest 0. 1 g. Orientation of abalone in the containers was. where possible, in a vertical plane with the leading shell margin facing downward. Each container was fitted with a plastic grill to allow water to drain away from the abalone. Forty- five abalone were randomly assigned to each of the following three storage treatments: 1. cool — involved refrigeration at .^ C immediately after weighing: 2. aiucool — required the abalone to remain al ambient tem- perature (24°C @ 1400 h) for 4 h before refrigeration (3'C @ 2000 h) for a further 20 h: and 3. ambient — the abalone remained at ambient temperature (24"C @ 1400 h to 20°C @ 2000 h) for 24 h. Abalone were reweighed in their containers after 1. 2. 4. 6. S. 10, and 24 h. At each weighing, the combined weight of the container, abalone. and drained abalone fluid was measured. Then the weight of the container and fluid w ilh the abalone lifted above. but not out of, the container was measured. As the abalone were weighed, observations were made regarding the degree of pedal adhesion and responsiveness to stimulation. Although relative humidity was not measured directly, evapo- rative loss of water was used as a de facto alternative. To test tor evaporative loss of water 10 1.0-L plastic containers tilled with about 350-mL tapwater were placed in each of the two storage environments and reweighed at the same time intervals as the abalone, and additionally at 5 h. Potential for fluid loss was investigated by measuring the vol- ume and weight of fluid that could be drained by gentle squeezing of the soft tissues of fifty abalone selected at random from the same aquarium. Statistical Analysis Weight loss in the first experiment was determined from the difference between the combined weight of the bin and abalone at f, and the initial weight at /„. In the second experiment, weight loss caused by evaporation and handling was calculated as the differ- ence between the weight of each abalone in its container at tiine ;, and at the commencement of the experiment at /„. The differences between the weight of an abalone in its container at f,, the weight of the container with the abalone suspended above it at r, and the weight of the empty container at /„ were used to determine the weight of fluid lost directly through physiological processes. For both experiments, the weight of fluid lost was then subject to analysis of variance using the ANOVA procedure in the SAS" statistical software package (SAS Institute 1990). Data from Ex- periment 2 were log (x -i- 1 ) transformed to reduced heterogeneity of variances ainong treatment combinations, eliminate a positive linear relationship between treatment means and their variances, and reduce non-normality in their distribution. The ANOVA tested weight loss as a response to the predictor variables of storage exposure (Treatment) and duration of expo- sure (Time). A two-factor ANOVA model with repeated measures on Time was specified. Because, in all instances, the interaction between storage treatment and time was highly significant {P < 0.01) simple investigations of levels of main effects were con- ducted. One-way ANOVA was performed on storage treatments for each level of time and then repeated measures one-way ANOVA on time was applied to each storage treatment. The Scheffe test for multiple comparison of means was used to detect significant differences among levels of each predictor variable. Volume of fluid lost by squeezing was regressed against the weight oi this fluid for 50 additional abalone. RESULTS Experiment I: Weight Loss from Bins of Abalone Significant weight loss (df = 10.120, F = 3.04. P = 0.0001), ranging from 8-149^, occurred through drainage from bins of aba- lone over 17 h. Although there were no significant differences among treatments (df = 2,12, F = 0.07, P = 0.5168), probably because of low statistical power (Table 1 ), continuous storage in the cool room produced the largest loss; whereas, covering with a moist cloth at ambient temperatures produced the least (Fig. 2). Further analysis showed that weight loss increased significantly over time for all three treatments (df F.^ = 24.46, F. = 5.89. P = 0.0001). Multiple com- parison of means (Table 2) showed that weight loss of abalone stored under cool conditions was not significant over the first 5 h but was significantly greater between 5 and 17 h. Weight loss under ainbient conditions differed significantly over time intervals greater than 3 h throughout the experiment. There were significant differences between the weight loss during the first 1 .5 h and losses over the subsequent 15.5 h. The lack of significant contrast in the means is almo.st certainly a result of the low power of the analysis. Minimum significant differences (MSDs) over time were greater than 50% of the grand mean for all treatments (Table 1). The low power was a consequence of the use of only five repli- cates per treatment under circumstances where differential drain- age through holes in the bases of the plastic bins caused substantial within-treatment variation. Experiment 2: Weight Loss from Individual Abalone Fluid loss from individual abalone varied significantly between cool (3''C) and ambient (21°C) storage (df = 2.132. F = 6.16. P = 0.0028) and overtime (df = 6,292, F = 905.17, P = 0.0001). After 24 h, fluid loss from ambient (15% loss) and cool (14% loss) did not differ significantly, but fluid loss from those abalone trans- ferred from ambient to cool conditions (20% loss) was signifi- cantly greater than in the other two treatments (Table 3). The TABLE L Minimum signilkunt differences (MSDl in weight loss from hins of abalone, among treatment means and as a proportion of the grand mean for each time level (a = 0,05). Time (h) Grand Mean MSD MSD/Grand Mean (%) i).5 0.92 1.48 I.O 1.53 2.33 1.5 2.02 2.83 2.0 2.59 2.91 2.5 2.96 3.06 3.0 3.54 3.29 3.5 3.97 3.54 4.0 4.43 3.84 4.5 5.02 4.32 5.0 5.30 4.29 7.0 I0.S3 8.15 161 152 140 112 103 93 89 87 86 81 75 798 GORFINE -Cool Ambient - Moist Figure 2. Mean cumulative weight loss {%) over time for commercial sized bins of blacklip abalone (H. rubra) subjected to three different conditions of storage (error bars are standard errors). largest difference between cool and ambient treatments occurred after 4-6 h, and the greatest difference among all three treatments occurred at 8 h (Fig. 3). Most fluid loss occurred during the first 6-8 h of exposure (Table 4). Abalone with the most weight loss at each time of measurement tended to have the poorest pedal adhe- sion, consequently strong pedal adhesion was least apparent among abalone undergoing cool storage. Differences in evaporative loss of tapwater between cool and ambient temperatures (df = 1.18. F = 8.61./' = 0.0089) and over time (df = 7.126. F = 137.95. P = 0.0001) were significant but small (<2% of treatment means). For each treatment, the losses were relatively small, ranging from \.-i-\.5% (95% confidence) in the cool room and 1.5-1.7% (95% confidence) in the ambient room after 24 h. Evaporation under ambient conditions was ini- tially less than in the cool room during the first 4 h (Table 5). After 4 h had elapsed, the two curves converged when the ambient temperature rose to a maximum for the experiment of 24°C (Fig. 4). From 4-10 h. there was no significant difference between treat- ments and after 24 h, ambient evaporation was slightly greater than in cool room (Table 5a). Evaporative loss occurred significantly throughout the time period, although the difference in loss between 5 and 6 h under ambient conditions was nonsignificant (Table 5b). Evaporative losses of the more viscous abalone fluid from the containers in which the abalone were housed over the same period were somewhat greater than the tapwater evaporation. The 95% confidence ranges for the evaporation of abalone fluid were 2.2- 2.4% for cool, 4.2^.4% for ambient and 3.0-3.3% for abalone transferred from ambient to cool. Potential for loss by squeezing abalone showed fluid contents ranging from less than 1-15%, with 95% of individuals containing between 5-7% of fluid. Regression of volume versus weight of fluid showed close correspondence (/i = 50, R- = 0.92) sufficient to assume 1.0 g = 1.0 niL fluid. DISCUSSION Exposure of prosobranch gastropods to air occurs naturally for those living in the intertidal zone. The greater the height of exis- TABLE 2. Scheffe multiple comparison of differences among mean weight losses from bins of abalone over time for each treatment level (underlining bar ^—^^ identifies means that «ere not significantly different, a = 0.05). Treatment Cool 0.5 OIQ 1.0 04.^ Time (h) 1.5 2.0 1.44 .M 3.0 3.5 }.0'-> 4.0 .Vxs 4.5 5.0 3,96 17.0 13.73 AmhienI 1.17 I »(-! ,74 3.06 }.M) 4,13 4,62 .S.37 5.66 1041 Moist 141 .7.S 4 i:^ 4.71 6,1 1 6,:s S,36 The Implications of Post-Harvest Weight Loss 799 TABLE 3. Simple comparison ot treatment means of lluid loss from individual abalone exposed to air for each time level (a ; (X + 1) transformed). 0.05; data viere log Time Ih) df K P Scheffe Comparison Grand Mean .VISD 1 2. 1 32 1.S3 0.1638 Nonsignificant 1.40 0.22 2 2,132 2.29 0.1048 Nonsiiinificant 1.58 0.23 4 2,132 18.27 0.0001 Amcool s ambient < cool 2.04 0.25 6 2,132 19.11 0.0001 Ambient < cool = amcool 2.55 0.24 8 2,132 41.27 0.0001 Ambient < cool < amcool 2.77 0.20 10 2.132 32.42 0.0001 Ambient < cool < amcool 2.84 0.20 24 2.132 12.14 0.0001 Cool = ambient < amcool 3.05 0. 1 7 tence above the low-water mark, the greater the adaptation to prolonged air e.xposure. Resistance to desiccation is of prime im- portance to these intertidai marine snails, and. in the extreme, some species can survive in their nattiral habitat for months without seawater contact. Loss of body water is the direct cause of death from desiccation and losses within the range of 10-309}- water content have resulted in mortality in several species (Hyman 1967). Abalone are predominantly subtidal and are, therefore, less adapted to survive desiccation, hypoxia, and thermal stress from prolonged air exposure. Although the abalone mantle cavity has a moderate water storage capacity, the ability to retain that water during emersion is limited by having a large shell aperture. My results show substantial loss of water and body fluid through drain- age from the mantle cavity within 24 h of air exposure. Newell (1979) provides an example where water loss overtime at 21°C for the lower shore limpet Patella cochlear ranged between 25-35Vf at 5 h and 35^0% after 50 h. These ranges of water loss are about twice that of the second experiment in my study. Losses of 5-20<7r were sufficient to cause substantial morbiditv and mortality within the first 10 h and almost lOO'/f mortality after 24 h with some notable exceptions. A couple of large abalone held at ambient temperature seemed exceptionally healthy at the con- clusion of the second experiment. Newell (1979) cites a study of six species of intertidai prosobranchs from the Cape Peninsula, South Africa, that showed median mortality ranged from 16-33% among species, depending on the period of emersion and duration of exposure to low relative humidity. Time to 50% mortality ranged from 1-12 days at temperatures between 3.5-20°C with generally shorter survival at the higher temperatures. As stored water is lost from the mantle cavity and body tissues, thermal tolerance will diminish and osmotic concentration of body fluid may tend toward lethal levels (Newell 1979). The larger percentage of losses in the second experiment as compared with the first possibly arose because of drainage losses from bins occurring aboard the research vessel while at sea during the first experiment. These losses were not measured: whereas, all losses during the second experiment were accounted for in the estimates. Another plausible reason for the difference between the two experiments may have been retention of lost fluid in the first experiment that failed to drain through the bases of the bins during the postlanding period. The retained fluid would have been in- cluded in the weights of the bins of abalone. The high initial MSDs relative to the mean weights during the first experiinent possibly reflect within-treatment variability in the amount of water retained in each bin. This variability is likely to have reduced over time, because movement of bins during repeated weighing caused pro- 20 15 S 10 5 - Ambient - Amcool -Cool 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (ii) Figure 3. Mean cumulative fluid loss (%) over time for individual blacklip abalone (//. rubra) subjected to three different storage conditions (error bars are standard errorsi. 800 GORFINE TABLE 4. Simple comparison of differences in fluid loss from individual abalone exposed to air over time for each treatment le\el (a = 0.05; data were los (x + I) transformed). Treatment df Scheffe Comparison Grand Mean MSD Ambient 6,264 261.39 n.OdOl Amcool 6.264 434.39 0.0001 Cool 6.264 264.45 0.0001 111 < 2h < 4h < 6h < Sh = lOh = 24h Ih < 2h < 41i< 6h< 81i = lOh = 24h Ih = 2h< 4h< 6h = 8h = lOh = 24h. & 6 < 24 2 IS (1.16 2.43 0.20 2.34 0.26 gressive loss of retained water. In other words, the statistical power to detect significant weight loss increased during the course of the experiment. Although the use of bins accurately portrays industry practice, the use of individual containers more accurately reflects the potential for weight loss. There is some evidence that contraction of the pedal muscle is associated with water loss. Hyman ( 1967) cites a study by MoitIs showing that when the pedal muscle of Nactica species is stimu- lated to contract, fluid is emitted from the mantle cavity in a volume similar to the capacity of the shell. The initial rapid in- crease in weight loss when abalone were transferred to the cool room inay have arisen from pedal muscle activity that caused more rapid water loss than occuired in those abalone stored under am- bient conditions. Poor pedal adhesion, particularly among those abalone subjected to cool storage conditions, may have been re- lated to water loss from the pedal muscle, or reduced activity levels to conserve energy. Alternatively, it may have been a consequence of a hypoxia-induced reduction in muscle tone. Failure to maintain strong adhesion to the substrate is a clear sign of stress or mor- bidity in abalone. Under normal conditions, large abalone have been observed to resist a force of one-half ton (Hyman 1967). The shell morphology of abalone is also less suited to thermal stress during air exposure than the shells of more globose-shaped prosobranchs. such as the turbinids and trochids, that have the additional advantage of an operculum. Temperature is the single most important factor in controlling the quality and shelf life of fresh seafood. Consequently, it is usual for most seafood to be stored at low temperatures before processing. Both bacterial growth and biochemical reactions will be increased when har- vested abalone are stored at elevated temperatures. Although cold storage will suppress increases in hosted bacterial populations, paradoxically, it may depress biochemical processes that otherwise defray the negative effects of hypoxic stress due to air exposure. Abalone of several species subjected to hypoxic stress caused by air exposure have been shown to produce increased quantities of the glycolytic metabolites D-lactate and tauropine rapidly, and to use arginine phosphate as an additional source of anaerobic energy (Donovan etal. 1999. Tjeerdeina et al. 1991, Watanabe et al. 1992. Wells & Baldwin 199,'i). Wells and Baldwin (199.S) refer to specu- lation that the accumulation of tauropine as an additional pyruvate reductase end-product may be an adaptation in haliotids to protect their tissues during hypoxic episodes. Their studies indicate that lactate is the preferred product during such environmental hypoxia as air exposure and tauropine during functional hypoxia arising from exercise. They suggest that abalone may be adapted to con- tinue metabolizing during air exposure associated with commercial handling, processing, and shipping, despite an inability to irrigate their gills. Unlike my results, most of the abalone [H. iris Martyn and H. aiistralis Gmelin) exposed to air during their study main- tained good pedal adhesion and muscle tone for 24 h. The capacity to use anaerobic glycolysis during hypoxia is likely to vary among abalone species (Donovan et al. 1999). It is possible that H. nihra is more dependent on aerobic metabolism than species studied to date, but this demands further investigation. The greater weight loss and apparently higher morbidity in abalone initially exposed to ambient air for several hours before refrigerated storage may be related to disruption of these biochemical processes caused by substantial reductions in temperature. This may partly explain why abalone held at ambient tempera- ture lost less fluid than those placed directly in cold storage (<5'C). However, the poorer perfonnance of those transferred to cold storage after 4 h seems likely to have resulted froin the ad- ditional stress imposed by multiple temperature changes. The vari- ability in differences among treatments over time may reflect the effects of thermal stress interfering with biochemical responses to hypoxic stress. Although I initially suspected that fluid loss may involve active physiological processes directed at preventing gill desiccation, this does not seem to have extended to active excretion of water to irrigate the gills, as has been reported for limpets (G. Parry pers. TABLE 5(a). Simple comparison of tapwater evaporation (%) treatment means for each time level (a = 0.05; data were log (x -i- 1) transformed). Time (h) df F P Schefl'e Comparison Grand Mean MSD 1 1. 18 192.47 0.0001 Ambient < cool 0.20 0.03 2 1. 18 65.34 0.0001 Ambient < cool 0.37 0.04 4 1. IS 6.85 0.0174 Ambient < cool 0.47 0.05 5 1, 18 0.31 0.5838 Nonsignificant 0..54 0.05 6 1. 18 3.18 0.0914 Nonsignificant 0.60 0.05 8 1. 18 1.89 0.1865 Nonsignificant 0.68 0.04 10 1. 18 0.35 0.5636 Nonsignificant 0.75 0.04 24 1. 18 6.91 0.0170 Ambient > cool 0.93 0.04 The Implications of Post-Harvest Weight Loss 801 TABLE 5(b). Simple Cdiiiparison of differonces in tapwaler e\aporatlon (9f I over time for each level of treatment (a = 0.05: data were log (X + 11 transformed). Treatment df F P Scheffe Comparison Grand Mean MSD Ambient Cool 7. 63 7. 63 1234.46 1452 0.00(11 O.OOOI Ih < 2h < 4h < 5h = 6h < 81i< lOli < 24li 111 < 2li < 4h< 5h < 6h< 8h < lOh < 24h 0.77 0.S4 0.07 0.05 comm.). What probably occurred in this study was initial excretion of body fluids into the mantle cavity where necessary to maintain the gill tissues in a liquid en\ ironment. followed by a loss of body fluid and water retained within the mantle cavity once morbidity progressed to the extent that muscle tone was reduced and active water retention ceased. This would explain why lost fluid had a bluish hue consistent with the oxyhemocyanin respiratory pigment in abalone blood (Barnes 1986) and why I observed episodic losses from some abalone into their containers during weighing. Evaporative losses simulated using tapwater were small throughout the experimental period and between treatments. In contrast, losses of abalone fluid were greater than the tapwater losses and varied with each treatment temperature regime. Losses of around 4% or more of the fluid from abalone subjected to ambient temperatures warrant further consideration. I would have expected evaporation of this apparently higher \iscosity fluid com- prising seawater. blood, and mucous secretions to be less than for water. Nonetheless, evaporative loss was accounted for in the es- timates and does not affect the conclusions from this study. Results from this study have important implications for assess- ment and management of the Victorian abalone fishery. Post- harvest weight losses of 10-20'* equate to the removal of 6-12% (about 350.000 to 700,000 average sized abalone) more than ex- pected from the current TACs. Catches used to model the fishery will need to be adjusted to account for these quantities. Because exploitation of weight loss before notification is a recent phenom- enon in the history of the fishery, inclusion of the additional quan- tities in catch inputs will produce model outputs showing a higher risk of decline in population projections. If these risks are deemed to be unacceptable, then management decisions to reduce the TACs may follow that will cost the Victorian abalone catching sector several million dollars in income. Greatest losses occurred under conditions similar to tho.se most commercially harvested abalone are exposed to during fishing, transportation, and storage before processing. Although losses are inevitable aboard diver vessels while at sea (typically for periods of 4— 6 h). these can be offset by careful packing of bins, shielding from the weather, and periodically irrigating or deluging the bins with seawater. Losses estimated during the first experiment were determined after the abalone had been similarly exposed aboard a research boat. Clearly, there is an incentive for abalone processors to become adept at maximizing weight losses before notification for decrementing quota allocations. From a management perspec- tive, it is important to consider whether these practices are desir- able and equitable. Processors exporting live abalone are unable to take advantage of the regulations and must pay more per abalone. This may be offset by higher than normal wholesale prices for live product. Commercial abalone divers harvesting abalone destined for can- ning are at a financial disadvantage from the additional effort expended to attain quotas, unless they are compensated by higher than normal beach prices. However, the high rate of vertical inte- gration within the industry, where those in the catching sector are also involved in the processing sector, ensures this practice is at least tolerated by most divers. It seems that industry concerns expressed a decade ago ha\'e given way to acceptance of overnight 2.0 1.8 1.6 1.4 - g 1.2 c 0 I 1.0 0 I 0-8 0.6 0.4 - 0.2 0.0 r Ambient -Cool 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Time (h) Figure 4. Mean cumulative evaporation of tapwater over time in ambient and cool storage environments (error bars are standard deviations). 802 GORFINE drainage as slandard industry practice. That most of the weight loss in the second experiment occurred during the first 8-10 h may partly explain why processors do not exploit the full 24-h post- landing period available, preferring, instead, only to allow catches to drain overnight. Alternative explanations include greater risk of spoilage if abalone remain unshucked for too long and cost effi- ciencies associated with the engagement of casual factory workers. Evidently, the practice of maximizing weight loss through drainage does not seem to have a negative impact on the quality of canned or frozen abalone products produced in Victoria. Anecdotal information from persons with experience in the processing sector indicates that partial dehydration before canning may enhance the uptake of water into the meat. Some claim that increases of 20- 309^ are achievable, but this remains to be verified. However. Wells and Baldwin (1995) suggest the possibility that changes in metabolites produced during prolonged air exposure may affect the organoleptic qualities of abalone and cite several studies linking taste and meat quality to biochemical processes. This may be temperature dependent and less likely during refrigerated storage: however, it deserves attention. In the future, it is probable we will see the introduction of official electronic beach weighing stations that will automatically decrement abalone quotas as catches are landed at major Victorian ports. The impetus for this will come as much from the need to reduce fishery management costs as to ensure compliance. Such practices are already part of the federally managed Southeastern Trawl Fishery in Australia and discussions have commenced re- garding trial applications of electronic weighing for quota- managed fisheries under state jurisdiction. Until changes in the determination of official catch weights are implemented, post- harvest weight loss will need to be taken into account when setting annual catch quotas for the Victorian abalone fishery. ACKNOWLEDGMENTS This work was funded by the Abalone Assessment Sub- program conducted by the Marine and Freshwater Resources In- stitute on behalf of Fisheries Victoria. Cameron Dixon, David Forbes, Mark Fertier. and Bruce Waters are thanked for diving assistance in collecting the abalone, and Cameron and David are also thanked for their assistance with weighing and recording. Sonia Talman provided considerable assistance with weighing and recording throughout the second experiment and made construc- tive comments on an early draft of the manuscript, for which 1 am grateful. In addition. Noel Coleman deserves thanks for his helpful suggestions on an earlier draft. Comments from an anonymous reviewer also improved the manuscript. LITERATURE CITED Barnes, R. D. 1986. Invertehrate zoology, 5th ed. New York: Saunders. 89? pp. Donovan, D.. J. Baldwin & T. Carefoot. 1999. The contribution of anaero- bic energy to gastropod crawling and a re-estimation of minimum cost of U"ansport in the abalone. Haliotis kamlschatkana (Jonas). / £v/). Mar. Biol. Ecol. 23.5:273-284. Gardiner, M. S. 1972. The biology of invertebrates. New York: McGraw- Hill. 954 pp. Hickman. C, P. 1973. Biology of the invertebrates. Saint Louis: Mosby. 757 pp. Hyman, L. H. 1967. The invertebrates: mollusca I. vol. VI. New York: McGraw-Hill. 792 pp. Newell, R. C. 1979. Biology of intertidal animals. 3rd ed. Kent: Marine Ecological Surveys. 781 pp. SAS Institute. 1990. SAS/STAT user's guide, version 6, 4th ed.. vol. 1. Carey. NC: SAS Institute. Tjeerdema. R. S.. R. J. Kauten & D. G. Crosby. 1991. Interactive effects of penlachlorophenol and hypoxia in the abalone Haliotis nifcwens as measured by in-vivo "P NMR spectroscopy. Aqua. Toxicol. 21:279- 294. Walanabe. H.. H. Yamanaka & H. Yamakawa. 1992. Post-mortem bio- chemical changes in the muscle of disk ahalone during storage. Nippon Siiisan Gakkaishi. 58:2081-2088. Wells. M. G. & J. Baldwin. 1995. A comparison of metabolic sU"ess during air exposure in two species of New Zealand abalone, Haliotis iris and Haliotis australis: implications for the handling and shipping of live animals. Aqiiaciiltitre 134:361-370. Journal of Shclljhh Research. Vol. 20, No. 2. 803-SI 1, 2(X)I. TRIGGERS AND TARGETS: WHAT ARE WE AIMING FOR WITH ABALONE FISHERIES MODELS IN AUSTRALIA? HARRY K. GORFINE,* BRUCE L. TAYLOR, AND TERRY I. WALKER Marine '• ^ y. h y. h. s y. h. z y. h, z, s y. h. z. s. y" y. h, z. s, y* \. It. z. -s. v" 0.075 0.102 0. 1 62 0.144 0.225 0.168 0.286 0.-307 0.352 0.369 0.411 0.398 0.426 The variables are abbreviated as follows: v = year (1980-1998), \ = season (1-4). ; = zone/subarea (A. B, CNP. CP. D). /; = entitlement holder (1-84). 820 Plaganyi et al. UJ o 1 5 05 oS^ ckV o> o$3 o$3 r^ cxl^ rx^ ^# ^c^% ^CJ* ^C5* ^c* ^ ^ ^ .# .<# FISHING YEAR Figure 4. Comparison between the standardized catch-per-unit-effort (CPUE) trends for the "poached" (CP) and "nonpoached" (CNPl sub- areas of zone C. Values are derived from an iterative effort weighted GLM (see text). this period (Tarr 1993). It is to be expected from a relatively slow-growing long-lived resource afforded a respite as high and unsustainable historic catch levels are reduced substantially to be- low the then current sustainable yields. If we assume that the CPUE trend in z,one C is a reasonable index of stock abundance, this suggests that; whereas, the zone C commercial take during at least some of the 1970s and 1980s was below sustainable yield (SY) levels, total catches during the 1990s have again exceeded such yields, resulting in a concomitant decline in CPUE values over this most recent period. This recent decline in CPUE is fully consistent with catch data only if the latter are considered to in- clude the considerable poaching component of the over-all catches over recent years as estimated by the model. The similarities in the catchability coefficient estimates ^cpuE^f^p^ and q^'''-"' [CNP] (Table 7) for the two subareas are as expected, given the approximately equal habitat areas and the simi- larities in the standardized catch rates over much of the 1980s (Fig. 4). The difference in the pristine spawning biomass estimates B;f[CNP] and B;;'[CP] (Table 7) are, in the main, attributable to the partitioning of the historic zone C catch data between the two subareas. The estimated poaching catches are substantial in all of the scenarios investigated in this analysis (Table 7), despite the fact that the recent steep decreases in the FIAS data are not adequately reproduced by any of the model scenarios. It is important to note that the inability of the model to simulate the trends in the FIAS data successfully, particularly the steep decline in subarea CP. suggests that pending further analyses, model estimates are likely to be positively biased. The inadequate fit of the model to the FIAS data (Fig. 5) may. in part, be explained by the fact that since 1994, there has been a dramatic decrease in young abalone (."i— 10 mm) found during div- ing surveys in zones C and D (TaiT et al. 1996). This has been ascribed to a decrease in the abundance of sea urchins (in turn caused by an increase in the abundance of the rock lobster — an important predator on urchins). Young abalone shelter and feed under sea urchins: hence, it has been proposed that these multi- species interactions are resulting in substantially reduced abalone recruitment (Tan' et al. 1996). This factor is ignored in the current model, but needs to be investigated further (see Future Work), ll should be noted, however, that only animals older than approxi- mately 5 y are sampled on the FIAS transects; thus, there is at least a 5-y time lag before the effects of a recruitment failure are mani- fest in the data. Comparisons of the log-likelihood contributions for the base- case and scenarios I-IV (Table 7) suggest that the main conflict in the model tits is between the tit to the CPUE data for the poached subarea and the catch-at-age data for the recreational sector, FIAS (poached subarea in particular, see, e.g.. Fig. 9) and poaching sector to a lesser extent. In these instances, an improved fit to the CPUE data (i.e., a decrease in the log-likelihood and & values) results in a sacrifice to the fit to the catch-at-age data. The fact that the CPUE (CP) and age data are pushing in different directions requires further consideration. The base-case estimates of the ratios B;i'(9S)/Blf{53} are 0.85 and 0.17 for subareas CP and CNP, respectively, but these esti- mates varied substantially across the various model scenarios. This stresses the importance of refining the estimate of li. for example. as well as investigating further the model's sensitivity to the as- sumption that natural mortality is age independent. The use of an age-dependent mortality function versus an age-independent mor- tality value is particularly important because of. among other things, the sensitivity of the estimated MSY values to this assump- tion (Table 7). The results discus.sed above follow only given the CPUE data, but the appropriateness or otherwise of using CPUE data as an index of abalone stock abundance has in the past been questioned both in the South African context and elsewhere in the world. There are a number of reasons why CPUE may be a reasonable index of abalone stock abundance in the South African context but not necessarily elsewhere in the world. The major South African abalone fishery is located in shallow kelp bed areas, relatively close inshore, along a stretch of coastline measuring less than a hundred kilometers and with relatively easy access to most of the areas (R. Tarr pers. comm.). Furthermore, the 10 or so launching sites currently in use along this coastline are exactly the same sites that were in use during the 1930s when the fishery began; hence, fishers have fished fairly evenly throughout the area. Even if very large aggregations existed in the past, the resource was so heavily fished during the 1960s that these would since have been reduced substantially; hence, the effect of spatial patchiness dampened to a large extent. Moreover, fishing effort is spread over most of the major fishing zones so that the general pattern of diver behavior is not one in which new aggregations are sought out and then se- quentially depleted so that catch rates remain fairly stable, despite the resource abundance declining, as has been demonstrated for fisheries off Australia (Shepherd & Baker 1998), California (Teg- ner et al. 1989). and Mexico (Prince & Guaman del Proo 199.^). It is now well established that in such areas as South Australia and Tasmania, the spatial variation in density is such that, coupled with the added complexities of diver behavior. CPUE cannot be con- sidered a reliable indicator of stock abundance (Breen 1986, Prince 1992. Shepherd et al. 1992. Keesing & Baker 1998. Worthington & Andrew 1998). As proposed by Dichmont et al. (2000) one option available to resource managers with some confidence in CPUE as an index of abundance for their fishery, is to use a dynamic model tuned to both CPUE data and also a survey index. The FIAS data series not only provide a valuable independent index of abundance but are also useful, because they include in- formation on abalone from age 5 upward compared with the CPUE data, which captures information on animals aged approxi- mately 8 y or more. This means that the FIAS abundance index is Age-Structured Stock Assessment Model 821 TABLE 7. Best fit esliiHiitcs of the pre-expldilatiim spiiwning bioiiiass /?;,'' for the •■piiached" CP and "nonpoached" CNP areas of zone C. the estimated natural nicirtalit> estimate .1/ and llie poaching niaxinmm Cl'„,.,^. »hert the proportion of C7'„,^^ from area CNP is fixed at 0.1 in the base-case. III Sens. I 11 Age-Dependence IV Base-Case h /'i,„.„i, of A/ (No Downweighting) h 0.7 0.6 0.7 0.7 0.7 Ppouch 0.1 0.1 0.3 0.1 0.1 Wt Age 0.1 0.1 0.1 0.1 1 Bi" [CNPJ I.SS4 1745 2689 6123 1601 B'o" [CP] 3358 3457 .3422 6303 3404 ^CPUE jCNP] 2.21 X 10"" 1.85 X 10"" 0.98 X iO-'' 1.04 X 10"" 2.03 + 10"* ^CPUE |cp] 1.28 X I0-" 1.55 X 10^" 1.34 X 10-^ 1.72 X 10-" I.IO X 10-" M 0.1.54 0.174 0.17 M(a) = 0.23 + 0.3/a + 1 0.24 CP,„„ 242 250 359 242 203 a (CS) 8.983 9.047 8.857 9.654 9.404 a (RS) 9.376 8.796 8.997 9.195 9.020 fl(PS) 6.588 6.622 6.386 5.533 6.750 a (FS) 5.549 5.489 5.466 5.617 5.513 kL (CS) 0.007 0.006 0.006 0.005 0.009 |X(RS) 0.005 0.006 0.(»5 0.004 0.006 M.(PS) 0.141 0.148 0.120 0.127 0.136 |x(FS) 0.059 0.048 0.052 0.063 0.073 8(CS) 2.189 1.435 17.41 1.391 1.265 S(RS) 1 .097 2.510 1.143 2.053 2.433 6 (PS) 1.212 1.145 1.228 1.149 1.477 8(FS) 1.032 I.IIO 1.516 1.758 1.727 -InZ. CPUE (CNP) [d-^'^l -27.863 [0.14] -27.958 10.14] -27.437(0.14] -25.996(0.15] -27.413(0.14] -In/. CPUE (CP) la*^^] -38.477 [0.08] -37.490 [0.08] -37.791 [0.08] -30.408(0.12] -30.992(0.12] -InZ. FIAS (CNP) -4.729 -4.686 -4.873 -4.575 4.585 -InZ. FIAS (CP) 13.661 13.830 13.748 14.186 14.896 -Inl age CS (CNP) [d^l -6.671 [0.12] -6.691 [0.12] -5.793 [0.13] -7.127(0.12] -63.759(0.13] -InZ. age CS (CP) [df ] -11.898 [0.09] -11.553 10.09] -11.677(0.091 -9.280(0.11] -I05..598 10.I0] -InZ. age RS [it] (combined) -1.495 [0.14] -1.481 [0.14] -0.899(0.16] -3.1 II [0.09] -38.441 (0.08] -InL age RS (CP) |al 0.246 [0.221 0.239 [0.22] 0.099(0.21] 0.186(0.21] 1.413(0.21] -InL age FIAS (CNP) [ct| -1.941 [0.11] -1.781 [0.12] -1.816(0.12] 0.299(0.21] -19.152(0.11] -In/, age FIAS (CP) [&] -2.461 [0.14] -5.023 [0.091 -1.600(0.16] -6.409 [0.07] -73.650 [0.06] -InL age PS |.t] -0.792 [0.08] -0.600 [0.09] -1.II7(0.05( -0.411 [0.12] -2.602(0.14] -In/. (Total) -82.421 -83.195 -79. 1 56 -72.645 -.349.815 B;-;^ (98)/K ICNP] 0.85 0.58 0.64 0.39 0.69 B,r (98)/K [CP] 0.17 0.35 0.39 0.23 0.62 MSY [CNP] 136 153 231 376 193 MSY [CP] 289 304 295 387 410 Minimum values of the negative of (he log-likelihood function and estimated current depletion are also shown. The estimated selectivity parameters are shown for the commercial sector (CS). recreational sector (RS). poaching sector (PS) and FIAS (F). Values of a are given in square brackets alongside the corresponding log-likelihood values. more sensiti\e than the CPUE index (cf. Figs. 2 and 4) to changes in the younger age classes. This is particularly itnportant in zone C, where extensive poaching of sublegal ubalone individuals as well as a hypothesized multispecies effect (TaiT et al. 1996) have caused a rapid decline in the nuinbers of sinall individuals. Consequently, the FIAS data have indicated a more negative scenario than have the CPUE data and are, thereby, providing valuable clues as to the factors to consider in modeling the re- source. Finally, we reiterate that the inferences above are based on preliminary model results only. Coefficients of variation and con- fidence intervals remain to be calculated for the model estimates, and the sensitivity of the model to its assumptions and parameters must be examined. This paper is not, and does not pretend to be a thorough and complete analysis. Rather it is intended as a descrip- tion of the stock assessment approach being applied in South Af- rica, motivated by the fact that it currently differs from approaches in practice elsewhere in the world for abalone resources. Planned Future Work Some of the issues to be addres.sed in future work are outlined below. 822 Plaganyi et al. (a) NONPOACHED SUBAREA CNP liJ 3 Q. O ■D 0) in •a c ■♦-' to o& oO- o5> o& oft a^ c* ^ ^ ^ ^ ^ ^ ^ ^ .^ ,^ ^ '^ •^ FISHING YEAR (b) POACHED SUBAREA CP UJ Q. O ■D V tf) re ■a c re 55 25 1 5 05 OBSERVED -MODEL OBS_FIAS (95% CI.) -MOD FIAS c*^ Sr ^ ^ .50%' and increased substantially on one remaining productive reef. An overfishing hypothesis best explains the long-term decline in the catch of many reefs and the transfer of fishing to distant or remaining reefs. The history of management is reviewed and shows that, over the lifetime of the fishery, management has moved from control by government with little consultation with industry in 1967 toward self-management with endorsement by government from about 1995. Since 1995, management has been unresponsive to declines in the fishery, and possible reasons are explored. We advocate application of the precautionary approach by management at the scale of the metapopulation to arrest the declines and the establishment of recovery plans for areas in serious decline. KEY WORDS: abalone fishery, serial decline. Hiilioris laevifiuui. metapopulation, larval retention, fishery indicator, precautionary approach, fishery management INTRODUCTION Focus on the causes and prevention of collapse of fisheries is a direct consequence of increasing emphasis on the conservation of exploited stocks and their sustainable management. The spectacu- lar collapses of anchovetta. cod. and crustacean stocks in recent decades (Steele 1981. Hutchings & Myers 1994. Orensan/. et al. 1998) have served to catalyze deep concern and critical analysis of the causes in the hope that appropriate lessons can be learned. Many abalone fisheries have also collapsed worldwide (Shep- herd & Baker 1998). except in Australia, where they have appar- ently been sustained (reviewed in Prince & Shepherd 1992), and where envious eyes have often turned to learn the touchstone of success. The collapse of abalone fisheries has inany causes under- lying the obvious proximate cause of overfishing. Basic problems of assessment are: ignorance of stock-recruitment relations and practical difficulties and high cost of obtaining fisher-independent data, such as abundance and recruitment estimates for many inde- pendent stocks. Problems confronting managers are: absence of proved fishery indicators and tendencies to maldistribution of ef- fort. Yet, despite their reputation, all is not well with the Australian *Corresponding author. E-mail; shepherd.scoresby@saugov.sa.gov.au abalone fisheries. In particular, the fishery for greenlip abalone (Hallotis liievigiira Donovan) has suffered declines of varying se- verity in several states (Shepherd et al, 2001 ), despite the fact that possibly more is known of the biology and population ecology of this species than any other abalone species. This paper seeks to discover an underlying pattern in the declines and why they are continuing in South Australia, despite the abundant knovsiedge and the recent warnings of decline. We focus our attention on the greenlip fishery, although blacklip abalone. H. rubra Leach, is taken on most reefs, although in different habitat. First, we summarize the diverse evidence, both biologic and historic, showing a decline in the fishery, and undertake a prelimi- nary analysis of factors associated with that decline. The recent (1995) availability of catch data at a fine (metapopulation) scale (Keesing & Baker 1998) has provided a catch history of some 49 independent greenlip stocks in the Western and Central Zones (see Shepherd & Brown 1993): our analysis is limited to these zones. We consider the evolution of management from government con- trol toward self-inanagement in the context of the changing legal and administrative structures. We then describe the history of key management decisions and explain the long-term declines in terms of failure to apply a precautionary approach. We conclude with recommendations for the improvement of management and the 829 830 Shepherd and Rodda restoration of the stocks. Our conclusions are relevant generally to management of exploited, sedentary invertebrates with multiple stocks. METHODS AND MATERIALS Historical Information and Catch and Effort Data Historical data were extracted from the minutes of the Abalone Advisory Committee and later of the Abalone Management Com- mittee (AMC) and the records of the first author, who has been involved in research and management of the fishery continuously since 1968. Catch and effort data, reported on monthly logbooks that recorded daily fishing excursions, have been collected since the inception of the fishery in 1967. Early data on catch and hours fishing were collected on a grid scale of -37 x 37 km. and, from 1979. on a fine population scale with map code numbers (Fig. 1 ) allotted to individual reefs, islands, and sections of coast, reflecting the actual distribution of abalone populations. Thus, the Western and Central Zones have 153 map codes embracing about 164 dis- tinct abalone metapopulations. Excluding the far west region north of Pt. Brown, which has -34 greenlip populations, there are 138 reef codes embracing -74 greenlip and a similar number of black- lip metapopulations. The number is approximate, because we have assumed that isolated reefs and islands haxe distinct metapopula- tions and that mainland populations separated by unsuitable habitat of >-20 km are distinct (see Shepherd & Brown 1993). The temi population in this paper (names are followed by map code numbers in brackets for easy cross reference) means an infeiTed metapopu- lation, or in some cases, two or more juxtaposed metapopulations, following the detailed genetic work of Brown and Murray ( 1992). The populations considered and their coiTcsponding map codes are listed in Tables 1, 2. Our study is confined to 49 map code areas, including Franklin I, but excluding the seldom visited, very small populations at Greenly I., west Spencer Gulf and Sir Joseph Banks Group in the Western Zone, and Covvell Grounds and Marion and Stansbury Reefs in the Central Zone with a total mean cumulative annual catch of -I t. Catch and effort data at the scales described are given by Kees- ing et al. (2000). with partial analyses of specific features by Keesing and Baker (1998) and Shepherd and Baker (1998). Pend- ing a full analysis of the data, we have here undertaken a prelimi- nary analysis to identify important features deserving detailed study. We estimated rates of decline in the catch of individual populations by simple regression analysis, over the full time period (usually 19 or 20 y from the commencement of intense fishing) where the decline was monotonic, or with a broken stick model, where the catch fell sharply and then stabili^^ed at a low level; in the latter case, we estimated the rate of decline by regression from the start of fishing to the year in which the catch had declined by 95%. For a few areas, mostly in the Central Zone, catches were initially low and peaked during the mid- to late- 1980s; in these cases, the regressions to estimate rates of decline commenced at the point where catches first peaked, our object being to estimate decline rates from the time of peak catches rather than in relation 0 100 Scale (km) Far West Streaky Bay ""^■'^i;, Venus Bay sp SA Waterloo Bay I SCjH V> SF lOA"-, lOC Western Zone .y Spencer Gulf HI J ,4^95 0.8 8F,G.N Elliston Cliffs 2.4 3.1 (80) 0.3 0.20*** (19 y) 99 8P,9E,F Flinders 1 (Gem) 15.7 16.4(80) 11.2 0.35 NS 41 9A.B Ward 1. 13.6 13.5 (80) 13.1 0.07NS(19y) 12 0.45 lOA.B Tungketta 0.3 2.6(83) 0.1 0..36* (10 y) >99 IOC Pearson 1. 0.7 1.0(82) 0.3 0.12** (17 y) 97 0.25 llA Sheringa 4.1 3.9 (80) 3.1 0.13* 63 IIB.C Kiana 3.4 4.2 (80) 0.5 0.25*** (17 y) >99 12A,C Drummond 20.6 23.7 (80) 2.6 1.53*** (17 y) >95 0.59 13A,C Frenchman 3.7 5.3(80) 0.3 0.24*** 98 <0.01/m- 13D-F Reef Head 13.9 13.2(80) 2.9 2.09** (14 y) >89 14A.B Pt. Whidhy 4.4 4.7 (80) 0.9 0.49** (14 y) >95 14C.D Misery Bay 27.8 30.6 (80) 7.8 1.58** 95 14E Whidhy Is. 0 13.7(83)*' 1.2 0.30 NS (11 y) 91 15A Avoid Bay 49.1 47.5 (80) 2.2 3.17*** >95 0.60 15B.16A D'Anville Bay 5.8 3.8(80) 1.6 0.15** 75 16C Fishery Bay 18.3 12.8(80) 4.4 0.48*** 79 18C-F Thorny Passage 40.0 44.2(80) 37.5 0.30 NS 13 1.0 19A-C 19D,E Dangerous Reef 0.4 2.4(83) 0.8 0.10 NS (16 y) 49 22A,24A Hardwicke Bay 2.7 29.7 (84)' 12.4 4.0** (12 y) 87 23 SW Yorke 5.3 4.6(80) 1.8 0.17NS(10y) 48 25A Backstairs 6.2 6.6(80) 0.5 0.40 NS (15 y) 76 0.41 26 West Bay 1.0 4.0(85) 2.2 0.27 NS 29 28 Vivonne Bay 1.8 2.2 (82) 0.7 0.20** (17 y) 98 29 Gantheaume 4.3 3.9 (80) 3.4 0.09 NS 37 30 DEstree Bay 4.5 4.8(82) 2.5 0.20NS(17y) 42 31 Willoughby 3.7 3.6(82) 2.5 0.59* (10 y) 90 32 NE Kangaroo 1 2.0 2.0 2.0 0.14 NS (lOyl 50 Historic catch is the mean annual catch (t) for 1971 to 1978. early catch is the mean annual catch for 3 y commencing with the year shown in brackets, current catch is the mean annual catch (t) for 1996-1998. rate of decline (t/yr' ) is for the period 1979 to 1998 or a lesser period where indicated, percentage decline is the present productivity as a percentage of original production (see te.xt). Z is the instantaneous total mortality rate from catch curve analyses, and survey densities. D. are in numbers/ni". A blank indicates absence of data. Statistical significance is shown thus; * = P < 0.05; **■ = P < 0.01; *** = P < 0.001; NS = not sianificant. "closed to fishina I9S2-I9S6 and 1995-1999. ''reef discovered 1983. 'reef discovered 1984. to any specific fixed starting point. The 50 and 8U'/( points of decline and total percentage declines were determined from the regressions. The mean annual catch for 1971 to 1978 (Table 1 ) is given to provide an estimate of productivity early in the fishery after removal of the virgin stock from 1967 to 1970. The "early catch" mean annual catch data in Table 1 are the mean annual catches over 3 y in the early 1980s after stabilization of the catches following granting of the right of divers to sell their licences in 1980. The mean annual catches were then only very slightly lower (<0.2%) than the mean annual catches for 1971 to 1978 and only 12.6% higher than the mean annual catch for the first 4 y after quota introduction in 1985. The current catch data given (mean annual catch for 1996 to 1998) are estimates of the present pro- ductivities of populations. In this paper, productivity is estimated by the annual catch of a stable population averaged over a number of years and assumed to be more or less fully exploited (Kesteven 1996). All catch data are given In total (in-shell) weights (TW), Esti- mates of the total mortality coefficient, Z, are from; Shepherd and Baker (1998) for Waterloo Bay (8H-M). Sceale Bay (4D.E). Ward I. (9A.B). Thorny Passage ( I8C-F) and Tiparra (21 ): Shepherd et al. (2001 ), for Avoid Bay ( 15A) and Backstairs Passage (25A): and O'Loughlin and Shepherd (unpublished data) for the remaining sites. Mean density estimates are unpublished data. 832 Shepherd and Rodda TABLE 2. Greenlip abalone populations in the Western and Central Zones with stable or increasing catches. Historic Catch (t) Current Catch (t) Map Code Name (1971-1978) Early Catch (tl (1996-1998) Rate of Change % Up Z or D yc.D Hotspot 23.0 26.2(79) 26.1 +0.01 ns 1.3 0.30 9G.H Flinder^ SE 7.1 8.3 (80) 11.1 +0.13 ns 40 17A,B Sleaford 2.4 1.4(80) 2.4 +0.02 ns 26 18A.B West Pt. 8.8 5.3(79) 6.3 +0.17 ns 23 18G-L Thistle 1. 8.5 9.8(80) 8.4 +0.008ns 3 19F Wedge I. 2.9 3.7 (80) 5.3 +0.08 ns 88 21 Tiparra 45.1 67.8 96.4 +2.16 ns 11 0.72 2IH C. Elizabeth 7.4 25.4(79) 11.7 +0.03 10 27A Hanson Bay 1.1 0.5 (79) 2.2 +0.04 87 See caption to Table 1 for explanation of headings. Estimates of the spatial extent of fishing areas in 1978 to 1979 and in 1995 to 1999 were made during individual diver interviews by one of us (SAS) when divers marked on maps at a scale of 1 km = 1 cm (twice that scale for Watchers (8C.D) and Fishery Bay ( 16C)) the spatial extent of the places they fished. Hence, the data show only the gross extent of greenlip habitat, including inter- spersed sandy patches, rather than the precise extent of rocky substratum actually occupied by abalone. The areas shown here were drawn from the common area indicated by a minimum of four divers in 1978 to 1979 but only three divers in 1995 to 1999, because fewer divers fished areas that contracted spatially. In the multiple regression analysis, the dependent variable, the proportional rate of decline of catches for individual reefs, was the absolute rate of decline calculated as described above divided by the initial production that was the y-intercept of the regression of catch versus time. The independent variables examined were: ini- tial production (as defined above), distance from port, depth, and a "Closure Index" of the population. Distance from port was the distance from the mid-part of the reef in question to the nearest mainland boat ramp along the sea route used by divers; depth was the average depth of the population as then exploited as supplied by divers during the 1978 to 1979 diver interviews; and the "Clo- sure Index" was the proportion of the perimeter of the areal extent of the fished reef bordered by land or emergent reef. For this purpose, reef maps prepared from the diver interviews described above were used to calculate the index, which took a value be- tween 0 and 1 . History of Management The history of the fishery has been reviewed by Prince and Shepherd (1992) with later detail added by Keesing and Baker ( 1998). Here, we summarize briefly the salient features of fishing controls and management. The fishery is controlled by input mea- sures, size limits, and licence limitation, and by output measures, and indisidual quotas. After the virgin stocks were fished down by 1970 (Shepherd & Baker 1998, Keesing et al. 2000) and the number of divers reduced to 21 in the Western Zone and 5 in the Central Zone by 1971. the fishery was stable for the next 9 y. In 1973 to 1975 there was slight evidence of stagnation in the fishery as divers aged until three more divers were admitted in 1976. In 1980. licences became saleable, and catches increased somewhat until 1984, when growth overfishing became marked in the Western Zone and various con- trols were introduced to reduce the level of fishing (see below). In 1985. individual quotas were introduced in the Western Zone (12 t for greenlip) and in 1989 in the Central Zone (23.7 t for greenlip). Greenlip quotas were reduced by 3 t (25%) in the Western Zone in 1989. During the 1990s, as inshore catches progressively declined in the Western Zone, competitive behavior between divers became evident as divers at the beginning of the fishing year competed to be the first to fish the fishing grounds and maximize their catch rates. Management has evolved progressively since 1967. From 1967 to 1 978 the Director of Fisheries managed the fishery, under del- egation from the Minister of Fisheries. In practice, decisions were made on scientific advice and/or in response to representations by the divers' association. Such key decisions as licence limitation and their early reduction in number, and the imposition of zones in the fishery were made in accordance with ministerial policy and sometimes in conflict with the wishes of divers. This may be called the "instructive" period of government manageinent (Fig. 2) after the terminology of Sen and Raakjaer Nielsen (1996). In 1978, following a prolonged dispute in which divers resisted the addition of more licences to the fishery (Chattenon & Chatterton 1981 ), an advisory committee was established, comprising divers and gov- ernment representatives in about equal numbers. The committee had no authority, but made recommendations on research and man- agement. Matters, some of them contentious, considered by the committee in the first 7 y included: the issue of rewards by indus- Govemment management Cooperative Advisory Figure 2. Shift in comanagement arrangements over time in the South Australian abalone fishery. The years in brackets indicate approxi- mately the suggested point of transition ahmg the spectrum of change. Decline oe H. lkevigata and Review of Management 833 try for evidence of poaching: relief diver days (awarded during sickness of licensee); landing abalone in the shell; permit sharing; research closures; increase in size limits; and a buy-back scheme. On major issues, such as the admission of more divers and the right to sell licences, divers dealt directly with the government and bypassed the committee, but on other issues, divers' views were taken into account in the final decision. During this period, the arrangement could be termed consultative in terms of mode of influence (Hersoug & Ranes 1996) for major issues and coopera- tive for minor issues (Fig. 2). Thus, major issues were resolved by the Director or Minister and minor ones in a cooperative manner within the committee. By 1985, divers were showing increased responsibility in terms of resource conservation and management following evidence of overfishing in 1981-1983 (Lewis et al. 1984). Divers urged a 4-mo closure of the Western zone during 1984 to reduce the catch from the overfished region, and agreed to an increase in size limit from 130 to 145 mm SL for that zone (advocated by researchers since 1979). They also proposed closures during the spawning season, increased the reward to $5,000 for infonnalion about poaching, investigated the feasibility of quotas, and then voluntarily assumed quota restrictions before promulgation of the amending regula- tions, assented to the abolition of relief diver days, and offered assistance with research. In 1988, an abalone management committee (AMCl was for- mally established to "advise on management, harvesting strategies, and on policy and legislation." Divers instigated the landing of abalone in the shell in the Central and southern Zones (later modi- fied), and the introduction of quotas in those zones in 1988 and 1989. In 1989, the greenlip quota in the Western Zone (see Fig. 6) was reduced by 3 t TW (25%) upon evidence of declines in inshore populations and in partial recompense for the loss, an increase in the blacklip quota of 0.75 t TW was negotiated by the divers from 1990. A decline in the catch in Backstairs Passage was reported by research in 1989, and an increase in size limit for greenlip on Kangaroo Island was proposed, but not accepted. Thus, during the 1980s, divers took the initiative and were pro-active on many, but not all. measures directed toward the conservation of the stocks. In 1989, the president of the Abalone Divers Association was appointed chairperson of the management committee in tacit ac- knowledgment of the industry's constructive participation in man- agement. From this time, it might be claimed that management was joint with government, although formal acknowledgment of the de facto joint management arrangements did not occur until 1995. when the requisite regulations establishing the AMC became law. By 1992, the right of permit holders to employ divers to take their catch was recognized, and the practice began and soon became widespread, for licences to pass into the control of companies, in which one or more divers, under contractual or share fishing ar- rangements, were nominated to take abalone. The implication of this was that, as time went on, divers' representatives on the man- agement committee became increasingly likely to be business per- sons with little first-hand know ledge of abalone abundances in the field. The 1995 regulations essentially passed control of the fishery to divers, because only industry members had voting rights. Four government obser\ers on the committee represented administra- tion, management, research, and enforcement sections of govern- ment; subsequently other observers representing the umbrella as- sociation of all professional fishers, recreational, and conservation interests were added. The Director of Fisheries retained the power to set quotas, but in practice, quotas were recommended by the government manager, endorsed by the AMC, and approved by the director. Under the regulations, the Minister reserved the power to override decisions of the AMC, but this has not happened. By 1995. government policy was explicitly one of cost-recovery; i.e.. the industry was required to pay all costs of administration, man- agement, enforcement, and research in the fishery. In consequence, the AMC increasingly involved itself in the scrutiny of govern- ment expenditure on research and management and "demanded the types and levels of service thai offered the greatest value for money" (Geen & Nayar 1988). In 1997, an independent chairman of the management commit- tee was appointed, in accordance with government policy to have chairpersons with no financial interest in the fishery. Thus, from 1995. responsibility and power have increasingly passed to indus- try, and go\ernment observers base increasingly been limited to an advisory role. Overall, as shown in Figure 2, a transition from government management with little consultation, toward self- management with endorsement by government has occurred from 1967 to the present. At the same time, we may observe an increas- ing acceptance of responsibility by divers who, during the 1980s up till 1995. proposed many of the measures actually adopted to conserve stocks. In 1995. the AMC was first advised of declines in eight green- lip stocks (a subset of those in Table 1 ) and a few blacklip ones in the Western Zone, and workshops were held in 1995, 1996. and 1999 with industry to consider them. Researchers advocated clo- sures and an increase in size limits, but these were resisted, and no action resulted. Annual stock assessments and other reports to the AMC from 1995 to 1999 reiterated the .serious character of the declines and assessed the impact of a size increase. In 1999, in- du.stry accepted that there were problems in the fishery, and a stock recovery plan was initiated for four sites: Avoid Bay (15A). Frenchman (13B,C), Anxious Bay (6E,7A,B), and Sceale Bay (4D,E). Unfished, stunted abalone subpopulations occurred near these sites, and researchers and industry cooperated in the trans- location of adults into the fishing grounds to increase larval settle- ment. RESULTS Catch Declines The Western Zone is subdivided into two subzones. A and B. Subzone B (the far west) has low productivity, and the annual quota is correspondingly low (1.8 t per diver unallocated between blacklip and greenlip); greenlip catches have fluctuated cyclically but have shown no long-term downward trend. We do not consider them further in this study. The total catch for the two zones (Fig. 3) has remained stable save for a slight reduction in catch in the Western Zone with the 1989 quota reduction. The zones comprise many spatially sepa- rated greenlip populations of different producti\ities geographi- cally distributed on exposed coasts and near the gulf entrances, except for Tiparra (21) and Hardwicke (22,24) within Spencer Gulf (Fig. 1 ). The size distribution of the.se populations, in terms of their 1971 to 1978 mean annual productivity values (see Table 1) for each zone (Fig. 4) shows that the majority of reefs had productivities of <10 t and that the majority in each productivity size class in each zone have suffered declining catches since 1979. Examples of the declines in catch in individual populations are 834 Shepherd and Rodda Ul 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 Year Figure 3. Global grcenlip catch (T\V) and effiirl in the Western and Central Zones from 1968 to 1998. Data are for rinancial years, July to June; i.e., 68 = 1968 to 1969. Hollow bars (total catch) and dotted line (total effort) are for the Western Zone and black bars and continuous line are for the Central Zone. given in the companion paper (Sheplierd el al. 2001 ) and in Shep- herd and Baker (1998). Historic and present catches of declining populations, rates iif decline, and the percentage decline (derived from regression analy- ses), and estimates of the total moilality coefficient. Z, and recent density estimates are summarized in Table 1. Historic and early annual catch data (Table 1 ) are taken to indicate the approximate productivity of reefs over a decade or more during the 1970s and early 1980s, and current annual catch data are taken to indicate the present productivities of those reefs. In all, 39 reefs show declines in catch of 12 to >99% (mean % decline 71%) in the two zones (Table 1) and six reefs show no decline or increasing catches (Table 2). These percentage decline values represent the declines from maximum catches in 1980 to 1982, equivalent to the mean annual catches during the 1970s after removal of the virgin stocks. Annua) Producli\it> |l) Figure 4. Frequency distribution of the annual productivity (t) esti- mated from the mean catch of greenlip-bearing reefs in 1971 to 1978 in the We.stern (W) and Central (C) Zones. Reefs whose catches have declined by >50% are shown by crosshalching. Next, we examined the temporal pattern of the decline in catch in the Western Zone by plotting the frequency distribution over time of populations reaching their respective ."iO and 80% decline points after 1980. The results (Fig. 5) show that the number of declining reefs (extracted from the data in Table 1 ) at the 50% decline point peaked in 1989 to 1990, and the number at the 80% decline point (a subset of the same reefs) peaked from 1995 to 1996. Changes in Fishing Pattern The pattern of fishing gradually shifted from 1979 to 1998 in both zones. In the Western Zone, as near-shore catches declined, fishers increasingly sought their catch from more distant reefs. This long-term shift in the source of the catch is shown by the gradual proportional decline of the catch in 23 near-shore popu- lations from -65 to 70% in 1979 to -34% in 1998, and the cor- responding proportional increase in catch at 10 distant reefs from -26% to -50% over the same period (Fig. 6). Distant reefs are 30-50 km from the nearest port, except Venus (6A-D), Baird Bay (5B-F), (which are close to port but distant from Port Lincoln where most fishers live) and Sleaford (17A.B). The remaining reefs cumulatively contributed about the same proportion to the total catch over time. In the Central Zone, the picture is simpler. As catches declined on nine reefs around Kangaroo I. and Hardwicke Bay (22,24) (Table 1) they progressively increased at one reef. Tiparra (21), by a massive 77%. To determine whether the spatial shift in fishing in the Western Zone was related to more intense fishing on near reefs than on distant ones, we examined Z for 1 1 reefs (see Tables 1,2). A plot of Z versus distance (D) from port (Fig. 7), shows a decline in Z with increasing distance. We fitted the linear regression equation; Z = 0.71 - 0.008D (R- = 0.56; P < 0.01) For this regression, we used the earliest Z estimates available for Avoid Bay (1987) and Backstairs Passage (1980); whereas, the Decline of H. laevigata and Review of Management 835 D50% ■ 80% -| 1 hr™ Figure 5. Distribution o\ er lime of number of greenlip populations in the Western Zone that reached ^i)"t (blank) and 80% (striped) decline points during their decline trajectory. remaining estimates were obtained since 1995 (Shepherd & Baker 1998, O'Loughlin & Shepherd, unpublished data). Rate of Decline in Calch The relation between the proportional rate of decline of the catch in declining populations, and the variables: initial production (IP), distance (D). depth (De), and the "Closure index" (C) for populations in the Western Zone was examined by multiple re- gression analysis. The constant and the two variables, D (P = 0.002) and C (P = 0.01) were significant, but the variables IP and De were not significant and were removed from the regression. The regression equation (with standard eiTors in brackets) is: R = 9.4(1.6) -0.07(0.02)0 -9.2(3.5 )C (R- = 0.31: P = 0.0006) Plots of the proportional rate of decluie of reefs versus the vari- ables IP. D. and C. but not De (which shows little of interest), are shown in Figure 8. Spatial Contraction of Fishing Areas Plots of the recorded spatial contraction of fished reefs in six greenlip populations are shown in Figure 9. The respective spatial extent of areas originally and presently fished are given in Table 3. with the calculated percentage contraction of fishing areas. Table 79 so 81 82 8.1 84 85 86 87 Figure 6. Proportional change in of the greenlip calch for twenty- three near-shore and ten distant reel's in the Western Zone from 1979 to 1998. The distant reefs are \ enus Bay, Ward I., Hotspot. Flinders I. (both sides), Baird Bay. Sleaford, Thistle I., Wedge I., and southern Thorny Passage, and the near-shore reefs are the remainder of those listed in Table 1. other than Central Zone reefs. 1 ♦ ♦ Western Zone O Central Zone 0.8 * 0,6 ••r^^^~---_,^_^ 0,4 o • ♦ ^^ 0.2 0 30 40 Distance (km) Figure 7. Plot of total mortality coefficient, Z, for greenlip populations in the Western and Central Zones versus distance from port in km. 3 also includes the relevant data for the two populations. Avoid Bay ( 15A) and Backstairs Passage (25A). whose declines are de- scribed in greater detail by Shepherd et al. (2001). The spatial declines have been groundtruthed for four of the eight populations (see Table 1 and Shepherd et al. 2001), and they show very low densities in the areas of decline. Such low abundances are not surprising, because divers fish the areas in motorized scooters (Prince & Shepherd 1992) and cover the bottom systematically and over distances of >10 km per dive. The low relief calcrete (lime- stone) substratum, sometimes interspersed with sandy patches (see o 10 • • • a 8 • h \ • • • 4 • • ~^^,^^^^^^» •^~~~~~~~--- y--0.0672x+5.6l28 r" = 0 : 1 1 7 (1 • • •" •> 20 411 60 Distance (km) 80 0.4 Closure Index O 14 a: g_io « 01 8 O s 01 89 90 91 9; 9.1 94 95 9h 97 98 Q, 2 a • • C y = 0.7907x + 3.2214 • r- = 0.0123 • • • • • • • •• • • • t- • • •• • • • • -0.5 log Initial Production Figure 8. Graphs of the proportional rate of decline of greenlip popu- lations in the Western Zone versus a: distance: b: closure Index; and c: log (initial production). 836 Shepherd and Rodda ' ^^ \ d Hall^Bay ,' .. , /.■yy/y?Z'/ 1 :M ( Point Drummond t 0 30 ---v.q. v'^IJij^x Km """^--^ Coles p/oint N t Frenchman Bluff f Ig z^ I n'W -my-: 1! T /^\ 5^--/- 'My-y'' msliery Bay X^ /' ,- ^ - 'y \y±. ■■■■■■■■ // %•. 20 : X::/ 30 V mi 1 Sleaford Bay mi 1 \ \ /y.-y:: 1 30 m). such as Hotspot (9C.D). Thorny Passage (I8C-F). Ward I. (9A,B). and Pearson I. (IOC), of which only the last has declined, may find a partial refuge in depth and be more resilient to fishing (Karpov et al. 1998). However, some divers do occasionally fish to 30-m depths and say that deep-water populations have poor recruitment and recover more slowly than shallower populations after fishing. The question remains open until methods are found to survey deeper water. The significance of D (distance from port) is not surprising, given the gradient in Z with increasing distance from port, and is confirmatory evidence of the maldistribution of effort in the fish- ery (see above). However, the variable of special interest is the closure index C. This estimates the boundedness of a population by land (or emergent reef) and. by implication, the extent of larval retention within the population. Apart from a very low value of 0.15 for Whidby Isles (14E). where the population collapsed after 6 y fishing, values ranged from 0.30-0.74. We hypothesize that coastal topographic features restrict dispersal of larvae and tend to maintain recruitment or decrease its variability. The hypothesis has substantial support. Wolanski and Hainner (1988) reviewed the effect of coastal topography, such as headlands and islands, on larval transport and noted that topographically generated fronts aggregated eggs and larvae. In empirical studies on this abalone. Shepherd et al. (1992) found that larvae were retained in near- shore eddies and bays, and Shepherd et al. (2001) found that in Avoid Bay (I5A) and Backstairs Passage (25C) recruitment was stronger close to shore and weaker off-shore in open habitats, where currents were stronger. Furthermore, the pattern of spatial contraction toward shore around headlands and in bays, found for these populations as well as for six additional populations (Fig. 9). reinforces the evidence in favor of a strong effect of coastal to- pography on the resilience of greenlip populations. This hypothesis could be tested experimentally by measuring density of larval .set- tlers in collectors in populations with different values of C (see McShane 1995. Rodda et al. 1997). The hypothesis also explains the hitherto puzzling absence of greenlip populations from the many very small islets and sea-mounts that abound within the geographic range of greenlip abalone. Their C values may simply be too low for larval retention. Spatial contraction of stocks is known for declining fish stocks (e.g.. Rose & Kulka 1999). but has not previously been docu- mented for exploited molluscan stocks. The phenomenon may be an example of a much more general principle. Metapopulation models predict that larger metapopulations have higher local abun- dances than smaller ones, because recruitment rates are higher and extinction rates lower (Hanski & Gyllenberg 1997). The converse; namely, that reductions in density in local populations will reduce the metapopulation range but to a less extent, is expected to follow if recruitment rates declined. In the eight populations in Table 3 the mean percentage decline in catch (90%) is significantly more than the mean percentage spatial decline (76%) (t = 2.4; P < 0.05), suggesting that, on average, density declined more than the spatial area occupied, as predicted by theory. However, other models not based on metapopulation theory and with other assumptions may predict the same result (Rosenweig 1991. Holt et al. 1997. & see Lawton 2000). Allernatirc Explanations for the Declines The major underlying assumption of the above analysis is that the total catch reflects abundance. Jamieson ( 1993) noted the ten- dency in capture fisheries for divers to move to new reefs only when exploited ones are exhausted, and McShane ( 1998) observed the preference of divers to fish reefs closer to home. Keesing and Baker ( 1998) relied on these arguments in their earlier analysis of this dataset to interpret trends in abundance of local stocks. Shep- herd et al. (2001 ) reviewed diver behavior in the context of abalone fisheries and concluded that the diversity of diver exploratory be- haviors and the single dominant motivation of all divers; namely, maximization of catch rates, ensured complete coverage of every reef and efficient removal of abalone aggregations. Hence, the total catch should crudely reflect stock abundance in fully or near fully exploited stocks, although the relation may not be linear where spatial contraction of the stock occurs. Empirical evidence from Avoid Bay ( 15A) and Backstairs Passage (25A) show a cur- vilinear relation between survey densities and catch (Shepherd et al. 2001 ). However, in stable populations, there is no such corre- lation between catch and density (e.g. Tiparra Reef (21); r = -0.10. n = 18; Thorny Passage (I8C-F); r = -0.39, ;; = 11; Ward Island (9A,B); r = -0.07, ;; = 18; Hotspot (9C,D): r = 0.28, n = 17— see Shepherd et al. 1999). This result is not sur- prising, because surveys are not coordinated with fishing that may occur before or after an annual survey (Shepherd & Baker 1998). Thus, there is an asymmetric relation between catch and abun- dance in which the catch reflects abundance but only in declining populations. Nevertheless, it cannot be concluded that a declining catch is. of itself, proof of declining abundance. This is the well- known fallacy of qffininnf> the consequent. In formal terms, de- clining catch is a necessary, but not a sufficient, indicator of de- clining abundance. The alternative, testable hypothesis — that a population is stable but the catch declines for reasons unrelated to abundance — must now be considered. Two versions of this hypothesis have been proposed by indus- try and managers to account for the observed trends in the catch. First, it is suggested that changes in management since 1979 could have caused the patterns seen. The notable changes were: saleabil- ity of licences, size limit increase, and short-term clo.sure (Western Zone only), quota imposition, and quota reduction (Western Zone only), (see above). The argument ignores history and fails to dis- tinguish cause and consequence. The declines in catch from 1984 to 1988 in the Western Zone were real and pervasive and led to strong pressure from divers to introduce quotas and increase size limits and also led to increased government concern in 1988, which was followed by the quota reductions in 1989. Size limit increases could not have caused catch declines, because they were largely neutral in respect to yield (Shepherd & Baker 1998). Fur- thermore, if management measures up until 1989 caused the catch declines, they would not have continued unabated into the 1990s (see Fig. 5), but would have stabilized or oscillated within stable limit cycles as observed in the Far West (Keesing et al. 20(J0). The second version of the hypothesis is that the pattern of fishing has gradually changed over time, not as a response to declining abundance, but for reasons as diverse as: occasional presence of sharks, higher catch rates at offshore islands and its variant, the use of Global Positioning Systems, and preference for clear rather than dirty water. Decline of H. l\i:vig.\t.\ and Review of Management 839 We term both versions the undert'ishing hypothesis, because a necessary implication of each is that uihli'ifishiiii; of populations with declining catches is occurring. If a population is stable with a surplus available for capture, then the capture of less than the surplus will lead to an increase in population size. Moreover, the greater the decline in catch and the longer it has been declining, the greater would be the increase in numbers and biomass. until even- tually, the population would approach virgin densities, and Z would approach the natural mortality rate M. This hypothesis can be variously tested by research diver surveys, evidence of spatial contraction, and estimation of Z. from catch-curve analysis. The combined cumulative evidence in this paper of low densities from survey data, spatial contraction of populations and/or high Z values for twehe problematic reefs with declining catches (see Table 1) together point to widespread overfishing, inconsistent with the underfishing hypothesis. A satisfactory explanatory hypothesis must be able to withstand tests of internal coherence, external consistency, and predictive accuracy (Giere 1999). Since 1993 when major declines were first detected, overfishing of specific reefs has been conceded as data have accumulated, but only as exceptions to an otherwise stable fishery. A hypothesis is incoherent if every contradiction consti- tutes an (/(/ lux- exception to it. Is the underfishing hypothesis externally consistent with other statements? Historically, surveys of divers, increasingly accepted as a valuable source of information at a local scale (see Neis et al. 1999). have provided valuable evidence on divers" perceptions of the status of reefs. Surveys of divers of the Western Zone during 199.3 to 1999 (see Shepherd 1996 for a progress report) showed that divers" perceptions were cumulatively in unanimous agree- ment with inferences of decline drawn from catch data for every reef from Fishery Bay (16C) north to Sceale Bay (4D.E). except for Sceale Bay itself where opinions were divided. (Each diver was fainiliar with only a small subset of the reefs.) Divers judged the productivity of reefs by the frequency of their visits, a fact that has long been known. In an early perceptive submission to government the divers' association wrote: "When abalone start to gel hard to find in an area or a diver knows of somewhere he may do better, he moves on. and in this way rotates the areas and allows them to rest. The divers find that they are rotating their areas faster each \ ear. . . . Even though they are able to maintain their catch rate the areas are not supporting the previous year's effort."" (Letter from .Abalone Divers Association to Minister for Agriculture and Fish- eries dated 18 October 1976.) Thus, from the relatively early days of fishing, divers recognized that decision thresholds for ceasing to fish were a response to low abundance (see Shepherd et al. 2001 for further discussion). Last, if low abundance is not the basis of the decision to cease fishing a reef, no other credible cause has been proposed that coordinates the behavior of 29 idiosyncratic divers to underfish or cease fishing reefs in entire geographic regions for a decade or more. Thus, the underfishing hypothesis is counterbalanced or refuted by many other statements of divers and lacks credibility. Last, does the underfishing hypothesis have predictive accu- racy? As shown above, the hypothesis has been tested at twelve reefs and failed. Although testing the hypothesis at other reefs is an ongoing program, the underfishing hypothesis, so far, has failed every test of coherence, consistency, and predictive accuracy. It is. moreover, value laden, because it is proposed on behalf of those with a strong economic interest to maintain quota levels. A third hypothesis is that long-term environmental changes have affected population persistence in this abalone. Coastal wa- ters of South .Australia are intluenced by seasonal upvvellings from January-March each year, but this occurs well after the spawning season in October-November (Shepherd & Laws 1974) so is an unlikely agent directly affecting recruitment. Even if it affected adults, the upvvellings extend up to 60 km from the coast (Griffin et al. 1997) so the selective decline of populations in two zones and very few in the far west are difficult to explain. Moreover, the evidence of long-term temperature changes in the region is slight. Hence, this hypothesis seems untenable. In contrast, the overfishing hypothesis provides a common ex- planatory mechanism for the decline of catch on many disparate reefs and also suggests new areas for further research. One is that coastal topography influences population persistence via mecha- nisms of larval retention. This will be a useful criterion in the search for. and protection of, larval source habitats and in deter- mining sustainable exploitation rates of reefs. Management in Review The historical trend in management from government to fisher control, as exemplified in this fishery, is, in the view of Scott ( 1993). part of a broad endogenous movement, not unique to fish- eries, whose endpoint. at least for territorial stocks, is local self- management by fisher groups and cooperatives. Within fisheries, a general theory of government-community comanagement is emerging from the reviews of Jentoft (1989). Scott (1993). Pink- erton ( 1994). Wilson et al. (1994). Jentoft and McCay ( 1995). Sen and Raakjaer Nielsen (1996) and Hutton and Pitcher (1998). Al- most all of these reviewers argue that it is the role of government to set quotas and of the management committee to "decide the how. when, and where of fishing"" (Kesteven 1995). The present self-management system for the abalone fishery displays many of the positive features, such as flexibility, responsiveness, equity, and homogeneity of interest reducing internal conflict, considered by these authors and also lacks many negative features, such as deficits of information and excessive numbers of fishers. Fixing quotas is properly reserved to government. However, are there deficiencies in the present system as it has moved toward self- government? The intluence of government in the AMC seems to have de- clined relative to the power of the major interest groups since 1995. especially in matters relating to the conservation of stocks. Management has not applied the precautionary approach (PA) as mandated in the management plan (Zacharin 1997) nor addressed the maldistribution of effort by closures and other means. It might be argued that management has been unwilling to make hard de- cisions or that too much power has passed to an industry increas- ingly dominated by economic interests. It might be claimed, in defence of inaction, that the inherent uncertainty of assessment, the absence of satisfactory reference points (Shepherd et al. 2001 ). or the insufficiency of the evidence have not required action even under the PA. These issues are complex and controversial and will be considered elsewhere. Our purpose, here, is to highlight the basic problem and suggest a way forward. The problem of possible conflict of interest in the AMC could be met structurally by more effectively divorcing day-to-day management from the fixing of quotas. Refertal of the quota-setting function to an independent quota committee of experts, as operates in New South Wales, would facilitate independent and objective appraisal of the evi- dence. The problem of uncertainty of assessment mu.st be ad- 840 Shepherd and Rodda dressed by accuniulating more and stronger evidence, developing an agreed set of indicators for the fishery (Richards & Maguire 1998. Shepherd et al. 2001) and applying it to stocks at risk. In the meantime, application of the PA. incorporating understanding achieved from monitoring known, collapsed stocks, could mini- mize the risk of further overfishing. Precaulioiiary Approach (PA) Although this study does not establish conclusively the serial decline of the greenlip fishery as a whole, is the evidence sufficient to invoke the PA? The PA (reviewed by Foster et al. 2000) re- verses the onus of proof and states (Principle 15 of Rio Declaration as applied to fisheries) that where there are threats of serious damage, lack of full scientific certainty shall not be used as a reason for postponing measures to prevent overfishing. For data- poor fisheries (such as this one in which nothing is known of >90% of the stocks), respective FAO and U.S. national guidelines for application of the PA ( Garcia 1994, Garcia 1996. Restrepo & Powers 1999) advocate conservative measures, including: 1. setting catch limits based on qualitative judgements about stock status and stock rebuilding strategies for stocks be- lieved to be overfished, within time frames of 10 y (our italics); and 2. establishing provisional reference points by analogy to simi- lar and better known stocks. They also recommend that in situations of doubt, scientists analyzing management options should systematically analyze and highlight the most pessimistic scenarios. Dayton ( 1 998) argues that reversal of the onus of proof is the only mechanism by which action can be achieved, given the penchant for regulators to pro- crastinate by "creating imaginative, alternative explanations" or simply by asserting "that the data are incomplete" (Dayton 1998). Given the coherent. cumulati\'e evidence for a pervasive overfish- ing scenario there seems a proper case for application of the PA to this fishery. In the terms of Orensanz et al. (1998). overfishing "has to be considered as the default working scenario, even before being tested as a scientific hypothesis." Signs of serial decline of this greenlip fishery were first noted in 1995 (Shepherd 1996. Keesing and Baker 1998). and a mecha- nism for decline propo.sed by Shepherd and Baker (1998). The apparent continuing decline toward commercial extinction of many stocks finds a parallel in the decline of about a third of the Canadian salmon stocks (Riddell 1993). As emphasised by Pink- erton (1994). dependence of a fishery on fewer stocks is risky, because effort is increasingly focused on the remaining stocks and will increase the risk of their decline. On the other hand, conserv- ing more stocks is a buffer against future recruitment failure. which could simultaneously affect many stocks. For the conservation of sedentary invertebrate stocks. Jamieson (199.3) advocated: 1. very high minimum size limits; 2. unexploited refuges; 3. maintenance of high densities in many areas; and 4. specific management of metapopulation units. In this fishery, high size limits have been advocated by Shepherd and Baker (1998), creation of refuges and management of metapopulations by Keesing and Baker (1998), who also sug- gested multiple size limits and temporal closures to control yields. The establishment of small fishery reserves (Baker et al. 1996, Nowlis & Roberts 1999) in areas where relict subpopulations per- sist, supplemented by adult transplants, would allow aggregations of adults to accumulate and provide a source of larval recruits. A multiplicity of measures applied at a fine scale over large areas will be necessary to restore the former productivity of the fishery. The desired outcomes of comanagement are that sustainability, efficiency, and equity in managing the resource will be improved (Sen & Raakjaer Nielsen 1996). However, if the first of these fails, the rest no longer matter. Although management of this fishery may have enjoyed some success on the last two outcomes, the first is now at risk. ACKNOWLEDGMENTS We thank the many divers who patiently provided information during diver interviews from 1978 to 1999. Annette Doonan helped with the recalcitrant abalone data management database. Greg Howe. Dr. P. A. Breen, Dr. M. J. Tegner. and an anonymous referee kindly criticized early versions of the manuscript. The Fisheries Research and Development Corporation provided funds for the abalone database management system. REFERENCES Baker, J.. S. A. Shepherd & K. Edyvane. 19%. The use of marine fishery reserves to manage benthic fisheries, with emphasis on the South Aus- tralian abalone fishery. In: R. Thackway. editor. Developing Austra- lia's representative system of marine protected areas: criteria and guidelines for identification and selection. Canberra: Department of Environment, pp. 103-113. Brown. L. D. & N. D. Murray. 1992. Population genetics, gene flow, and stock structure in Haliotis nibni and Hiiholis Uievigiiki. In: S. A. Shep- herd. M. J. Tegner & S. A. Guzman del Proo. editors. Abalone of the world: biology, fisheries and culture. Oxford, UK: Blackvvell. pp. 24- 33. Chatterton. L. & B. Chatterton. 1981. How much political compromise can fisheries management stand? premiums and policy in closed coastal fisheries. Mm: Policy 5:1 14-134. Dayton, P.K. 1998. Reversal of the burden of proof in fisheries manage- ment. Science. 279:821-822. Foster. K. R.. P. Vecchia & M. H. Rapacholi. 2000. Science and the pre- cautionary principle. Science. 288:979-980. Garcia, S. M. 1994. The precautionary approach to fisheries with reference to straddling fish stocks and highly migratory fish stocks. FAO Fish- eries Circular No. 871. 76 pp. Garcia. S. M. 1996. Precautionary approach to capture fisheries and species introductions. FAO Fisheries Technical Papers No.350/2. .54 pp. Geen. G. & M. Nayar. 1988. Individual transferable quotas in the southern bluetln tuna fishery: an economic appraisal. Mar. Res. Econ. 5:365-87. Giere. R. N. 1999. Science without laws. Chicago: University of Chicago Press. Griffin. D. a.. P. A. Thompson. N. J. Bax. R. W. Bradford & G. M. Hal- legraeff. 1997. The 1995 mass mortality of pilchard: no role found for physical or biological factors in Australia. Mar. Freshwater Res. 48: 27_)2. Hanski, I. & M. Gyllenberg. 1997. Uniting two general patterns in the distribution of species. Science 275:397^00. Hardin, G. 1968. The tragedy of the commons. Science 162:1243-1248. Hersoug. B. & S. A. Ranes. 1996. What is good for the fishermen is good for the nation' Comanagement in the Norwegian fishing industry in the 1990s. Devolved and regional management systems for fisheries. EU Fisheries Project AIR-2-CT93-1392-PGXIV SSMA. 38p. (cited in Hut- ton & Pitcher 1998). Decline of H. iaevigata and Review of Management 841 Holl. R. D.. J. H. Lawton, K. J. Gaston & T. M. Blackburn. 1997. On the relationship between range size and local abundance: back to basics. Oikos. 78:183-190. Hatchings, J. A. & R. A. Myers. 199S. What can be learned from the collapse of a renewable resource? Atlantic cod. Gadus inoiinui. of Newfoundland and Labrador. Can. J. Fish. Aquat. Sci. 51:2126-214f). Hutlon. T. & T. J. Pitcher. 1998. Current directions in fisheries manage- ment policy: a perspective on comanagernent and its application to South African fisheries. In: S. C. Pillar, C. L. Moloney. A. I. Payne & F. A. Shillington, editors. Benguela dynamics. 5. Afi: J. Mar. Sci. 19: 471-+S6. Jainieson, G. S. 199.1. Marine invertebrate conservation: evaluation of fish- eries overe.\ploitation concerns. Ainer. Zoiil. 33:551-567. Jentoft. S. 1989. Fisheries comanagernent: delegating government respon- sibility to fishermen's organizations. Mar. Policy. 13:137-154. Jentoft, S. & B. McCay. 1995. User participation in fisheries management, lessons drawn from international e.xperiences. Mar Policy. 19:227- 246. Karpov. K. A.. P. L. Haaker. D. Albin. I. K. Taniguchi & D. Rusher. 1998. The red abalone. Halioii.s nifesccn.s. in California: importance of depth refuge to abalone management. J. SlicllfLsli Re.\. 17:863-870. Keesing. J. K. & J. L. Baker. 1998. The benefits of catch and effort data at a tme spatial scale in the South Australian abalone (Haliotis laevigata and Haliotis rubra) fishery. In: G. S. Jamieson & A. Campbell, editors. Proceedings of the North Pacific symposium on invertebrate stock assessment and management. Can. Si'ec. Puhl. Fish. Aqiiat. Sci. 125:179-186. Keesing. J. K.. K. C. Hall. A. M. Doonan & S. A. Shepherd. 2000. A data management and reporting system and historical analysis of catch records in the South Australian abalone fishery. FRDC Report. Project Number 94/167. South .Australian Research and Development Institute. Henley Beach. South Australia. Kesteven. G. L. 1995. Rejoinder, chaos, complexity, and community man- agement. Mar. Policy. 19:247-248. Kesteven, G. L. 1996. A fisheries science approach to worid fisheries or: three phases of an industrial revolution. Fisheries Res. 25:5-17. Lawton, J. H. 2000. Community ecology in a changing world. Oldendorf: Ecology Institute. Lewis. R. K.. S. A. Shepherd. P. Sluczanowski & G. Rohan. 1984. An assessment of the South Australian abalone resource. Adelaide: S. A. Department Fisheries. 59 pp. McShane, P. E. 1995. Recruitment variation in abalone: its importance to fisheries management. Mar. Freshwater Res. 46:555-570. McShane. P. E. 1998. Assessing stocks of abalone (Haliotis spp.l: methods and constraints. In: G. S. Jamieson & A. Campbell, editors. Proceed- ings of the North Pacific Symposium on Invertebrate Stock Assessment and Management. Can. Spec. Puhl. Fish. Aquat. Sci. 125:41^8. Neis, B., D. C. Schneider, L. Felt, R. L. Haednch. J. Fischer & J. A. Hutch- ings. 1999. Fisheries assessment: what can be learned from interview- ing resource users? Can. J Fish. Aquat. Sci. 56:1949-1963. NowlLs, J. S. & C. M. Roberts. 1999. Fisheries benefits and optimal design of marine reserves. Fish. Bull. 97:604-616. Orensanz. J. M., J. Armstrong, D. Armstrong & R. Hilborn. 1998. Crus- tacean resources are vulnerable to serial depletion — the multifaceted decline of crab and .shrimp fisheries in the Greater Gulf of Alaska. Rev. Fish Biol. Fisheries 8:117-176. Pinkerton. E. W. 1994. Local fisheries comanagement: a review of inter- national experiences and their implications for salmon management in British Columbia. Can. J. Fish. Aquat. Sci. 51:2363-2378. Prince. J. D. & S. A. Shepherd. 1992. The Australian abalone fisheries and their management. In: S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo. editors. Abalone of the world: biology, fisheries, and culture. Oxford, UK: Blackwell. pp. 407^26. Restrepo, V. R. & J. E. Powers. 1999. Precautionary control rules in U.S. fisheries management: specification and performance. ICES J. Mar. Sci. 56:846-852. Richards. L. J. & J-J. Maguire. 1998. Recent international agreements and the precautionary approach: new directions for fisheries management science. Can. J. Fish. Aquat. Sci. 55:1545-1552. Riddell. B. 1993. Spatial organization of Pacific salmon: what to conserve? In: J. G. Cloud & G. H. Thorgaard. editors. Genetic conservation of salmon stocks. Proceedings of a NATO Advanced Study Institute held June 2.3-July 5, 1991, Moscow. Series A. Life Sciences (USA) Vol. 248:23-11. Rodda, K. R.. J. K. Keesing & B. L. Fourcur. 1997. Variability in larval settlement of abalone on artificial collectors. Molluscan Res. 18:253- 264. Rose, G. A. & D. W. Kulka. 1999. Hyperaggregation offish and fisheries: how catch-per-unit-effort increased as the northern cod {Gadus iiiorhua) declined. Can. J. Fish. Aquat. Sci. 56 (Suppl. 1):1 18-127. Rosenweig. M. L. 1991. Habitat selection and population interactions: the search for mechanisms. Am. Nat. 137 (Suppl. ):S5-S28. Scott. A. 1993. Obstacles to fishery .self-government. Mar. Res. Econ. 8:187-199. Sen. S. & J. Raakjaer Nielsen. 1996. Fisheries comanagement: a compara- tive analysis. Mar. Policy. 20:405—118. Shepherd. S. A. 1996. Abalone stock assessment report: Western and Cen- tral Zones. West Beach. South Australia: SARDl Aquatic Sciences. 62 PR- Shepherd. S. A. & J. L. Baker. 1998. Biological reference points in an abalone (Haliotis laevigata) fishery. In. G. S. Jamieson & A. Campbell, editors. Proceedings of the North Pacific symposium on invertebrate stock assessment and management. Cuii. Spec. Puhl. Fish. Aquat. Sci. 125:235-246. Shepherd. S. A. & P. A. Breen. 1992. Mortality in abalone: its estimation, variability, and causes. In: S. A. Shepherd. M. J. Tegner & S. A. Guz- man del Prcio. editors. Abalone of the worid: biology, fisheries, and culture. Oxford, UK: Blackwell. pp. 276-304. Shepherd, S. A. & L. D. Brown. 1993. What is an abalone stock? Impli- cations for the role of refugia in conservation. Ccm. J Fish. Aquat. Sci. 50:2001-2009. Shepherd. S.A. & H. M. Laws. 1974. Studies on southern Australian aba- lone (genus Haliotis). II. reproduction of five species. Au.'it. J. Mar. Freshwater Res. 25:49-62. Shepherd. S.A. & D. Partington. 1995. Studies on southern Australian abalone (genus Haliotis). .\V1. recruitment, habitat, and stock relations. Mar. Freshwater Res. 46:669-680. Shepherd. S. A.. D. Lowe & D. Partington. 1992. Studies on southern Australian abalone (genus Haliotis). XIII. larval dispersal and recruit- ment. / E.'cp. Mar. Biol. Ecol. 164:247-260. Shepherd. S. A., K. R. Rodda & K. M. Vargas. 2001. A chronicle of col- lapse in two abalone stocks with proposals for precautionary manage- ment. J. Shellfish Res. 20:843-856. Shepherd, S. A., K. R. Rodda. T. Karlov. P. A. Preece & H. Williams. 1999. Abalone: fishery assessment report to PIRSA for the abalone fishery management committee. S.A. Fisheries Assessment Series 99/ 02. 21 pp. Figs 1-5. Steele, J. H. editor. 1981, Special issue. Living marine resources. Bio- scieiue. 3 1 . Wilson, J. A„ J. M. Acheson, M. Metcalfe & P. Kleben. 1994. Chaos, complexity, and community management of fisheries. Mar. Policx. 18:291-305. Wolanski, E. & W. M. Hamner. 1988. Topographically controlled fronts in the ocean and their biological influence. Science. 241:177-181. Zacharin. W. 1997. Management plan for the South Australian abalone fishery. South Australian fisheries management Series Paper No. 27. PISA and SARDI. Adelaide. 33 pp. Joiinial of Shellfish Research. Vol. 20. No. 2. ,S4.VS.Sf,, 2(101. A CHRONICLE OF COLLAPSE IN TWO ABALONE STOCKS WITH PROPOSALS FOR PRECAUTIONARY MANAGEMENT S.A. SHEPHERD,' * KATE R. RODDA,' AND KELLY M. YARGAS' '' 'South Australian Research ami Dcveloimient Institute. PO Box 120. Henley Beach. South Australia .ABSTltACT Two populations of grecnlip abalone. Hulioiis lacvificila. m Backstairs Passage and Avoid Bay. collapsed during 30 years of fishing. The former population was monitored from 1980 to 1999. and the latter from 1987 to 1999. Both populations showed strong .spatial contraction as their respective subpopulations in open habitat failed first compared with those in inlets, bays and around islands, where recruitment was higher. Stock-recruitment curves are gently sloping or indeterminate, indicating weak density- dependence in the one case and high variability in the other case. From data gathered during the decline of these populations we review \ arious fishery indicators which may be useful for monitoring greenlip abalone populations and for predicting collapse. Total catch was correlated with relative spawner abundance from survey data, and thus appears to be a useful indicator of abundance in this fishery. Other useful indicators are densities and size compositions, from which annual recruitment strength and total mortality can be derived, and survey evidence of spatial contraction of fished areas. Egg-per-recruit analyses are also useful to estimate the extent of departure from predetermined safe levels of egg production. Given the fuzzy knowledge of the status of populations, several indicators are best used in concert to determine management strategies. We propose a "troublespot thermostat" approach in which a suite of indicators, sequentially triggered off, call for increasingly severe management responses. In this way declines can be detected and arrested, and timely steps taken to avoid collapse. KEY WORDS: population ecology, stock-iecruilnient. spatial contraction, fishery collapse, fishery indicator, biological reference point, abalone fishery. Hallolis ku'vigaui egg-per-recruit INTRODUCTION While a legislative framework to achiexe the ecological sus- lainability of fisheries has been in place for some years in Aus- tralia, management has either been weak or has lagged seriously (Nicholls & Young 2000). The principle risks degenerating to empty rhetoric unless rigorous criteria are developed to detect and arrest fishery declines at an early stage. In particular robust bio- logical reference points to indicate declines have been developed for very few fisheries. The fishery for greenlip abalone (Haliotis laevigata Donovan I started in about 1966 in southern Australia and was controlled in the different States initially by input measures and then by quotas progressixely introduced in the 19S0s (Prince & Shepherd 1992). Greenlip populations have been slowly declining in the last two decades, first in Victoria and progressively in Tasmania and South Australia (Officer 1999, Shepherd & Rodda 2001). Management response to the declines has been tardy for many reasons. The absence of clear fishery indicators has promoted uncertainty, con- tinued high catch rates have engendered skepticism about over- fishing concerns, and the propagation of the notion that a stock- recruitment relation in abalone "is largely uncorroborated dogma" (McShane 199.S) has favored complacency. Catch data from the South Australian abalone fishery were collected froin 1 968 at a coarse scale of 37 x }il km, and from 1 979 at a finer (metapopulation) scale (Sluczanowski 1986, Keesing & Baker 1998, Shepherd & Rodda 2001 ). Data collection at the fine scale enabled researchers from about 1995 to follow long-term trends in individual populations. Annual monitoring of commer- cially exploited populations started in 1980 (Shepherd & Baker 1998) at a few sites and was gradually extended to many others. Two of these populations, monitored during part of their fishing *Corresponding author. E-maii address: shepherd. .scoresby@saugov.sa. gov.au "Present address: Rua Izidoro Chanoski. 81. 80820-580 Curitiba. Parana. Brasil. history, collapsed during the 1990s and this article describes their decline. We document the spatial contraction of the stocks, provide length compositional data and estimates of the total mortality co- efficient, Z, obtained during the decline phase, and the decline in density of spawners and recruits. We present spawner-recruit curves and evaluate various fishery indicators of possible use in abalone fisheries. In the absence of unambiguous indicators of overfishing, we suggest a protocol, which we term a trouble spot thermostat approach, by which population declines can be detected early, data assembled, and declines arrested by appropriate man- agement measures before final stock collapse when recovery be- comes problematic. In this article we use the terms population and stock to mean an infened metapopulation (Shepherd & Brown 1993. Shepherd & Rodda 2001). following the detailed genetic studies of Brown and Mun'ay ( 1992) on greenlip abalone. MATERIALS AND METHODS Site Descriptions Backstairs Passage Backstairs Passage lies at the entrance to Gulf St Vincent (see map in Shepherd & Rodda 2001 ). Greenlip abalone populations occur discontinuously on boulder or reef bottom over an area of -2 km" from several km east of Blowhole Creek to a little north of Cape Jervis on the Fleurieu Peninsula, South Australia (Fig. la). Rocky substratum is Cambrian greywackes and schists, tilted at a sharp angle providing microhabitat for juvenile abalone. Further offshore, boulders lie scattered on sandy or rocky bottom (Keesing et al. 2000). Abalone habitat is narrow at the southern end but broadens toward Fishery Bay and Cape Jervis where rocky bottom extends continuously or interspersed by seagrass patches to nearly 1 km from shore to a depth of 12-13 m. Highest densities occur near the entrances to the numerous inlets along the coast, in Fish- ery Bay and north of Lands End. The Passage is swept by tidal currents, which reverse twice daily due to diurnal and semi-diurnal components (Bye 1976, 843 844 Shepherd et al. Figure 1. (a) map of greenlip abalone hubilal in Backstairs Passage shi>Hing fislied areas in 1978 (liglit shading), and in 1999 (dark shading). Survey sites are marked with arrows, (b) map of greenlip abalone habitat off Point Avoid showing fished areas in 1978 (light shading) and in 1999 (dark shading). Very dark areas are intertidal reef platforms. Survey sites 1-7 are shown by crosses, and numbered 1 to 7 from left to right. CuRONicLt OF Collapse in Two Abalone Stocks 845 Bowers & Lennon 1990). During neap tides, whicli occur every 14 days, when tidal movement is minimal, water velocities are 35 cm sec"', increasing to 1.4 m sec"' seven days later during the spring tide (Carvallio & Bye 1996). The ma.ximum tidal excursion is > 6 km during a neap tide and > 20 1cm during a spring tide. Close inshore on the coast of Fleurieu Peninsula currents are slower, and in the inlets from Blowhole Creek to Fishery Bay and north of Lands End (Fig. la) there are reverse cunents or eddies. There is a mean annual residual flow out through Backstairs passage of -10 cm sec"', suggesting a net transport of larvae southeasterly, unless the net flow is reversed by southerly winds. Fixed sampling sites were established in 19S0 at the southern, central and northern sections of the population, namely at Blow- hole Creek in the south. Spaceship and Cooalinga Bays in the center, and in Fishery Bay and Lands End in the north. Rocky bottom was continuous at the sampling sites and divers either followed a constant direction or in the case of the southern sites followed the sand-rock interface. Avoid Bay Abalone habitat of gneissic or calcrete substratum extends over an area of about 14.7 km" from the 30 m isobath south of Pt Avoid between Price and Golden Islands mainly at depths of 15-25 m northwards toward Pt Avoid, with smaller local populations noi'th- west to northeast of Pt Avoid and northeast of Golden L (Fig. lb) (Keesing et al. 2000). From drogue studies at similar coastal sites we estimated that maximum tidal currents which reverse twice daily with a north/ south flow are -20 cm sec"', and that the maximal tidal excursion is ~4 km. These estimates inay be extreme due to the dampening effect of swell on water currents (McShane et al. 1988). Seven survey sites were established at 250-m intervals between Price and Golden Islands at 17-21 ni depth (Fig. lb). Moil iloriiig Procedures Monitoring of greenlip populations started in Backstairs Pas- sage in 1980 and in Avoid Bay in 1987. For each population we chose representative sites according to the known distribution of the populations and subject to time and depth constraints. The sampling was systematic to minimize spatial variability, and sam- pling sites were fixed by reference to land marks in Backstairs Passage and by radar or Geographic Positioning System (GPS) in Avoid Bay. We used the free-swimming, timed swim technii|ue (Kenchington 1978, Shepherd 1985) to measure the relative abun- dance and size distribution of greenlip in situ. Divers measured all abalone encountered with an underwater measuring gauge, in which data points of shell length (SL) are imprinted onto plastic (Shepherd 1985), within a swathe of Im for 10 min. Surveys were done in the summer and fall from November to May. The number of replicate 10-min swims per site varied accord- ing to the conditions and number of divers available. In Backstairs Passage the mean replication was 5.3 (range 3-9) in the southern section, 8.1 (range 3-16) in the central section and 6.7 (range 2-12) in the northern section. In Avoid Bay there were four rep- licate lO-minute swims per site. For Fishery Bay we lacked data for the years 1982, 1985 and 1988. For the purpo.se of calculating mean values for Fig. 6 we estimated recruitment (i.e. the density, D,, of the 3+ age class) in those years from the density, Dj, of the 4-1- age-class in the following year at the site from the equation D, = 0.44 Dj' " (range of D4 values 0-1.9/100 m") derived empiri- cally over 10 y from all sites combined in the central section where we had the most data. For the same years we interpolated the density of the spawning stock as the mean of the values for the preceding and succeeding year. The above equation has no impli- cations about survival of the D, age-class due to the effect of changing sighting probabilities of recruits over time (Shepherd 1990). Data Sources and Analysis Catch data, presented in total (in-shell) weights, (TW) are de- rived from the catch and effort monitoring system in place since 1968. Backstairs Passage is in the Central Zone where there are six licensed divers and Avoid Bay is in the Western Zone with 23 divers. Some poaching in Backstairs Passage was reported occa- sionally during the 1980s but not since. Backstairs Passage has been predominantly fished by only one or two divers since the mid 1980s. About half the catch from Avoid Bay was graded by size during processing. The three grades, in terms of number of meats per unit weight, are approximately (after metric conversion): 1: <4/kg; 2: 4-6/kg: 3: >6/kg. The data are given in full in Keesing et al. (2000). Juvenile H. Uicvii;citii are cryptic to an age of two and a half years then gradually emerge with size (Shepherd 1990, Shepherd & Partington 1995), so we used the i+ age-class as an index of 82 84 86 YEAR Figure 2. Catcli/efforl and survey historv of the Backstairs Passage greenlip fishery showing: (a) change in total mortality coefficient, Z, over time; (b) mean adult and recruit (3+ age-class) densities for sur- vey sites from 1979-1998; (c) change in coefficient of variation of mean recruitment at survey sites over time, with a trend line fitted to the data: (d) total catch (t TW) and catch rate in kg/hr from 1968-1999. 846 Shepherd et al. recruitnient strength, and estimated its size range at the date of sampling from the known growth rate. The growth rate of H. laevigata is hnear during the first 4 y of life in Backstairs Passage (21.3 mm/yl and at Avoid Bay (22.2 mm/yr) (unpublished data. Shepherd & Triantafillos 1997) and we assumed a uniform birth- date on 1 December each year. We used knife-edge separation from the 2-i- and 4+ year classes to estimate their abundance in length-frequency samples. In order to estimate density from the timed swim data we used the equation relating area covered, A. to swimming time. T, num- ber of abalone measured, N. measuring time. /;;, and diver power, ;-. given by Shepherd (1985): A = Ti-- inrN Swell and algal cover were low at all our sampling sites so we modified the original equation given by Shepherd (1985) to ex- clude effects of swell and algal cover on diver power. As the individual power and measuring time of some of the divers used in the surveys were not calibrated (although all were experienced in abalone survey work) we used mean values of r = 20 m~ min"', and )n = 4 sec (see Shepherd 1985) to calculate the area covered per 10 minute swim. Mean recruitment density (D,) and adult densities were estimated for each site and the mean for all sites gave the mean annual population values. Recruitment variability was measured as the coefficient of variation (the standard dexiation of the mean annual recruitment values for all sites divided by the mean) and e.xpressed as percent- age values (Pimm 1991). Estimates of the orisinal reef area fished for abalone were obtained from information supplied by commercial divers in 1977 to 1979 to one of us (SAS). We drew maps during interviews with three divers for Backstairs Passage and four for Avoid Bay. and the resultant map was the common area described by the divers for Avoid Bay, but a composite map for Backstairs Passage. v\hich included our own survey data. Total mortality rates, Z. were obtained from catch-curve analy- sis. Commercial samples of shells (>450 shells per site), taken throughout the fishing area, were aged (O'Loughlin & Shepherd unpublished data) with the shell-aging technique of Shepherd and Triantafillos (1997) and an age-length key prepared for each site. Length-frequency distributions were then converted to age- frequency distributions and Z estimated from the Chapman and Robson (CR) (I960) estimator, preferred by Dunn et al. (1999) to regression methods. The equation used is; CR = Ln(l 4-a- l/n)/a where a is the mean age (above the recruitment age) and n is the sample size. The analysis assumes a stable age-length structure. This seems reasonable for this abalone. given the absence of evi- dence for inter annual variability in growth rates (Shepherd & Heam 1983). To estimate the density of the spawning stock in Backstairs Passage for the missing year, 1996. we used the empirically de- rived curvilinear relation between total catch and adult density gi\en in Results. Similarly we estimated the density of the spawn- ing stock in Avoid Bay for 1988 from the respective equation for Avoid Bay (see Results). In this paper the spawning (or adult) stock is the number of LENGTH (mm) Figure 3. Backstairs Passage. Length frequency data from annual research surveys for all sites combined. The 1993 survey was interrupted by bad weather. Chroniclh of Collapse in Two Abalone Stocks 847 individuals s4 years of age i.e. >85 mm for Backstairs Passage and >9() mm for Avoid Bay. RESULTS liaikMairs I'msage Tlie calc'li liistory of the Passage shows high initial catches in the first three years, typical of a virgin stock, followed hy a sharp decline in the catch, and then irregular pulse fishing until the population collapsed after about 1990 (Fig. 2). The catch rates remained high, with occasional falls until 1998 when it fell sharply to a low rate. The fall was due. according to the main diver then working there, to his replacement by an inexperienced diver. Survey data from 1980 to 1999 provide more detail about the mechanism of decline. Length-frequency data pooled for all sites within years (Fig. 3) showed little change during the period of monitoring, and estimates of Z derived from thein also varied little around a mean of 0.37 (s.e. 0.01, range 0.32-0.41 ), with no down- ward trend as density declined (Fig. 2a). Mean adult density de- clined more or less continuously until 1993. and then le\eled out. but by 1999 had fallen even lower (Fie. 2). Recruit densilv varied in an apparently cyclic manner but showed a long-term downward trend, and recruitment variability increased, as shown by the sig- nificant ((- = 0.65; F < O.O.S). long-term, upward trend in the coefficient of variation (CV). Local populations at survey sites did not decline uniformly (Fig. 4). The Fishery Bay subpopulution failed first following a steep decline in 1985 to 1987. and by 1993 greenlip abalone were virtually extinct there, except on inshore reefs where small num- bers still persisted even in 1999. The spatial contraction of the subpopulation in Fishery Bay and that east of Blowhole Creek (Fig. la) was -115 ha. The remaining populations declined more or less linearly, at a mean annual rate of 3.8% for the southern sites and 5.8% for Lands End. The difference in decline rates was significant (/ = 3.4; P < 0.05). Spatial contraction also occurred in these subpopulations as they retreated toward the coast, but we have not attempted to estimate the areas involved. The total catch. C in tonnes TW. smoothed with a 3-y running average, shows a curvilinear relation with adult density. D in num- bers/m- (Fig. 5). The equation of best fit is: C = 10.03 -(-2.46 In D (R- 0.36; />< 0.05) 1986 N = 329 ^n-n„n[]nn[][]nn LENGTH (mm) > 1987 N = 410 I': nnnoli Q. 2 _ nnnnnnnnf] 1 L_ LENGTH (mm) 1991 z '* N = 1698 s '° — r ■ q: 4 -^nnfinOnn 1 1 On 1 j, J, LENGTH (mm) LENGTH (mm) Figure 3. Continued. 848 Shepherd et al. 0.8 _ 0.7 E 2 0.6 >- t 0.5 w z g 0.4 O < 0.: 0.1 0 0.5 0 * Blowhole + Cooalinga + Spaceship • Fishery Bay a Lands End 11 D Blowhole + Cooalinga + Spaceship ■ Fishery Bay Q Lands End in 80 s: «4 8(1 88 m 94 111 W Figure 4. Backstairs Passage. Survej densities of (a) adults in nuni- bers/ni", and (bl recruits in numbers/lOdnr. from 198(1-1999 In three sections of the population: Blowhole Creek to Spaceship Bay, Fishery Bay, and Lands Fnd. Assuming that the Backstairs Passage populaliiin is isolated from other populations i.e. without e.xtemal larval sources, and that larvae from component subpopulations enter a common larval pool (see Discussion), we can examine recruitment in relation to spawner abundance. Stock-recruitment (SR) curves for (a) data from ail sites combined and (b) for the Fishery Bay population alone (Fig. 6) show lower curves for Fishery Bay compared with that for all sites, consistent with the more frequent recruitment failures at Fishery Bay. We used the same annual densities of spawning stock for both curves on the assumption that the Fishery Bay subpopulation was linked by larval dispersal to other sub- populations in Backstairs Passage. The equations of best fit are: R,„= 1.(37 -t- 0.28 In D (R' = 0.H)) for all sites combined and: R, = 0.77-l-0.28LnD (R- = 0.16) 6X1 O o o Figure 5 vs adult 9 8 7 6 • • • • • • 5 ^.^^ • y = 2.457SLn(x) + 10035 4 r'- 0.3552 3 • 2 1 n 0.1 0.2 0,3 0.4 0.5 Adult Density (m" ) for the Fishery Bay subpopulation. In these equations R^,,, and R, are the mean density of the 3-1- age-class for all sites, and for Fishery Bay, respectively, in numbers/m" and D is the mean adult density for all sites in numbers/m~. The data points for 1977 to 1999, for which density was estimated from the catch, are shown in Figure 6 as hollow symbols. These estimated points were ex- cluded from the regression analyses, but are included to show that strong recruitment from the earlier years of the fishery may have produced more elevated or even dome-shaped curves. Note that no strong recruitment occurred at mean densities of <0.1.'i /m". Avoid Bay We do not have data at a reef scale for the catch from 1968- 1970 when virgin stocks were removed, but the subsequent mean annual catch in Avoid Bay for years 1971 to 1978 was 49.1 t (Keesing et al. 2000). From 1979 to 1983 the annual catch was on average {4% higher than the previous eight years and then de- clined sharply from 1984 until 1998 when the catch was ~1 t (Fig. 7), giving a mean annual rate of decline of 5.8% since 1979. Catch rates in Avoid Bay remained more or less constant, save for a slight decline in 1996. Surveys of the Bay started in 1987 when the catch was already well in decline. Length-frequency data (Fig. 8) show little change over the period, and estimates of Z similarly are within a narrow range of 0.50 to 0.67 with a slight but non-significant downward trend (see Fig. 9a). Adult densities, initially two to three times higher at the inshore Sites (1—4) than at the offshore sites (5-7). declined more or less continuously at both, so that from 1996 offshore densities were only -0.05. m"' and too low to be worth fishing. Density of recruits showed only a weak downward trend over time at Sites 1-4. while at Sites 5-7 recruitment was low, variable, and not detectable after 1995. The coefficient of variation of recruitment (Fig. 7b) increased until 1993 and thereafter fluc- tuated widely. Although survey data are lacking for the most pro- ductne period of the fishery (1971-1983) it is clear that adult populations failed at the offshore sites earlier than they did at the inshore ones. Spatial contraction of the stock was very marked (Fig. lb), with a total loss of fishable habitat of-14 km". The onl\ places now fished are those close to Price and Golden Islands and near Pt Avoid. Plots of changes in proportional abundance of three size grades o o 01 Q U as 18 1.6 1,4 1,2 1 0,8 0,6 0.4 0,2 0 y"0.2809Ln(x)+l 068 1^ = 0.1851 Fishery Bay y = 02831 Ln(x) + 0 7728 r' =0 1634 0,1 0,2 0,3 0.4 0.5 Adult Density (m" ) . Backstairs Passage. Plot of total greenlip catch in tonnes T\V density in numbers/m" with curve of best fit. Figure 6. Backstairs Passage. Stock-recruitment curves for all sites combined and for Fishery Bay alone with curves of best fit. The mean annual density of recruits (3+ age-class) is plotted against the mean density of adults 3 y earlier. Recruit densities are: solid triangles {all sites), solid circles (Fishery Bay) and hollow symbols for the years 1977-9. Chroniclh oi- Collapse in Two Abalone Stocks X4y I ] recruits Siles 5-7 \/^^\ recruils sites 1-4 — A — adults sites 1-4 — ■ — adults sites 5-7 recruits sites 1-4 0 3 § >- 0= S 79 80 81 82 83 84 85 86 8? 89 90 91 92 93 94 95 96 97 98 YEAR Figure 7. Catch histon of grcenlip abalone in \\mA Ba> shoHing: (a) mean adult (number.s/m-| and reiruit (numbers/100 nrl densities for 1979-1999; (b) cbange in coetncient of variation of mean recruitment for sites 1-1; (c) total catch (t T\\ ) and catch rate in kg/hr for 1979- 1999. over time (Fig. 9b) show that the largest grade (Grade 1 ) increased significantly (P < 0.001). and the smallest (Grade 3) decreased significantly (P < 0.05) over the 21 y, but this may have been contributed to by the size limit increase in 1 984. If we exclude data prior to 1 984 then the same trends in grades are apparent but they are not significant (P > 0.05). The total catch. C. smoothed with a 3-y running mean, shows a curvilinear relation with adult density. D (Fig. 10). The equation of best fit is: C = 20.57 + 6.76 In D (R" = 0.63; P < 0.05) Plots of mean recruitment density for nearshore and offshore sites vs mean adult density (Fig. 1 1 ) show very variable recruitment in relation to adult density at nearshore sites and generally poor re- cruitment for the few offshore data points. The shape of any SR relationship appears indeterminate so we simply show the line of best fit for each group of sites forced through the origin. The curves imply weak density-dependence. DISCUSSION The two studies provide vital information on the process of decline of these abalone populations. Two effects are conspicuous: the increase in recruitment variability and the spatial contraction of the populations. Increasing recruitment variability implies that a population will be more susceptible to overfishing during years of poor recruitment (Myers & Pepin 1994, Shepherd & Baker 1998). Hence this effect may itself contribute to the spatial contraction of a population because subpopulations with poor recruitment will run a higher risk of increasing depletion, fragmentation and ex- tinction. However, independendy of any effect of recruitment \ari- ability, spatial contraction may occur through differential recruit- ment or juvenile mortality throughout the habitat of the original population. We have no data on juvenile mortality but the topo- graphic features defining population boundaries suggest a me- chanism for differential recruitment acting at the pre-settlement stage. Shepherd et al. ( 1992) noted that topographic coastal or benthic features, such as inlets and rock pinnacles, increased local recruit- ment strength of greenlip abalone. McShane (1995) reviewed the effects of hydrodynamics on larval transport and concluded that in open waters abalone larvae may be transported large distances but near reef of high relief can be retained near the natal site. Shepherd and Rodda (2001 ) reviewed other studies showing that islands and bays generate sheer zones and eddies that concentrate larvae. The studies of Shepherd et al. (1992) and Rodda et al. (1997) both suggested transport distances for greenlip larvae of several km from their natal site in places of tidal currents. Studies of the behavior of greenlip larvae (Madigan et al. 1997, Madigan 2000) show that during the trochophore and early veliger stage, together lasting about four days. lar\ae tend to swim upward. Hence they are likely to be distributed throughout the water column and trans- ported passively in wind-driven and tidal currents for several days before settlement (Sasaki & Shepherd 1995). Larval behavior at settlement is also important. McShane (1992). McShane (1995). Nash (1992) and Shepherd and Parting- ton (1995) all showed a significant inverse relation between re- crtiitment strength and water movement. McShane (1992) consid- ered that water turbulence could keep larvae away from suitable sites, thus preventing them from making contact with the substra- tum and responding to settlement cues. Thus several hydrody- namic factors and larval behavior can interact to cause spatial differences in recruitment in abalone habitat. In greenlip abalone spawning is believed to occur mainly dur- ing October to November at around the neap tide (Rodda et al. 1997. Shepherd & Daume 1996. and K. Rodda pers. comm.). so that larvae would remain pelagic to within two days of the begin- ning of the spring tide when currents are maximal. Thus tidal currents would transport larvae for much of the tidal cycle from neap to flood tide. Hence we may conclude that both in Backstairs Passage and in Avoid Bay. where tidal excursions are at about the same spatial scale as the abalone populations, larvae originating from any subpopulation could settle in any other subpopulation. and are likely to be concentrated by hydrodynamic forces near inlets, headlands and islands. Wind-driven currents and swell could bias up or down any estimated distances of larval transport by tidal currents. Our recruitment data are consistent with the abo\e predictions. In Backstairs Passage (Fig. la), the sites of strongest recruitment, at Lands End, Blowhole Creek. Cooalinga Creek and Spaceship Bay were all in near-shore eddies out of the strong current flowing through the Passage. In Fishery Bay, where recruitment failed earliest, the habitat was of low relief and water currents were very strong. In Avoid Bay (Fig. lb) populations persisted near the head- land of Point Avoid and around Price and Golden Islands, but largely disappeared from the low-relief habitat with stronger tidal currents between the islands. In summary, the three effects, in- creasing recrtiitment variability, concentration of larvae by topo- graphic features and an inverse relation between settlement strength and water velocity, can together explain the differential declines observed. 830 Shepherd et al. 1987 N = 696 ^JkJlll \k Dl CD O — I- to ci 1969 N = 636 -^ lhw>i t^ CO O 1990 N = 1156 f>jrrnlll 14 1992 2 12 N = 398 a; 10 0 , n WV 1 1 UK 5 5 z 12 I 10 ^ 1991 N = 423 g 12 u] 2 10 jJllllW r^ CO O — 6 IT) T T f- OD o m 1993 N = 269 JilllilL d in ti in T T V Figure 8. Avoid Bay. Length-frequtnc> data from annual research surveys for all sites combined. Catch-Ahundaiice Relations The relation between total catch and abundance is complex and can be confounded by sampling contingencies, spatial contraction of a stock, and diver behavior. We consider them in turn. Survey densities will vary according to the time of survey in relation to fishing and according to the representativeness of the survey sites in relation to the population response to fishing. Re- searchers tend to choose sampling sites at shallower, more acces- sible places close to shore, where subpopulations may be more resilient to fishing and hence may not be representative of the whole population. Spatial contraction will also confound a direct catch-abundance relation. As abundance is a function of density x area, if the area occupied by the population decreases, overall abundance will de- cline faster than a decline in density alone at fixed sites, unless the sites are systematically located throughout the spatial extent of the population. Lastly, divers operate in a visual fishery and a range of human responses may influence their catch. The dominant motivation of commercial divers is to maximi/e their catch rates (Keesing & Baker 1998. Prince & Hilborn 1998). Divers use their knowledge to fish the largest aggregations, and then progressively smaller ones, and finally cease fishing when densities are low. The deci- sion threshold prompting departure from a reef is principally de- termined by the catch rate, but other factors such as distance from homeporl and depth also play a role (Shepherd & Rodda 2001 ). Divers accept lower catch rates close to homeport and in shallower water (Keesing & Baker 1998: McShane 1998). Divers also have different spatial strategies, and include: risk-takers, divers who fish more reefs and take risks lookmg for new concentrations of abalone:/o//ouer.s, who are content to revisit familiar grounds: and those with intennediate strategies (Prince & Hilborn 1998, Mc- Shane 1998). Surveys of South Australian divers by one of us (SAS) have shown that most (-70%) divers tend to fish a group of 5-10 populations with which they are most familiar. But a few. -20% in the sample, (the risk-takers) show more exploratory be- havior and fish up to 20 populations. The remainder has interme- diate strategies. The different diver strategies ensure that all ex- ploitable stocks are periodically visited and fished. The implication from these behaviors is that, in the long term, a stable or fluctu- ating total catch implies stability and a declining total catch, as also declining effort, a declining abundance. While other factors, such as price, weather, and diving conditions, may also affect the be- havior of divers there is no evidence that they do other than in- crease the noise, without changing the underlying relation. The catch-density curves for the two sites support the hypoth- esis that the total catch crudely reflects abundance. The relation is noisy and curvilinear as predicted, and suggests that total catch is likely to show hyperstability during the early decline of a popu- lation, but when abundance has substantially declined the total catch will decline more steeply. Stock-Recruitment (SK) Relations Caddy ( 1999) noted that in spatially structured populations of sedentary species stock recruitment relations depended on local densities not stock size. Hence SR relations can strictly only be fully described by a family of curves covering every local area with a different recruitment pattern (see Wing et al. 1998. Oren- Chronicle of Collapse in Two Abalone Stocks 851 14 5,2 t 6. 1994 N = 533 firi jYi_ SI 2 0 .^.tfwrif fii s s 14 -1 u ut 12 - f4 10 - u. t- H ■ UJ 2 - a 0 I 1996 N = 750 ■V 14 LENGTH (mml 1999 if; 12 N = 280 S'O lYl ^ 8 ■ 1- 6 z P.1 rfilll Ilimuw„ LENGTH (mm) Figure 8. Continued. sanz & Janiieson 1998 for a discussion of SR patterns in spatially structured stocks). Given the spatial liinitations of our data set and the small time series for Avoid Bay. the pair of .SR curves pre- sented for each of the two sites can at best be taken as vaguely representing (a) resilient abalone habitats where larvae are retained tsj 1978 1980 1982 1984 1986 1988 1990 1992 YEAR 1994 1996 1998 Figure 9. .Avoid Bay. (a) change in total mortality coefficient, Z, from 1987-19W; (b) change in proportions of three grades of meat weight in the total catch, expressed as a percentage, from 1978-1998. Grade 1 is < 4 meats/kg: grade 2 is 4-6 meats/kg; grade 3 is >6 meats/kg. by hydrodynamic effects, and (b) vulnerable habitats where larval settlement and/or survival are less successful. As discussed previ- ously the models assume a common larval pool. Whether habitats are vulnerable because of intrinsic factors such as their hydrodynamics or predation, or because of attractant biotic factors, such as the presence of conspecifics, is speculative. On the latter possibility there are four field studies in addition to some laboratory experiments on abalone which suggest that the presence of conspecifics plays an important role in attracting set- tling larvae. Tutschulte (1976) removed adult abalone from ex- perimental areas of 25 m~ and found that recruitment in those areas failed. Prince et al. (1987) and Prince et al. (1988) did similar experiments with the same results. Recently Andrew et al. ( 1998) did the reverse and placed adult abalone in crevices previously without them. Thev found a 10-fold increase in recruitment. In 0 15 0 2 0 25 ADlLTDENSrr\ (N.m •) Figure 10. .Avoid Bay. Plot of total catch (running mean) vs mean adult density with curve of best fit. 852 Shepherd et al. Adiilt Density (Nos per m ) Figure 11. Plots of mean annual recruit density vs adult density for Avoid Bay. Recruit densities are: solid circles for Sites 1—1 and solid squares for sites 5-7 and hollon symbols for years in which adult densities were estimated from the catch. Trend lines for Sites 1—1 and 5-7 are show n. See caption to Figure 6 for 3-y lag time procedure used in the plots. Western Australia T. Adams (pers. comm.) did a similar experi- ment with greenlip abalone at a site from which abalone had dis- appeared some years before and later recorded good recruitment there. Conversely Shepherd et al. (1992) did both additions and removals of adult greenlip at experimental sites with moderate to strong tidal currents but found no effect of local adults (but a strong effect of topography) on recruitment. More recently Daume et al. ( 1999) showed experimentally that H. hwvigata settled gre- gariously in the presence of older conspecifics consistent with the slime trail hypothesis suggested by earlier workers (McShane 1992). Whatever the mechanism for the failure of recruitment, the SR curves demonstrate the complexities of spatially structured stocks. Unlike the single isolated population in Waterloo Bay with its restricted larval dispersal (Shepherd & Partington 1995). these populations are not internally homogeneous in terms of recruit- ment strength. Hence SR relations will vary according to habitat and any analysis based on a few survey sites must be a gross over-simplification. We do not pretend that the SR curves presented here represent the true shape of any underlying SR relation. As argued by Hilborn (1986). Underwood and Fairweather (1989) and Koslow (1992), the concept of a deterministic relation between stock size and recruitment is not easily applicable to invertebrate stocks with high fecundity and correspondingly high, strongly density-dependent mortality during the early life history. Nevertheless, the lower and likely flatter. SR curves for vulnerable habitats at each site suggest the Achilles heel of these populations and the mechanism of their spatial contraction. Flat SR curves imply weak density dependence such that even small increases in exploitation reduces the recruit- ment rate below the replacement rate and thus leads eventually to population collapse within that habitat (Shepherd & Gushing 1980). Prince and Guzman del Proo (1993) modeled populations of declining Mexican abalone stocks and found that a similar flat SR relationship best explained the decline of those stocks. Perusal of the SR curves suggest that recruitinent overfishing in Backstairs Passage must have occurred about 1 980. because no strong recruit- ments occurred after that year. In Avoid Bay overfishing probably occurred before 1987 when surveys began, but. in the absence of information about ttie rate of spatial contraction, conclusions about the timing of overfishing events are speculative. Maintenance of productivity of these greenlip populations at the catch levels ex- isting up till about 1985 will thus depend on maintaining abalone densities in vulnerable habitats at much higher levels than those required for resilient habitats. Fishery Indicators (FI) The concept of FI and the related one of biological reference points (B.R.P. ) have emerged in recent years to enable managers to continuously evaluate the state of fished stocks and employ strat- egies to prevent their collapse (Myers et al. 1994). The approach is precautionary and accepted both in this fishery (Zacharin 1997) as in international instruments (Caddy 1998, Caddy 1999). Recent developments in this approach include improved decision-making under uncertainty, and consensual adoption of decision rules once liinits are approached or exceeded. Nowhere is this more urgent than in abalone fisheries which have depressingly collapsed in nearly all major fisheries in the world (Shepherd & Baker 1998), and now partially in southern Australia (Shepherd & Rodda 2001 ). Empirical data from these two stocks pro\'ide a testing ground of potential indicators for this fishery. With modification we be- lieve it may be widely applicable to other abalone fisheries. First we note that abalone must be classified as data-poor fisheries, because they comprise many independent stocks for which good fisher-independent data even in the best conditions are seldom available. So what is needed are very simple indices that can be derived "with only fuzzy knowledge about stock levels and recruit- ment curves' (May et al. 1978). Such indices are to be sought from ( 1 ) historical catch and effort data provided by fishers to govern- ment agencies, (2) survey data, and (.3) parameters such as the total mortality coefficient, Z, and dependent analyses such as egg-per- recruit (EPR). Historical Catch and Effort Data It is almost trite that catch rates (CPUE) offer little or no help as an FI (Breen 1992, McShane 1998. Orensanz et al. 1998). This is because abalones have a clumped distribution, and after fishing they tend to re-aggregate in favored habitats (Shepherd & Parting- ton 1995). Because divers search for aggregations and give up searching at low densities, searching time is not proportional to overall density and catch-rates show hyperstability (Sluczanowski 1986). CPUE thus depends very largely on diver behavior as de- scribed previously. Shepherd & Partington (1995) noted that as the fishery in Waterloo Bay collapsed CPUE increased, because divers ceased searching the whole area of the bay and concentrated their effort in the few remnant places where abalone aggregated. A similar behavior may explain the constancy of CPUE in Backstairs Passage (except 1998-1999) and Avoid Bay despite the long-term decline in abundance. Although CPUE is of dubious value, we propose that a time series of total catch data is a crude but useful indicator of abun- dance in declining populations, providing it is collected at a fine scale (Shepherd & Brown 1993, Keesing & Baker 1998). The proviso is important because, if the catch data are collected at too coarse a scale, serial depletion occurring at a finer scale is masked. The use of total catch itself is the concept of maximum constant yield (MCY) currently used in New Zealand (Annala 1993, Francis 1993). It is "the maximum constant catch that is estimated to be sustainable, \\ ith an acceptable level of risk, at all probable future Chronicle of Collapse in Two Abalone Stocks 853 levels of biomass". Shepherd and Baker ( 1998) used the constancy of historic catches for specific metapopulations as empirical evi- dence on which to base estimates of safe fishing mortality rates (see later in text). Shepherd and Rodda (2001 ) examined declines in the catch of many other greenlip populations in this fishery and showed that alternative explanations for their decline were so im- plausible that in this fully exploited fishery, a declining catch should be accepted as evidence of declining productixity. Changes in the size grading of catches, like other measures based on the size structure, probably reflect changes in the exploi- tation rate (Andrew et al. 1997, Worthington et al.l998) and may therefore be useful to monitor. However, Keesing et a!.(2()()0) noted great ambiguity in the signal, due to changes in diver be- havior and spatial contraction of a stock. The exploitation rate, Z, may remain constant or even fall during the decline of a stock, (Shepherd & Baker 1998), so that neither it nor measures related to size structure are definitive indicators. As in the two examples in this article, populations may collapse with little or no change in CPUE, size grading, length-frequency distributions or Z. Survey Data If changes in the spatial extent of declining abalone populations are common, as our data for other declining populations suggest (Shepherd & Rodda 2(X)I ). then surveys need to be designed ex- plicitly to cover the spatial extent of the populations to be moni- tored. If the coverage is partial (as at both sites described here) or biased, then the results are, to that extent, diminished. In research surveys other valuable information such as data on density, the size composition of the emergent population, and the aggregational structure (Shepherd & Partington 1995, McShane 1998) can also be obtained. While spatial information and density estimates to- gether clearly provide the best FI of a stock, cost, especially in deeper water, prevents their widespread or optimal use. Data from Backstairs Passage and Waterloo Bay show a high risk of recruit- ment failure below mean densities of 0.2 (±0.05 l/nr. The Avoid Bay data suggest even higher densities of -0.3/nr should be main- tained, given that the population was already in decline when surveys started. However, a collapse threshold density was only evident for the Waterloo Bay SR curve, possibly due to high larval retention and the more constant recruitment there. The existence of vulnerable habitats, with more variable recruitment, complicates the picture and indicates the need for even higher densities as a buffer against recruitment failure. Size compositional data are valuable because they can be used to estimate recruitment either in length-based inodels (Worthing- ton et al. 1998) or in age-based analyses. According to Shepherd el al. (1984) a time series of recruitment is the most unambiguous indicator of abundance in a population. Lastly, there are measures based on the frequency and size of aggregations in surveys. Such measures certainly change as a stock collapses, and they may prove to be \aluable, as implied by McShane (1995), and shown by Shepherd and Partington (1995) for the Waterloo Bay population. Egg-Per-Recruit Analyses and Mortality Coefficients The problem for precautionary management is to ensure that sufficient spawning biomass is conserved to minimize the risk of recruitment failure. One approach, termed a constant relative es- capement strategy, is to set a safe maximum fishing mortalitv, F (Mace & Sissenwine 1993, Garcia 1996). This leaves a variable spawning biomass from year to year depending on annual recruit- ment strength. Shepherd and Baker ( 1 998 ) proposed this for green- lip abalone and suggested F^,,,. for small stocks and Fjoc; for large stocks, the difference being due to the observed greater vulner- ability of small slocks to overfishing. Here the notation F,,,.,. for example, means that level of F which reduces egg production in an EPR analysis to 50% of the unfished value. This suggestion was based on the analysis of a number of exploited greenlip stocks combined with a global review of nine other abalone fisheries. During the major period of fishing the egg production conserved was 32% in Backstairs passage and 37% in Avoid Bay (O'Loughlin & Shepherd unpublished data). Both were clearly inadequate, and demonstrate the need to reduce F substantially. The recovery of a stock is very unlikely to occur if fishing is permitted to continue in depleted stocks, because divers will con- tinue to fish down the shrinking remnant subpopulation until near extinction. The failure of other collapsed greenlip stocks to recover after 15-20 y (Shepherd & Brown 1993, unpublished data) sug- gests that depensation at low stock densities may occur. An alternative strategy is to have a constant absolute escape- ment, as advocated by Caputi ( 1992) for a prawn fishery, a strategy that would require routine estimation of annual recruitment strength. This would enable better exploitation of strong year- classes, and may be well suited to short-lived species like prawns with highly variable recruitment. For longer-lived species like aba- lone in which the catch typically comprises individuals aged six to ten years (Shepherd & Avalos-Boija 1997, Officer 1999), and recruitment strength is very costly to measure, this strategy would confer little advantage and in any event would be practically im- possible to implement. Mace ( 1994) suggested a threshold biomass based on the shape of the SR relationship i.e., that biomass at which recruitment is one half of the maximum possible. While this is readily determined from the Ricker SR curve for Waterloo Bay (Shepherd & Parting- ton 1995), it is indeterminate for populations like Backstairs Pas- sage and Avoid Bay where SR curves vary spatially and may show a continually ascending curve as in vulnerable habitats. Any cho- sen threshold from such a curve is dubious because gains in re- cruitment can always be expected from every increase in spawning stock. Besides, an SR curve will be unknown for all but the very few stocks that have been monitored to collapse. Other BRPs and FIs are reviewed by Caddy (1998) but either require better data sets than are usually available for abalone or are not appropriate for sedentary stocks. Use of Multiple Indicators in a Fuzzy System The above analysis suggests that three kinds of measures, catch data, density, and possibly patch size, data, and F4o_5qc5, values, are the best candidates for use as FIs in the greenlip fishery. They need to be evaluated by simulation and, if show n robust, should then be incorporated into the management protocols described below. Given the lack of information available for indi\ idual populations, except for catch data, but the capacity to quickly gain more as required in specific trouble spots, we propose a 'trouble spot ther- mostat' approach (Caddy 1998. Caddy 1999). First, populations with declining catch (trouble spots) are identified from historic catch data and targeted for assembly of archival information and field survey. Because total catch data lend to be hyperstable early 854 Shepherd et al. TABLE 1. Trouble spot thermostat protocol for decision making after detection of early declines in individual greenlip abalone populations. Read table from bottimi to top bv analogy to a rise in temperature. The table uses a suite of fishery indicators (left), ordered from levels I to 4 according to increasingly severe criteria in a troubled population. Kach level leads by an arrow in column 2 to an agreed management response in column 3, and triggers either a demand for more detailed research information or a sterner management response. The ultimate measure is a closure of the troubled population if the criteria are not met. Metapopulation Fishery Indicator Agreed Management Response Close fishery. Establi-.h recovery plan. 4 Total catch decline >6U'*' OR Recruitment decline >20% over last 4 years 3 F > F4„^,-F„|,; according to site" -^ Take corrective measures to reduce F (reduced season, closure or increased size liniit) 2 Estimated spatial decline >50% o OR Mean survey density <0.25 m"- Do EPR analysis. Commence annual recruitment survey. OR Z>0.4 I Total catch decline >309r ' ^ Survey spatial e.xtent of population and compare with historical records. Do catch curve analysis. ' Catch decline per map code area since 1985 (Western Zonel. I9S8 (Central Zone) and 1989 (Southern Zone), being the years in which quotas were introduced. " F409t for populations of initial productivity >I0 t/y'. F,,,,,.; for populations with initial productivity <10 t/v ' (from Shepherd & Baker 1498). Even higher % values of F should he conserved in populations already in serious decline. ' Adopt this alternative if corrective measures at level 3 have not been taken; otherwise adopt the second alternative. in a population decline we propose a small decline (30%) as a trigger for initial action. As the indicators are generated and trig- gered at level 2 (Table 1 ), management action is required to com- mission more data on the trouble spot in question. Then, if the indicator, Fj,,,;. for example, is triggered, a more severe tiianage- ment response is dictated. The alternative indicators at several steps reflect the uncertainty inherent in any single indicator. The ultimate response is to close the troubled population to fishing and initiate a stock recovery program. The consequences of overexploitation of a poorlv managed stock are decline and ultimately collapse, so the risks in procras- tination are great and will generate large economic losses. A ther- mostat approach would stimulate timely coirective action, and restore the fishery at a small present cost but with great long-term benefit. ACKNOWLEDGMENTS This paper is dedicated to the memory of Kevin L. Branden who competently and enthusiastically led the dive team for 12 years during these surveys. We also thank other participant divers: A. Bunnell, P. Clarkson, M. Clark, S. Clarke, B. Davies, A. Dal- getty, C.H. Deane, B. Foureur, N. Holmes, J.K. Keesing, D. Lowe. I. McGraith. M. Miller, A. Mower. P. Preece. S. Rodda, J. A. Turner, J. R. Turrubiates, J. Van der Broek, G.W, Wright. Com- mercial divers helpfully provided information on their fishing ar- eas. The RV Ngerin crew assisted divers and manned the diving dinghies in Avoid Bay. G. Wright prepared Figure 1 and Dr P. A. Breen made many constructive comments. Dr S.A. Guzman del Proo and an anonymous referee criticized the manuscript. The Fisheries Research and Development Corporation provided funds during the early surveys. LITERATURE CITED Annala. J. 1993. Fishery assessment approaches in New Zealand's ITQ system. In: G. Kruse, editor. Proceedings of the International Sympo- sium on management strategies for exploited fish populations. .Anchor- age, 21-24 October 1992. pp. 791-806. Andrew, N. L.. D. G. Worthington & N. L. Brett. 1997. Si/e structure and growth of indiv iduals suggest high exploitation rates w ithin the tlsherv lor blacklip abalone. Haliolis nihra. in NSW. Australia. Molliiscan Res. IS:27.s-287. Andrew, N. L., D. G. Worthington. P. A. Brett & N. Bentley. 1998. In- teractions between the abalone fishery and sea urchins in New South Wales. Final Report to Fisheries Research and Development Corpora- tion. Sydney: N.S.W. Fisheries Research Institute. 64 pp. Bowers, D. G. & G. W. Lennon. 1990. Tidal progression in a near-resonant system: a case study from South Australia. Est. Coastal Shelf Sci. 30:l7-.34. Breen, P. A. 1992. A review of models used for stock assessment in abalone fisheries. In: S. A. Shepherd. M. J. Tegner & S. A.Gu/nian del Proo, editors. Abalone of the world: biology, fisheries and culture. Oxford: Blackwell. pp. 253-275. Brown. L. D. & N. D, Murray. 1992. Population genetics, gene flow, and stock structure in Haliotis rubra and Haliolis laevigata. In: S. A. Shep- herd. M. J. Tegner & S. A. Guzman del Proo. editors. Abalone of the world: biology, fisheries and culture. Oxford: Blackwell. pp. 24-33 Bye. J. A. T. 1976. Physical oceanography of Gulf St Vincent and Inves- tigator Strait. In: C. R. Twidale, M. J. Tyler & B. P. Webb, editors. The Natural history of the Adelaide region. Adelaide: Royal Society of S.A. pp. 143-160 Caddy. J. 1998. A short reviev\' of precautionary reference points and some proposals for their use in data-poor situations. Rome: FAO Fisheries Technical Paper 379. 29 pp. Caddy. J. 1999. Fisheries management in the twenty-first century: will new paradigms apply? Rev. Fish Biol. Fish. 9:1^3. Caputi, N. 1992. Spawning stock-recruitment relationships. In: D. A. Han- cock, editor. Recruitment Processes. Australian Society for Fish biol- ogy workshop. Hobart. 21 August 1991. Bureau of Rural Resources Proceedings No. 16. Aust Canberra: Govt Publishing Service, pp 142- 146 Carvalho. O., Jr. ct J. Bye. 1996. The first recorded successful Backstairs Passage swim. 5. Aust. Geographical Journal 95:70-74. Chronicle of Collapse in Two Abalone Stocks 855 Chapman. D. G. & D. S. Robson. 1%U. The analysis of a catch curve. Biometrics 16:354-368. Daume. S, M., W. J. Woelkerling & S. Brand-Gardner, mw. .Settlement of abalone larvae {Halions laevigata Donovan) in response to non- geniculate coralline red algae (Corallinales. Rhodophstai. ,/. £v/). Mar. Biol. Ecol. 234:125-134. Dunn. A.. R. I. C. C. Francis. & L J. Noonan 1499. The sensitivity of some catch estimators of monality to stochastic noise, error, and selectiv ity. N.Z. Fisheries assessment Research Document 99/5. 23 pp. Francis. R. I. C. C. 1993. Monte Carlo evaluation of risks for biological reference points used in New Zealand fishery assessments. In: S. J. Sinith. J. J. Hunt & D. Rivard. editors Risk evaluation and biological reference points for fisheries management. Can. Spec. PkIjI. Fi.sh. Aquat. Sci. 120. pp.221-230 Garcia. S. M. 1996. Stock-recruitment relations and the precautionary ap- proach to management of tropical shrimp fisheries. Mar. f/f.v/ni'. Rc.s. 47:43-58. Hilbom, R. 1986. A comparison of alternative harvest tactics for inverte- brate fisheries. In North Pacific Workshop on Stock Assessment and Management of Invertebrates, editors G. S. Jamieson & N. Bourne. Can. Spec. Piihl. Fish. Aquat. Sci. 92:313-317. Keesing, J. K. & J. L. Baker. 1998. The benefits of catch and effort data at a tine spatial scale in the south .Australian abalone fHaliotis laevigata and H. rubra) fishery. In Proceedings of the North Pacific Symposium on Invertebrate Stock Assessment and Management. G. S. Jamieson & J A. Campbell, editors. Can. Spec. Publ. Fish. Aquat. Sci. 125:179- 186. Keesing. J. K.. K. C. Hall A. M. Doonan & S. A. Shepherd. 2000. A data management and reporting system and historical analysis of catch records in the South Australian abalone fishery. Report of South Aus- tralian Research and development Institute to FRDC. Report No. 94/ 167. Kenchington. R. A. 1978. Visual surveys of large areas of coral reefs. In: D. R. Stoddart & R. E. Johannes, editors. Coral reefs: research methods UNESCO. Paris, pp 149-161. Koslow, J. A. 1992. Fecundity and the stock-recruitment relationship. Can. J. Fish. .Aquat. Sci. 49:210-217. Mace. P. M. 1994. Relationships between common biological reference points used as thresholds and targets of fisheries management strate- gies. Can. J. Fish. Aquat. Sci. 51:110-122. Mace. P. M. & M. P. Sissenwine. 1993. How much spawning per recruit is enough? In: S. J. Smith, J. J. Hunt & D. Rivard, editors. Risk evaluation and biological reference points for fisheries management. Can. Spec. Publ. Fish. Aquat. Sci. pp. 101-118. Madigan, S. 2000. The biology of larval and juvenile abalone. Ph.D. thesis. Flinders University of South Australia. Adelaide. Madigan, S.. P. Hone & J. Havenhand. 1997. Genetic variation in the sw imming heha\ lour of Haliotis laevigata and H. rubra abalone larvae. Third International Abalone symposium. California Sea Grant. Ab- stract. May. R. M., J. R. Beddington & J. R. Horwood. 1978. Exploiting natural populations in an uncertain worid. Math. Biosc. 42:219-252. McShane, P. E. 1992. Eariy life history of abalone: a review, pp. 120-138. In: Abalone of the worid: biology, fisheries and culture. S. A. Shep- herd, M. J. Tegner &. S.A. Guzman del Proo. editors. Oxford: Black- well. McShane. P. E. 1995. Estimating the abundance of abalone: the importance of patch size. Mar. Freshw. Res. 46:657-662. McShane, P. E. 1998. Assessing stocks of abalone iHaliotis spp. i: methods and constraints. In: G. S. Jamieson & J. A. Campbell, editors. Proceed- ings of the North Pacific Symposium on Invertebrate Stock Assessment and Management. Can. Spec. Publ. Fish. Aquat. Sci. pp. 41-48. McShane. P. E.. K. P. Black & M. G. Smith. 1988. Recruitment processes in Haliotis rubra (Mollusca: Gatropoda) and regional hydrodynamics in south east Australia imply localised dispersal of larvae. J. E.xp. Mar. Biol. Ecol. 124:175-203. Myers, R. A. & P. Pepin. 1994. Recruitment variability and oceanographic stability. Fish. Oceanogr. 3:246-255. Myers, R. A., A. A. Rosenberg. P. M. Mace, N. Barrowman & V. R. Restrepo. 1994. In search of thresholds for recruitment overfishing. ICES J. mar. Sci. 5 1 : 1 9 1 -205 . Nash, W. J. 1992. An evaluation of egg-per-recruit analysis as a means of assessing size limits for blacklip abalone [Haliotis rubra) in Tasmania. In: S. A. Shepherd, M. J. Tegner & S. A.Guzman del Proo. editors. Abalone of the worid: biology, fisheries and culture. Oxford: Black- well, pp. 318-340. Nicholls. D. & T. Young. 2000. Australian fisheries management and ESD-the one that got away? Env. and Planning Law Journal 1 7:272- 293. Officer, R. A. 1999. Size limits for greenlip abalone in Tasmania. Univer- sity of Tasmania. Technical Report No. 5. pp. 46 Orensanz. J. M.. A. M. Parma & M. A. Hall. 1998. The analysis of con- centration and crowding in shellfish research. In: G. S. Jamieson & A. Campbell, editors. Proceedings of The North Pacific Symposium on Invertebrate Stock Assessment and Management. Can. Spec. Publ. Fish. Aquat. Sci. pp. 143-157. Orensanz. J. M. & G. S. Jamieson. 1998. The assessment and management of spatially structured slocks: an overview of the North Pacific Sym- posium in Invertebrate Stock Assessment and Management. In: G. S. Jamieson & A. Campbell, editors. Proceedings of The North Pacific Symposium on Invertebrate Stock Assessment and Management. Can. Spec. Publ. Fish. .Aquat. Sci. pp. 441^59. Pimm. S. 1991. The balance of nature? Ecological issues in the conserva- tion of species and communities. Chicago: University of Chicago Press. 434 pp. Prince. J. D. & S. A. Guzman del Proo. 1993. A stock reduction analysis of the Mexican (Haliotid) fishery. Fisheries Res. 16:25^9. Prince. J. D. & R. Hilborn. 1998. Concentration profiles and invertebrate fisheries management. In: G. S. Jamieson & J. A. Campbell, editors. Proceedings of the North Pacific Symposium on Invertebrate Stock Assessment and Management. Can. Spec. Publ. Fish. Aquat. Sci. pp. 187-196. Prince. J. D. & S. A. Shepherd. 1992. The .Australian abalone fisheries and their management. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo. editors. Abalone of the world: biology, fisheries and culture. Oxford: Blackwell. pp. 407-426. Prince. J. D.. T. L. Sellers. W. B. Ford & S. R. Talbot. 1987. Experimental e\idence for limited dispersal of haliotid larvae {Haliotis: Mollusca: Gastropoda). J. E.xp. Mar Biol. Ecol. 106:243-263. Prince. J. D., T. L. Sellers, W. B. Ford & S. R. Talbot. 1988. Confirmation of a relationship between the localised abundance of breeding stock and recruitment for Haliotis rubra Leach (Mollusca: Gastropoda). J. E.xp. Mar Biol. Ecol. 122:91-104. Rodda. K. R., J. K. Keesing & B. L. Foureur. 1997. Variability of larval settlement of abalone on artificial collectors. Molluscan Res. 18:253- 264. Sasaki. R. & S. A. Shepherd. 1995. Lar\al dispersal and recruitment of Haliotis discus hannai and Tegula spp. Mar Freshw. Res. 46:689-695. Shepherd, J. G. & D. H. Gushing. 1980. A mechanism for density- dependent survival of larval fish as the basis of a stock-recruitment relationship. / Cons. Int. Explor. Mer. 39:160-167. Shepherd. J. G.. J. G. Pope & R. D. Cousens. 1984. Variations in fish stocks and hypotheses concerning their links with climate. Rapp.P-v. Cons. Int. Explor Mer. 185:255-267. Shepherd, S. A. 1985. Power and efficiency of a research diver, with a description of a rapid underwater measuring gauge: their use in mea- suring recruitment and density of an abalone population. In: C. T. Mitchell, editor. La Jolla: Diving for Science. American Academy of Underwater Science. Pp. 263-272. Shepherd, S. A. 1990. Studies on southern Australian abalone (genus Hali- otis). XI. Long-term recruitment and mortality dynamics of an unfished population. .Aust. J. Mar Freshw. Res. 41: 475—192. Shepherd. S. A. & M. Avalos-Borja. 1997. The shell microstructure and 856 Shepherd et al. chronology of the abalone Hallalis cdrnigarti. Mulliisccin Res. 18:197- 208. Shepherd, S. A. & J. L. Baker. 1998. Biological reference points in an abalone (Haliotis laevigata) fishery. In: G. S. Jamieson & A. Campbell, editors. Proceedings of the North Pacific Symposium on Invertebrate Stock Assessment and Management. Can. Spec. Publ. Fisli. Aquat. Sci. pp. 235-245 Shepherd. S. A. & L. D. Brown. 1993. What is an abalone stock' Impli- cations for the role of refugia in conservation. Can. J. Fish, .\qiiat. Sci. 50:2001-2009. Shepherd, S. A. & S. Daume. 1996. Ecology and sui^ival of juvenile abalone in a crustose coralline habitat in South Australia. In: Y. Wa- tanabe, Y. Yamashita & Y. Oozeki, editors. Survival strategies in early life history stages of marine resources. Rotterdam: Balkema. pp. 297- 313. Shepherd. S. A. & W. S. Beam. 1983. Studies on southern Australian abalone (genus Hatiolis). IV. Growth of H. laevigata and H. rubra. .Aiist. J. Mar. Freslnv. Res. 34:461-475. Shepherd, S. A. & D. Partington. 1995. Studies on southern Australian abalone (genus Haliotis). XVI. Recruitment, habitat and stock rela- tions. Mar. Freshw. Res. 46:475-492. Shepherd, S. A. & K. R. Rodda. 2001. Sustainabilily demands vigilance: evidence for serial decline of the greenlip abalone fishery and a review of management. J. Shellfish Res. 20:829-841. Shepherd, S. A. & L. Triantafillos. 1997. Studies on southern Australian abalone (genus Haliotis). XVII. A chronology of H. laevigatii. Mol- liiscan Res. 1 8:233-246. Shepherd, S. A., D. Lowe & D. Partington. 1992. Studies on southern Australian abalone (genus Haliotis). XIII. Larval dispersal and recruit- ment. J. Exp. Mar. Biol. Eeol. 164:247-260. Sluczanowski, P. R. 1986. A disaggregate model for sedentary stocks: the case of the South Australian abalone. In: G. S. Jamieson & N. Bourne, editors. North Pacific workshop on stock assessment and management of invertebrates. Can. Spec. Publ. Fish. Aquat. Sci. pp. 393^01. Tutschulte, T. C. 1976. The comparative ecology of three sympatric aba- lones. Ph.D. thesis. San Diego: University of California. Underwood, A. J. & P. G. Fairweather. 1989. Supply-side ecology and benthic marine assemblages. Trends in Evolution and Ecology 4: 16-20. Wing, S. R., L. W. Botsford & J. F. Quinn. 1998. The impact of coastal circulation on the spatial distribution of invertebrate recruitment, with implications for nianagement. In: G. S. Jamieson & J. A. Campbell, editors. Proceedings of the North Pacific Symposium on Invertebrate Stock Assessment and Management. Can. Spec. Publ. Fish. Aquat. Sci. pp. 285-294 Worthington, D. G., R. C. Chick, C. Blount, P. A. Brett & P. T. Gibson. 1998. A final assessment of the NSW abalone fishery in 1997. NSW Fisheries. Fishery Resource Assessment Series No. 5. 67pp. Zacharin, W. 1997. Management plan for the South Australian abalone fishery. South Australian fisheries management series paper No. 27. PISA and SARDl. Adelaide. 33 pp. .Iciinial III Sluiljhh Ri'scarch. Vol. 20. No. 2, X57-.S(il. 2001. QUANTIFYING THE PHYSICAL AND BIOLOGICAL ATTRIBUTES OF SUCCESSFUL OCEAN SEEDING SITES FOR FARM-REARED JUVENILE ABALONE (HALIOTIS MIDAE) S. DE WAAL* AND P. COOK Zoology Department. University (if Cape Town. Private Bag. Rondehosch 7701. Cape Town. South Africa .ABSTRACT In short term experiments of under 3 mo, carried out in McDougail's Bay. on the northwest coast of South Africa, sur\ i\ al of seeded juvenile abalone (Haliolis inidae) has been shown to be directly linked to both the size of the abalone at seeding and to the physical attributes of the seeding site. The presence of the sea urchin ( Parenchinu.s angulosus) has been shown to be insignificant in terms of short-term survival of juvenile H. midae. Within the context of selecting the right seeding site attributes, short-term survival was shown to be up to a minimum average of 59% for animals with an average size of 26.61 mm (SD ± 1.33mm), compared to a minimum average of 24% for animals with an average size of 13.87 mm (SD ± 1.73 mm). The fact that the presence of urchins played no significant role in the survival of juvenile abalone, ranging in size from 13 to 27 mm, changes the idea of what the characteristics of successful seeding sites might be. A positive correlation has been established between habitat consisting of stacked boulders, of diameter less than 30 cm. and abalone survival. KEY WORDS: seeding size, urchin exclusion, site characteristics, habitat, abalone INTRODUCTION Seeding experiments using juvenile abalone. Haliotis inidae. carried out by Sweijd et al. ( 1998) on the Nortliwest coast of South Africa showed that some potential existed for the successful in- troduction (seeding) of this species along this stretch of coastline. To optimize success, it is necessary to understand both the recruit- ment and ecological dynamics of the species and to determine the life cycle stages that would be most successful in terms of both survival and growth (Tegner & Butler 1989. Davis 1995). One of the most fundamental aspects of the successful recruitment of H. Midae is the association between the juveniles of the species and the urchin, Pareclunus angulosus. This association is considered so strong that the presence of the urchin is considered a driving factor in the population dynamics of this abalone species (Day 1998, Tarr 199.5, Tarret al. 1996). The association between urchin species and abalone species is reported worldwide, and is not restricted to the South African species (Inoue 1987, Kojima 1981, Rogers-Bennet & Pearse 1998, Tegner 2000). Indications are that naturally occurring H. midae. up to the average size of 35 mm, are found almost exclusively under urchins (Day 1998, Tarr et al. 1996). While there are also indications that long-term survival of juvenile abalone is linked to the utilization of urchins as refuge, the relationship between artificially seeded H. midae and their as- sumed habitat has not been investigated or quantified. A range of long-term experiments was designed to begin addressing these issues. Initial results, collected over a period of 6 mo. using aba- lone ranging in size between 8 and 18 mm. showed very low- survival rates, ranging from 0 to 12%. This was in contrast to the results of the previous experiment (Sweijd et al. 1998). in which animals ranging in size between 8 and 14 mm showed an average 6-mo survival rate of iO'Jc. seeded in what was regarded as similar seeding environments. While survivors did show a very high as- sociation with sea urchins, it became clear that the precise physical attributes of the seeding site, in terms of shelter and protection against both possible predation and against adverse ocean condi- tions, appeared to be an important factor in survival (Emmett & Jamieson 1988, Schiel 1993). The initial results were confusing; *Correspondiiig author. E-mail address schaIkdO@freemail.absa.co.za the effects of rough sea conditions could not be differentiated from the effect of abalone seeding size or the influence of the physical or biological attributes of the seeding site. To address these issues, a series of short-term experiments were carried out in McDougalTs Bay. Port Nolloth. seventy kilometers south of the Orange River at 29"'17'S and 16"52'E. The experimental area is in a bay which, while protected from direct wave action by a reef, is exposed to strong cunents and surges. An important advantage of this site was that it allowed constant access by divers to the seeded abalone. The average depth of the experimental area was between 2 and 4 m. with the physical substratum on which the animals were seeded, being predominantly loose, overlapping boulders, of differing sizes. The experiments were designed to address three basic ques- tions: 1. the importance of sea urchins to the survival of seeded ju- venile abalone. 2. the influence of abalone size on short-term survival. 3. the type of habitat which offers juveniles short-term protec- tion and shelter. METHODS The seeding procedure used for all experiments was the same. The traps used for seeding were those described by Sweijd et al. 1 1998). consisting of lengths of PVC piping of decreasing diam- eter, cut in half and placed inside each other. A gap of approxi- mately 1.5 cm was left between pipe sections, and the pipes were then mounted on a flat plastic base. This base was then mounted on a steel plate, which acted as a v\'eight that kept the trap stable on the ocean floor, allowing the animals to leave and return to stable refuge (Schiel 1993). The juvenile abalone were gently brushed into the semi-cylindrical tubes making up the trap. The traps were then closed at both ends with perforated plastic covers that allowed a through-flow of water, and placed into well-aerated, seawater holding tanks. After approximately 30 niin. the traps were placed in plastic bags, sealed and placed in Styrofoam containers with sponge-wrapped freezer blocks. The containers were closed and transported to the seeding sites. The trip normally took approxi- mately 20 min. On arrival at the seeding sites, the traps were immediately taken out of the containers and placed into rock 857 858 DE Waal and Cook TABLE 1. Urchin densilies and juvenile abalone survival. Urchins Vt Found 9f Found Period % Density Under Under Site (days) Survival Per M- Urchins Boulders 1 17 5.67 6.79 35.29 64.71 -) 17 5.67 0 0 100 3 17 15.67 8.91 27.66 72.34 4 17 3.33 0 0 100 5 23 28.67 0 0 100 6 23 25.33 4.51 39.47 60.53 7 23 28.33 0 0 100 8 23 38 8.59 29.82 70.18 Total number of urchins were counted in a 3 ni radius around each trap. Period in days indicates time between seeding and sampling. relative percentage cover of different size boulders and the amount of exposed area between boulders. Percentage cover by encrusting con'aline algae was estimated for the experimental area. From four of these sites, all the sea urchins, in a radius of three meters from the trap, were removed. At the other four sites, all those urchins collected from the first four sites were placed directly surrounding the traps. The average density of urchins around each of these traps was 7.20 m" (SD ± 2.02). Each trap contained 300 animals with an average size of 12.24 mm (SD ± 1.62 mm) The traps were moni- tored every few days to make sure that those from which the urchins had been removed remained clear of urchins. The sites were sampled in two groups, after 17 days and 23 days respec- tively, with each group having two sites with urchins and two sites without urchins (Table 1). On sampling, the number of urchins in a 3 m radius around each trap was counted and the percentage utilization of urchins and boulders by abalone juveniles was cal- culated for each site, together with abalone survival. pools or in shallow water. The traps were then moved underwater to the seeding sites. The traps used in the experiment were kept closed until they were in position, and then opened. All the traps were monitored for approximately 20 min after they were opened. Other than a few- animals, which were sitting on the plastic covers, no animals ex- ited from the traps and no predators were attracted to the traps in the time immediately following opening. Traps were also observed on a daily basis, and this revealed that by far the greatest majority of animals did not begin to leave the traps until well into the night of the first day. and in many cases not until the second or third day . Diver efficiency and sighting probabilities were considered. However, due to the fact that only one person was responsible for the searches, the very small numbers of experimental animals used, and the restricted experimental areas, they were disregarded. The sighting probabilities mentioned by Shepherd (1498) were used to weigh the survival figures for the different sized abalone used in the size experiments. Vrchin Exclusion Expiriment To test the short-term effects of the presence of sea urchins on juvenile abalone the following experiment was carried out. Eight experimental sites were selected in typical west coast kelp beds in which there was an abundance of loose boulders, with a biota including local species of crustose corraline algae and sea urchins. The sites were chosen to resemble each other as closely as pos- sible, both in terms of physical appearance and in depth (2-4 m). In a 3-m radius around each trap, an estimate was made of the The Effect of Abalone Size at Seeding The same procedure as above was followed, except that no urchins were removed from any area. Along with percentage cover by encrusting corraline algae (e,stimated per square meter for the experimental area), urchin densities were calculated around each site. Eight experimental sites were again selected to resemble each other as closely as possible. 200 animals were placed in each trap. 100 larger ones, with an average size of 26.61 mm (SD ± 1.33 mm), and 100 smaller ones, with an average size of 13.87 mm (SD ± 1.73 mm). Placing both the large and small animals together in the traps meant that all the animals were exposed to exactly the same conditions on release. The traps were sampled in two groups of four, the first group after one month, and the second group after two months (Table 21. Quantifying the Physical Attributes of the Seeding Sites In an attempt to quantify the physical attributes of each site, a very basic approach was taken. All the experiments mentioned above were treated in the following way. In a 3-ni radius around the trap the percentage cover of the surface was estimated accord- ing to the following categories: 1. boulders with a diameter of greater than .50 cm. 2. boulders with a diameter of between 30 and 30 cm. 3. boulders with a diameter of less than 30 cm. 4. exposed area, i.e. flat rocky or sandy area that provides little or no protection for juvenile abalone in terms of cracks or crevices that can be used as a microhabitat. TABLE 2. Seeding size and juvenile abalone survival. Site Period (days) Size (mm) Std. Dev (mm) Survival % Size (mnil Std. Dev (mml Survival % 1 30 28.07 3..56 62 12.25 1.63 26 2 30 24.64 ■^ t 55 12.25 1.63 29 3 30 24.64 3.2 68 12.25 1.63 33 4 30 25.53 2.9 54 12.25 1.63 5 5 60 27.5 3.47 66 15.44 2.14 36 6 60 27.5 3.47 59 15.49 2.14 23 7 60 27.5 3.47 50 15.49 2.14 19 8 60 27.5 3.47 33 15.49 2.14 24 Attribl'tks o(- Abalone Seeding Sites 859 One problem with the categories listed above is that boulders are obviously not completely round and while the estimate is car- ried out on two-dimensional basis, the site is three-dimensional. In an attempt to compensate for this potential problem, the same person did all the estimates and, as a result, there is a degree of consistency throughout the series of experiments. Survival was correlated with the different classifications of site description at the end of each experiment, along with percentage cover of crus- tose corraline algae species, uichin density, percentage utilization of urchins by abalone per site, abalone seeding size and the ratio between percentage boulder cover to percentage exposed area. RESULTS Urchin Exclusion Experiment At all the sites, sampling was performed around the traps until no more animals were found. During destructive sampling, all boulders were picked up. all the urchins were removed and all crevices and cracks were inspected. The majority of the animals were found within a radius of 2 m of the trap and very rarely were any animals found further than 3 m from the trap. Survival rates give an indication of minimum survival, because some of the surviving animals may not have been found. Destructive sampling can. hov\ever. be regarded as effective enough to compare experi- mental results between sites. The first four traps were placed on a day in which sea condi- tions were becoming progressively worse, and the last four traps were seeded during a sustained period of calm sea conditions. This is reflected in the survival rates, with the second batch of animals all surviving in much greater numbers than the first batch. The first four sites were sampled after 17 days and the second four after 23 days. Using the Mann-Whitney U Test, no significant difference was found in the survival of abalone between those sites from which all the urchins had been removed and those sites from which they were still present, p = 0.67 (Fig. 1). The Effect of Abalone Size at Seeding The first group of four sites was sampled after one month and the second group after two months. Again destructive sampling was used (Table 2). Using the Mann-Whitney U Test, it was found that there was a significant difference in survival between animals that were below 18 mm in length and those that were greater than 24 mm in length (p = 0.00136, Fig. 2). This difference in survival D a ~r ±Std Dev I I ±Sld Er D Mean URCHIN ABUNDANCE 60 < 50 S 40 30 D : j. I ^ : O ~r iStd Oev I I ±Sia Err o Mean 1 = ave 26 6 mm ( SD 1 33 mm I - ave 13 87 mm ( SD 173 irmi ) Fifjuri' 1. Results Ircim a Mann-Whitnev U Test, showing that tliiTc is no significant dllTerence in abalone survival between sites with and without urchins. Figure 2. Results from a Mann-Whilney U Test, showing that seeding size makes a significant difference in the survival of juvenile abalone. was calculated within the context of the similar seeding site char- acteristics for the different sized animals used in each experiment. While there was a very strong positive correlation between size of seeded abalone and an increase in short-term survival (significant at a 99% confidence level), there was no significant difference between survival after one month and survival after two months, for either size class. Seeding Site Attribute Classification Within the context that the substrate is a three dimensional environment, made up of layered boulders and open spaces, the strongest positive correlation (significant at a 95% confidence level) between survival and boulder size is that between survival and boulders with a diameter of smaller than 30 cm. A negative correlation exists between survival and exposed area. DISCUSSION The fact that the abalone used in the urchin exclusion experi- ment were shown to gain no significant benefit in terms of sur- vival, from the presence of urchins, was in direct contrast to ex- pectations (Day 1998, Tarr et al. 1996, Kojima 1981. Scott et al. in prep. Rogers-Bennet & Pearse 1998). These are. however, short term results (Fig. 1 ). While the advantage, in terms of access to drift kelp, is abundantly clear when observing the masses of drift kelp trapped by urchins, the advantage of the shelter provided by urchins becomes negligible, indicated both by the numbers of un- utilized urchins and the extent to which the abalone are hidden in the cracks and crevices provided by rock piles and boulders. Sur- vival rates of juvenile abalone may have been influenced by the fact that abalone not living under urchins would have had to seek other sources of nutrition, thus exposing themselves to predation. Abalone living in close association with urchins have been shown to expose themselves less than animals living in cracks and crev- ices (Day 1998). Table 3 shows the marginally significant positive correlation between survival and the presence of urchins. The ex- periments mentioned above, suggest that the juvenile abalone were not relying on urchins for survival. These results are not contra- dictory to those from the urchin exclusion experiment. This is also reinforced by the insignificant correlation between the percentage of successful abalone that did utilize urchins as refuge. The cor- relation between the presence of urchins and abalone survival cotild be interpreted as indicating an area in which there was good habitat for both urchins and abalone (Day 1998). In other words, what is a good habitat for urchins may also be a good habitat for 860 DE Waal and Cook TABLE 3. Spearman Rank Order Correlations between site attributes and abalone survival. Categories N R t(N-2) p-Level Boulders with >50 cm diameter 24 0.003089 0.01449 0.988569 Boulders <.'iO >30 cm diameter 24 -0.109507 -0.51ft74 0.610492 Boulders with <30 cm diameter 24 0.407414 2.09248 0.048149 Exposed area 24 -0.32226ft - 1 .59675 0. 1 24587 Ratio of boulders over exposed area 24 0.344539 1.72143 0.099207 % Cover corraUne algae sp. 24 -0.132326 -0.62617 0.537647 Urchin density 24 0.402841 2.06441 0.050966 Abalone size 24 0.700911 4.60928 0.0001-^6 % Abalone utilizing urchins as refuge 24 0.095773 0.451289 0.656198 abalone. If this is so. the relationship between abalone and urchins may be coincidental. One of the common features found among the sites discussed above was that the abalone were found clumped in very close proximity to the traps, by far the majority being within a radius of two meters, with as many as .30 or more found under single boul- ders (Saito 1984. Tegner & Butler 1985). There was very little evidence that the juveniles had moved out of the experimental areas. This is in contrast to results on juveniles of the species Haliotis rufescens, which seemed to move very quickly (Rogers- Bennet & Pearse 1998). Observations at seeding sites, specifically chosen to be unsuitable at the preliminary stage of experimenta- tion, did show some abalone staying in the trap for up to 4 mo. In those cases, however, it was obvious that the trap provided the only suitable local shelter. Movement from the trap would obvi- ously expose the juveniles to predation. Not finding the abalone in subsequent sampling would not clarify the question of whether they had moved or whether they had been eaten by predators (Tegner & Butler 1985). The choice of seeding site obviously influences behavior extensively, determining whether an abalone will move and how far it will move (Saito 1984). From the results of the current experiments, it was clear that, if the habitat was suitable, movement was very restricted. Results also showed that the smaller the abalone, the higher the mortality rates resulting from moving into an exposed area. The link between long-term survival of artificially reared juvenile abalone and urchins still needs to be quantified. Whether a habitat of layered boulders could replace any long-term benefit that abalone might derive from the presence of urchins, still needs to be established. This will influ- ence the selection of commercial seeding sites. In defining the attributes of a successful seeding site, the com- plexity of a three dimensional habitat made up of layered rocks, and the potential refuge that these may provide, must be kept in mind. The smaller the rocks, the higher the surface area and the more cracks and crevices available per unit area. This may explain the positive correlation between survival and boulders with a di- ameter less than .30cm, compared to the bigger boulders. While not statistically significant, there did appear to be some type of nega- tive correlation between survival and exposed space. This is to be expected because, in these areas, no shelter was provided for the juvenile abalone. The ratio of area under boulders to area exposed (Table 3). showed a positive correlation with sur\ival (Schiel 1992, Schiel 1993, Shepherd 1998). While this has not proven to be signifi- cant, the relationship between sheltered and exposed area was a good, simple measure of habitat suitability. Although the relation- ship between the amount of crustose corraline algae and survival is shown as not significant in these experiments, naturally occurring juvenile abalone and urchins are found predominantly on encrust- ing corralines. The movement of sand and general substratum as a result of strong currents is a major and obvious cause of mortality in natural and seeded abalone (Tegner 1992, Tegner 2000). The pre.sence of corraline algae could be an indicator of reef or rock surfaces that are relatively clean of sand, providing abalone with secure footing and space to move, as well as a source of nutrition. Observations in Mc Dougalls Bay showed the ease with which smaller animals were washed into the water column when boulders were disturbed. Many of the juvenile abalone released themselves vs ithout being touched by divers. Many of the sites showed up to 609f losses within one week. Initial mortalities were \'ery high and probably accounted, to a large extent, for much of total mortality. It seems that mortalities in smaller seeded animals inay be caused by a lack in the ability to handle strong water currents and distur- bances. Farm-reared animals are not exposed to the same extremes of water movement as natural abalone. and animals may need time to adjust to these conditions. It is not clear whether this is a behavioral process or whether it is a muscular strengthening pro- cess. Based on this, the size-test experiments were designed (Fig. 2. Table 2). The larger animals used in the experiments showed a completely different response when being sampled. They were much more resilient to being disturbed and, in some cases, remov- ing them caused shell damaged because they were lodged so se- curely. Increasing seeding size, and thus both the muscle and shell strength of the juveniles, may take pressure off the behavior ad- aptation phase that the juvenile has to go through after seeding, (Olla et al. 1998, Munro & Bell 1997, Tarr et al. 1996). While there are some differences between the effects of size on survival of seeded Haliatis rufescens and H. fulgens from the USA (Tegner 2000), the current results on H. midae. are similar to those of both Japanese and Australian researchers. While Japanese research has suggested seeding a minimum size of 30 mm. combined with the correct choice of refuge or seeding site for Haliotis discus ( Masiida & Tsukamoto 1998. Munro & Bell 1997), Australian research has shown that both size and the abundance of predators play an im- portant role in the survival of seeded Haliotis sp. (Shepherd 1998). There was also an indication of size playing a role in the experi- ments of Sweijd et al. ( 1998). Two different size batches of ani- mals were placed in the same site at two different times. The larger 14-mm animals showed a 39.27^ survival compared to a 27.87r survival figure for 8-mm animals, over a six-month period. While the traps were deployed at two different times, resulting in two entirely different sets of environmental conditions acting on the animals, some site attributes would have been the same. The relationship between seeding size and survival is not a simple one. The effects of size on survival have many facets. Not only does shell thickness increase, but so also does muscle strength. The ability to assimilate survival techniques and experi- ence may also be affected. Seeding is a traumatic event, with high initial mortality. In the case of H. midae. it may be possible to address this through correct site selection and seeding size selec- tion. The transition from short-term survival to long-term survival, growth, movement and the environmental factors influencing this need to be investigated. These experiments suggest that differences Attributes oh Abalone Seeding Sites 861 exist in the behavior of natural and seeded populations of juvenile abalone. Understanding these differences may lead to a better un- derstanding of the factors that affect sur\ ival of seeded juvenile abalone. .ACKNOWLEDGMENTS Research funding was provided by both Port Nolloth Sea Farms and the South African Foundation of Research and Development. LITERATURE CITED Davis. G. E. 1995. Recruitment of juvenile abalone (Halioris ,«/)/). i Mea- sured in artificial habitats. Mar. Fresim: Res. 46:549-554. Day, E. 1998. Ecological interactions between abalone (HuUotis midael juveniles and sea urchins {Parechinus cmgulosus). off the southwest coast of South Africa. Ph.D. Thesis. University of Cape Town. Emmett, B. & G. S. Jamieson. 1988. An experimental transplant of north- ern abalone, Haliotis kamtschalkana. in Barkley Sound, British Co- lumbia. Fish. Bull. 87:95-104. Inoue, K. 1987. Growth and survival of artificial abalone seed released in Shijiki Bay, Japan. Seikai National Fisheries Research Institute. No. I0I-I03. Kojima, H. 1981. Mortality of young Japanese black abalone Haliotis cUscus discus after transplantation. Ball. Japan. Soc. Sci. Fish 47(2): 151-159. Masuda, R. & K. Tsukamoto. 1998. Stock enhancement in Japan: review and perspective. Bull. Mar. Sci. 62(2):337-358. Munro, J. L. & J. D. Bell. 1997. Enhancement of Marine Fisheries Re- .sources. Rev. Fish. Sci. 5(2): 185-222. Olla, B. L., M. W. Davis & C. H. Ryer. 1998. Understanding how the hatchery environment represses or promotes the development of be- havioral survival skills. Bull. Mar. Sci. 62(2):53I-550. Masuda. R. & K. Tsukamoto. 1998. Stock enhancement in Japan: review and perspective. Bull. Man Sci. 62(2):337-358. Rogers-Bennet, L. & J. S. Pearse. 1998. Experimental seeding of hatchery- reared juvenile red abalone in Northern California. / Shell. Res. 17(3): 877-880. Saito, K. 1984. Ocean ranching of abalones and scallops in Nonhern Japan. Acjuacullure 39:361-373. Schiel. D. R. 1993. Experimental evaluation of commercial-scale enhance- ment of abalone Halioris iris populations in Neu Zealand. Mar. Ecol. Prog. Ser 97:167-181. Schiel. D. R. 1992. The enhancement of paua {Haliotis iris Martyni popu- lations in New Zealand. In: S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world. Fishing News Books. Oxford: Blackwell. pp. 475-484. Scou, J. A., N. A. Sweijd, P. A. Cook & R. P. Smullen. Sea urchins as a refuge for seed-sized juvenile abalone (Haliotis midae) from predation- implicalions for abalone ranching, (in prep.). Shepherd, S. A. 1998. Studies on Southern Au.stralian abalone (Genus Haliotis) XIX: long-term juvenile mortality dynamics. J. Shell. Res. l7(3):813-825. Sweijd, N., Q. Snethlage. D. Harvey & P. Cook. 1998. Experimental aba- lone iHaliolis midae) seeding in South Africa. J. Shell. Res. 17(3):897- 904. Tarr, R. J. Q. 1995. Growth and movement of the South African abalone Haliotis midae: a reassessment. Ma. Fresim: Res. 46:583-590. Tarr. R. J. Q.. Williams, P. V. G. and Mackenzie, A. J. 1996. Abalone, sea urchins and rock lobster: a possible ecological shift that may affect traditional fisheries. S. A. J. Mar Sci. 17:319-323. Tegner, M. 2000. Abalone enhancement in California: What we've learned and where we are going from here. In: A. Campbell, editor. Workshop on Rebuilding Abalone Stocks in British Columbia. Can. Spec. Publ. Fish. Aquat. Sci. pp. 130. Tegner, M. J. & R, A. Butler. Abalone Seeding. 1989. In: K. O. Hahn, editor. CRC Handbook of culture of abalone and other marine gastro- pods. Boca Raton, FL: CRC Press, pp. 157-182. Tegner, M. J. & R. A. Butler. 1985. The survival and monality of seeded and native red abalones, Haliotis rufescens. on the Pales Verdes Pen- insula. Calif. Fish Game. 7I(3):150-I63. Tegner. M. J. 1992. Broodstock transplants as an approach to abalone stock enhancement. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo. editors. Abalone of the world. Fi.shing News Books. Oxford: Blackwell. pp. 461-473. Jniiival of Slicllfi.sh RfSfiinh. Vol. 20. No. 2. SA.^-S6(i, 2001. USE OF A SPREADSHEET MODEL TO INVESTIGATE THE DYNAMICS AND THE ECONOMICS OF A SEEDED ABALONE POPULATION S. DE WAAL* AND P. A. COOK Zoology Department. University of Cape Town. Private Bag. Rondehoscli 7701. Cape Town. Sontli Africa ABSTRACT Using a simple spreadsheet model it was possible to investigate both the dynamics of a seeded abalone population and the econoiiiic implications of seeding such a population. The two variables used to drive the population model were the initial survival after seeding-induced mortality, and age-specific differential mortality. In the model, initial survival or initial population size was the variable that played the dominant role in determining the potential yield of any seeded population. In reality, seeding size plays a significant role in initial survival. Within the same experimental context. 14-mm animals showed a minimum survival of 24% while 26-mm animals showed a minimum of 56%. an exponential increase in potential yield. To address the issue of potential yield on a commercial scale, differential mortality amongst seeded populations of abalone needs to be researched. A cost-benefit analysis has shown that high production costs will only allow ranching to be economically viable in areas where the potential differential mortality regime is very favorable, where survival per age class increases as the population ages. A'£')' WORDS: Haliuiis muiac. cost-benefit analysis, commercial seeding, abalone INTRODUCTION The possibility of introducing the abalone species Haliotis ini- dae to the Northwest coast of South Africa has resulted in the development of an experimental commercial seeding program along a 60 km stretch of coast between Kleinzee and Koingnaas. 120 km south of the Orange River. There are no e.xisting stocks of abalone in this area, though there are fossil records of an extinct abalone species H. saldanhae. dating back to more than 1 million years ago (Sweijd et al. 1998). The success of sticii a commercial venture will be based on a number of criteria, including the fol- lowing: 1. a very specific understanding of the ecology of the species, especially of the juvenile stage of its lifecycle. 2. the quantification of the mortality of the seeded populations. 3. the long-tenn economic viability of such a venture. To do a cost-benetlt analysis of the potential project, based im realistic experimental data. iHeppel & Crowder 1998. Hilborn 1998. Schiel 1993). a simple spreadsheet model was used as a means to consolidate biological data collected from seeding ex- periments. The model was developed after reference to current literature, expert opinion from abalone research scientists and data collected from the present experimental seeding program. As aba- lone age. the probability of survival to the next age class changes; this is age-specific differential mortality or survival. Experimental data has shown that seeding induced mortality, be it due to han- dling stress or behavioral adaptation problems, is size related (de Waal & Cook 2001 . Schiel 1993. Shepherd et al. 2000). One of the underiying themes of the model was that a trade-off existed be- tween initial mortality and survival of seeded juveniles and the cost of producing juvenile abalone of different size classes. One of the goals in developing the model was to attempt to quantify this relationship. While differential mortality figures do exist for some species (Masuda & Tsukamoto 1998). differential mortality figures do not exist for Haliotis inidae. Although these figures are not known, commercial ventures are taking place in South Africa. Being able to explore the ecological *Correspondmg author. E-mail address schalkdo 't 3 CO Age-classes Figure 1. (al Difftrential mortality rate regimes used in Simulations A, B, and C. Simulation C is the one selected as being most realistic with survival increasing per age class to a maximum of 9n'7c. (bl Survivorship curves generated by Simulations A, B, and C. These curves show the cumulative effect of differential mortality rates. Al- though the lines converge over time, the differential mortality rates from Figure la were chosen to be completely different. This shows the importance of knowing the actual mortality rates. METHODS Description of the Model To determine population frequency per age class (;;) at a spe- cific time period (t), the population frequency in the previous time period ()) at t - 1). was multiplied by the age-specific percentage survival fiiiure from the year (t) in the formula: ^(",- X C/c survival per age class), To determine the percentage survival per age class at a specific time period (t), the percentage survival per age class for the pre- vious time period (t - 1), was multiplied by a factor equal to. greater than or smaller than 1. If the factor is equal to one the percentage survival per age class remains the same for the follow- ing year. If the factor is greater than one, survival per age class increases per age class. In other words, the older an animal be- comes the less chance it has of dying. If the factor is less than one. survival per age class decreases and the percentage chance of dying increases with age. Age class frequencies were calculated as an average at the end of a cycle (one run), which for simulation purposes could be up to .^0 y. the maximum life span for H. iiiiilac (Tan' 1995). Age class periods were equal in length of time to the first time period. All simulations in this case were done in time periods of 1 y. The following are assumptions made to develop the model. I . Differential mortality rates were developed using the as- sumption that percentage mortality per age-class either in- creased or decreased by a constant factor in consecutive age classes (Fig. la). 2. Differential mortality rates included all limiting factors on survival, density dependence, senescence, carrying capaci- ties etc. 3. Maximum survivorship of any age class could be 9()9'f of the previous age class. (Average annual mortality is taken to be approximately 11% for all size classes of natural abalone stocks by the Marine and Coastal Management Institute of South Africa (pers. com. C. Moloney). 4. There was no natural recruitment taking place in the seeded population. Iiivesligatiiig the Range of Potential Differential Mortality Regimes While a range of differential mortality rates may exist for any species in a range of habitats, it was necessary to restrict the investigation to only three experimental options. In the following simulations, an initial survival figure of 24% was used. Simulation A A basic approach to modeling a differential mortality regime is to assume a constant mortality rate per age class throughout the entire population, regardless of age. This is a simplistic approach and there are no indications that H. midae does behave in this way (Fig. la). Simulation B In this case, it was assumed that mortality increased per age class by 5% per year from a maximum survival rate of 90%. However unrealistic, this allowed a broader investigation into the possible survival options and behavior of the model (Fig. la). Simulation C In this case it was assumed that mortality decreased per age class, survival per age class doubled to a maximum of 90% (Fig. I a). The shape of this survival curve seems to be the most realistic option, although the exact age or size specific differential mortality rates and the gradient of the curve are not known. (Tegner 2000, de Waal & Cook 2001, Shepherd et al. 2000. Schiel 199.^). Figure lb shows survivorship per age class generated by the percentage survival per age class curves from Figure la and the initial population number of 240.000 (based on a 24% post seeding survival figure). Figure lb shows the cumulative survival percent- age or the probability of an animal reaching a certain age. This was the figure that indicates the yield per recruit of a seeding operation in terms of age class frequency distributions. Sensitivity Analysis of the Model with Regards to Survival Regime and Initial Survival It was necessary to test the behavior of the model in relation to the two driving variables, seeding induced mortality and differen- tial mortality. To do this, the model was run with stochastic re- cruitment generated at the beginning of each year. Two hundred runs were done for each scenario analysis. A random number generator between 1000 and 100 million was used to simulate recruitment at the beginning of every year. Four general scenarios were developed to test the sensitivity of the model to the two variables (Fig 2). In Group 1, differential mortality rates were manipulated over time while initial survival figures, or initial population sizes, were kept constant. In Group 2 differential mortality rates were kept constant while initial popu- lation sizes were manipulated over time. Economics of Abalone Seeding 865 150 12 3 4 5 Time Figure 2. Sensiti\ity analysis results in "hicli litlur initial population numbers were manipulated "hile differeMitial mortality rales per age class were kept constant o\er lime. (Jroup 2. Alternati>el.v. differential mortality rates were manipulated while initial population size was kept constant. Group 1. The period of time is conceptual and is only used to show the difference in total population numbers as a result of manipu- lating the two driving variables. Cost-Benefit Analysis A simple cost-benefit analysis was developed using experimen- tal initial survival figures for both I4-mm and 26-mm seed sizes, a range of differential mortality regimes, and the present market price for abalone. A purchase price of R I per animal for 14-mni animals and R 4 per animal for 26-mm animals, and a selling price of R210 per kg for 300 g abalone, were assumed. The exchange rate used was R 6 (SA) to S! (US). Prices were quoted by Port Nolloth Sea Farms. A series of differential mortality rates were used in which sur- vival per age class ranged from being constant to increasing by 60% per age class as the population ages. Initial population figures were based on experimental data (de Waal & Cook 2001 ). 24 9^ for animals with an average size of 14/mm. and 56% for animals of average size 26/mm. One initial seeding batch was allowed to run the course of the model to a market size of 300 g. Using an average growth rate of 1 .5 mm per month, and a mass-length curv e mea- sured from animals that have been at sea at Kleinzee for 13 mo. this size would have been reached after approximately 6.5 y {un- published data). RESULTS Sensitivity Analysis Both driving variables, seeding induced mortality and differ- ential mortality, had an influence on the model output in terms of the rate of population decrease. Over the same time period, in those simulations in which initial population size was manipulated. Group 2 (Fig. 2), the proportional impact was far greater than in those simulations in which differential mortality rates were ma- nipulated. Group 1 (Fig. 2). While model output was generated by a combination of the effects of both variables, it is clear that initial survival after seeding-induced mortality had stabilized, played the dominant role. Ciist-Benefit Analysis With an average growth rate in the range o\ 1 .5 mm per nio measured in Port Nolloth for farm-reared juveniles, it would take approximately 9.5 mo for animals to reach 14/mm, and a further 7.5 mo to reach 26 mm. a total of 17 mo. The animals seeded at size 14 mm would reach harvesting size after 6 y at sea. The animals seeded at 26 mm reached that size after 5 y. On average however, harvesting for both seeding scenarios could take place in the seventh year after spawning. Growth rates are variable and this must be kept in mind when interpreting analyses. For example, the estimated time to legal size for experimental animals seeded at Port Nolloth was estimated to be anywhere between 3.5 y and 6 y (Sweijd et al. 199S). DISCUSSION The model was designed to investigate the dynamics of a single population of abalone. based on contemporary knowledge of the dynamics of abalone in general. The ultimate goal would be to generate experimental data to the point where modeling can be based on accurate differential mortality and initial survival figures. At this stage, this was not possible. However, a choice based on scientific facts could be made between realistic and unrealistic options. Indications from this project, and from contemporary lit- erature, are that differential mortality decreases with age, i.e. sur- vival increases until senescence sets in. Davis (1995) estimated survival of juveniles in artificial habitats to be 32% after 1 y and 24% after 2 y. In other words, in the second year survival had increased to 75% compared to 32% in the first y. Seeding-induced mortality would account for a large percentage of the first y mor- tality, and differential mortality would account for the decrease in mortality of the seeded population in the second year (Shepherd et al. 2000). Seeding-induced mortality is clearly the most important factor to consider when seeding. The percentage survival and the survivorship curves, in Figures la and lb. showed very clearly that it is quite possible for theo- retical curves to converge at some point. If only one point value is known, very little could be determined about the dynamics of the model. To model production accurately, it would be necessary to have a good idea of the differential mortality regime acting in that specific habitat. The goal of objective research should, in this case, be to determine the most realistic mortality regimes in any given situation. Models used with appropriate experimental data will allow analysis of the feasibility of projects before any serious financial input is made and should be used and developed as more biological data become available (Munro & Bell 1997. Wilson et al. 1998. Heppel & Crowder 1998. Hilbom 1998). It is absolutely essential to understand the concept of differential mortality and survival before any predictions can be made about the success of seeding programs. The model showed that, with the quoted purchase price of seed, it is not economically viable to seed juvenile abalone of either size if the survival per age class does not increase by at least 10% per age class. If the survival per age class increases by more than 40% per age class, it would be viable to seed even the smaller animals. The model showed very clearly that there was an exponential increase in potential yield for larger seed animals compared to smaller seed animals (Table 1 ). This is a reflection of what hap- pens under natural conditions (Shepherd at al. 2000. Schiel 1993. Tegner 2000). Experimental results by De Waal & Cook (2001) have shown an exponential increase in survival of juvenile abalone with an increase in seeding size. This could possibly be attributed to an increase in muscle and shell strength of the juvenile H. midae. Both behavior and the ability to adapt could also play an important role in this differential survival. According to the model, this difference in potential yield was caused by both the strong effect that initial survival has on the model output, and by the speed at which high initial survival causes the survival per age class value to reach a maximum of 90%. It is important to note that while experimental evidence suggests (unpublished data) that both 866 DE Waal & Cook TABLE I. Cost-benefit results for abalone harvested after approximately 6.5 years at sea. Survival Per Seeding Harvesting Buying Selling Profit Age Class Size mm Mass kg R (mill) R (mill) R (milll Remains constant 14.00 13.76 1.00 0.00 -1.00 26.00 92.52.30 4.00 1.94 -2.06 Increases by 10% 14.00 101.81 1.00 0.02 -0.98 26.00 38568.32 4.00 S.IO 4.10 Increases by 20% 14.00 633.00 1.00 0.13 -0.87 26,00 80243.65 4.00 16.85 12.85 Increases by 30% 14.00 2641.08 1 .00 0.55 -0.45 26.00 86416.24 4.00 IS. 15 14.15 Increases by 40% 14.00 5463.39 1 .00 1.15 0. 1 5 26.00 92588.83 4.00 19.44 15.44 Increases by 50% 14.00 8264.97 1.00 1.74 0.74 26.00 98761.42 4.00 20.74 16.74 Increases by 60% 1 4.00 11145.12 1.00 2.34 1..34 26.00 99202.32 4.00 20.83 16.S3 adults and ju\eniles may survive well in the sanie habitat, this ina\ not always be the case. Seeding must be done in an area in which both initial mortality and long term mortality are as low as pos- sible. In any potential ranching and stock enhancenient project, pro- duction costs that are too high do not allow continuation. This model has allowed an insight into the limitations set by natural processes before the production process has begun (Hilbom 1998, Moksness et al. 1998. Schiel 1993). Production costs could be much lower than the quoted purchase prices and better potential income could be obtained by producing low-cost abalone seed rather than purchasing it. A trade-off exists, not only between supply and demand in terms of the market price of abalone, but also between niarket price and the seeding strategy and site char- acteristics of the project concerned. In abalone ranching, produc- tion costs must be seen to include both the negative effect of initial seeding mortality and the negative effect of the potential differen- tial mortality rates that exist. The importance that these two vari- ables have on the potential econotnic viability of a ranching proj- ect, and the lack of experimental data concerning these variables, emphasizes the need for additional research in this field. ACKNOWLEDGMENTS This research was funded by both Port Nolloth Sea Farms and the South African Foundation of Research and Development (T. H.R.I. P.) fund. We thank both Professor John Field and Dr. Colleen Moloney for their advice on both the model and the manuscript. LITERATURE CITED de Waal, S. & Cook, P. A. 2001. Quantifying the physical and biological attributes of successful seeding sites for farm-reared juvenile abalone (haliotis midae): / Shellfish Res. 20(2):857-861. Davis. G. E. 1995. Recruitment of juvenile abalone {Hulunis .v/>/>. ) mea- sured in artificial habitats. Mar. Freshwuter. Res. 46:549-554. Heppel. S. S. & L. B. Crowder. 1998. Prognostic evaluation of enhance- ment programs using population models and life history analysis. Bull. Mar. Sci. 62(21:495-507. Hilbom, R. 1998. The economic performance of marine slock enhancenient projects. Bull. Mar. Sci. 62(2):661-674. Masuda. R. & K. Tsukamoto. 1998. Stock enhancement in Japan: review and perspective. Bull. Mar. Sci. 62(2):337-358. Moksness. E.. R. Stole & G. van der Meeren. 1998. Profitability analysis of sea-ranching with Atlantic Salmon (.Salma saltiri). Arctic Charr iSalveliiius alpinus). and European lobster {Homarus ^aiiimariis) in Norway. Bull. Mar. Sci. 62(2):689-699. Munro, J. L. & J. D. Bell. 1997. Enhancement of marine fisheries re- sources. Reviews in Fisheries science 5(2): 185-222. Schiel, D. R. 1993. Experimental evaluation of commercial-scale enhance- ment of abalone Halioris iri.'. populations in New Zealand. Mar. Ecol. Prog. Ser. 97:167-181. Sweijd, N., Q. Snethlage, D. Harvey & P. Cook. 1998. Experimental aba- lone (Haliotis midae) seeding in South Africa. J. Shell. Res. 17(3):897- 904. Shepherd. S. A.. P. A. Preece & R. W. G. White. 2000. Tired nature's sweet restorer 7 Ecology of abalone [Halions spp.) stock enhancement in Australia. In: A. Campbell, editor. Workshop on Rebuilding Abalone Stocks in British Columbia. Can. Spec. Publ. Fish. Aquat. Sci. pp. 130. Tarr. R. J. Q. 1995. Growth and movement of the South African abalone Haliotis midae: a reassessment. Mar. Freshw. Res. 46:583-590. Tegner. M. 2000. Abalone enhancement in California: What we"ve learned and where we go from here. In: A. Campbell, editor. Workshop on Rebuilding Abalone Stocks in Brilish Columbia. Can. Spec. Publ. Fish. Aquat. Sci. 130. Wilson. J. A.. R. W. Langton & C. Van Orsdel. 1998. A model for the preliminary analysis of the economic feasibility of Atlantic Cod enhancement in the Gulf of Maine (USA). Bull. Mar. Sci. 62(2):75- 687. Joiinuil of Shellfish Research. Vol. 20, No. 2. 867-874, 2001. TRANSMISSION OF THE RICKETTSIALES-IJKE PROKARYOTE ^'CANDIDATUS XENOHALIOTIS CALIFORNIEMSIS'' AND ITS ROLE IN WITHERING SYNDROME OF CALIFORNIA ABALONE, HAUOTIS SPP. JAMES D. MOORE,' * THEA T. ROBBINS,' RONALD P. HEDRICK," AND CAROLYN S. FRIEDMAN' - 'California Dcpaitincut of Fisli and Game, UC Bodega Marine Laboratory. 2099 We.stside Road, Bodega Bay CA ^4^2 J: ^Department of Medicine and Epidemiology. UC Davis School of Veterinary Medicine. Davis. CA 95616 ABSTRACT Withering syndrome (WS) is a chronic, wasting disease responsible for mass mortality in southern California popula- tions of black abalone Haliotis ciacherodii and responsible for significant losses of cultured red abalone H. nifesccns. Ongoing studies in our laboratory indicate that a recently described gastrointestinal Rickettsiales-like prokaryote, "Candidatiis Xenolialioiis califomi- ensis" (RLP) is the etiologic agent of WS. Here we describe attempts to experimentally transmit the RLP and demonstrate its role in WS. In two preliminary experiments. RLP-infected black abalone postesophagus homogenate (IPEH) was injected into the foot or orally administered to RLP-free black abalone. No RLPs were detected 8 wk after pedal injection. Low rates of transmission were observed 8 to 12 wk after oral inoculation, although an RLP-positive animal was also detected in a negative control group inoculated with filtered seawater. In a separate, 16-wk study, RLP infections were detected in red abalone that received effluent from a tank of infected red abalone while control animals that received direct-source seawater remained RLP-free. A fourth, long-term study delivered IPEH or seawater to RLP-free red abalone with either bath exposure or intra-digestive gland injection. An additional treatment to test a potential viral etiology for WS consisted of intra-digestive gland injection with a 0.1-0.2 p,m filtrate of IPEH and subsequent treatment with antibiotics. Each treatment was administered six times over a 16-wk period. Cohabiting RLP-infected black abalone with RLP-free red abalone provided a positive control treatment. At wk 63 post-initiation, untreated, saline injected and IPEH filtrate injected groups had low cumulative mortality (0-20%), while mortality in the IPEH bath. IPEH injection and cohabitation treatment groups was 70-90%. There were statistically significant relationships between experimental treatment, RLP burdens and signs of WS. Low-level RLP infections of uncertain origin were observed in one each of the duplicate tanks of the negative control and the IPEH filtrate-injected animals. An absence of WS signs in recipients of the IPEH filtrate provides strong evidence that the agent of WS is non-viral. Collectively these studies provide solid evidence that the RLP is the etiologic agent of WS. KHY' WORDS: HiiliiHis nifeseens. ciaclienidii. Rickettsials, withering syndrome, transmission, abalone INTRODUCTION The bluck abalone HoUntis cnicherodii is one of several hali- otid species found on the Channel Islands and mainland of south- em and central California, USA. This species lives in the intertidal zone and is considered less desirable for human consumption than several subtidal species, but became an important component of the commercial fishery in the 1970s when stocks of subtidal spe- cies declined, presumably due to increasing recreational and com- mercial fishing pressures (Davis et al. 19921. Resource managers and abalone fishermen first noticed large numbers of dead and dying black abalone on the Channel Islands in the mid-1980s (Haaker et al. 1992). The term Withering Syndrome was coined to describe the affected abalone. characterized by lethargy and a greatly reduced pedal mass (Haaker et al. 1992). Southern Cali- fornia black abalone populations have essentially collapsed, with mortality greater than 907f in many areas (Haaker et al. 1992. Richards & Davis 1993). Stocks of subtidal red (//. nifeseens) and pink (A/, corrugala) abalone also declined during the 1980s (Davis et al. 1992) but the role of fishing pressure, climatic events and WS in those declines remains unclear. Although the appearance and severity of WS outbreaks ap- peared to be enhanced by warm water temperatures (Steinbeck et al. 1992, Tissot 1995). the epizootiology of the di.sease indicated invohemeiit of an infectious agent (Lafferty & Kuris 1993). WS spread throughout the Channel Islands followed by movement to *Corresponding author. E-mail address: jimmoore(<"ucdavis.edu the California coastline, where it advanced northward during the early 1990s (Lafferty & Kuris 1993. Alstatt et al. 1996). An initial investigation of nine black abalone with WS and five without signs of the disease reported the presence of several parasites and a gastrointestinal Rickettsiales-like prokaryote (RLP), although none occurred in all affected abalone while being absent in healthy-appearing abalone (VanBlaricom et al. 1993). The RLP formed large intracellular inclusions in gut epithelium, causing hypertrophy but little other cytopathology. RLPs are common en- dosymbionts in molluscs, often being unassociated with disease (Sparks 1985. Elston 1986). Gardner et al. (1995) examined black abalone from bt)th where WS occurred and where it did not. They found that the RLP was present only in abalone from the WS- endemic site, and noted pathological changes in the digestive gland that could be due to the RLP. Subsequently. Friedman et al. ( 1 997) examined starvation, temperature, the renal coccidian Pseiidoklos- siii luiliotis and the gastrointestinal RLP as causes of WS. While eliminating the coccidian. starvation and elevated temperature as being directly responsible for WS. complex relationships between the RLP and temperature warranted further investigation. Fried- man et al. (this volume) transmitted WS and the RLP between black abalone. Groups of initially RLP-free black abalone that contracted the RLP by cohabitation with RLP-positive animals had a higher proportion of animals with signs of WS and higher pro- portion that died than did unexposed RLP-free control animals. However, the authors also observed a lack of significant correla- tions between WS signs and RLP infection intensity, further indi- cating a complex relationship in black abalone. The RLP is not restricted in host .species to the black abalone. 867 868 Moore et al. The RLP is present in most or all abalone aquacultiire facilities in California, which raise red abalone to a market size of approxi- mately 9 cm in land and cage-based operations throughout the state (McBride 1998). This is due to its increasing range in wild popu- lations and the widespread distribution of infected seedstock prior to it being recognized as a potentially significant pathogen (C. S. Friedman, unpublished observations). During the 1997-1998 El Nino, red abalone aquaculturists in the southern and central por- tions of the state began observing sharply elevated frequencies of animals showing WS signs in association with elevated water tem- peratures. Recently. Moore et al. (2000) demonstrated that farmed red abalone with RLPs can show little or no signs of WS at lower temperatures (I4.7°C). but elevation of temperature to IS.S^C caused elevated mortality, expression of WS signs and increased RLP burdens. With increasing evidence for its role in WS. Fried- man et al. (2000) recently provided a description of the RLP, designated ^'Candidatus Xenohaliotis catiforniensis." The term "Candidalus" in the taxon indicates that the species was described largely on morphological and DNA sequence-based data and still requires the serological and biochemical analyses that would be performed if the species could be cultured in vitrti. The coinciding geographic boundaries and conelation between WS and the RLP do not alone demonstrate that this agent causes the disease. An alternative hypothesis suggests that the RLP pro- liferates in host animals debilitated by WS. itself caused by an unidentified, possibly viral agent. Our current studies, while lim- ited by the difficulty of obtaining and maintaining RLP-free aba- lone, were designed to experimentally transmit the RLP and ex- amine links between severity of RLP infection and expression of WS signs. MATERIALS AND METHODS The experimental designs of the four consecutive experiments conducted are shown in Figure I. Experiments 1-3 were designed to investigate transmission of the RLP. Animals were therefore sacrificed at time periods of several weeks to several months, generally before expression of WS occurred. Experiment 4 was designed to allow sufficient time for development of WS and was terminated following extensive mortality in treatment groups 15 mo post-initiation. Animals and Animal Husbandry All experiments were conducted in the Pathogen Containment Facility at the Bodega Marine Laboratory, from which all effluent is treated with 10 mg/L chlorine for two hours and dechlorinated with sulfur dioxide before release. Lidded, 8-L tanks with outflows situated so that each lank held 5 L were used in Experiments 1-3. In Experiment 4. lidded ."iS-L tanks with standpipes situated to hold 50 L were used. All tanks received sand-filtered, aerated seawater and were supplied with kelp (Macrocystis pyrifera) several times per month. Care was taken to prevent cross contamination between tanks, including submersing gloved hands, measuring tools etc. in a tamed iodine solution (Prepodyne) and spraying surfaces with 70% ethanol as necessary. Animals were tagged by inserting either a numbered stainless steel washer on stainless steel wire or a numbered vinyl sleeve (Floy Tag, Seattle) on a cable tie through the first and second most recently formed respiratory pores. Ben- zocaine (Sigma. 40 mg/L, 15 min baths) was used as an anesthetic to dislodge animals for transfer or treatment. • • • •o 1 M injection Oral ! M inject & oral IPEH IPEH FSW o# Oral Oral Oral Oral Oral Cohabitation IPEH IPEH IPEH IPEHtillrate FSW dO, 14 dO, 14, 64 d56, 64 dO. 14, 64 d 0. 14, 64 d 0-84 • o Watertome Waterborne RLP + water RLP ■ water • •• • DG injection Bath DG injection DG injection Untreated Cotiabitation IPEH IPEH IPEHfltrate FSW Figure 1. Schematic diagrams of the experimental units in Experi- ments 1 through 4. Circles represent the numbers of tanks in each treatment group, with experimental tanks shaded, negative control tanks white and positive control tanks hiack. The first line of text indicates the method of delivery of the material shown in the second line of text. For FXperiment 2 the third line of text shows the days on which animals were treated: the first group was sacrificed al day 28 and the remainder on day 84. IPF^H: RLP-infected black abalone postesophagus homogenate; FSW: Filtered seawater: DG: Digestive gland. Histology: Quantification of RLP Infection Intensity and WS Signs Animals that died during each experiment and all survivors at experiment termination were processed for histological detection of RLP burden. Animals in Experiment 4 were also assessed for WS-associated pathomorphological changes in the digestive gland and foot. Tissue sampling and histological processing was per- formed as previously described (Moore et al. 2000). Davidson's- fixed (Shaw & Battle 1957), hematoxylin and eosin-stained 5 jim paraffin tissue sections containing postesophagus, digestive gland, foot muscle, kidney and gonad were prepared for each animal, and slides were encoded to prevent bias during assessment. RLP in- fection intensity was estimated for the postesophagus and digestive gland (the two tissues in which RLPs can be found at high densi- ties. Friedman et al. 2000). RLP burdens were quantified in each tissue using the scale of Friedman et al. (19971 based on the average number of RLP inclusions per 200x magnification field of view: absent = (0), 1-10 = (1), 11-100 = (2), or greater than 100 = (3). Four different disease signs of withering syndrome were assessed using integral scales from 0-3, modified from Moore et al. (2000). For all four parameters, (0) represented a normal healthy appearance. For body shrinkage, (I I, (2) and (3) represented slight, moderate and severe shrinkage respectively: for digestive gland metaplasia or digestive gland atrophy, ( I ). (2) and (3) denote 5%-\0'7c. 1 19^-25%. and greater than 25% of the di- gestive gland being comprised of transport duct epithelia or con- nective tissue, respectively. Foot degeneration scores of (I), (2) and (3) denote muscle fibers comprising 76'7r-90%, 5 1 %-75% and less than 50% of the foot muscle, respectively. A condition index, CI = total weight, g/( shell length. cm)'\ was also used to assess body shrinkage. The presence of the renal coccidian Pseudoklossia haliotis (Friedman et al. 1995) was also noted. RlCKETTSIALHS-LlKK PrOKARYOTE AND WITHERING SYNDROME 869 Preparation of Infecled Postesopliagiis Homogeiiale (IPEH) Experimenls 1. 2 and 4 attempted transmission of the WS-RLP using several methods of delivering homogenized, RLP-infected postesophagus tissue excised from adult black abalone (IPEH). Black abalone donor animals were collected from Vandenberg Air Force Base, California, where WS has been endemic since ap- proximately January 1994 and RLP prevalence was approximately 90% during the collection period (C. S. Friedman & J. D. Moore, unpublished observations). Animals were collected several times per year during 1996-1998, held in 9-l3°C seawater and fed kelp (ih lihiliiiii. Postesophageal tissue (including the "crop" as de- scribed by Bevelender, 1988) from three to eight animals was pooled and minced in 0.2 ixm filtered seawater (FSW) on ice. The tissue was homogenized with a 7-ml Tenbroeck homogenizer. The tissue homogenate contained large pieces of connective tissue and other debris which were removed by gentle centrifugation ( 100- 250 X g for 2-5 min). The clarified homogenate was transferred to a new tube and diluted with FSW as necessary to obtain required volumes as described below. Experiment I: RLP Transmission hy Intramuscular Injection or Oral Inoculation In July 1997, black abalone were collected from Sobranes Point (Carmel), California, a location where WS had not been observed. One month after collection, the animals were randomly distributed to 5 tanks (4-5 abalone per tank) and acclimated from 14''C to I9°C over 1 wk. The treatment groups are shown in Figure 1. The IPEH consisted of 1 .745 g of postesophagus from four black aba- lone diluted to 15 ml. The IPEH was administered at a rate of 6.8 p.1 IPEH/ g whole animal weight to four animals in each of two duplicate tanks by either intramuscular injection in the foot with a 23-gauge needle or oral delivery using a 23 gauge animal feeding tube inserted into the mouth. A single tank (n = 5 animals) served as a negative control; animals were treated with both oral delivery and intramuscular injection of FSW. All animals were sacrificed on day 60 post-initiation and examined for the presence of RLPs by histology. Experiment 2: RLP Transmission by Oral Inoculation In August 1997, black abalone were collected from Carmel Point (Carmel), California, where WS had not been observed. In March 1998 the abalone were randomly distributed into six tanks (/] = 6 animals/tank). Temperatures during the experiment ranged from 1I.5''C-15'C until day 40, after which temperatures were elevated to 17°C-20°C, with an overall mean ± s.d. of 15.6°C ± 2.8°C. Treatments (Fig. 1) consisted of multiple oral inoculations with IPEH or a 0.2 |jim filtrate of IPEH at a rate of 3.0 p.l/g whole animal weight on days 0, 14 and 64 post-initiation. The IPEH on each day of treatment consisted of 5.2-7.5 g of postesophagus from three or four black abalone brought to a volume of 15 ml with FSW. To prepare IPEH filtrate. IPEH was centrifuged (400 x g. 5 min) and then filtered twice through 0.2 (xm syringe filters. As a positive control, three black abalone from Vandenberg Air Force Base were added to one tank of the Carmel Point abalone and maintained until termination of the experiment. On day 56, after discovering negative results in the IPEH treated group that was sacrificed at day 28, a previously untreated group was inoculated with a more concentrated IPEH (2.8 g postesophagus in 5 ml FSW without centrifugation). One IPEH treated group was sacri- ficed at day 28 and all other groups were sacrificed on day 84 post-initiation and examined for the presence of RLPs by histol- ogy. Experiment J: Walerborne RLP transmission In March 1998, adult red abalone were collected by scuba from Caspar Cove and Mill Cove, Mendocino County, California where WS had not been observed. In August 1998. 14 animals were divided equally into two groups and placed in 8-L tanks. Each of the two tanks was connected to its own intermediate 8-L tank by approximately 25 cm of tubing. The inflow to the intermediate tank of one treatment was connected to an 18°C water source ( = control). The inflow to the other intermediate tank was connected to the outflow of a 30-L tank that contained approximately 60 RLP-positive red abalone 2-10 cm in length and was supplied with the same seawater source. The animals were sacrificed 1 1 1 days post-initiation and examined histologically for the presence of RLPs. Experiment 4: IV.S' Development Following RLP Transmission by Digestive Gland Injection or Bath Exposure Experiment 4 utilized animals from the same collection as Ex- periment 3. In July 1998 one hundred animals were tagged, equally distributed among ten tanks and allowed to acclimate for 29 days, during which period the water temperature was raised from 15°C to 18°C. Animals were treated on days 0, 21, 35, 56, 91 and 109 of the study with the treatments shown in Figure 1. On each treatment date, IPEH was prepared from 6.5-7.9 g of postesopha- gus tissue excised from six to eight black abalone and brought to a final volume of 160 mL with FSW (FSW/3'7r heat-inactivated fetal bovine serum on day 0). The IPEH filtrate was prepared by successive filtration of IPEH through 8 |j.m, 0.8 (xm, 0.2 ijtm and 0.1 p.m filters, except on day 0, when 0.1 p.m filters were unavail- able and the 0.8 p,m filtrate was put through 0.2 |j.m filters twice. For all digestive gland injection treatments (FSW, IPEH, IPEH filtrate), animals were injected with 3.6 |J.L/g whole animal weight in the posterior ventral portion of the digestive gland using a 23 gauge needle. Recipient animals ranged in weight from 221 g to 1,360 g resulting in injection volumes ranging from 0.8 ml to 4.9 ml. For the bath treatment, tanks were drained and brought to a volume of 30 L, to which 25 mL of IPEH was added. Tanks were then maintained statically with aeration for two hours, after which flowing seawater was returned. The cohabitation positive control replicates were each supplied on day 0 with three Vandenberg black abalone showing variable degrees of body shrinkage. Black abalone that died were replaced until day 104. at which time all black abalone were removed. After the discovery that an animal in the IPEH filtrate treatment that died on day 91 was infected with the RLP, a decision was made to inject these animals with antibiotics. This was based on the design of this treatment to allow a potential viral pathogen to infect recipients in the absence of the RLP. These abalone were injected in the foot with oxytetracycline (10 mg/ml in 29c saline) at a rate of 2 1 mg/kg tissue weight. Injections were made every 48 h for a total of three injections, followed by two to three weeks without injections and repeating the series two more times for a total of nine injections over 47 days. Five feeding trials were conducted to measure the amount of kelp consumed by abalone in each treatment. An amount of kelp equal to 15% of the total body weight in each tank was added with the amount remainine after 24 h being reclaimed and weighed. 870 Moore et al. The experiment was terminated on day 44fi post-initiation wlien mortality in the bath treatment approached levels seen in the co- habitation and IPEH injected groups. Statistical analysis of the results from Experiment 4 included a Chi Square contingency table analysis comparing IPEH injected. IPEH bath, cohabitation and IPEH filtrate-injected treatment groups to the negative control group (a = 0.05). Data for WS signs (body shrinkage, digestive gland metaplasia, digestive gland atrophy or foot degeneration) were condensed into low (score of 0-1) or high (score of 2-3) categories of severity and 2 x 2 contingency tables compared observed vs. expected frequencies in each treatment group and the negative control group. Similarly, frequencies of low vs. high lev- els of each WS sign were compared in animals with low-level RLP infections (RLP intensity score 0-1 ) vs. high level infections (RLP intensity score 2-3) for the postesophagus and digestive gland separately. When the frequency in any cell was less than five observations a Fisher Exact test was employed. RESULTS Experiments I and 2: RLP Transmission by Intramuscular Injection or Oral Inoculation In Experiment 1. 60 days after a single injection or oral inocu- lation with infected black abalone postesophagus homogenate (IPEH) or filtered seawater, one individual that received IPEH orally contained a single focus of two RLP inclusions in the poste- sophagus (RLP intensity = 1). In Experiment 2. no RLPs were seen among the six animals sacrificed at day 28. Upon termination at day 84, RLPs were detected in postesophagi of two of the six individuals orally inoculated with non-clarified IPEH on days 56 and 64, and in one individual in the FSW-injected negative control group. The RLP inclusions in these animals were small and very infrequent (RLP intensity = 1 ). Five of the six cohabitation posi- tive control animals had low RLP intensity scores (RLP intensity = 1 in postesophagus. digestive gland, or both) and in one animal no RLPs were detected. Experiment J: Waterhorne RLP Transmission Histological analysis of survivors in each tank at day 111 re- vealed the RLP to be present in 100% (6/6) of the animals receiv- ing eftlucnt from the RLP-positive tank, while those receiving direct-source seawater had no RLP infections (0/6). One animal in each tank died prior to termination and in each case tissues were too necrotic for histological analysis. Intensities of postesophagus and digestive gland infections in the six RLP-positive animals ranged from relatively mild ( 1 ) to severe (3). Experiment 4: WS Development Following RLP Transmission by Digestive Gland Injection and Bath Exposure No significant difference in condition index was observed be- tween tanks at day 0 (ANOVA, p = 0.29) and only one individual (in one of the replicates to receive IPEH by bath treatment) showed slight body shrinkage. Mortality in the group injected in the di- gestive gland with IPEH began during wk 27 and reached 90'/; at termination (Fig. 2). Mortality began during wk 33 in the tanks to which infected black abalone were added (positive control) and cumulative mortality also reached 90%. Mortality in the IPEH bath treatment group began during week 35 and steadily increased to 70% by termination of the study. Nearly all of the animals that died in the positive control, IPEH injected and IPEH bath treated groups 100- Negative controls IPEH filtrate injection IPEH bath IPEH injection Cohabitation 20 30 40 50 Weeks post-imtiation Figure 2. Cumulative mortality in Experiment 4. Treatment groups consist of duplicate tanl 0.6, (2) slightly shrunken abalone with a CI from 0.55 to 0.59. and (3) severely atrophied animals with a CI < 0.55. The inner and outer shells were checked for epibionts under a dissecting microscope. The prevalence of symbionts was estimated as the number of infested abalone/nuniber of abalone examined xIOO. Seipulid infestation was qualified, according to the percent- age of shell surface covered, as low (I to 25%), medium (26 to 50'7f), or high (51 to 75%). Shell blisters caused by burrowing worms {.Polydora) were also counted. The visceral mass of the abalone was fixed whole in Davidson fixative (Shaw & Battle 1957) for at least 24 hours. Four transverse sections containing portions of the digestive tract (posterior esophagus included), kidney, muscle and gills, were processed for paraffin histology. Deparaffinized. Spim sections were stained with iron hematoxylin, eosin, fuchsin acid and aniline blue (Gray 1954). The degree of renal coccidian infection was estimated at 200x magnification and ranked in accordance with Friedman et al. 875 876 Caceres-Martinez and Tinoco-Orta TABLE 1. Diseases and sjmbionts of abalone species (modifled from Sliields & Perkins 1997). Abalone species Svnibiont Interaction Disease Reference Virus Nordotis discus Bacteria Halioris spp. H. discus hannai H. rufescens. H. kamlscliatkanna, H. midae H. midae Fungi H. sieboldii H. iris, H. ausrralis, H. virginea virginea H. cracherodii Protozooans H. craclwrodii Haliolis spp. H. kaniisiharkttiKi. H. rufescens H. laevigata, H. rubra, H. cyclohates. H. scalaris Sponges H. cracherodii. H. rufescens, H. tuherculaui Flatworms H. discus hannai H, tuberculata H, fulgens H. corrugala, H. fulgens, H. craclwrodii Snails and Clams H. roei Haliolis spp. H. fulgens, H. corrugata H. cracherodii, H. rufescens Annelid Worms Haliolis spp. H. diversicolor aquatilis, H. kamtschalkana. H. tuberculata Haliolis spp. Crustaceans H. asinia, H. coccoradiala. H. sqamata H. fulgens, H. corrugata. Haliolis spp. from California Virus? Intracellular prokaryote Similar to Rickettsiales "Candidatus Xenohaliolis califoniiensis" Vibrio fluvialis-ll Vibrio alginolylicus Clostridium liluseberense Haliphlhoros milfordensis Fungi Debaryomyces hansenii Flabellula cliala and Clydonella rosenfieldi Margolisiella f= Pseudoklossiaj haliolis Labyrinthuloides lialiolidis Perkinsus ol.^cni Cllona celalii calilorntaiui Proctoeces sp. Trematode Opecoelidae Echinocephalus pseudouncinalus Patelloida nigrosulcala Pyraniidellid snails Lilliophaga arislala Lithophaga plwnula Penitella conradi Terebra.sabella Heterounciniala Polydora armata, P. websleri, P. flavala oriemalis, Polydora spp. Broccardia knoxi Pea crabs Copepods Belaeus harfordi -10 0 -/o -/o -/o -/o 0 0 0 Tumor-like Possibly Witherine syndrome Pustule disease Juvenile vibrosis Clostridium infection Mycosis Fungi infection Coccidiosis Infection by L. lialiolidis Perkinsiosis Infestation by Cliona Trematode parasitism Trematode parasitism Red disease Nematode parasitism Pyraniidellid infestation Boring clams infestation Peniiella infestation Sabellid pest Mud worm or blister worm infestation. Mud worm or blister worm infestation. Harada et al. 199.1 Gardner et al. 1995. Alsatt et al. 1996. Alvarez-Tinajero et al. 1999. Friedman et al. 2000. Taiwu et al. 1998 Elston and Lockwood 198.1. Di.xon and Hecht 1991 Dixon and Hecht 1991 Hatai 1982 Friedman 1997. Grindley et al. 1998 Gardner et al. 1995 Gardner et al. 1995 Friedman et al. 1995 Dresser and Bower 1997 Bower 1987a. Bower 1987b Goggin and Lester 1995 Hansen 1970. Clavier 1989 Shimazu 1970 Crofts 1929 Romero 1996 Milleman 1951. Milleman 1963 Scheibling et al. 1990 Roberts and Orr 1961 Alvarez-Tinajero et al. 2000 Hansen 1970. Alvarez- Tinajero et al. 2000 Oakes and Fields 1996 Fitzhugh and Greg 1999 Hansen 1970, Kojima and Imajima 1982, Clavier 1989. Blake 1996. Handley and Berguist 1977 Geiger and Martin 1997 Alvarez-Tinajero et al. 1999 Chace and Abbott 1980 Interaction level: - negative, 0 neutral, ? unknown. -10 negative or neutral depending on the infestation degree. Red Abalonh Symbionts from Baja California 877 (1997): (1+) nococcidia, (2+) 1-10 coccidia per field of view. (3+) 1 1-100 coccidia per field. (4+) 101-1000 coccidia per field and (5+) >1000 coccidia per field. Similarly. RLP infection intensities were ranked according to the number of bacterial foci into the following categories: (1 + ) no RLP. (2+) 1-10 RLP. (3+) I 1-100 RLP. (4+) 101-1000 RLP and (5+) > 1000 per field of view at 200x magnification. The appearance of the digestive gland was also quantified following similar criteria to Friedman et al. ( 1997). A score of (1+) was given if tissues were normal (Bevelander 1988). (2+) if moderate degeneration of the tissues was observed, and (3+) for severe degeneration of the tissues. Protozoa in the e.sophageal pouch and gill branchiae were counted from each his- tological preparation. RESULTS The mean shell length (±SE) of apparently healthy abalone was 34.4 mm (±1 .03 ). whereas that of moribund animals was 25.9 mm (±0.67) and empty shells 24.0 mm (±0.2). An ANOVA comparing the means indicated no significant difference in shell length be- tween moribund abalone shells and empty shells, although these groups' shells were significantly smaller than those of apparently healthy abalone (Student-Newman-Kleus. P < 0.001 ). Apparently healthy abalone were active, responded to tactile stimulation. while the color and consistency of their foot muscle were consid- ered normal. Moribund abalone were weak, retracted and had a flaccid foot muscle. They also showed slow tactile responses. Analysis of the shell revealed the presence of two polychaete symbionts. one belonging to the family Serpulidae (spirorbids). and the other to the Spionidae, probably from the genus Polydora. The spirorbids formed calcareous tubes on the external surface of the shell. Respiratory holes were not occluded and neither defor- mation of the shell, nor perforations toward the internal side of the shell were caused by these worms. They were present on all ob- served shells (in all groups), shell coverage ranging from medium to high. The prevalence of spionids in healthy abalone was 10% with an intensity of 2 worms per host: in moribund abalone. preva- lence was 1% with an intensity of 1.5 (±SE 0.12). and in empty abalone shells prevalence was 13% with an intensity of 2.5 (±SE 32). The histopathological study revealed the presence of macroga- metes (Fig. la) and microgametes (Fig. lb) of a renal coccidian parasite in the left and right kidneys. Coccidia were present in 72% of moribund and lOVr of healthy abalone. The degree of infection in the moribund group ranged from moderate to heavy (4-5+) whereas it was moderate (A+) in apparently healthy abalone. RLPs were found in the epithelial cells of the digestive tract, including the posterior esophagus, digestive diverticuli (Fig. 2a) and intestine. They were present in all moribund animals and 849;- of the apparently healthy abalone. Apparently healthy abalone in- fected by RLP showed no detectable clinical signs of the WS. The degree of infection in moribund animals ranged from mild to heavy (3-5+). and 64% of them showed degeneration of the di- gestive gland ranging from moderate (2+ in 28%) to severe (3+ in 72.2%). In heavily infected animals, expelled inclusion bodies were found in the lumina of the digestive tract (Fig. 3b). In con- trast. 31% of the apparently healthy infected abalone presented only mild digestive gland degeneration (2+). Apparently healthy abalone weighed significantly more than moribund abalone (6.62 Figure 1. lai \l;Riipj;amilts of MargoUsiella (Pseudoklosia) haliotis (large arrovvl in tht left kidney of a healthy abalone. Small arrows are showing hemocite around the parasite, (b) Microgametes of ,W. haliotis < arrows I in the right kidney of a moribund abalone. Scale bar = 20 ^m. g SE ± 0.54 vs. 1.84 g SE ± 0.1 1 respectively, r-test, P < 0.001 ). The condition indices of apparently healthy abalone (0.6) were higher than those recorded for moribund abalone (0.34). but were similar between apparently healthy without and with (0.6) RLP infections (Kruskal Wallis test. H = 31.6. P < 0.001 and Dunn's pairwise multiple comparison method P < 0.005). These condition values con'elated with degeneration of the digestive gland ob- served in moribund animals. A ciliate protozoan with a granular endoplasm and large vacu- oles was found in the lumen of the esophageal pouches (Fig. 3a. b). The unidentified ciliate was pleomorphic and measured 40 x 12 to 90 X 19 (j.m. The ciliate was occasionally observed in contact with the epithelial cells of the esophagus, but was generally free in the lumen of the esophageal pouches. Prevalence in apparently healthy abalone was 78%. compared to 57% in moribund abalone. The maximum intensity of the protozoan in histological sections was 330 (mean 50.9 ± SE 18. S) in apparently healthy abalone and 120 (mean 24 ± SE 10.1) in moribund abalone (Mann-Whitney Rank sum test. P > 0.05). Other ciliates of different sizes were observed in the branchial cavity and among the lamellae of the branchia (Fig. 4a). These organisms contained a granular cytoplasm with large vacuoles (Fig. 4b). Their prevalence in healthy abalone was 17% compared to 88% in moribund abalone. The maximum intensity of this pro- 878 Caceres-Martinez and Tinoco-Orta Figure 2. i:ii Uikillsiiilis-liki- |)riikaryotes in the post-esopliagus epi- thelial coluninar cells of an infested morihund abalone. Note rupture of some columnar cells, (arrow in the middlel, expelled RIJ' (arrow), and great quantity of mucus and hemocjtes in the lumen of the post- esophagus (hig arrows). Scale bar = 125 (im. (b) A close up of the infected epithelium and some expelled bacterial foci (big arrow) and rupture in the epithelium (small arrow ). Scale bar = 20 pm. tozoaii in healthy abalone was 21 (mean 10.2 ± SE 4.1 ). compared to 85 (mean 30.3 ± 6.7) in moribund abalone, although the differ- ences were not significant (/-test. P > 0.05). DISCUSSION Although symbionts of wild and cultured abalone populations in California have been widely documented (Friedman 1991. Haaker et al. 1992. Lafferty & Kuris 1993. Davis 1993. Oakes & Fields 1996. Friedman et al. 1997. Shields & Perknis 1997). this is the first health evaluation of cultured red abalone from Baja Cali- fornia. We found the presence of a disease and several symbionts similar to those recorded in abalone from California. This is in accordance with the fact that the culture of red abalone in Baja California depends on the importation of brood and seed stock from California. Comparison of the diseases and symbionts of the cultured abalone with local red abalone populations in the wild was not possible due to the scarcity and difficult accessibility of these natural populations. These difficulties explain the need for importing red abalone from California. There is no information on the possible effect of Polydora spp. in small and thin shells of young red abalone. In oysters, mussels and clams the polychaete complex. Polydoia-Broccardia. results Figure 3. (a) Protozoans (arrows) in a healthy abalone's esophagous pouch lumen. Note that there is apparently no reaction of the host. .Scale bar 80 pm. (b) Close up of a protozoan touching the epithelial cells of the esophagous pouch. Scale bar = 20 pni. in loss of condition, indirect high mortality and economic losses (Lauckner 1983. Handley & Berguist 1997). Several members of the complex have been observed in Halio!is spp. (Hansen 1970. Blake 1 996. Handley & Berguist 1 997 ). Contrary to Clavier (1989) who reported that H. tuhenuhita smaller than 50 mm were not infested by Pnhdoni spp.. we found burrowing worms in indi- viduals smaller than this size. Serpulids are commonly attached to pilings, floats, algae, mussels (Abbott & Reish 1980) and infested abalone, although they apparently do not represent any risk. There was no evidence of presence of the sabellid polychaete Terebrasa- hella heterouncinala which Fitzhugh and Greg (1999) report in- fests abalone in California. Margolisiella ( = Pseudoklossia) lialiotis has been recorded in the black abalone H. cracherodii. red abalone H. nifesceus. blue- green abalone H. fulgens and pinto abalone. H. kamschalkana (Friedman et al. 1993. Friedman et al. 1995). This coccidian has a broad geographic range that includes Bahi'a de Todos Santos Mexico and the California coast. It does not seem to be lethal to these host species. This coccidian has been discounted as a causal agent of WS (Friedman et al. 1993. Friedman et al. 1995. Friedman et al. 1997). The prevalence of coccidia recorded in moribund abalone (72%) was similar to those found in abalone from central and southern California (Friedman et al. 1997). A lower preva- lence and intensity of infection in apparently healthy abalone com- pared to moribund abalone may be coincidental. However, further Red Ahaloni-; Symbionts from Baja California 879 . .'^'W^ * >'•' .CAi: Figure 4. (a) Several ciliated prolo/iia (thin arroHsl among tlie l)ran- ihial niaments ot a mi>ril)und abalone. Note the necrotic /.ones in the branchial lllaments (wide arrow). Scale bar = 11)0 yun. (hi A close up of ciliated protozoa around the branchial filaments of a moril)und abalone. Note the large nucleus and vacuoles in the protozoan. Scale bar = 20 pm. Studies on alterations in abalone pliysiology in relation with renal coccidia are needed. The presence of RLP in cultured Mexican stocks of red abalone must be considered in the aquaculture facilities management and sanitary practices. The majority of the observations of WS in the red abalone were reported from laboratory/aquaculture facilities in California (Moore et al. 2000). Friedman (1995) reported that dur- ing a survey of abalone for WS in California, from 1 to 4% of the red abalone from San Miguel Island had visual signs of WS. Mori- bund abalone in our studies were characterized by retracted vis- ceral tissues, discoloration and shrunken appearance of the foot muscle, low condition, presence of RLP in the digestive gland and some degeneration of the digestive gland, suggesting that these abalone were affected by WS. A similar situation has been re- ported by Moore et al. (1999) who associated dramatic mortality of red abalone from aquaculture facilities with the WS. Withering syndrome in black and red abalone has a long incubation period {^-1 months and 7-8 months respectively), before clinical signs are visible (Friedman el al. 1997. Moore & Friedman unpublished results cited by Friedman et al. 2000). Hence, RLP infected indi- viduals are undetectable by visual condition. The degree of infec- tion in asymptomatic but infected abalone was similar to those found in moribund abalone and could be explained by suscepti- bility differences among individuals and incubation period. Our sampling was earned out during the 1997 El Nino when tempera- ture was 2°C above normal in the area. High water temperatures have coiTesponded with increased mortality of RLP-infected black abalone (Friedman et al. 1997). Undoubtedly, more thorough and detailed studies (using large samples) concerning RLP infection in different abalone species, and studies focusing on the identification and differentiation of RLP species or strains associated with aba- lone species are needed. The similar prevalence and total niunbers of ciliates from the esophageal pouches in healthy and moribund abalone found in the present study suggest a commensal symbiosis. Further taxonomic and biological studies are needed to determine the species of this protoz.oan and the nature of the sytnbiosis. Vanblaricom et al. (1993) recorded the presence of suctorian ciliates in gill squashes and mantle scrapings of healthy black abalone and those apparently aftlicted by the WS. Presence of these ciliates did not relate to clinical signs of the disease in the host abalone. A considerable number of ciliates has been associ- ated with marine bivalves and the majority of these appear to be harmless commensals (Lauckner 1983). In this study, similar prevalence of ciliates in apparently healthy and moribund abalone suggested that these organisms are not related to the moribund condition of some of the abalone examined. An increase in total numbers of protozoans on the gills coincides with host weakness and death (Lauckner 1983) which may be due to the saprozoic habits of some species (Kudo 1982). Abalone from Baja California have almost the same symbionts species as those in California. This highlights that it is necessary to maintain a continuous monitoring program to avoid introduction of new potentially dangerous symbionts to the region. ACKNOWLEDGMENTS Rebeca Vasquez-Yeomans and Sergio Curiel Rami'rez helped us in processing the samples and in data analysis. James D. Moore, from Bodega Marine Laboratory sent us a histological slide of a red abalone in order to compare our results. This project was supported by project number 623 1 06-CICESE. LITERATURE CITED Abbott. D. P. & D. J. Reish. 1980. Polychaeta: the marine annelid worms. In; R. H. Morris. D. P. Abbot & E. C. Haderlie. editors. Interiidal Invertebrates of California. Stanford: Stanford University Press, pp. 44S-489. Altstatt, J. M., R. F. Ambrose, J. M. Engle, P. L. Haaker, K. D. Lafferty & P. T. Raimondi. 1996. Recent declines of black abalone Halioiis crach- erodii on the mainland coast of central California. Mar. Ecol. Prog. Ser. 142:185-192. AIvarez-Tinajero, M., J. Caceres-Marti'nez, J.G. Gonziilez-Aviles & CD. Tinoco-Orta 1999. Analisis histopatologico del abiikin a/ul Haliotis fulgens y amarillo Hiiliolis corrugata de Isia de Cedros, Baja Califor- nia. Mexico. IV Congreso Latinoamericano de Malacologi'a. Co- i|Liimbo, Chile, fi-IO Septiembre. Abstracts.zx AIvarez-Tinajero, M. 2000. Simbionles de los abulones amarillo {Haliotis cormgata) y azul (Haliotis fulgens) de Isia de Cedros e Islas San Benito, Baja California, Mexico. Tesis de Maestria. Centre de Inves- tigacion Cienti'fica y de Educacion Superior de Ensenada. 107 pp. Bevelander. G. 1988. Abalone gross and fine staicture. Pacific Grove: The Boxwood Press. 80 pp. Blake. J. A. 1996. Family Spionidae. Gruhe. IS.SO. In: Taxonomic Atlas of 880 Caceres-Martinez and Tinoco-Orta the Santa Maria Basin and Western Santa Barbara Channel, vol. 6. The Annelida. Part 3. Polychaeta: Orbiniidae to Cosuridae. Santa Barbara: Santa Barbara Museum of Natural History, pp. 81-223. Bower, S. M. 1987a. Labyrinthuloides haliotidis n. sp (Protozoa: Laby- rinthomorpha). a pathogenic parasite of small juvenile abalone in a Bnstish Columbia mariculture facility. Can. J. Zool. 65:1996-2007. Bower, S. M. 1987b. Pathogenicity and host specificity of Labyrinthu- loides haliotidis (Protozoa: Labyrinthomorpha), a parasite of juvenile abalone. Can. J. Zool. 65:2008-2012. Chace, F. A. & D. P. Abbot. 1980. Caridea: The Shrimps. In: R. H. Moms. D. P. Abbot & E. C. Haderlie, editors. Intertidal Invertebrates of Cali- fornia. Stanford: Stanford University Press, pp. 567-576. Clavier. J. 1989. Infestation of Haliotis lubercidata shells b\ Cliona celaia and Polydoia species. Proceedings of the 1st International Symposium on Abalone. La Paz, Mexico, pp. 16-20. Crofts. D. R. 1929. Haliotis. Liverpool Mar. Biol. Comin. Mem. T\p. Br. Mar. Plants Anim. 29:1-174. Culver, C. S. & J. B. Richards. 1992. Black abalone mortality: establishing a research agenda. Summary of a workshop sponsored by the Sea Grant Extension Program, University of California, Cooperative Extention. Report No. T-CSGCP-024. 31 pp. Davis. G. E. 1993. Mysterious demise of southern California black aba- lone, Haliotis cracherodu Leach, 1814. / Shellfish Res. 12:183-184. Dixon. M. G. & T. Hecht. 1991. Identification and treatment of a Clostridium and Vibrio infection in South African abalone. Haliotis midae L. / Fish Dis. 14:693-695, Dresser, S. S. & S. M. Bower. 1997. Margolisiella kabata gent et sp, n (Apicomplexa: Eimeriidae), a parasite of native littleneck clams, Pro- tothaca stamiiiea. from British Columbia, Canada, with a taxonomic revision of coccidian. Folia Parasit 44:241-247. Elston, R. & G. S. Lockwood. 1983. Pathogenesis of vibriosis in cultured juvenile red abalone, Haliotis rufescens Swainson, J. Fish Dis 6:1 1 1- 12S. Fitzhugh. K. & G. W. Greg. 1999. A remarkable new genus and species of fan worm (Polychaeta: Sabellidae: Sabellinae) associated with marine gastropods. Imertebr. Biol 118:357-390. Flores, A. R.. J, C. Vivanco, M. Villalobos, E. Arroyo & E. Vazquez. 1995. Abulones Cultivados: the first commercial scale abalone farm in Mexico. Abst. Aquaculture 95 World Aquaculture Society, pp. 1 18. Friedman, C. S. 1991. Coccidiosis of California abalone, Haliotis spp. / Shelfish Res. 10:236. Friedman, C. S., W. Roberts, G. Kismohandaka & R. P. Hedrick. 1993. Trasmissibility of a coccidian parasite of abalone, Haliotis spp. / .Sliell- fish Res. 12:201-205. Friedman. C. S. 1995. Current status of withenng syndrome. .Abahme Focus 1:1-2. Friedman, C. S.. G. R. Gardener, R. P. Hedrick, M. S. Stephenson. R. J. Cawthom & S. J. Upton. 1995. Pseiidoklossia haliotis sp. n. (Apicom- plexa) from the kidney of California abalone, Haliotis spp. (Mollusca). J. Invertebr. Path. 66:33-38. Friedman, C. S., M. Thorson, C. Chun, P. L. Haaker & R. P. Hedrick. 1997. Withering syndrome of the black abalone. Haliotis cracherodii (Leach); water temperature, food a\ailahilily. and parasites as possible causes. / Shellfish Res. 16:403-41 1. Friedman, C. S.. K. B. Andree, K. A. Beauchamp, J. D. Moore, T. T. Robbins, J. D. Shields & R. P. Hedrick. 2000. ■Candidatus Xenohali- otis califomiensis' . a newly described pathogen of abalone Haliotis spp., along the west coast of North America. Iin. J. of Syst. and Evol. Microbiol. 50:847-855. Gardner, G. R.. J. C. Harshbarger, J. L. Lake, T. K. Sawyer, K. L. Pnce. M. D. Stephenson, P. L. Haaker & H. A. Togstad. 1995. Association of prokaryotes with symptomatic appearance of withering syndrome in black abalone Haliotis cracherodii. J. Invertebr. Path. 66:111-120. Geiger, D. L. cS: J. W. Martin. 1997. Symbiotic Pea crabs (Decapoda: Pinnotheridae) in abalone. Monterey: Biology. Fisheries and culture ot the third International Abalone Symposium, pp. 80. Goggm. C. L. & R. J. G. Lester. 1995. Mar. Freslmal. Res. 46:639-646. Gray. P. 1954. The microtomist formulary and guide. Blakiston, New York. 749 pp. Grindley. R. M.. J. A. Keogh & C. S. Fnedman. 1998. Shell lesions in New Zealand W(i/;(>//,v spp. (Mollusca. Gastropoda). J. Shellfish Res. 17:805- 811. Haaker, P. L., D. V. Richards. C. S. Friedman. G. E. Davis. D. O. Parker & H. Togstad. 1992. Mass mortality and withenng syndrome in black abalone Haliotis cracherodii in California, pp. 214-224, In: S.A. Shep- ard, M. J. Tegner. & Guzman del Proo, editors. Abalone of the World, Oxford: Blackvvell Scientific. Handley, S, J. & P. R. Berguist. 1997. Spionid polychaete infestations of intertidal Pacific oysters Crassostrea gigas (Thumberg), Mahurangi harbor in northern New Zealand. Aquaculture 153:191-205. Hansen, J. C. 1970. Commensal activity as a function of age in two species of California Abalone. Veliger 13:90-94. Harada, T.. N. Okazaki, S. Kamiya. Y. Otoishi, Y. Hayakawa & S. S. Kubota. 1993. Tumors in nervous tissues of Abalone, Nordotis discus. J. Invertebr. Pathol. 62:257-261. Hatai, K, 1982. On the fungus Haliphlhorus milfordensis isolated from temporarily abalone [Haliotis sieboldii). Fish Pathol. Tokyo 17:199- 204. Kojima, H. & M. Imajima. 1982. Burrowing polychaetes in the shell of the abalone Haliotis diversicolor aquatilis chiefly on the species of poly- dora. Bull Japan. Soc. Sci. Fish 48:31-35. Kudo, R. R. 1982. Protozoologia. Ed. Continental. 905 pp. Kuris, A. M. & C. S. Culver. 1999. An introduced sabellid polychaete pest of cultured abalone and its potential spread to other California gastro- pods, (in press). Lafferty, K. D. & A. M. Kuris. 1993. Mass mortality of abalone Haliotis cracherodii on the California Channel Islands: test of epidemiological hypotheses. Mar. Ecol. Prog. Ser. 96:239-248. Lauckner. G, 1983, Diseases of Mollusca: Bivalvia. In: O. Kinne. editor. Diseases of Marine Animals, Vol. II. Introduction to Scaphopoda. Bi- ologische Anstalt Helgoland, Hamburg, pp, 520-603. Milleman, R. E. 1951. Echinocephalus pseudouncinatus n. sp.. a nematode parasite of the abalone. J. Parusitoi 37:435—439. Milleman, R. E. 1963. Studies on the taxonomy and life history of echi- nocephalid worms (Nematoda: Spiruroidea) with a complete descrip- tion of Echinocephalus pseudouncinatus Milleman, 1951. J. Paras. 49:754-764. Moore, J. D., T. T. Robbins & C. Friedman. 1999. Withering syndrome in farmed red abalone. Haliotis rufescens. 91" Annual Meeting of the National Shellfisheries Association. April 18-22, Halifax, Nova Scotia. Abstracts, pp. 37. Moore, J. D.. T. T. Robbins & C. S. Friedman. 2000. Withering syndrome in farmed red abalone. Haliotis rufe.uens: thermal induction and asso- ciation with a gastrointestinal Rickettsiales-like prokaryote. ./. Aquat. Anim. Health 12:26-34. Oakes, F. R. & R. C. Fields. 1996. Infestation of Haliotis rufescens shells by a sabellid polychaete. Aquaculture 140:139-143. Roberts, R. & V. Orr. 1961. Review of pyramellid host with notes on an Odostomia parasitic on a chiton. Nautilus 74:85-95. Romero, H, L. S. 1996. Descripcion morfologica y abundancia de los parasitos en gonada de abulon azul (Haliotis fulgens. Philippi) de Isla Maedalena. B. C. S., Mexico. Tesis de licenciatura. Universidad Au- tonoma de Baja California Sur. Mexico. 67 pp. Scheibling. R. E., T. Evans, P. Mulvay, T. Lebel, D. Williamson & S. Holland. 1990. Commensalism between an epizoic limpet. Patelloida nigrosulcata. and its gastropod host, Haliotis roei and Patella laticos- Red Abalonh Symbionts from Baja California lata, on intertidal platforms off Perth, Western Australia. Ausi. J. Mat: Freshwul. Res 41:647-645. Shaw, B. L. & H. I. Battle. 1957. The gross microscopic anatomy of the digestive gland of the oyster CrassosUea virginicii (Gmelin). Ciiiuu/iiiii J. Zitol. 35:325-346, Shields, J. D, & F. O, Perkins. 1997. Parasitological examination of wast- ing disease in black abalone, Hulioris cracheroJii. National Marine Fisheries Service. Final Report. 22p. Shimazu, T, 1970. A metacercaria of the aenus Prticlocccs (Fellodistomi- dae: Trematoda) from abalone, Hctliotis discus lumimi. of Rebun Island, Hokkaido. Bull Jap. Soc. Scieni. Fish. 38:813-816. Taiwu, L. D.. Z. Mingjin, X. Jian, Jianhai & L. Ruiyu. 1998. Studies on the pustule disease of abalone Haliotis discus hannai Ino. On the Dalian cost. J. Shellfish Res. 17:707-711. Vanblancom, G. R., J. L. Ruediger, C. S. Friedman. D. D. Woodard & R. P. Hedrick. 1993. Di.scovery of the withering syndrome among black abalone Haliotis craeherodii Leach, 1814, populations at San Nicolas Island, California. / Shellfish Res. 12:185-188. Jininuil oj Shcllfisli Research. Vol. 20, No. 2. 88.V88S. 2{)()1. LIFE HISTORY OF AN EXOTIC SABELLID POLYCHAETE. TEREBRASABELLA HETEROUNCINATA: FERTILIZATION STRATEGY AND INFLUENCE OF TEMPERATURE ON REPRODUCTION' CARL A. FINLEY." TIMOTHY J. MULLIGAN,' AND CAROLYN S. FRIEDMAN" '* ^ Hniuboklt State University. Areata. CA 9552I-S299; -Department of Medieinc and Epidemiology, Seliool of Veterinary Medieine. University of California, Davis, CA 95616: California Department of Fisli and Game and Bodega Marine Lahoratoiy. P.O. Box 247. Bodega Bay. CA 94923 ABSTRACT The California ahalonc aquaculliirc industry has hccn struggling to nd itsell of an exotic sabellid. Teiflva.saljella lu'tenmncimua following its inadvertent introduelion from South Alrica in the late lySOs. The developnieiU of an effective manage- ment strategy is dependent upon understanding the life history of this sabellid. including its fertilization strategy (e.g. self-fertilization) and its generation time. In the present study, red abalone. Haliolis mfescen.s Swainson 1822. with single sabellid infestations were isolated in containers at 18°C (single host and sabellid per container). This first, parental generation (P). was held in isolation until individuals produced F, larvae. The F, larvae were subsequently isolated until individuals produced a second. F2. generation. In a separate study, uninfested red abalone were exposed to sabellid infested abalone at 1 1.2°C. 15.6°C, and 20.9^C, temperatures typically encountered in California. The larvae were subsequently observed as they developed to specific life stages: the initiation of feeding, the development of all 1 1 setigers (which closely relates to sexual maturation) and the completion of their life cycle as recognized by the production of motile, infesting, larvae. Approximately 50% of the sabellids examined at 1 1 .2°C. I5.6°C. and 20.9°C had developed the ability to feed by day 6, 5 and 4 (P < 0.001 ), had developed all 1 1 setigers by day 83, 68 and 48 [P < 0.001 ) and had produced larvae by day 298, 165 and III (f < 0.001 ), respectively. This research demonstrates that isolated individuals do pose the threat of producing fullv functional offspring and that the generation time of T. Iicwnnmcinahi is significantly temperature dependent. KEY WORDS: sabellid. abalone. life history, self-fertili/ation. TeiclvasalH'lla INTRODUCTION The California abalone aquaculture industry is piescntly strug- gling with two serious diseases, one of which is an introduced sabellid polychaete, recently described as Terehrasahclta hctcr- ouneinata (Fitzhugh & Rouse 1999). This sabellid was cited by California growers as their most serious "problem and constraint," reporting that the reduced growth rates and negative public per- ception associated with the worm were having a substantial nega- tive impact on the industry (Mcbride 1998). Terehrasahetla het- eroimeiiuita is believed to have been introduced from South Africa in the late 1980s, via the importation of infested research abalone (Kuris & Culver 1999. Ruck & Cook 1998). Concern exists re- garding the threat that the sabellid poses to the California cultured abalone industry and to native gastropod populations if It were to become established within the state's intertidal ecosystem (Kuris & Culver 1999). Presently, there is one documented case of T. hetermmciuuta infestations in gastropods (snails and limpets) ad- jacent to a culture facility (Kuris & Culver 1999). Broad-scale removal of infested hosts, combined with clean-up efforts at the culture facility may have curtailed the sabellid from becoming permanently established in the intertidal enviionment (Kuris & Culver 1999). Infestations occur by the unique ability of T. helerouneinata to overcome the host abalone's defenses and settle upon the under side of the leading edge of the shell. Normal shell deposition is disrupted as the host attempts to cover the irritant with nacre. In heavily infested abalone. the deposition of the prismatic shell layer all but ceases, resultina in the domed shell found in association *Corresponding author. E-mail address: carolynt'@u. washington.edu. Cur- rent address: School of Aquatic & Fishery Sciences, University of Vv'ash- ington, PC Box 355020, Seattle, WA 98195. "Contribution No. 2140, Bodega Marine Laboratory, University of Cali- fornia at Davis. with the reduced growth rates (Culver et al. 1997. Kuris & Culver 1999. Ruck& Cook 1998). Attempts to control T. lu'tcniuuciuatu m commercial settings have included manipulations of water temperatures, the coating of abalone shells with wax, quarantine of infested stocks, and im- proved sanitation practices (Oakes & Fields 1996, Oakes et al. 1995. Leighton 1998). Many of these techniques have been met with limited success, allowing low-level infestations to persist. Suggested new treatments include novel therapeutic delivery sys- tems using micro-encapsulation (Ruck & Cook 1998. Shields et al. 1998) and the use of ultrasound (Loubser & Domiehl 2000). Un- fortunately, none of these techniques, to date, have been success- fully applied in a culture setting. The initial description of T. hctcroiiuciuaici noted that individu- als were siimiltancous hermaphrodites, but did not indicate if they were functional hermaphrodites, capable of self-fertilization. Upon settling at the 7th setiger stage, the sabellids are covered with nacre by the host resulting in the passive formation of what becomes the adult's tube. Development is characterized by the formation of a branchial crown and four additional setigers. As the 1 1th setiger is formed, an alteration of the type of uncini on the 6th setiger also occurs, both events occur as individuals are recognized as sexually mature (Fitzhugh & Rouse 1999), Spermatogenesis occurs in the 8th setiger and oogenesis occurs in both the 9th and 10th setigers. The mature adult broods several embryos within its tube for an unknown, temperature dependent, amount of time. The embryos eventually mature and develop into motile larvae that subsequently emerge from the tube of the adult and seek a suitable area of attachment (Fitzhugh and Rouse. 1999). In the following study we examined the fertilization strategies and the life history of T. helerouneinata. We investigated if an isolated sabellid is capable of reproduction through two genera- tions. We assessed the reproductive capabilities of T. heternunei- nata by quantifying generation times at three water temperatures that reflect the range of temperatures encountered in California. 883 884 FlNLEY ET AL. MATERIALS AND METHODS Animals and Husbandry Commercially reared red abalone. Haliotis iKfesceiis Swainson 1822, that measured 15-25 mm in shell length, were purchased from or donated by culturists in California. All animals were held in the Pathogen Quarantine Facility at the Bodega Marine Labo- ratory, Bodega Bay, California. All effluent produced at this fa- cility is chlorinated (>10 ppm CI" for 2-3 h) and dechlorinated prior to release. Animals were reared in tlow-through. full strength, sand-filtered seawater. E.xperiments were conducted at average mean temperatures of ambient (12.6^C). heated (15.6'C, ~18°C and 20.9°C) or chilled ( W.l'C) seawater. All abalone were fed Macrocystis pyrifera Agardh 1820, ad Uhitiim and remained under artificial light on a 12-hour dark. 12-hour light cycle for the duration of these experiments. Strict sanitation protocols were fol- lowed to insure that no cross-contamination occun'ed. Single Infestations Trial I Fifty uninfested red abalone were placed into a 10-1 container \\ ith five red abalone, heavily infested with T. hcternuncinata, for 24 hours. The newly infested abalone were inspected with a dis- secting microscope 48 hours post-exposure and infestations were quantified. Fourteen abalone with single sabellid infestations were removed and individually isolated in 200-ml containers. Four aba- lone, that had multiple sabellid infestations per abalone, were iso- lated in a similar manner to serve as positive controls. Four unin- fested abalone were isolated in a similar manner to serve as nega- tive controls. All 22 containers received ambient seawater. -12.6°C. and were randomly placed on a wet table. Abalone were sampled every 14 days by gently pushing aside the epipodium and mantle tissue and inspecting the leading edge of the shell for the presence of recently settled larvae. Following the discovery of new larvae, treatments were sampled every seven days. Uninfested abalone were added to any container in which the original infested host abalone had died, providing live hosts for larval settlement. The abalone were observed for a total of 32 weeks, at which point the sabellids that had not produced larvae were carefully dissected from their tubes and observed under dis- secting and compound microscopes to determine if signs of repro- duction were visible. Single Infestations Trial 2 Forty uninfested red abalone were exposed to T. heteroimcinata infested abalone as de.scribed in Trial 1. Abalone were inspected 48 hour post-exposure and the infestations were quantified. Four abalone with single parental (P) generation infestations (Fig. 1) and four uninfested abalone (negative control) were removed and individually isolated in 7()0-ml containers. Animals were visually inspected weekly. Sabellid larvae settle preferentially on live hosts (C. A. Finley & C. S. Friedman, unpublished observation). To allow F| larvae to settle only on recipient, sabellid-free abalone. infested (P host) abalone were sacrificed after 60 days post- exposure and a single uninfested abalone (larval recipient) was placed into each tank with the infested shucked shells. Once a recipient abalone became infested with an F, larva, the newly infested animal was removed and isolated (single abalone with single sabellid) in a 700-ml container. Additional uninfested aba- lone were added to the containers with the P generation sabellids to attain multiple individuals with single F, infestations (Fig. 1). r ^ is ~) r •) r -) Figure 1. Individual ahalone from single infestation trial 2. Two suc- cessive generations of sabellids and the number of days until offspring «ere produced is Illustrated. Superscripts ( 1— 4| indicate ( 1) tbe paren- tal generation. (2) the number of days until larvae were observed, (3) the F, larvae produced b> the parental bv the parental generation and (4) the F, larvae produced bj the F, generation. Host abalone containing F, larvae were visually inspected for signs of F, larvae. All containers received heated seawater (~18°C) for the duration of this trial. The abalone were observed for a total 497 days, at which point any sabellids that had not produced larvae were carefully removed and observed for signs of reproduction. Life History Trial Terehrasabi'lla lu'leroiiiuiiuilci infestations were achieved by commingling 200 uninfested and 20 infested (10 live and 10 shucked) red abalone in a 10-1 container for a 24 hour period. Following the exposure the 200 newly infested abalone were di- vided into three replicates tanks. Exposures were conducted at 1 1.2°C, 15.6°C, and 20.9X. Infestations were quantified 48 hours after the end of the exposure period. The removal of the 20 heavily infested donor abalone marked time zero and every subsequent 24-hour period was regarded as an additional day of development. Development to the following three life stages was recorded: the ability to feed, the development of the full complement of setigers and the production of infestive, motile larvae. Replicates were sampled every 24 hours for the first eight days, followed by weekly sampling. At each sampling point, six sabellids were re- moved from each replicate by sacrificing the host abalone and gently breaking up the shell with a scalpel. The ability to feed was determined by providing the sabellids with suspended stained lipid beads (LBs). Following the protocol of Shields et al. (1998) microencapsulated, sudan black-stained, tripalmatin lipid beads that ranged in size from 3 to 30 p.m were produced. Several abalone, with a total of six sabellids, were re- moved from each replicate tank and placed into three 200-ml containers to which 0.3 g of the lipid beads were added. The seawater-lipid bead suspension was stirred every 10 minutes. After 30 minutes, the abalone hosts were sacrificed and the sabellids were excised and examined for the presence of LBs in the diges- tive tract using a light microscope. We enumerated the number of feeding sabellids, terminating the LB exposures when all sabellids in a given replicate were able to feed. To identify the development of all 1 1 setigers, six sabellids Sabellid Life History 885 Figure 2. Abalone shell showing both the F, and F, generation. Bar = Smni. were removed weekly from each tank and the number of setigers was enumerated using a hght microscope at 600x magnification. When 100% of the sabellids sampled from each replicate had developed a full complement of setigers, sampling was reduced to once every I— t weeks depending on the number of sabellids re- maining and the temperature under investigation. The completion of the life history of T. hetennmcinata was defined by the production of either a motile, infesting larva or by a recently settled larva. Sabellids were excised as above and the presence or absence of motile or newly settled larvae was re- corded. The experiment terminated when 100% of the sabellids inspected from each replicate completed their life cycle. Statistical Analysis With the assumptions of an ANCOVA not being met. an alter- native way to analyze the data was needed. Linear regression lines were calculated for each replicate. The number of days that it took 509(- of the sabellids to reach a dexelopinental stage (T^o). was then calculated from the regression line using the formula Y = bX -I- a. setting the dependent variable. Y. equal to 50% and solving for X. The T,|| estimates were then compared using one-way analysis of variance (ANOVA). The data from the replicates at each temperature were also pooled and the number of days that it took for 50% of the sabellids to reach a developmental stage was calculated. The T^,, estimates from the pooled replicates at a given temperature were then pre- sented as the best estimate for the number of days post-settlement required for 50% of the sabellids to reach a given developmental stage at each temperature. RESULTS Single Infestations Trial I. Larval production by six of the 14 indi\idually isolated sabel- lids began during weeks 26 to 32 (Table I ). Of the sexen sabellids that had not produced any larvae, three contained well-developed internal eggs, one contained both internal and external eggs and embryos, and the remaining three tubes were devoid of sabellids at the termination of Trial 1 . One abalone was contaminated bv the effluent from another tank and was discarded. Abalone with multiple sabellid infestations (positive control) also began producing larvae during week 26. Sabellids on 3 of the 4 abalone produced larvae (Table 1 ) and only vacant sabellid tubes were found on the fourth animal at the termination of Trial 1. No sabellid infestations were found on the negative control animals (Table 1). Single Infestations Trial 2. Three of the four P generation sabellids produced an F, gen- eration. The first larva was observed on day 168 (Fig. 1). Subse- quently, two of the P generation sabellids produced additional F, larvae (three and four, respectively). Six of the F, generation pro- duced an Ft generation. The first larva appeared 1 17 days after the F| parent had settled (Fig. 1). Figure 2 illustrates a host shell with both F| and F^ sabellids. Life History Recently settled larvae that had fed on stained lipid beads were easily distinguished from individuals that had not fed (Fig. 3). The amount of time required for the sabellids to develop the ability to feed increased significantly as the temperature decreased (ANOVA, p < 0.001 ). The pooled replicates indicated that 50% of the sabellids reached this developmental stage at 20.9°C. 15.6 °C. and 11.2'C by 3.7. 4.9 and 6.0 days post exposure, respectively (Fig. 4). The rate at which the sabellids developed a full complement of setigers was also affected by temperature. The number of days required for the sabellids to develop all 1 1 setigers required sig- nificantly more time as the temperature decreased (ANOVA, p < O.(Wl). Analysis of the pooled data indicated that 50% of the newly settled larvae developed a complete complement of setigers at 20.9X. 15.6°C. and 1 l.2°C by 47.7. 68.4 and 83.1 days post exposure, respectively (Fig. 5). The majority of the sabellids ob- served had developed both eggs and sperm in conjunction with the development of all 1 1 setigers. The number of days required for the sabellids to produce mo- tile, infesting larvae or newly settled larvae required significantly more time as the temperature decreased (ANOVA, p < 0.001). Analysis of the pooled data indicated that 50% of the newly settled larvae had completed their life history at 20.9°C. 15.6^C. and 1 1.2°C by 1 10.8. 165.0 and 297.8 days post exposure, respectively (Fig. 6). DISCUSSION Self-fertilization is not considered the predominant mode of fertilization for many simultaneous hermaphrodites due to the del- Figure 3. A juvenile sabellid with lipid-stained beads \isible in the digestive tract (arrow). Scale bar = KM) (jni. 886 FiNLEY ET AL. 100 ♦ y o) 75 V / r2=0.83 T3 4 S 50 LL / /* 55 25- n i y / 1 — 1 1 — 20.9°C 1 1 100 . / /" c 75 '/' r2=0.88 T3 50 X 0) / /• ^ 25 n - /■ 1 1 15.6°C 1 1 12 3 4 5 6 7 Days Post Exposure Figure 4. Percentage offeedinji sabellids in pooled replicates at each of three temperatures investigated. Regression lines indicate that 3.7, 4.9 and 6.0 days post-exposure, 50% of the sabellids had developed the ability to feed at 20.9 C, 15.fi C, and 1I.2C, respectively. eterious effects believed to be associated with inbreeding (Heath 1977. Beckwitt 1982, Michiels & Streng 1998). There are, how- ever, several examples of successful self-fertilization among ma- rine invertebrates (Ghiselin 1974, Beckwitt 1982, Knowlton & Jackson 1993, Scharer & Wedekind 1999), some of which include members of the class Polychaeta (Ghiselin 1969, Hsieh 1997). Hsieh (1997), for e.xample. found no deleterious effects of self- fertilization in the simultaneous hermaphrodite Laoncmc iilhiciii- gillum. also a sabellid polychaete. Most polychaetes broadcast spawn small ova that are capable of being distributed great distances from the parent. In contrast. T. heteroitnciiuita. produces a relatively small number of large eggs that are brooded within their tube and lack a pelagic stage. Com- bined, these characteristics make T. heteruimcinata a likely can- didate for self-fertilization (Knowlton & Jackson 1993). In addi- tion, adult sabellids are sessile and rely solely upon their short infesting motile stage for the distribution of progeny to other hosts. Given this life history, one would predict that if the benefits of reproductive assurance associated with self-fertilization out- weighed the deterrents of ■"selfing," the requirement of cross- fertilization would be selected against (Jarne & Charlesworth 1993). For a sessile parasite such as T. heternimcinata. that invests a relatively large amount of energy into a limited number of eggs combined with a lack of control over accessing a partner, self- fertilization would be a successful strategy to assure that a found- ing sabellid on a host abalone would remain capable of producing progeny. Self-fertilization is one plausible explanation for the production of the F| and F, generations that we observed, though alternative feililization strategies, such as parthenogenesis, cannot be ruled out. These experiments were designed to investigate whether a single isolated sabellid has the ability to produce functional prog- eny or if a minimtmi population number is necessary to assure cross-fertilization. Our research indicates that a single sabellid is capable of founding a viable population. Following its accidental introduction in the late 1980"s, T. het- erounciiiaki was able to quickly spread throughout the state's farmed abalone industry. This rapid spread was due, in part, to the interdependency within the industry on seed transfers, in conjiuic- tion with the early misidentification of the sabellid as a native polychaete. Following the discovery of the sabellid as an exotic species, the Calit\irnia Department of Fish and Game (CDF&G) recognized the organism as a threat to the state's natural resources and aquaculture industry. In an attempt to control the further spread of the polychaete, the CDF&G, consulting with the indus- trv-initiated Abalone Sabellid Worm Advisory Committee, devel- oped a policy that all abalone transfers by farms be inspected and deemed to be sabellid-free (Aquaculture Disease Committee, July 6, 1995; Thoesen 1994). This sampling regime was selected, in part, becau.se a minimum population density of sabellids was be- lieved to be required to establish a viable population. Our research demonstrates that low-level infestations (or even a single founding 17 31 45 59 73 87 101 115 Days Post Exposure Figure 5. Percentage of sabellids that had developed a full comple- ment of setigers (n = II) in pooled replicates at each temperatures investigated. Regression lines indicate that 47.7, 68.4 and 83.1 days post-exposure, 50% of the sabellids had developed 11 setigers at 20.9 C, 15.6 C, and U.2°C, respectively. Sabellid Life History 887 1UU t c o T3 O 1— 75 */ r2=0.76 50 25 ♦ / ♦ 20.9°C ^ o^ n -♦^♦^ 1 100 75 o 3 ■D n 50 £1 O. a: 25 S5 0 100 en C) 75 D T3 O 50 CI 25 • %•% r2=0.75 •••• / • , •— y ^ r- 15.6°C 1 ' 1 So % % % ^9^ ^^0 ^^e ^^e ^<2? ^^^ Days Post Exposure Figure 6. Percentage of sabellids that had produced motile, infestive larvae or ne«lv settled lar\ae in pooled replicates at each temperature investigated. Regression lines indicate that after 1 1(1.8, 165.(1 and 297.8 days, 50% of the sabellids had completed their life history at 2().9 C, 15.6°C, and 1I.2C, respectively. sabellid) could pose a significant threat not only to the aqiiacnl- turist. but to the marine intertidal ecosystem. Many polyehaetes demonstrate positive relationships between environmental temperature, within the physiological ranges, and the rate at which they reach maturity (Cha et al. 1997, Olive et al. 1997. Qiu & Qian 1997. Qiu & Qian 1998). TerehnisuhcUa Iwt- erouncinata appears to follow this trend. Previous research has reported that reproductive maturity in 7". heterouncinata can occur in approximately one month (Kuris & Culver, 1999) or in three months (Ruck & Cook 1998) when animals were held at undefined ambient temperatures. Our controlled laboratory studies found that the most rapid development occurred at the highest of the three temperatures investigated (2().9°C), where average development of a full complement of setigers (n = 11) took 47.7 days. This decrease in time to maturity in association with elevated seawater temperatures would be expected to result in increased infestation levels, an observation that agrees with what has been experienced by California growers (B. Beede pers. eomm.). Higher infestation levels are typically observed in the warmer waters of the southern part of the state, particularly in association with El Nifio events. Previous studies on polyehaetes have demonstrated that at lower temperatures (again within physiological ranges) develop- ment would be prolonged, but that one would expect to observe a significant number of embryos and larvae developing into mature adults (Qiu & Qian 1997). Our research indicates that the devel- opment of all 1 1 setigers and the age at inaturity is indeed pro- longed at lower temperatures and would be expected to result in lower infestation le\'els. Tcichnisahella heterouncinata does, however, appear to be capable of completing its life history at the lower temperatures encountered in the northern part of the state where annual averages range from 8-13°C (Mcbride 1998). This portion of the state should not be viewed as a thermal refuge from T. Iieteionncinata, although its life history will take longer to complete and less obvious, low level infestations, may result. ACKNOWLEDGMENTS We appreciate the editorial comments of James D. Moore. This work was supported, in part, by the Saltonstall-Kennedy Programs of the National Oceanic and Atmospheric Administration. U.S. Department of Commerce and the National Sea Grant College under grant numbers NA96FD0206 and NA66RG0477. Project No. R/A-32PD. respectively. Additional support was provided by the California State Resources Agency. California Department of Fish and Game and the University of California, Davis. The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or any of its sub agencies. The U.S. Government is authorized to reproduce and distribute this work for governmental purposes. LITERATURE CITED Beckwitt. R. 19X2. Electrophoretic evidence for selt-tcrtill/atidii in two species of spirorbid polyehaetes. Bull. Smithem Calif. Acad. Sci. 81: 61-68. Cha. J.-H.. D. Martin & M. Bhaud. 1997. Effects of temperature on oocyte growth in the Mediterranean terebellid Eupolymnia ntiebulosa (Anne- lida: Polychaeta). Mar. Biol. 128:433^39. Culver. C. S.. A. M. Kuris & B. Beede. 1997, Identification and manage- ment of the exotic sabellid pest in California cultured abalone. A pub- lication of the California Sea Gram College System. La Jolla: Univer- sity of California. Publication No. T-04I. 29 pp. Fitzhugh. K. & G. W. Rouse. 1999. A remarkable new genus and species of fan worm (Polychaeta: Sabellidae: Sabellinae) associated \«.i(h smnc marine gastropods. J. Invertehr. Biol, 118:357-390. Ghiselin, M. T. 1969. The evolution of hermaphroditism among animals. Quart. Rev. Biol. 44:189-201. Ghiselin. M. T. 1974. Love's labor divided: Hermaphroditism. In: The economy of nature and the e\ olullon of sex. Berkeley: The University of Cahfornia Press, pp 108-129. Heath. D. J. 1977. Simultaneous hermaphroditism; cost and benefit. J. Theorel. Biol. 64:363-373. Hsieh. H. 1997. Self-fertilization: a potential fertilization mode in an es- tuarine sabellid polychaete. Mar. Ecol. Prog. Ser. 147:1433-148. Knowlton. N. & J. B. C. .lackson. 1993. Inbreeding and outbreeding in marine invertebrates. In: N. W. Thornhill. editor. The natural history of inbreeding and outbreeding: theoretical and empirical perspectives. Chicago: The University of Chicago Press, pp. 200-249. Kuris. A. M. & C. S. Culver. 1999. An introduced sabellid polychaete pest infesting cultured ahalones and its potential spread to other California gastropods. J. Imerichr Biol. 118:391-403. Leighton. D. L. 1998. Control of sabellid infestations in green and pink abalone. Haliotis fulgeits and H. corrugata, by exposure to elevated water temperatures. J. Shellfish Res. 17:701-705. 888 FiNLEY ET AL. Louhser. N. C. & N. Dorniehl. 2000. The use of ultrasound in the treatment of sabeUid infestations in South African abalone. / Shellfish Res. 19: 524. Mcbride, S. C. I'WS. Current status of abalone aquaculture in the Califor- nias. J. Shellfish Res. 17:59.V600. Oakes, F. R. & R. C. Fields. 1996. Infestations of Haliolis riifescens shells by a .sabellid polychaete. Aquaculnire. 140:139-143. Oakes, F. R.. R. C. Fields, P. F. Arthur & G. A. Trevelyan. 1995. The use of wax to control the shell parasites of red abalone. Haliotis Rufescens. ,/. Shellfish Res. 14:273. Olive. P. J. W.. J. Fletcher. S. Rees & G. Desrosiers. 199S. Interactions of environmental temperature with photoperiod in determining age at ma- turity in a semelparous polychaete Nereis (Nean(hes) Vireiis Sars. / Thermal Biol. 22:489-197. Qiu, J-W. & P-Y Qian. 1997. Combined effects of salinity, temperature and food on early development of the polychaete Hydroides eleguns. Ma- rine Ecology Progress Series. 152:79-88. Qiu, J-W. and P-Y Qian. 1998. Combined effects of salinity and tempera- ture juvenile survival, growth and maturation in the polychaete Hy- droides elegans. Marine Ecology Progress Series. 168:127-134. Ruck. K. & P. Cook. 1998. Sabellid infestations in the shells of South African molluscs: Implications for abalone mariculture. J. Shellfish Res. 17:693-699. Scharer. L. & C. Wedekind. 1999. Lifetime reproductive output in a her- maphrodite cestode when reproducing alone or in pairs: a time cost of pairing. Evohit. Ecol. 13:381-394. Shields, J. D., M. A. Buchal & C. S. Friedman. 1998. Microencapsulation as a potential control technique against sabellid worms in abalone culture. J. Shellfish Res. 17:79-83. Thoesen, J. C. 1994. Suggested procedures for the detection and identifi- cation of certain finfish and shellfish pathogens. 4"' ed.. Version 1. Fish Health Section, American Fisheries Society. Jimmal of Shellfish Research. Vol. 20. No. 2. 889-893. 2001. SHELL BORING CLAMS IN THE BLUE ABALONE HALIOTIS FULGENS AND THE YELLOW ABALONE HALIOTIS CORRUGATA FROM BAJA CALIFORNIA, MEXICO MA. DEL CARMEN ALVAREZ-TINAJERO,' JORGE CACERES-MARTINEZ,' AND JOSE GUADALUPE GONZALEZ-AVILES" 'Ci'iHro dc Im-esiit^acion CieiUificu y dc Ediicacldn Superior de Ensenada. Lahoratorio de Patologi'a de Moliiscos. Dc'pariamenlo de Acuicullura. Ensenada, Baja California. I 2732 Mexico: 'Sociedad Cooperativa de Prodiiccidn Pesquera. Pescadores Naeionales de Ahiilon. Ensenada, Baja California. Mexico ABSTRACT The blue abalone Haliotis fulgens and yellow abalone Halioris conugata support Ihe major portion of the abalone fishery in Baja California, Mexico. Abalone shells are marketable according to the quality of their shell, which is determined by the quantity and kind of epibionts, such as boring clams. In order to determine the e.\tent of infestation by .shell boring clams in these abalone species, shell samples from the commercial catch were obtained from Isla de Cedros. Baja California, and analyzed. Three shell boring species. Lilhophaga arisuita. Liihophugu pliinuila and Penhellu conmdi. were found infesting the blue abalone. The yellow abalone was infested by L ari.suihi and L. pluimihi. The prevalence and density of these two species were greater in the yellow abalone than in the blue abalone. The density of boring clams increased on older and consequently larger abalone shells (greater shell surface), although this trend needs confirmation. X-ray images revealed the severity of shell damage in heavily infested animals. The identi- fication of different fishing areas that provide different quality abalone shell, due to their boring clams load, opens the possibility for detemiining management regulations for obtaining better shells for marketing. A'£}' WORDS: boring clams, Lilhuphaga sp.. Peniiella sp., Haliotis fulgens. Halioris cornigata. abalone INTRODUCTION The abalone fishery is one of the most impoilant economic activities in Baja California, Mexico. This is because of the great value of the foot muscle of the abalone as a delicacy for human consumption, and because of the beauty of its shell which is used in jewelry. Two species, the blue abalone Haliotis fulgens (Philipi 1845) and yellow abalone Haliotis corntgata (Gray 1828). support this fishery. Abalone shells from Baja California are marketable according to their quality, w hich is closely related to epibionts load and shell appearance. Top quality shells are those that have less than 507f of shell suiface covered by epibionts, and no dark spots and blisters on the inner surface. These shells are exported for jewelry to Taiw an. Korea and China, reaching a value of about L'SS 1 .500 per ton. Shells that have 509r to 809^ of their external surface covered with epibionts. and blisters on their inner surface are worth around US$1,000 per ton. Shells that have >80'7f of their external and internal surface affected by epibionts are not exported, but are marketable locally at a low scale, reaching a value of US$750 per ton. The abalone shell is an adequate substrate for a great variety and quantity of epibionts to settle on. Of these, the boring clams may produce severe damage, as their burrows may penetrate the inner shell of the abalone. causing the host to seal the hole inter- nally with a leathery layer of conchiolin. Moreover, the abalone may secrete more nacreous material inside the bore site, forming a blister (Abbott & Haderiie 1980). Occasionally, a great blister may appear to be an irregular half-pearl, however, this is empty and of no marketable value. On the other hand, abalone may produce true pearls between the visceral mass and the inner shell as a reaction against foreign particles. These abalone pearls are sold locally as valuable and expensive souvenirs, but are not exported because they are rare. There have been no studies investigating the possible associa- tion of boring clams with the abalone, even though these epibionts damage the abalone shell and thus negatively affect their market- ability. In order to determine the identity and degree of infestation by shell boring clams in the blue and yellow abalone at Isla de Cedros, samples were obtained from a commercial catch. Infesta- tion rate as a function of abalone age and shell surface area was examined. MATERIALS AND METHODS In November 1997. 38 blue abalone and 48 yellow abalone were collected commercially from the northern end of Isla de Punta Norte 20- 10- San Agustin 28 °00 'i 15 0 2Q . I 11,5° 10- *Corresponding author. E-mail address: jcaceres@cicese.mx Figure 1. Sampling sites on the Isla de Cedros, Baja Calit'ornia, Mexico. 889 890 Alvarez-Tinajero et al. ■p^ ^H^K|||^||M^ HL.'. ■-•■^'-.j^lflmJ^M ^^I^^^^Qfl ^^^^^^^g^H ^HpP^VH ^^^^^^K^ ^^^^^1 ^Bkl d ^^^^^B^ ^^1 ^^^HHl ^ ' - ''' ^J^^^^^^m ^^^^^^^^^^^^^ ^^A^jM^^I ^HMkL_^;^ii^(^^H Figure '. (B). Abalone Shell ot'Haliotis cornigata (A) and Haliotis fiilgens Figure 4. Lithophuga plumiila, the most conspicuous character is the plume-like sculpture on the upper and posterior part of the valves. Cedros (Punta Norte). A further 13 blue abalone and 21 yellow abalone were obtained from the commercial catch at San Agustin near the southern end of the island in June 1998 (Fig. 1 ). Abalone foot muscle and viscera were detached from the shell, and the shell was numbered, cleaned and dried. The shells (Fig. 2) were exam- ined under a stereoscopic microscope for the presence of boring clams, including blisters and spots of conchiolin deposition. All clams were extracted from the shell with the aid of dissecdon needles. Larger clams were measured with an electronic caliper; and the smallest, with a micrometer placed on the stage of a stereoscopic microscope. All clams were stored in dried vials and identified (Soot-Ryen 1955, Keen & Coan 1974. Haderlie & Ab- bott 1980). The clams were grouped by species and size class to describe the composition of the epibiont clam population in ihe shell of each abalone species. Prevalence was estimated as the number of infested abalone/number of abalone examined xlOO. Density (in- tensity) was recorded as the number of clams per cm'' of abalone shell surface. The surface area of each cleaned abalone shell was estimated in the following way: aluminum foil was made to con- form to the shape of abalone shells and all their irregularities. The foil was carefully cut. dried and weighed. Pieces of aluminum foil were weighed, obtaining the following least squares regression equation: where X = y -0.0019/0.046 y = foil weight (g) and X = area (cm"); (r" = 1 ) Figure 3. Lithophugii arislala. easilj recognized bv the crossing pro- jections of the sculpture on the posterior part of the \alves. Figure 5. Peiiitella coiiradi. oval shaped, the shell gaping vvidel) ante- riorlv. Shell Boring Cl.'WIS on Baja California Abalone 891 TABLE 1. Mean, miniiiium. muxinuiiii den,sitj and pre>alenci- of .shfll horinj; clani.s in abalones collected in Punta Norte, isia dc Cedros B. C. Holioli\ corrugala Haliolh fulgens Mean Sd Min/Max Prev. Mean Sd MinAlax Prev. L. arislala L. plumula P. conradi 0.116 O.OW) 0 0.2 1 7 0.152 0 0-1.057 0-0.663 0 43.7 25.1) 0 0.077 0.017 0,005 0.351 0.073 0.023 0-2.121 0-0.396 0-0.138 13.1 5.3 5.3 Density. -No. Clams/ciir ol abakme shell. Standard deviation (Sd). Prevalence (Prev.) % of ahalone infested with borina clams. TABLE 2. Mean, minimum, maxinuim density and prevalence of shell boring clams in abalones collected in San Agustin. Isia de Cedros B. C. Halititis corriigata Haliolis fulf>eiis Mean Sd Min/Max Prev. Mean Sd Min/Max Prev. L. aristata L. plumula P. lonnuli 0.400 0.180 0 0.602 0.260 0 0-2.198 0-0.800 0 57.1 42.8 0 0.048 0.062 0.017 0.099 0.226 0.027 0-0.288 0-0.8 1 8 0-0.066 23.1 7.7 7.7 Density. -No. Clams/cnr of abalone shell. Standard deviation (Sd). Prevalence (Prev.) 9c of abalone infested with horini; clams. San Agustin CH L anstala [H L plumula r>4i ^ XcL. l^ L arislala L plumula P conradi L arislala L plumula P conradi Figure 6. Burrowing clam densitv in abalone Haliolis curnigata and H. fulgens from the localities studied. TABLE 3. Comparison of total density of boring clams Lilhophaga per abalone species. Punta Norte San Agustin L. aristata L. plumula L. aristata L. plumula H. carruiJata H. fulgens Mann-Whitney U (P) 0.12 0.10 0.01 0.06 0.02 0.12 0.40 0.(J5 0.05 0.18 0.07 0.12 Punta Norte San Agustin LH L. aristata CH L. plumula L fhrfifll 'I 1 ]^h,^ ]IL 0- ]- 6 1- i;- 15- I!. 21- 24. 0. J- 4 1- 12. IS- 18- 21- 2< - ,Q 29 59 89 119 149 179 209 2J 9 269 40 ., 29 59 89 119 149 179 209 2)9 269 >. 30- a c ^.r^.^. ,^ .- Huiw 0- }- b 9. 12- 15- 18- 21 - 24- 29 59 89 119 149 179 209 219 269 0- )- 6 9- 13- 15- 18- 21- 24- 29 59 89 119 149 179 209 23 9 269 Size (mm) Size (mm) Figure 7. Mean size class distribution (±SE) for boring clams on the abalone Halintis cnrrugata (,\) and H. fulgens (B) from localities stud- ied. 892 Alvarez-Tinajero et al. TABLE 4. Spearman correlation coefficient (r) between boring clam density, area and age of H. carnigalu and H. Julgeiis sheWs. Punta Norte San Agustin H. cor iigata H fiilgens H. corrugata H. fiilgeiis Area Age Area Age Area Age Area Age r P 0.33 0.02 (I.2S 0.04 0.4S <0.0() 0.40 0.02 0.25 0.26 0.13 0.fi2 0.34 0.04 0.17 0.8S The surface of the abalone shells was calculated using the weight of the aluminum molds obtained. The age of the abalone was estimated by polishing the spire zone of each shell and counting the prismatic growth lines (dark lines 1 in the spire under a stereoscopic microscope, according to the method of Shepherd and Avalos-Borja ( 1997) and Shepherd et al. ( 1995). A conelation analysis was used to determine the asso- ciation between the age and shell surface area of each abalone shell, as well as the intensity of clam infestation. A comparison was made between the density of the infestation determined by extraction of the clams from the abalone shell and that observed on x-ray plates. RESULTS Three shell boring clams. Lithophaiia arisiala Dill\\>n 1S17. Lithophaga phwuiki Hanley 1843 (family Mytilidae) and Penitella cimnuU Valenciennes 1S48 (family Pholadidae) (Fig. 3. Fig. 4. Fig. 3) were found infesting blue abalone shells. Only Litliopluiiia Figure 8. X-ray plate from Haliotis fulgens. Note the boring clams holes and boring clams inside (large arrow (. Small dots indicate per- forations produced by Cliuna sp. or poljchactes (small arrow). iirislala and Lithopliaga phiinuhi infested the yellow abalone. The holes made by the three boring clams were similar and some reached the inner side of the host shell. A leathery layer of con- chiolin of inegular and variable size often formed a pustule that occluded the perforation. Nacreous material occluded old perfora- tions forming irregular half-pearls. Table 1. Table 2 and Fig. 6 show the mean, minimum and maximum boring clam densities and prevalence found per abalone species and locality. Lithophaga arisuua and L. phimithi were the more abundant boring clams; Penitella coiiraji was \ery scarce. Table 3 shows a comparison (Mann- Whitney U test) between density per abalone species and locality. In general. Haliotis corrugata had a higher density of boring clams than Haliotis fulgens. The mean size distribution of boring clams is shown in Fig. 7. Distribution of mean size class at Punta Norte indicates that the si/.e distribution of the burrowing clam population at that site is stable. In San Agustfn, however, the mean size class distribution suggests a young population with recent recruitment. In Punta Norte, older Haliotis fulgens individuals (with greater shell suiface area) are infested with more boring clams per unit area than are younger abalone (with smaller shell surface area). The same pattern holds for Haliotis corrugata (Spearman corre- lation coefficient. Table 4). Figure S shows the holes made by shell boring clams. Some clams remain in the holes and other epibionts. such as the boring sponge Cliona sp., make small holes in their shell. Only 81% of the infestation intensity obtained from direct count and extraction of the clams were revealed on X-ray plates, as boring clams on the edges of the abalone shell were not detected on the latter. DISCUSSION The differences in prevalence and density of infestation by boring clams between the abalone species from the same locality may be related to several factors, including age, shell area, abalone species, specific shell characteristics, vertical distribution of bor- ing clams and particular environmental conditions. The total sur- face available for colonization by epibionts and the time spent underwater (age) by the substrate (abalone shell) are factors that determine the extent of colonization by benthic fauna. The depth distribution of the genus Lithophaga is similar to that of Haliotis corrugata. while Penitella conradi lives close to the lower inter- tidal zone where Haliotis fulgens is present (Soot-Ryen 1955. Keen & Coan 1974. Abbott & Haderlie 1980. Haderlie & Abbott 1980). A similar association was recorded in the red abalone Hali- otis rufescens and black abalone Haliotis cracherodii in relation to the distribution of P. conradi. in California (Hansen 197(.)). The depth range of H. rufescens. which is primarily sublittoral. corre- sponds better to that of P. conradi than does that of the black abalone H. cracherodii. Shell Boring Clams on Baja California Abalone 893 The settlement of several mytilids. other bivalve species, and other invertebrates may be favored by filamentous and rugose substrates (Caceres-Martfnez et al. 1994). The rugosity of the yel- low abalone's shell may thus present a better opportunity for settlement of mytilids and other invertebrates than the smoother shell of blue abalone. .Shell chemistry may also inlliience boring clam settlement and penetration. Another consideration that could affect the density of epibionts on abalone shells is the differential growth rate of abalone from different localities. Some wild stocks of dwarf yellow abalone in Islas Benito have been recorded. These have a greater quantity of epibionts than those of normal growth (unpublished data). Differences in boring clam prevalence and densities between the abalone from San Agusti'n and Punta Norte could be due to differential environmental conditions, since samples were taken on different dates (Punta Norte in November 1997. and San Agustin in June 1998). A direct comparison is thus difficult. Comparisons of growth rates and epibiont infestation between island abalone and mainland abalone in the area do not exist. The trend of more shell boring clams in older than in younger abalone, and in abalone with greater shell surface area than in those with smaller shell surface area must be investigated over a broad range of age and environment (mainland populations versus island populations). To the best of our knowledge, this is the first record of the association of LUhophaga aristata and L pliimiila with blue and yellow abalone. The genus Lithophaga has been widely recorded in association with corals (Scott 1986. Krumm & Jones 1993). submerged construction structures, sedimentary rocks (Kleeman 1994) and bivalves such as the pearl oyster (Doroudi 1996). Peni- tella coiinidi has been reported on shells of at least four abalone species on the Pacific coast, but most often in the red abalone (Haderlie & Abbott 1980). Differences in the size distribution of bunowing clams from the studied localities indicate differences in recruitment and survival of these species. This may be associated with environmental con- ditions and collection dates. However, repeated monitoring of size class distribution is necessary to obtain information on the popu- lation dynamics of the burrowing clams. Moreover, it is necessary to investigate such environmental factors as currents, temperature and food availability to see if these are related to the presence and size class distribution of burrowing clams. There are advantages and disad\antages to the two techniques used to determine tlie intensity of infestation by boring clams. The X-ray technique, although practical and non-destructive, does not allow identification of the boring clams and does not note the presence of all clams, missing those at the edges of the shell. On the other hand, extraction of the clams and direct counting is time consuming and shell destructive, although exact. Economically, the results of this study highlight that it is pos- sible for producers to detect certain zones that provide shells of different quality. It could thus enable managers to set regulations to protect abalone of these zones from the epibionts of other zones. For example, transplanting abalone from heavy burrowing clam- infested areas to those considered less infested could be controlled. In conclusion, the yellow abalone had more boring clams, mainly of the genus Lithophaga. than the blue abalone. There is a trend of more boring clams in older and larger abalone (shell surface) than in younger and smallest abalone. More studies are needed to confirm the observed trend and to compare the density of boring clams between zones, and between mainland and island abalone populations. Additionally, studies investigating the possi- bility of managing different areas to obtain top quality abalone shells must be carried out. ACKNOWLEDGMENTS We thank Alejo Ojeda Ibarra from Pescadores Nacionales de Abulon for providing information about the abalone shell market. Ulises Nijfiez Mi'guez and Hector Serrano Parra, for collaboration in shell processing. This study was supported by project 021PN-I- 297 CONACyT Mexico and the agreement between PNA and CICESE in Mexico. LITERATURE CITED Abbott. P. D. & E, C. Haderlie. 1980. Prosobranchia: Marine .Snails. In: R. H. Morris. D. P. Abbott & E. C. Haderlie. editors. Inlertidial Inverte- brates of California. Stanford University Press, pp. 231-307. Caceres-Martfnez, J., F. J. A. Robledo & A. Figueras. 1994. Settlement and post-larvae behavior of Myiihts galloprovincialis: field and laboratory experiments. Mar. Ecol. Prog. Ser 112:107-1 17. Doroudi. M. S. 1996. Infestation of pearl oysters by burrowmg and louling organisms in the northern Persian Gulf Ind. J. Mar. Sci. 25:I6S-169. Haderlie. E. C. & P. D. Abbott. 1980. Bivalvia: The Clams and Allies. In: R. H. Morrris. D. P. Abbott & E. C. Haderlie. editors. Intertidial In- vertebrates of California. Stanford University Press, pp. 354-411. Hansen. J. C. 1970. Commensal activity as a function of age in two species of California abalone. The Veliger. 13:90-94. Keen. A. M. & E. Coan. 1974. Marine molluscan genera of Western North America. Stanford University Press. 208 pp. Kleeman. K. 1994. Biocorrosion by bivalves. In: P. C. Dworschak. M. StachowiLschy & J. A. On. editors. Influences of organisms on their environment. Berlin: Blackwell. pp. I4.'i-L'iS. Krumm. D. K. & D. S. Jones. 1993. New Coral-Bivalve Association (Ac- tinastrea-Lithophaga) from the Eocene of Florida. J. Paleont 67:945- 95 1 . Scott. P. J. B. 1986. A new species of I.ithtiphaga (Bivalvia:Lithophaginae) burrowing corals in the Caribbean. ./. Moll. Stud. 52:55-61. Shepherd. S. A.. M. Avalos-Borja & Q. M. Ortiz. 1995. Toward a Chro- nology oi Haliotisfulgens, with a Review of abalone Shell Microstruc- ture. Mar. Freshwater Res. 46:607-615. Shepherd. S. A. & M. Avalos-Borja. 1997. The shell microstructure and chronology of the abalone Haliotis corrugata. Moll. Res. 18:197-207. Soot-Ryen, T. 1955. A Report on the Family Mytilidae (Pelecypoda) Allan Hancock Pacific Expeditions. The University of Southern California Press. 249 pp. ,/,)((;-;(((/ ofSlwlll'ish Reseanh. Vol. 20. No. 2, 895-904, 2001. PARALYTIC SHELLFISH POISONING IN THE ABALONE HALIOTIS MIDAE ON THE WEST COAST OF SOUTH AFRICA GRANT C. PITCHER,' * JOSE M. FRANCO." GREGORY J. DOUCETTE,^ CHRISTINE L. POWELL,' AND ANNA MOUTON^ \\hiriiu' ami Coastal Management. Private Bag X2. Rogge Bay. 8012. Cape Town. South Africa: 'Instituto de Investigaciones Marinas. Eduardo Cabello. 6. 36208 Vigo. Spain: ^Marine Biotoxins Program, Center for Coastal Environmental Health & Biomolecular Research. NOAA/National Ocean Sen'ice. 219 Fort Johnson Road. Charleston. SC 29412: *Arc-Onderstepoort Veterinary Institute, Private Bag X05. Onderstepoorl 01 10. South Africa ABSTRACT In April 1999. monitoring on two West Coast abalone farms provided evidence of the presence of Paralytic Shellfish Poisoning (PSP) toxins in the cultured abalone Hiilioiis miihw. Subsequent analyses of wild annuals collected from the West Coast also revealed the accumulation of PSP toxins ni these gastropods. The toxicity of individual animals as measured by the AOAC mouse bioassay showed considerable variation, ranging from below the assay detection limit to a maximum of 1609 (ig STXeq 100 g^'. Initial observations found PSP toxins in abalone to be coincident with blooms o( Alexandriuin catenella indicating that this dinotlagellate was the probable cause of abalone toxicity. Subsequent detection by receptor binding assay, of toxicity in abalone on the South Coast, an area considered free oi A. calenelta blooms, casts some doubt as to the source of the toxins. The toxin composition in the abalone as determined by HPLC was dominated by STX. and differed significantly from the toxin profile oi A. catenella and the co-occurring iTiussel, MviUhs galloproviiicialis. Either these findings indicated a high capacity for biotransformation of PSP toxins by abalone or that A. catenella was not the source of the toxin. Investigation of the anatomical distribution of toxins revealed that they were not evenly distributed throughout the abalone tissues, but appeared to concentrate in outer layer tissue. The muscular foot made a disproportion- ately low contribution to the total toxin content of the mollusc, whereas the epipodial fringe, although comprising a small proportion of the abalone total weight, contributed substantially to the total toxin content. The epipodial fringe is typically included with the muscular foot as that part of the animal marketed for human consumption. The negative impacts of PSP contamination on abalone spawning and larval survival are presented and the findings of this study are compared to observations of PSP toxins in the abalone Haliotis tuberculata on the Galician coast. The inability of abalone to detoxify or depurate accumulated PSP toxins below the regulatory level threatens the future of the established abalone fishery and the newly developed aquaculture operations on the West Coast of South Africa, A/-.')' WORDS: Paralytic Shellfish Poisoning, saxitoxin. abalone. Haliolis mielae. South Africa INTRODUCTION The abalone Huliotis inidae forms one of the oldest fisheries on the South African coast, with present-day operations including recreational, subsistence, and commercial activities. During the 1990s, land-based farming of this species has also been under development and has recently attained commercial scale produc- tion. In April 1999, monitoring on two West Coast abalone farms provided evidence of the presence of Paralytic Shellfish Poisoning (PSP) toxins in cultured abalone. Subsequent analysis of wild ani- mals also revealed the accumulation of PSP toxins in these gas- tropods. The accuniulalion of PSP toxins typically occurs when toxic dinoflagellates are filtered from the water by bivalves such as clams, mussels, oysters, or scallops, which then accumulate the algal toxins to levels that are potentially lethal to humans or other consumers. Considerable progress has been achieved in the struc- tural elucidation of PSP toxins. Following identification of the parent compound saxitoxin (STX), approximately two dozen natu- rally occurring derivatives have been described (Shimizu 1996). which vary greatly in their potency and are subject to inter- conversions and changes in relative proportions as they are trans- ferred through the food chain. PSP occurs off the West Coast of South Africa and has always been attributed to the dinoflagellate Alexandriuin catenella (Pitcher & Calder 2000). confirmed as the source of the toxin by Sapeika ( 1948). Although bivalve molluscs *Corresponding author. E-mail address: gpitcherCs'nicm.wcape.gov.za are the most common vectors of shellfish toxins, other vectors have been identified, including scavenging, predatory, and grazing gastropods (Shumway 1995). Abalone typically feed on kelp and other seaweeds. They are usually found attached to rocky sub- strates and feeding is accomplished by browsing on drifting sea- weed. This feeding behavior and diet would indicate that abalone are unlikely candidates for the ingestion and accumulation of PSP toxins derived from toxic dinoflagellates such as Alexandrium spp., which primarily occur in the plankton. This paper reports on investigations into the observations of PSP in abalone on the South African coast, including the incidence and distribution of toxic animals, the identiflcation of the toxins, intrapopulation variation in toxicity, the anatomical distribution of toxins, and the depuration of toxins. The possible source and mode of uptake of toxins and the potential impact on spawning and recruitment success are also considered. MATERIALS AND METHODS Animals analyzed for this study were collected from five aba- lone farms, (Farms A-E) and froin the field, at sites spread be- tween the northern and southern farm locations (Fig. I). All data presented here originate from animals collected between April 1999 and December 1999. A single wild animal collected in the vicinity of Jacobsbaai in April 1998 was also analyzed. The AOAC mouse bioassay was used in the routine analysis of PSP toxicity and a receptor-binding assay was utilized for comparative pur- poses. High performance liquid chrotnutography and mass spec- trometry were used to determine and confirm, respectively, the 895 896 Pitcher et al. 32°S 33°S 34°S 35°S 36°S 33°S 34°S 35°S 36°S (a) Elands ' Bav SOUTH AFRICA Cape J. Paternoster ■.oiumcNnc i, . l l - ^^Jacobsbaai WEST \ . Yzerfonlein COAST , ly^E RobbenI Island P.CAPETOWN Cape Peninsula L — ^ - Hermanus SOUTH COAST * > 1000 ugSTXeq 100 tt > 500 UgSTXeq/ I.-, 80 UgSTXeq/ lOOg . — c TV . — / I i\n.. (b) lOOg ',■• Farm A # ■6- Kr.Farm B ft ^ ou ug 3 1 A eq / I uug < 80 UgSTXeq/ lOOg < 40 ug STX eq / lOOg < 10 UgSTXeq/ lOOg IS^E 19°E 20°E 21°E 22°E 23°E 24°E 25°E 26''E 27°E Figure 1. (a) Map depicting sample locations and (b) the occurrence of toxic abalone on the South African coast as indicated by the maximum toxicity recorded at each locality between April and December 19^9 (toxicity >40 fig STX eq 100 g ' was determined by the mouse bioassay; toxicity <4(l fig STX eq 100 g ' was determined by the receptor binding assay). toxin composition in abalone. Tlie toxin composition ot the PSP- producing dinoflagellate. A. catenella. and the mussel, Mynlus galloprnvincialis. collected at Elands Bay during March 1999 were also determined for comparative purposes. Intrapopulation variability in toxicity was investigated by analysis of individual animals of the same size cultured under similar conditions or collected from the same locality in the wild. In most instances, the entire animal, once removed froiu the shell, was used in the analysis. In other cases, toxin analysis was caiTied out on specific body parts (e.g., the viscera, foot, epipodial fringe, digestive gland, gonad, mantle, or gills) to examine the anatomical distribution of toxins. Foot and epipodial fringe tissues were also analyzed following removal of epithelial tissue by scrubbing. Typically, analyses were performed on individual abalone. hut in some cases, animals were pooled for analysis. In assessing the rate of toxin depuration, 60 cultured animals were removed from Farm B on 13 May 1999 and 60 wild animals were removed from Paternoster on 29 June 1999, and placed under controlled conditions in a Cape Town laboratory. The animals were maintained on a diet of kelp in flow-through tanks supplied with filtered (18 p.m) seawater. Abalone were periodically sacri- ficed to assess the rate of depuration. The mouse bioassay was used to measure toxicity and two animals were sacrificed for each assay in order to reduce the effects of intrapopulation variability in toxicity. In an attempt to ascertain the source of the toxins, time-series data on the incidence of the PSP-producing dinofiagellate, A. catenella. at Elands Bay on the West Coast are presented. Data on the distribution of PSP toxin-contaminated bivalves, as monitored during the period 1989-1999 are also presented. These data were collected as part of a Harmful Algal Bloom monitoring program detailed by Pitcher and Calder (2000 1. The distribution of toxic shellfish is compared to the distribution of abalone catches, deter- mined by the Total Allowable Catch allocations, which generally reflect the distribution of abalone (R. Tan' pers. comm). When PSP contamination of cultured abalone was initially de- tected, problems with spawning and larval survival were experi- enced on affected West Coast farms. Samples of eggs and larvae from these farms were examined by light and scanning electron microscopy and compared to normal eggs and larvae of the same age from a farm on the South Coast. Samples of eggs were col- lected at spawning, at fertilization, and at regular intervals until hatching. Samples of larvae were collected at regular intervals from hatching until settlement. All samples were taken in dupli- cate, one sample placed in 4% gluteraldehyde in 0.2 M sodium cacodylate buffer and the other in Davidson's fixative (Austin and Austin 1989). Animals fixed in gluteraldehyde were processed for scanning electron microscopy. Samples were rinsed twice in so- dium cacodylate buffer, after which they were dehydrated through an ascending series of ethanol. The samples were then critical point dried from lOO'/r ethanol through liquid carbon dioxide in a Polaron Critical Point Drier. Dried samples were mounted onto viewing stubs and sputter coated with gold. Samples were viewed at 3 to 8 kV accelerating voltaee in a Hitachi S-2500 Scanning PSP IN Abalone on West Coast of South Africa 897 Electron Microscope. Animals fixed in either giiileruldehyde or Davidson's fixative were processed for liglit microscopy. Samples were rinsed twice in distilled water to remove adhering detritus. A modified double embedding technique (Austin & Austin 1989) was used prior to processing routinely for light microscopy. Sec- tions were cut at 6 (xni thickness and stained with haematoxylin and eosin. The following procedures were adhered to in the analysis of PSP toxins: Mouse Bioassay Net toxicity was measured by the standard AOAC mouse bio- assay (Association of Official Analytical Chemists, AOAC 1990). the method adopted worldwide to monitor the safety of shellfish PSP toxin levels for human consumption. Acidic aqueous extrac- tion of the tissue in 0. 1 M hydrochloric acid (heated to 100°C) was followed by intraperitoneal injection of 1 ml of the extract into each of three standardized mice. The time from initial injection to mouse death was recorded and the toxicity determined from Som- mer's table. HPLC-FD High performance liquid chromatography with tluorescence detection (HPLC-FD) was used to determine the concentration of individual PSP toxins (Franco & Fernandez -Vila 1993). The sepa- ration of toxins was carried out in a 5 \>.m Lichrospher 100 RP-18 column ( 12.'i x 4 mm i.d.). Two isocratic elutions were used, the first, for the separation of the carbamate toxins, neoSTX, dcSTX and STX, and the second for the separation of the GTX toxins. The first eluent was 1.5 niM octane sulfonate in 10 mM aminonium phosphate buffer (pH 7.2) plus 6% acetonitrile. at a rate of flow of 1 mL min~'. The second eluent was 2 mM octane sulfonate in 10 mM ammonium phosphate buffer (pH 7), at a rate of flow of 0.8 niL niin"'. The detector was a Waters 474 spectrofluorimeter. set at 330 nm ex., 390 nm em. Millennium software was used for recording and integrating peaks. Two pumps were used for deliv- ering postcolumn reagents. The postcolumn reaction was per- formed in a Teflon coil (10 m x 0.5 mm i.d) at 65' C. The first reagent was the oxidant, a solution of 7 mM periodic acid in 50 mM sodium phosphate (pH 9). and the second the acidifiant. a 0.5 M solution of acetic acid. Both were delivered at a rate of tlow of 0.4 niL min"'. The toxin reference standards for dcSTX and STX were supplied by the project EUR 18318 of the European Com- mission. The remaining reference standards for neoSTX and GTX 1.4 were purchased from the National Research Council of Canada. The C.,. GTX, and GTX,, toxins were identified and quan- tified by mild acid hydrolysis of the sample with an equal volume of 0.4 M HCl at 100°C for 15 min. This process transforms C, to GTX,. C, to GTX,. Cj to GTX,. C4 to GTX^. GTX, to STX and GTXft to neoSTX. The values of toxicity were expressed in p.g STX eq 100 g~' tissue thereby enabling comparison of HPLC and mouse bioassays. Values in p,g toxin g"' were deri\ed from the sum of the concentrations of all toxins detected by HPLC. These values were convened to jjig STX eq 100 g~' by means of molar specific conversion factors inferred from the values of specific toxicity given by Oshima (1995). Confirmation of the presence of STX was performed by mass spectrometry with an ion-trap mass spectrometer model LCQ (Finnigan. ThermoQuest, USA). Off-line "nanospray" ionization was canied out using disposable gold-coated capillary probes, as described previously by Marina et al. (1999). The specific detec- tion of toxins at high sensitivities was performed by examining for the appearance of daughter ions produced by the fragmentation of the coiTesponding precursor ions. Receptor Binding Assay The receptor binding assay used in the detection of PSP toxins exploits the highly specific interaction of these compounds with their biological receptor, the Noltage-dependent sodium channel, and is therefore based on functional activity (Doucette et al. 1997. Powell & Doucette |lii Press]). This assay has been compared with mouse bioassay-based toxicities of AOAC extracts of various bi- valve molluscs and results between methods agreed closely (Dou- cette et al. 1997). The assay protocol and calculations of sample values used herein follow those described by Doucette et al. ( 1997) and modified by Powell and Doucette (In Press). Briefly, the re- ceptor assay was performed in a microplate filtration fomiat and involved the incubation of a rat brain synaptosome preparation containing the receptors, with ["^Hj-STX and unlabeled PSP toxins contained in a standard or sample. After removal of unbound toxin by washing, the remaining bound radioactivity (i.e., [^H]-STX). which IS inversely and quantitatively related to the concentration of PSP toxins in a sample, was determined by scintillation count- ing on a Wallac MicroBeta model 1450. Toxin concentrations were expressed in terms of |j.g STX eq 100 g~'. RESULTS AND DISCUSSION Incidence and Distribution of Toxic Abalone In April 1999 monitoring of abalone at Farm B. for PSP toxins by means of the mouse bioassay. yielded positive results. Subse- quent tests at the nearby Farm A were also positive and many of these animals were reported to be paralyzed, in that they were no longer able to attach to a substrate and were unable to right them- selves. Brood stock were amongst the animals affected and some mortalities were recorded. At this time, divers reported the pres- ence of a large number of detached or paralyzed abalone in the kelp beds in the vicinity of Paternoster and analysis of these ani- mals also indicated the presence of PSP toxins. Subsequent to these initial findings, approximately 300 abalone. both cultured and wild, from five abalone farms and from .several sites in the field have been analyzed by mouse bioassay. Although toxin con- centrations varied considerably at each locality, the highest toxin concentrations were recorded on the northernmost farm. Farm A. where the single highest \alue was 1609 (jig STX eq 100 g"'. Toxin concentrations generally decreased southwards and the Cape Peninsula was the southern limit of toxic abalone as detected by the mouse bioassay (Fig. I ). On the two West Coast farms toxin concentrations exceeded the regulatory level of 80 (xg STX eq 100 g"' in the majority of animals, particularly on Farm A (Fig. 2a. b). Although toxins were detected in nearly all the wild animals tested on the West Coast, toxin concentrations were less than the regulatory limit in the majority of these animals (Fig. 2c). Of all the abalone tested on the South Coast by mouse bioassay, toxicity was detected only in animals collected on the western shores of False Bay (Fig. 2d). However, analysis of abalone from Farms C, D and E by means of the more sensitive receptor binding assay revealed the presence of PSP toxins at concentrations below those detectable by the mouse bioassay (Fig. lb). Assessment of all toxin analyses indicated a 898 (a) <80ug STXeq 100g-1 Pitcher et al. (b) > 80 ug STXeq lOOg-l £ 80 ug STXeq 100g-1 > 80 ug STXeq lOOg-1 Farm A Mean = 401 ug STXeq lOOg ', n = 26 Farm B Mean = 101 ug STXeq lOOg '. n » 167 (0 (d) . 80 ug STXeq 100g-1 < 80 ug STXeq lOOg-1 <80 ug STXeq IOOg-1 WEST COAST ■ WILD ABALONE Mean - 80 ug STXeq lOOg ', n « 69 SOUTH COAST - WILD AND CULTURED ABALONE Mean = 16 ug STXeq lOOg '. n = 22 Figure 2. The incidence of toxic abalone as detected by the mouse bioassay on (al Farm A (bl Farm B (c) from wild animals collected on the West Coast, and (d) from both cultured and wild animals collected on the South Coast. geographical gradient in abalone toxicity, toxin concentrations de- creasing from nortli to south. Analysis of a single wild abalone from Jacobsbaai collected in April 1998 by means of the mouse bioas.say provided a negative result. Although not conclusive, this single sample suggests that PSP toxins were not present in abalone. at concentrations detect- able by the mouse bioassay, during the previous year. Toxin Composition As the mouse bioassay is a non-specific lest and owing to the fact that abalone are unlikely candidates for the accumulation of PSP toxins, analysis by means of HPLC and mass spectrometry was required to identify and confirm the presence of PSP toxins. The toxin composition of abalone as determined by HPLC was rather unusual in that it was dominated by a single toxin, notably STX (Fig. 3a). The presence of STX was confirmed by mass spectrometry by the appearance of daughter ions (282 DA) pro- duced by the fragmentation of the corresponding precursor ions (300 DA). The specific toxicity and toxin composition of the PSP- producing dinoflagellate, Alexandrium catenella. present in bloom proportions in Elands Bay during March 1999, were also investi- gated. Different dinotlagellate strains vary greatly in their specific toxicity depending on environmental and growth conditions, and the single measurement of 1.75 pg STX eq celP' classify this dinoflagellate as a strain of low toxicity. The relative proportion of various PSP derivatives is a conservative property for a given isolate (Cembella et al. 1987) and the blooms in Elands Bay were found to be characterized by a high proportion of the less toxic N-sulfocarbamoyl toxins (C, and C,) and moderate proportions of the more potent carbamate (STX and GTX4) and decarbamoyl toxins (dcGTX,) (Fig. 3b). Simultaneous collection and analysis oi the mussel, Myiilus galloprovincialis. from Elands Bay during March 1999 provided toxicity values of 1 610 and 60 940 |jLg STX eq 100 g"' as deter- mined by the mouse bioassay and HPLC, respectively. A signifi- cant difference between the toxin profile of the causative di- notlagellate and the contaminated bivalve indicated a high degree of toxin transformation (Fig. 3c). It is well known that the toxin composition in bivalve tissues can differ significantly from that of the toxic dinotlagellates ingested (Bricelj & Shumway 1998) and as is often the case with bivalves, the mussels were found to contain reduced proportions of N-sulfocarbamoyl toxins (C, and Ct) and higher proportions of the carbamate toxins (neoSTX and STX). This observation is common when the algae ingested are rich in N-sulfocarbamoyl toxins which are more labile than other toxins and therefore more likely to be transformed to their more toxic carbamate analogs (Hall & Reichardt 1984). Although the precise in vivo mechanism of conversion between N-sulfo- carbamoyl and carbamate toxins in shellfish has not been deter- mined, it may lead to an increase in net toxicity (Bricelj & Shum- way 1998). The dramatically different toxin profile of H. inidae compared to A. catenella. indicates either a very high capacity for biotransformation of PSP toxins by abalone or that A. catenella was not the source of the toxin. Toxicity as measured by HPLC was notably higher than that measured by the mouse bioassay where v and x are toxicity mea- sured by HPLC and mouse bioassay respectively, V = 2.34v + 340.7, f = 12.1, r = 0.57, p < 0.007 Intrapopulalion Variation in Toxicity Toxicity as measured by the mouse bioassay showed consid- erable variation between individual animals, exceeding that ac- counted for by imprecision in the mouse bioassay (ca. -1-/- 20%, Adams & Furfari 1984). For example, on Farm A, the measured values of toxicity for whole animals ranged from 63-1609 |jLg STX eq 100 g"'. An understanding of the causes of this high intrapo- pulation variation in toxicity is crucial in the design of monitoring programs for the assurance of seafood safety (Bricelj & Shumway 1998). PSP IN Abalone on West Coast of South Africa 899 C2 C3 C4 DCGTX3 GTXl GTX2 GTX3 GTX4 NEOSTX STX C2 C3 04 DCGTX3 GTXl GTX2 GTX3 GTX4 NEOSTX STX (C) ~M CI C2 C3 C4 DCGTX3 GTXl GTX2 GTX3 GTX4 NEOSTX STX Figure i. Toxin coinposilion of (a) cultured and wild abalone Halioth midae, (hi a phvCoplankton sample with a high proportion of cells of Alexaiulriiim cuWiiillu and (c) the bivalve Myliliif, i^alUipnninciuli'' (the phytoplankton and bivalve samples were collected during March in the vicinity of Elands Bay). In this study, animals of a similar size and subjected to similar conditions of exposure to PSP toxins, such as farmed animals from a single basket or wild animals from a particular region, demon- strated marked vanabdity in toxicity (Fig. 4). The toxicity of 10 animals (size: 55-60 mm) sampled from a single basket on Farm B. ranged more than 5-fold from 57-307 |xg STX eq 100 g~'. Similarly, the toxicity of 10 animals (size: all 30 mm) sampled from a single basket on Farm A. also ranged 5-fold from 77-383 |xg STX eq 100 g"'. The toxicity of 52 animals collected in the vicinity of Robben Island (size: 1 15-170 mm) ranged more than 10-fold from 29-314 [j.g STX eq 100 g"'. This variability pre- vented the identification of any trends in toxicity associated with individual differences, such as body size, which arc considered to influence toxin accumulation rates. N = 10 max 383 in Mean»SD Mean-SD U3 Mean*SE MeanSE o Mean N=10 mm 57 n p a N.52 min29 max 314 a J — — [— — — Robben Island Figure 4. Intrapopulation variability in toxicity in abalone tested from a single basket at each of Forms A and B, and from abalone harvested in the vicinitv of Robben Island. Anatomical Dislributitm of Toxins Paralytic shellfish toxins are not evenly distributed throughout shellfish tissues, resulting in pronounced differences in the abso- lute toxicities of individual tissues. The contribution of each tissue to the total toxin body burden is therefore a function of both its absolute toxicity and relative weight contribution. Discarding of those tissues that selectively sequester PSP tox- ins (e.g. evisceration) may. in some cases, provide an effective tool to reduce the risk of PSP (Bricelj & Shumway 1998). In many instances only the foot of the abalone is marketed for human consumption, therefore the anatomical distribution of toxins is of considerable interest and importance. This study has revealed that toxins are not evenly distributed throughout the tissues of abalone (Fig. 5). Initial investigations divided the animals into three com- ponents, the viscera, foot, and epipodial fringe. Results indicated moderate toxicity in both the viscera and foot with dramatically higher levels of toxin in the epipodial fringe, a body component with a high surface area. Therefore, the muscular foot of abalone, despite contributing substantially to the total weight of the abalone, makes a disproportionately low contribution to the total toxin con- tent of the animal. The epipodial fringe, while making only a small contribution to the total weight of the abalone, contributes sub- stantially to the total toxin content and is typically included with the muscular foot in that part of the animal marketed for human consumption. The removal of epithelial tissue by scrubbing both the foot and epipodial fringe dramatically lowered the toxin levels in these tissues, suggesting that the toxins were concentrated in the epithelium (Fig. 5a). Scrubbing could therefore provide a possible means of reducing the levels of toxin prior to marketing the prod- uct for consumption. Further studies included examination of gill, mantle, digestive gland, and gonad tissue (Fig. 5b). These results varied somewhat from the initial investigations. Very high toxin levels were evident in the gill tissue, again indicating the accumulation of toxins in outer layer tissue. The toxicity of the digestive gland was also particularly high. Although toxin levels in the epipodial fringe were substantially less than those measured in the initial experi- ments, they were significantly higher than those in the foot and gonad. The capacity for in vivo biotransformation and the selective retention of individual PSP toxins, are major determinants of the differences in toxicity among tissues. However, establishment of 900 Pitcher et al. (a) (b) 1600 1400 1200 1000 800 600 400 200 0 1800 1600 1400 1200 1000 800 600 400 200 0 I I Mean+SE MeanSE E-Fringe i I Mean+SE Mean-SE r^3 E-Fringe Figure 5. Anatomical distribution of toxins as determined in la) the viscera, foot, and epipodial fringe lto\in concentrations were also determined folloHing removal of epithelial tissue from the foot and epipodial fringe by scrubbing (n = 12)) and lb) the fool, epipodial fringe, digestive gland, gonad, mantle, and gill tissue (n = 10). the source and mode of uptake of the toxins in abalone is required to understand the distribution of toxins and hov\ they may change over time. Depuration of Toxins In order to manage the problem of PSP in abalone it is impor- tant to establish the rate of depuration of toxins, as shellfish are known to vary markedly in their ability to detoxify accumulated toxins. Cultured and wild abalone were placed under controlled aquarium conditions and periodically sacrificed and analyzed for toxins in order to assess the rate of detoxification (Fig. 6). The results obtained over a 7-month period indicate that the prolonged retention of toxins is a characteristic of abalone. During this pe- riod, measures of toxicity ranged from below the regulatory level to 853 (jLg STX eq 100 g"' with no apparent declining trend, thus reflecting only the high intrapopulation variation in measured toxin levels. Incidence of Alexandrium catenella and PSP The only known source of PSP toxins in the Benguela is the dinoflagellate, A. calenelUi. which is usually associated with red o + Faim B CULTURED ABALONE 0 PATERNOSTER WILD ABALONE o + + O 0 + o + + - ■■■+-+ + + - o o o i a o tide on the West Coast of South Africa (Pitcher & Calder 2000). A seven year time-series obtained from daily monitoring at Elands Bay provides an indication of the interannual variability in the incidence of A. catenella in the plankton on the West Coast (Fig. 7). Although A. catenella is typically observed each year as a component of the plankton, usually in the latter months of summer, the detection of PSP toxins in abalone for the first time appears to have coincided with particularly high concentrations of A. catenella. Furthermore, at the time at which toxins were first de- tected in abalone on Farin B. examination of water samples from the farm revealed an abundance of A. catenella. Blooms of this dinoflagellate are common north of Cape Col- umbine and are usually advected southwaids. often as a flood event, during late summer and autumn. Blooms typically do not extend beyond Cape Point onto the South Coast (Pitcher & Calder 2000). Consequently the highest incidence of PSP is observed north of Cape Columbine while PSP has not been recoided east of Cape Point (Fig. 8a). Initial observations of PSP toxins in abalone therefore coincided temporally and spatially with blooms of A. catenella, indicating that this dinoflagellate was the probable cause of abalone toxicity. Subsequent detection, however, of low con- centrations of PSP toxins in abalone on Farms C. D and E. on the + + + 1000 + 400 + o 200 +t + " 4 M + *i. m + n s 10 h > / + ++ :5: + + + + + 4 + + + -H + + ♦ -1^ + + *+ 2 + + + + + + + ■Ht + + + +■ +f *■ + + * + + -»■ + ++ f- ++ + + ■»•■ -*i + Figure 6. A time .series of toxin concentrations in abalone removed from Farm B and from the wild at Paternoster and maintained under controlled laborator\ conditions. 1 Jan 93 1 Jan 94 1 Jan 95 1 Jan 96 1 Jan 97 1 Jan 98 1 Jan 99 Figure 7. Dailv time series of .Ale.iandrium catenella concentrations at Elands Bay for the period 1 July ly*.^ - M) .Inne 199'). PSP IN Abalone on West Coast of South Africa 901 \ (b) J J uo 0% \ noo-1 \ 34 00- 15% (3 Cape Town ^ 3500- 36 00- : : 85% Figure X. (al The incidence of PSP-contaminaled mussels, Myliliis i-alldpnnincialis. as monitored during the period l'*8')-l'>99 (as % of the total samples in which PSP toxins were detected In means of the mouse bioassa> ) and (b) the distribution of abalone catches. SiHilh Coast, an area considered to be free of blooms of A. catcnella. casts some doubt as to the source of PSP toxins in abalone at least in this region. I'lilinlial Kailiifiical Impact The contamination of bivalves is usually facilitated by their relative insensitivity to PSP toxins. However, it has become ap- parent that harmful algae may compromise survival and growth in some bivalve populations (Shumway 1990). although little is known about the ecological effects PSP toxins may have on field populations. Results of this study indicate that within the abalone fishery, recruitment may be severely influenced by PSP. The contamination of abalone with high levels of PSP toxins on Farm A was accompanied by clinical signs of paralysis in certain animals, including brood stock. Animals dropped off the sides of holding tanks and were unable to right themselves, and some mor- talities were experienced. Histological examination of affected ani- mals showed no significant abnormalities. Brood stock that recov- ered resumed spawning, but viable larvae were not produced. Light and scanning electron microscopy of eggs and larvae from Farm A showed irregular division of eggs and the formation of grossly abnormal larvae (Fig. 9). The larvae were characterized by irregular shapes and the appearance of cilia in clumps over their entire surface. Although most eggs containing such larvae hatched, the larvae were unable to swim and did not develop further. Prob- lems with larval development persisted for several months. It is not possible to say whether the larval abnormalities were directly re- lated to PSP contamination. It is likely that the reproductive prob- lems were caused by the deterioration in brood stock health and that the effect of the toxin on larval development is therefore only indirect. Similar, larval abnormalities could probably be caused by other brood stock stressors, but further research is needed to in- vestigate this possibility. In comparing the distribution of PSP toxin-contaminated mus- sels (Fig. 8a) to the distribution of abalone catches (R. TaiT. pers. comm) which generally retlect the distribution of abalone (Fig. Sb). we find them to be inversely distributed. That is. abalone are absent from areas where there is a high incidence of PSP, but abundant in areas not prone to PSP. Without any obvious physi- ological or ecological factors determining the northern limit to the distribution of abalone. these results suggest that the absence of abalone north of Columbine may be determined by the higher incidence of PSP toxins in that region. .4 Comparison to the Obsenatioiis of I'SP Toxins in Haliotis tuberculata on the Galician Coast The only other occurrence of PSP toxins in abalone has been reported in the species Haliotis tiihercidata on the Galician coast of Spain. In 1991 Martinez et al. (1993) detected PSP toxicity in samples of abalone collected in the Ria de Arousa, and the Gali- cian Coast was closed to the har\esting of abalone in 1993 as toxin le\'els exceeded the regulatory limit (Bravo et al. 1996). A number of comparisons may be made between the observations of PSP toxins in abalone on the Galician and South African coasts. The single greatest difference was in the toxin coinposition profile Fig. 10). Whereas H. midae was dominated by STX, H. tuberculata was dominated by decarbamoylsaxitoxin (dcSTX) with a small proportion of STX (Nagashima et al. 1995, Bravo et al. 1999). Similarities included the observation of highest toxicity in the epithelium of the H. tuberculata foot (Bravo et al. 1999), with the absolute toxicity of the epithelium being more than 300-fold greater than that of the gut or muscle. Another similarity was the slow depuration of toxins, as Bravo et al. (1996) observed no depuration of toxins in H. tuberculata maintained under controlled laboratory conditions for 3 mo. As is the case in H. tnidae. the source or origin of toxins in H. tuberculata remains uncertain, as the geographical range of the toxic abalone does not coincide exactly with the PSP-producing dinotlagellates known in the region. On the Galician Coast, no significant differences have been observed in the toxicity of aba- lone from different locations, despite a higher incidence of toxic dinoflagellates on the central and southern coasts. Since abalone attach to rocky substrates and browse on kelp and other seaweeds, the toxins may originate from these seaweeds. Other potential sources may include bacteria and bluegreen algae. If toxic di- notlagellates are the source of toxins in abalone. the mode of 902 Pitcher et al. Figure 9. la) A normal newly hatctied larva known as a trochophore stage and (b) an ahalone larva produced b> PSP-eontaminated brood- stock. The normal larva is able to swim and remains planktonic for approximately five days. It is regular in shape, possesses a clearly detlned protostonial tuft and a continuous and complete velum. uptake remains to be determined. It is possible tliat settled vegeta- tive cells or cysts may be taken up during feeding and observations in culture have indicated a high incidence of ^4. caieiwIUi cells associated with the mucilage layer of kelp fronds. The production of extracellular toxins by A. cateiiella may also serve as a possible source of PSP toxins and the adsorption/uptake of these toxins by abalone may provide an explanation for the accumulation of toxins in outer layer tissue. The observations of PSP toxins in abalone also show several similarities to those made on the butter clam Saxidomas giganteiis. Toxins were, as in H. midae. preferentially accumulated as STX in the outer surface of the siphons. STX was accumulated in clams feeding on an Ale.xandriuin isolate containing GTX and neoSTX but no STX. and ST.X accumulated in the siphon only after the depuration of GTX and neoSTX (Beitler & Liston 1990). Although immunohistochemical studies showed the location of STX to be in the columnar epithelium (Sniolowitz & Doucette 1995). the mechanism of tissue-specific retention of PSP toxins remains to be elucidated. Long-term retention, of high levels of STX has also been reported, typically taking several months to years to detoxify below the regulatory limit (Quayle 1969. Beitler & Liston 1990). Price and Lee (1971). and Price and Lee (1972) suggested that pH-dependent binding of STX to the melanin-containing tissues of the siphon was a mechanism of toxin retention in this species and Kvitek (1991), and Kvitek (1993) speculated that toxin compart- nientalization in butter clams had evolved as a chemical defense mechanism to reduce predation on clams. A similar mechanism of STX retention may exist in abalone which also possess melanin- containing tissue (Fox 1983). However, recent investigations by Bravo et al. (Submitted) have indicated that the toxins are associ- ated with secretory cells of the epithelium and not with cells con- taining granules of melanin-like pigment. CONCLUSION The findings of this study support the need for the inclusion of non-filter feeding molluscs in routine monitoring programs. The presence of PSP toxins in abalone now threatens the future of the abalone fishery and the newly developed aquaculture operations on the West Coast of South Africa. This threat is made more ominous by the slow depuration of toxins to below the regulatory level and by our limited understanding of the causes of the prob- lem. Priority areas of research should therefore focus on identifi- cation of the origin and establishment of the mode of uptake of the toxins. The concentration of toxins in the epithelium may provide the possibility for the development of a fairly simple treatment to reduce the toxicity to below the regulatory limit in the marketable product. Management of the problem of PSP toxins in abalone by such means needs to be explored further. ACKNOWLEDGMENTS We thank Cecilia Viljoen. Karen Whyte. Nick Loubser. Andre du Plessis and Paul Williams for supplying numerous abalone samples, and Desiree Calder for preparing these samples for toxin analysis. Components of the research were supported by the proj- ect MAR 95-1849. PSP IN Abalone on West Coast of South Africa 903 225 (a-i) 175 125 X 75 11 A 25 k V 1 100 75 50 25 0 -25 (b-i) Unknown Peak 20 30 225 (a- ) dcSTX 175 125 75 25 J Jli 100 75 50 25 0 -25 (b-ii) Unknown Peak 20 30 25 225 (a-iii) ?>175 S125 X \- o T3 X 1- u. 25 \ W ^v_ -oc; 100 a, 75 o § 50 I 25 ^ 0 -25 (b-iii) X X 1- o X ^ . 1- X ft CJ •- 11 O j\ JL IV Min. 10 ,,. 15 Mm. 20 25 Figure 10. A comparison of loxin profiles in ahaloni' on tlit South African and Galician coasts. Chromatograms of the (a) primary and (b) secondar) isocratic eiutions of (il Haliotis midae (ill Halitilis liiherculala and (iii) the toxin standards. LITERATURE CITED Adams. W. N. & S. A. Furfari. 1984. Evaluation of laboratory performani:e of the AOAC method for PSP toxin in shellfish. J. Assoc. Off. Anal. Chem. 67:1147-1148. Association of Official Analytical Chemists. AOAC. 1990. Official meth- ods of Analysis, l?"" edition. Arlington: Association of OtTicial Ana- lytical Chemists, pp. 881-882. Austin. B. & D. A. Austin. 1989. Methods for the microbiological exami- nation of tlsh and shellfish. Chichester. UK: Ellis Horwood Limited. Beitler. M. K. & J. Liston. 1990. Uptake and tissue distribution of PSP toxins in butter clams. In: E. Graneli. B. Sundstrom. L. Edler & D. M. Anderson, editors. Toxic Marine Phyloplankton. New York: Elsevier, pp. 257-262. Bravo. I.. E. Cacho. J. M. Franco. A. Miguez. M. I. Reyero & A. Martinez. 1996. Study of PSP toxicity in Haliotis liiherculala from the Galician coast. In: T. Yasumoto. Y. Oshima & Y. Fukuyo. editors. Harmful and Toxic Algal Blooms. Intergovernmental Oceanographic Commission of UNESCO, pp. 421-424. Bravo. I.. M. I. Reyero. E. Cacho & J. M. Franco. 1999. Paralytic shellfish poisoning in Haliolis luherculaia from the Galician coast: geographical distribution, toxicity by lengths and parts of the mollusc. .Ai/iiaiic To.xi- cology 46:79-85. Bravo. I.. J. M. Franco. A. Alonso. R. Dietrich & P. Molisl. Cytological study and immunohistochemical location of PSP toxins in the foot skin of the ormer. Haliotis tuberculata. from the Galician coast. Submitted. Bricelj. V. M. & S. E. Shumway. 1998. Paralytic shellfish toxins in bivalve molluscs: occurrence, transfer kinetics, and biotransformation. Reviews Fish. Sci. 6:315--^83. Cembella. A. D.. J. J. Sullivan. G. L. Boyer. F. J. R. Taylor & R. J. Anderson. 1987. Variation in paralytic shellfish toxin composition within the Protogonyanlax tamarensis/catenella species complex: red- tide dinotlagellates. Biochem. Sysl. Ecol. 15:171-186. Doucette. G. J.. M. M. Logan. I. S. Ramsdell & F. M. van Doiah. 1997. Development and preliminary validation of a microliter plate-based receptor binding assay for paralytic shellfish poisoning toxins. Tnxicon 35:625-636. Fox. D. L. 1983. Biochemistry of the Mollusca. In: P.W. Hochachka. editor. The Mollusca. Environmental Biochemistry and Physiology. New York: Academic Press, pp. 281-303. Franco. J. M. & P. Fernandez-Vila. 1993. Separation of paralytic shellfish toxins by reversed phase high performance liquid chromatography. with post column reaction and fluorimetric detection. Chro- matographia 35:613-620. Hall. S. & P. B. Reichardt. 1984. Cryptic paralytic shellfish toxins. In: E. P. Regalis. editor. Seafood Toxins. ACS Symposium Series 262. Wash- ington. D.C.: American Chemical Society, pp. 1 13-123. Kvitek. R. G. 1991. Paralytic shellfish toxins sequestered by bivalves as a defense against siphon-nipping t~ish. Mar. Biol. 1 1 1 :369-374. Kvitek. R. G. 1993. Paralytic shellfish toxins as a chemical defense in the butter clam {Saxiilonnis giganteii.s). In: T. J. Smayda. & Y. Shimizu. editors. Toxic Phytoplankton Blooms in the Sea. Amsterdam: Elsevier. pp. 407-412. Marina. A.. M. A. Garcia. J. P. Albar. J. Yagiie. J. A. Lopez de Castro & J. Vazquez. 1999. High sensitivity analysis and sequencing of peptides and protein by cuadrupole ion trap mass spectrometry. J Mass Spec- trom. 34: 1 7-27. Martinez. A.. J. M. Franco. I. Bravo. M. Mazoy & E. Cacho. 1993. PSP toxicity in Haliolis tuberculata from NW Spain. In: T. J. Smayda. & Y. Shimizu. editors. Toxic Phytoplankton Blooms in the Sea. Amsterdam: Elsevier, pp. 419-423. Nagashima. Y.. O. Arakawa. K. Shiomi & T. Noguchi. 1995. Paralytic Shellfish Poisons of ormer, Haliolis luherculaia, from Spain. J. Food Hyg. Soc. Japan 36:627-631. Oshima, Y. 1995. Postcolumn derivation LC method for PSP toxins. J.AO.AC Int. 78:528-532. 904 PiTCHHR ET AL. Pitcher. G. C. & D. Calder. 2000. Harmlul .Algal Blooms of the soulhern Qiiayle. D. B. 1969. Paralytic shelltish poisoning in British Columbia. Bull. Benguela Current: A review and appraisal of monitoring from 1989- F;.s7i. Res. Bd Can. 168:68. 1997. S. Afr. J. Mar. Sci. 22:255-271. Sapeika. N. 1948. Mussel Poisoning. S. Afr. Med J. 22:337-.138. Powell. C. L. & G. J. Doucette. 2000. A receptor hinding assay for para- Shimizu, Y. 1996. Microalgae metabolites: a new perspective. Ami. Rev. lytic shellfish poisoning toxins: recent advances and applications. Microbiol. 50:431-465. Natural To.xi?is. (in press). Shumway, S. E. 1990. A review of the effects of algal blooms on shellfish Price, R. J. & J. S. Lee. 1 97 1 . Interaction between paralytic shellfish poison and aquacullure. J. World Aqiiaciilntre Soc. 2 1 :65- 1 04. and melanin obtained from butter clam iSa.xidomit.s gii;anu-tt.'.} and Shumway. S. E. 1995. Phycotoxin-related shellfish poisoning: bivalve mol- synthetic melanin. J. Fish. Res. Bd. Can. 28:1789-1792. luscs are not the only vectors. Reviews Fish. Sci. 3:1-31. Price, R. J. & J. S. Lee. 1972. Paralytic shellfish poison and melanin Smolowitz. R. & G. Doucette. 1995. Immunohistochemical localization of distribution in fractions of toxic butter clam iSa.xidonius giganteiis) saxitoxin in the siphon epithelium ol the hullcr clam. Sa.xidonuis gi- siphon. J. Fi.Hi. Res. Bd Can. 29:1657-1658. ganreiis. Biol. Bull. 189:229-230. THE NATIONAL SHELLFISHERIES ASSOCIATION The National Shellfisheries Association (NSA) is an international organization of scientists, manage- ment officials and members of industry that is deeply concerned and dedicated to the formulation of ideas and promotion of knowledge pertinent to the biology, ecology, production, economics and man- agement of shellfish resources. The Association has a membership of more than lOOO from all parts of the USA. Canada and 18 other nations; the Association strongly encourages graduate students' mem- bership and participation. WHAT DOES IT DO? — Sponsors an annual scientific conference. — Publishes the peer-reviewed Jounuil of SItcllfish Resedrcli. — Produces a Quarterly Newsletter. — Interacts with other associations and industry. WHAT CAN IT DO FOR YOU? — You will meet kindred scientists, managers and industry officials at annual meetings. — You will get peer review through presentation of papers at the annual meeting. — If you are young, you will benefit from the experience of your elders. — If you are an elder, you will be rejuvenated by the fresh ideas of youth. — If you are a student, you will make useful contacts for your job search. — If you are a potential employer, you will meet promising young people. — You will receive a scientific journal containing important research articles. — You will receive a Quarterly Newsletter providing information on the Associatitm and its activities, a book review section, information on other societies and their meetings, a job placement section, etc. HOW TO JOIN — Fill out and mail a copy of the application blank below. The dues are 65 US $ per year ($35 for students) and that includes the Jounuil and the Newsletter! NATIONAL SHELLFISHERIES ASSOCIATION— APPLICATION FOR MEMBERSHIP (NEW MEMBERS ONLY) Name: For the calendar year: Date: Mailins address: Institutional affiliation, if any: Shellfishery interests: Regular or student membership: Student members only — advisor's signature REQUIRED: Make checks (MUST be drawn on a US bank), international postal money orders or VISA for $65 ($35 for students with advisor's signature) payable to the National Shellfisheries Association and send to Nancy Lewis, BiKikkeeper, PO Box 35tl V.I. M.S. Eastern Shore Lab. Wachapreague. VA 2348(J. USA. INFORMATION FOR CONTRIBUTORS TO THE JOURNAL OF SHELLFISH RESEARCH Original articles dealing with all aspects of shellfish re- search will be considered for publication. Manuscripts will be judged by the editors or other competent reviewers, or both, on the basis of originality, content, merit, clarity of presentation, and interpretations. Each article should be carefully prepaied in the style followed in prior issues of the Journal of Shellfish Research before submission to the Editor. Papers published or to be published in other journals are not acceptable. Title, Short Title, Key Words, Abstract: The title of the paper should be kept as short as possible. Please include a "short running title" of not more than 48 characters including spaces, and key words. Each manuscript must be accompanied by a concise, informative abstract, giving the main results of the research reported. The abstract will be published at the beginning of the article. No separate summary should be in- cluded. Text: Manuscripts must be typed double-spaced throughout on one side of the paper, leaving ample margins, with the pages numbered consecutively. Scientific names of species should be underlined or in italics and. when first mentioned in the text, should be followed by the authority. Common and scientific names of organisms should be in accordance with American Fisheries Society Special Publications 16 and 17: Common and Scientific Names of Aquatic Invertebrates fixmi the United States and Canada: Molhisks cmd CSNAIUSC: Decapod Crus- taceans, or relevant publications for other geographic regions. Abbreviations, Style, Numbers: Authors should follow the style recommended by the sixth edition (1994) of the Council of Biology Editors [CBE] Style Manual, distributed by the American Institute of Biological Sciences. All linear measure- ments, weights, and volumes should be given in metric units. Tables: Tables, numbered in Arabic, should be on separate pages with a concise title at the top. Illustrations: Line drawings should be in black ink or laser print and planned so that important details will be clear after reduction to page size or less. No drawing should be so large that it must be reduced to less than one third of its original size. Photographs and line drawings should be prepared so they can be reduced to a size no greater than 17.3 cm x 22.7 cm, and should be planned either to occupy the full width of 17.3 cm or the width of one column, 8.4 cm. Photographs should be glossy with good contrast and should be prepared so they can be reproduced without reduction. Originals of graphic materials (i.e., line drawings) are preferred and will be returned to the author. Each illustration should have the author's name, short paper title, and figure number on the back. Figure legends should be typed on separate sheets and numbered in Arabic. Digital Figures: Authors may provide digital figures (they are not required): they must be accompanied by hardcopy fig- ures of equal quality, which the printer will use for comparison and backup. If digital figures are supplied, please note the following instructions: • Each piece of art should be saved as its own file. • Files must be one of the follow iiig fonnats: TIF. EPS. or JPG. • Each file should be named according to its figure number and format (e.g.. ■"fig2b.tif). • Figures inust not be embedded in a word-processor or spreadsheet document: the printer cannot use images stored 111 Word. WordPerfect, Excel. Powerpoint, etc. • Resolution: line shots: 1 ()()() dpi: halftones/grayscales: 300 dpi if no lettering, 500 dpi if figure contains lettering. • Color figures: save the files as CMYK-encoded TIF images (preferred) or CMYK-encoded EPS or JPG images. Color figures have the same resolution requirements a BAV. above. Color illustrations will not be accepted unless the author agrees to cover the cost of associated reproduction and printing. Literature Cited: References should be listed alphabeti- cally at the end of the article. Abbreviations in this section should be those recommended in the American Standard for Periodical Title Abbreviations, available through the American National Standard Institute. 1430 Broadway. New York. NY 10018. For appropriate citation format, see examples below: Jouri^al: Walts, R. J.. M. S. Johnson & R. Black. 1990. Effects of re- cruitment on genetic patchiness in the urchin Echinometra mathaei in Western Australia. Mar. Biol. lO.'i: 145-151. Book: Claudi. R. & G. L. Mackie. 1994. Practical manual for Zebra Mussel monitoring and control. Boca Raton. FL: CRC Press. 227 pp. Chapter in Edited Book: Davio, S. R., J. F. Hewetson & J. E. Beheler. 1985. Progress toward the development of monoclonal antibodies to saxitoxin: antigen preparation and antibody detection. In: D. M. Ander- son. A. W. White & D. G. Baden, editors. Toxic dinoflagel- lates. Amsterdam: Elsevier, pp. 343-348. Page Charges: Authors or their institutions will be charged $100.00 per printed page. All page charges are subject to change without notice. A handling fee of $50 will be charged for all manuscripts accepted for publication. Proofs: Page proofs are sent to the conesponding author and must be coiTccted and returned within seven days. Alter- ations other than collections of printer's enors will he charged to the author(s). Reprints: Reprints of published papers are available at cost to the authors. Information regarding ordering reprints will be available from The Sheridan Press at the time of printing. Cover Photographs: Appropriate photographs for consid- eration for use on the co\ er of the Journal of Shellfish Research are encouraged. Black and white photographs and color illus- trations will be considered. Corresponding: An original and two copies of each manu- script submitted for publication consideration should be sent to the Editor, Dr. Sandra E. Shumway. Dept. of Marine Sciences, University of Connecticut, 1084 Shennecossett Rd.. Groton. CT 06340. Tel. 516-398-1539: Fax 860-405-9513. E-mail: sshumway(3' Southampton. liu.edu Membership information may be obtained from the Editor or the Treasurer using the form in the Journal. Institutional subscribers should send requests to: Journal of Shelljish Re- .search. P.O. Box 465. Hanover. PA 1 733 1 . JOURNAL OF SHELLFISH RESEARCH Vol. 20, No. 2 DECEMBER 2001 CONTENTS Peter Cook Preface 565 H. Roy Gordon and Peter A. Cook World ubalone supply, markets and pricing; Histoncal. current and future 567 Rodney Roberts A review of settlement cues for larval ahalone (Haliotis spp. ) 57 1 Ricardo Searcy- Bernal, Luis A. Velez-Espino and Casandra Anguiano-Beltrdn Effect of biotllm density on grazing and growth rates of Haliolis fnlgt-ns postlarvae 587 Saowapa Sawatpeera. E. Suchart Vpatham, Maleeya Kruatrachue, Yaowalak P. Chitramvong, Pichai Sonchaeng, Tanate Pumthong, and Jintana Niigranad Larval development in Haliolis asinina Linnaeus 593 Juan Gabriel Correa-Reyes, Maria del Pilar Sdnchez-Saavedra, David Alfaro Siqueiros-Beltrones and Norberlo Flores-Acevedo Isolation and growth of eight strains of henthic diatoms, cultured under two light conditions 603 J. A. Simental-Trinidad, M. P. Sdnchez-Saavedra and J. G. Correa-Reyes Biochemical composition of benthic marine diatoms using as culture medium a common agricultural lertilizer 611 Ryo Sasaki and Scoresby A. Shepherd Ecology and post-settlement survival of the ezo ahalone. Haliolis discus haniiai. on Miyagi coasts. Japan 619 D. Jackson, K. C. Williams and B. M. Degnan .Suitability of Australian formulated diets for aquaculture of the tropical abalone Haliolis asiniini Linnaeus 627 T. A. Shipton and P. J. Britz The partial and total replacement of fishmeal with selected plant protein sources in diets for the South African abalone. Haliolis inidae L 637 Victoria J. Allen, Islay D. Marsden and Norman L. C. Ragg The use of stimulants as an aid to wean fishery caught blackfoot abalone (Halioris iris) to artificial food 647 Sam J. Boarder and Muki Shpigel Comparative pertbrmaiices of juvenile Haliotis roei fed on enriched UIra rigida and various artificial diets 653 Susan C. McBride, Einat Rotem, David Ben-Ezra and Muki Shpigel Seasonal energetics of Haliotis fulgcns (Philippi) and Hal ions liihcniilata (L.) 659 Nuruddin M. J. Kabir, Mike F. Barker, Philip V. Mladenov and Brian E. Niven Spawning induction of yellowfoot abalone, Haliotis aiistralis using chemicals and ganglionic suspensions 667 Magdalena Litaay and Sena S. de Silva Reproductive performance indices based on physical characteristics of female blacklip abalone Haliolis rubra L 673 Stephen M. Hindrum, Christopher M. Burke, Stephen J. Edwards and Dean R. Johns Effects of combined exposure to elevated ammonia and low dissolved oxygen levels in greenlip {Haliotis laevigata Donovan) and blacklip (//. rubra Leach) abalone. 1, Growth and mortality data from simulated systems failure 679 Gen He and Kangsen Mai Ontogenetic trends of mineralogy and elements in the shell of abalone. Haliotis discus hamuli Ino 685 Elisa Serviere-Zaragoza, Alejandro Mazariegos-Villareal, German Ponce-Diaz and Silvia Monies Magallon Growth of juvenile abalone, Haliotis fulgens Philippi, fed different diets 689 Brian Drew, Dean Miller, Tes Toop and Peter Hanna Identification of expressed HSP's in blacklip abalone {Haliolis rubra Leach) during heat and salinity stresses 695 Scott Cummins, Amporn Thongkukiatkul and Peter J. Hanna Location of egg-laying hormone in reproductive structures and neurons of Haliotis rubra (Leach) using antibodies raised against recombinant fusion proteins 705 S. Sahaphong, V. Linthong, C. Wanichanon, S. Riengrojpitak, N. Kangwanrangsan, V. Viyanant, E. S. Upatham, T. Pumthong, N. Chansue and P. Sobhon Morphofunclional study of the hemocytes of Haliolis asinina 711 Somjai Apisawetakan, Malee Chanpoo, Chaiti Wanichanon, Vichai Linthong, Maleeya Kruatrachue, Edward Suchart Upatham, Tenate Pumthong and Prasert Sobhon Characterization of trabecular cells in the gonad of Haliolis asinina Linnaeus 717 CONTENTS CONTINUED ON INSIDE BACK COVER CONTENTS CONTINUED FROM BACK COVER M. Chanpoo, S. Apisawetakan, A. Thongkukiatkul, C. Wanichanon, V. Linthong, M. Kruatrachue, E. S. Upatham, T. Pumthong, P. J. Hanna and P. Sobhon Localization of egg-laying hormone in the gonads of a tropical abalone. Haliolis asbiina Linnaeus 725 A. Thongkukiatkul, P. Sobhon, E. S. Upatham, M. Kruatrachue, C. Wanichanon, Y. P. Chitramvong and T. Pumthong LUtrastruclure of neurosecretory cells in the cerebral and pleuropedal ganglia of Haliolis asinina Linnaeus 733 Konstantin A. Karpov, Mia J. Tegner, Laura Rogers-Bennett, Peter E. Kalvass and Ian K. Taniguchi Interactions among red abalones and sea urchins in fished and reserve sites of northern California: Implications of competition to management 743 Mia J. Tegner, Pete L. Haaker, Kristin L. Riser and L. Ignacio Vilchis Climate variability, kelp forests, and the Southern California red abalone fishery 735 Miguel A. del Rio-Portilla and Jose G. Gonzdlez-Aviles Population genetics of the yellow abalone, Haliolis conugata, in Cedros and San Benito Islands: A preliminary survey 765 Rickard A. Officer, Cameron D. Dixon and Harry K. Gorfine Movement and re-aggregation of the blacklip abalone, Haliolis rubra Leach, after fishing 771 Rickard A. Officer, Malcolm Haddon and Harry K. Gorfine Distance-based abundance estimation for abalone 781 Harry K. Gorfine and Cameron D. Dixon Diver behaviour and its influence on assessments of a quota-managed abalone fishery 787 Harry K. Gorfine Post-harvest weight loss has important implications for abalone quota management 795 Harry K. Gorfine, Bruce L. Taylor and Terry I. Walker Triggers and targets: What are we aiming for with abalone fisheries models in Australia? 803 Eva E. Plagdnyi, Douglas S. Butterworth and Anabela Brandao Toward assessing the South African abalone Haliolis inidae stock using an age-structured production model 813 S. A. Shepherd and Kate R. Rodda Sustainability demands vigilance: Evidence for serial decline of the greenlip abalone fishery and a review of management 829 S. A. Shepherd, Kale R. Rodda and Kelly M. Vargas A chronicle of collapse in two abalone stocks with proposals for precautionary management 843 S. de Waal and P. Cook Quantifying the physical and biological attributes of successful ocean seeding sites for farm-reared juvenile abalone {Haliolis midae) 857 S. de Waal and P. A. Cook Use of a spreadsheet model to investigate the dynamics and the economics of a seeded abalone population 863 James D. Moore, Thea T. Robbins, Ronald P. Hedrick and Carolyn S. Friedman Transmission of the rickettsiales-like prokaryote "Candidatus Xenohaliotis califomiensis" and its role in withering syndrome of California abalone. Haliolis spp 867 Jorge Caceres-Martinez and Gissel D. Tinoco-Orta Symbionts of cultured red abalone Haliolis rufescens from Baja California, Mexico 875 Carl A. Finley, Timothy J. Mulligan and Carolyn S. Friedman Life history of an exotic sabellid polychaete, Tercbrasabella heterouncinata: Fertilization strategy and influence of temperature on reproduction 883 Ma. del Carmen Alvarez-Tinajero, Jorge Caceres-Martinez and Jose Guadalupe Gonzdlez-Aviles Shell bonng clams in the blue abalone Haliolis fulgens and the yellow abalone Haliolis corrugala from Baja California, Mexico 889 Grant C. Pitcher, Jose M. Franco, Gregory J. Doiicette, Christine L. Powell and Anna Mouton Paralytic shellfish poisoning in the abalone Haliolis inidae on the west coast of South Africa 895 COVER PHOTO: A group of large, adult "perlemoen" (South African abalone: Haliolis inidae) against a backdrop of their primary food source, the kelp Ecklonia maxima. Photographed in a typical exposed habitat in 4 m water depth, Betty's Bay, South Africa, November 1988. (Rob Tarr) The Journal of Shellfish Research is indexed in the following: Science Citation Index®, Sci Search®, Research Alert®, Current Contents®/Agriculture, Biology and Environmental Sciences, Biological Abstracts, Chemical Abstracts, Nutrition Abstracts, Current Advances in Ecological Sciences. Deep Sea Research and Oceanographic Literature Review, Environmental Periodicals Bibliography, Aquatic Sciences and Fisheries Abstracts, and Oceanic Abstracts, JOURNAL OF SHELLFISH RESEARCH VOLUME 20, NUMBER 3 DECEMBER 2001 The Journal of Shellfish Research (formerly Proceedings of the National Shellfisheries Association) is the official publication of the National Shellfisheries Association Editor Dr. Sandra E. Shumway Depailment of Marine Sciences University of Connecticut Groton, CT 06340 Dr. Standish K. Allen. Jr. (2002) School of Marine Science Virginia Institute of Marine Science Gloucester Point, VA 23062-11346 Dr. Peter Beninger (2001) Laboratoire de Biologic Marine Faculte des Sciences Universite de Nantes BP 92208 44322 Nantes Cedex 3 France Dr. Andrew Boghen (2001) Department of Biology University of Moncton Moncton, New Brunswick Canada ElA 3E9 Dr. Neil Bourne (2001) Fisheries and Oceans Pacific Biological Station Nanaimo. British Columbia Canada V9R 5K6 Dr. Andrew Brand (2001 ) University of Liverpool Marine Biological Station Port Erin, Isle of Man Dr. Eugene Burreson (2001) Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Dr. Peter Cook (2002) Department of Zoology University of Cape Town Rondebosch 7700 Cape Town, South Africa EDITORIAL BOARD Dr. Simon Cragg (2002) Institute of Marine Sciences University of Portsmouth Ferry Road Portsmouth P04 9LY United Kingdom Dr. Leroy Creswell (2001) Harbor Branch Oceanographic Institute US Highway 1 North Fort Pierce. Florida 34946 Dr. Lou D'Abramo (2002) Mississippi State University Dept of Wildlife and Fisheries Box 9690 Mississippi State. Mississippi 39762 Dr. Ralph Elston (2001) Battelle Northwest Marine Sciences Laboratory 439 West Sequim Bay Road Sequim, Washington 98382 Dr. Susan Ford (2002) Rutgers University Haskin Laboratory for Shellfish Research P.O. Box 687 Port Norris, New Jersey 08349 Dr. Raymond Grizzle (2001) Randall Environmental Studies Center Taylor University Upland, Indiana 46989 Dr. Mark Luckenbach (2001) Virginia Institute of Marine Science Wachapreague, Virginia 23480 Dr. Bruce MacDonald (2002) Department of Biology University of New Brunswick P.O. Box 5050 Saint John, New Brunswick Canada E2L 4L5 Dr. Roger Mann (2002) Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Dr. Islay D. Marsden (2002) Department of Zoology Canterbury University Christchurch, New Zealand Dr. Tom Soniat (2002) Biology Department Nicholls State University Thibodaux, Louisiana 70310 Dr. J. Evan Ward (2002) Dept. of Marine Sciences University of Connecticut Groton, CT 06340-6097 Dr. Gary Wikfors (2002) NOAA/NMFS Rogers Avenue Milford, Connecticut 06460 Journal of Shellfish Research Volume 20, Number 3 ISSN: 0730-8000 December 2001 www.shellfish.org/pubs/jsr.htm .I.nimol of Shcllfisli Ri-.cunh. Vol. 20, No. 3, 905-912, 2001, ^ _^ ___^__ FIELD IDENTIFICATION OF THE EUROPEAN GREEN CRAB SPECIES: '(^AR'(Mi!.S'''^A'BmS'' AND CARCINUS AESTUARII SYLVIA BEHRENS YAMADA AND LAURA HAUCK Zoology Department. Oregon State University. Corvallis. Oregon FEB 2 6 2002 Wqo<3» heiM, (VIA OHii ABSTRACT Adults of the global in\adcis, Cincmus iiuifiuis and C aestuarii. can generally be distnigiiished in the field by three diagncstic characteristics: the shape of the coptilalory appendages (pleopods) in the male, the shape of the frontal area between the eyes, and the carapace width to length ratio. The pleopods of male C. nuienas are crescent-shaped and curve outward with the center of the crescents touching: those of C. aesiimrii are straight and parallel to one another. The frontal area of C, inaenas does not protrude and is bordered by three scalloped-shaped lobes with distinct bumps. The frontal area of C. aestiiarii is flat without distinct bumps and protrudes beyond the eyes. The carapace width lo length ratio of adult crabs is typically >1.29 for C iinu-iuis and < 1.27 for C. aestuaiii. KEY WORDS: Carcmus imiciuis. Cuniiiiis acstiKini. sibling species, invasive species, green crab INTRODUCTION The European green crab, a native to Europe and North Africa, is receiving nuich attention lately because of its success as a global invader. Established breeding populations outside its native range are currently found in Japan, South Africa. Australia. Tasmania, and on both the east and west coasts of North America ( Le Rou.x et al. 1990, Cohen et al. 1995. Grosholz & Ruiz 199.5. Furota et al, 1999). Twii forms of the green crab are now recogni/ed: the Atlantic Carciiiiis nniemis (Linnaeus. 1758) and the Mediterranean C, ues- tiiarii. Nardo, 1847 (old name Carciiiiis mediiemmeus Czerniav- ,sky, 1884) (Demeusy & Veillet 1953. Demeusy 1958, Holthuis & Gotdieb 1958. Almafa 1961. Zariquiey Alvarez 1968. Bulheim & Bahns 1996. Geller et al, 1997), The Strait of Gibraltar, with its high sill, acts as a partial barrier to larval and gene exchange. Consequently. MediteiTanean populations of many marine organ- isms, including the green crab, diverged from their Atlantic coun- terparts (Demeusy & Veillet 1953, Barsotti & Meluzzi 1968. Al- maija 1989. Quesada et all 995. Saavedra et al. 1995. Bulnheim & Bahns 1996. d'Udekem d'Acoz 1999). There is evidence, how- ever, of some mixing of Carcinus populations near the Strait of Gibraltar (Alma^a 1961. Clark et al. 2001. Armand Kuris. unpubl. observations). Within the last two centuries. Carcinus has been accidentally introduced into several regions outside its native range as a result of maritime commerce and ballast transport: C. muemis to the east and west coast of North America, Australia. Tasmania, and South Africa: and C, aesniani to Japan (Alma^a 1962, Geller et al, 1997), Since Carcinus spp, tolerate air exposure, starvation, and wide ranges in temperature and salinity, they are well adapted to survive ocean voyages and plane rides. Furthermore, the tree- swimming larvae have the potential to survive within the ballast tanks of cargo ships. Once released into a new environment, these ecological generalists can subsist on a variety of food organisms, including marsh vegetation, algae, crustaceans, mollusks. and fish (Cohen et al, 1995), Under favorable conditions, both species can reach sexual maturity within 1 y (Grosholz & Ruiz 1995. Furota et al. 1999. Behrens Yamada & Hunt 2000). The planktonic larvae produced bv colonists can travel on ocean currents and "seed" new breeding populations elsewhere (Behrens Yamada & Hunt 2000), Marine scientists, resource managers, shellfish growers, and mem- bers of the general public are concerned that these invaders may adversely affect marine communities by altering food webs, dis- turbing habitats, displacing native species, and preying on com- mercially important clams, mussels, oysters, and juvenile native crabs. When a green crab is sighted in a new geographical area, it is important to rapidly conftrm its identity, its source, and its mode of introduction. Once a vector has been identified, steps can be taken to prevent further individuals from arriving. If the numbers of a new invader can be kept below a critical threshold, a self- perpetuating population may not establish itself. Time is thus para- mount in identifying the invader's origin. While it may be possible to identify the most likely source of a new green crab invader with genetic tools (Bulnheim & Bahns 1996. Geller et al, 1997. Bagley & Geller 2000). these procedures take time and resources to imple- ment. We propose that a fast, simple field identification method be used to identify which Carcinus species is actually invading. Green crabs can be distinguished from other crabs in Europe and North Amenca by their fan-like shaped carapace, five promi- nent, sharp anterio-lateral teeth behind the eyes, and three lobes between the eye sockets (Fig, I; Crothers & Crothers 1988). Dis- tinguishing the Atlantic from the Meditertanean species, however, takes closer inspection. Various studies have focused on green crab species identification, but these references are not readily available to most English-speaking scientists and resource managers be- cause they are dated and/or published in Spanish, French, and Portuguese journals. Our goals are to list all the distinguishing characteristics described in the literature and to evaluate the use- fulness of each in a field setting. The latter was accomplished by examining preserved museum specimens from collection sites in the Mediterranean and Atlantic and fresh specimens from Oregon. MATERIALS AND METHODS We compiled the literature on green crab species identification and created a list of characters that could potentially be useful for distinguishing C. maenas and C. aestuaiii (Table I ), In addition. Dr. Toshio Furota kindly sent us his unpublished observations on the physical characteristics of C. aestuarii from Shinhama lagoon in Tokyo Bay Japan, In order to evaluate the usefulness of each morphological fea- ture in field identification, we obtained preserved specimens of similar size male C maenas and C. aestuarii from the Smithsonian Museum of Natural History, Collection sites included Vlissingen. Netheriands, Chausey Islands in the English Channel. Rabat. Mo- rocco. Tunis. Tunisia, and Marsala, Sicily (Fig, 2; Table 2), To supplement these samples, we collected live specimens from Yaquina Bay, Oregon using bailed and pitfall traps. For each 905 906 Yamada and Hauck ^^ ^^ B C D Figure 1. Distinguishing features of male specimens of C. maeiias and C. acsliiarii. Dorsal view of C. maeiias (A) and C. aestiiarii (B(. Note that the carapace of C. maenas is wider than that off. aesliiarii. Typically, the frontal area off. iiiaenas is scalloped-shaped with distinct "humps", while that of C. aesliiarii is flatter and protrudes. The posterior-lateral margin of the carapace in C. maenas is generally convex and the 5th anterio-lateral tooth appears to point forward. The posterior-lateral margin in C. aesliiarii is concave, while the 5th tooth is more elevated and appears to point outward. The legs of C. maenas appear shorter and thicker than those of C aestuarii. Ventral view of breastplates with abdomen removed show the male copulatory appendages (pleopods). Note that the male pleopods are crescent-shaped in C. maenas (O and straight and parallel in f. aestuarii (D). (Figure by Laura Hauck. reproduced with permission of Oregon Sea Grant.) specimen we prepared a data sheet and noted which of the char- acteristics in Table 1 could be used as a diagnostic tool in a field setting. For a feature to be useful, it had to be visible to the naked eye (or through a hand lens) and measurable with vernier calipers. RESULTS Demeusy and Veillet (1953) and Demeusy ( 1953. 1958) draw attention to a suite of characters that distinguish green crabs from the Atlantic and Mediterranean coasts of France (Table I ). They obtained specimens from Sete in the Mediterranean and from La Manche (the English Channel) and Roscoff on the Atlantic (Fig. 2). They note that Mediterranean crabs are generally smaller. hairier, and possess a narrower and thicker carapace than Atlantic crabs. Almaga (1961) expands this comparative study by adding populations from the Spanish Meditenanean and the Spanish. Por- tuguese, and French Atlantic coasts (Fig. 2). While some of the characteristics were conserved across the larger sampling area. others, such as the distribution of hair on the carapace, frontal areas, and claws, carapace texture, and the number of segments on the antennae, were not (Table 1). Almai;a (1972) observed cara- pace shape of Carcinus specimens from his former study areas us well as specimens from Australia and Maine and confirms that, in general, Atlantic specimens have wider and flatter carapaces than Mediterranean ones. He quantified these features as the carapace width (CW) to carapace length (CL) ratio (CW/CL) and the CW to carapace depth (CD) ration (CW/CD). For example. CW/CL is defined as the distance between the tips of the 5th anterio-lateral teeth divided by the distance between the central lobe of the frontal area and the posterior border of the carapace. CW/CL in C. maenas (>20 mm) varied from 1.27 to 1.35, while that in C. aestuarii varied from 1 .24 to 1 .27. The CW/CD for males >20 mm was 2.27 to 2.57 for C. maenas and 2.25 to 2.31 for C. aestuarii (Table 3). Zariquiey Alvarez (1968) provides a key for distinguishing the two species and includes photos of carapace shapes and male pleopods and a table of CW/CL for various size classes of both species and sexes (Table 4). The shape of the male pleopods is very diagnostic in separating adult Mediterranean from the Atlan- tic speciinens (Fig. 1; Tables 1 and 2). When the abdomen of a male crab is lifted, the copulatory appendages, formed from the first two pairs of pleopods. become visible. The pleopods of C. maenas curve outward in a crescent shape v\ ith the centers of the crescents touching each other. Thoseof C. aestuarii form straight, parallel lines and do not touch. While size and sex appears to affect lelative CW. the difference in the CW/CL between the species (for crabs >20 mm) is significant: 1.30 to 1.35 for C. maenas and 1.23 to 1.27 for C. aestuarii (Table 4). Rice and Ingle (1975) confirm Field lDf:NTiFicATioN of C. maenas and C aestuarii 907 TABLE 1. Morpholojjical featuri's useful for distinguishing ('. maenas and C. aestuarii in the field. 1, Demeusey and Veillet 1953; 2, Demeusy 1953; 3, DenHus\ 1958: 4. Almavu ''*<'•: ?• Zarii|uie> AlNare/, 1968; 6. Almava 1972: 7. Rice and Ingle 1975: 8. Rasmussen 1973: 9, Noel 1992: 1(1. Furota 1999; II. (,. Jensen and ( . Hieh, unpublished data: 12. Clark et al. 2(101: 13, this slud>. Feature C. maenas C. aestuarii Comments N:itive Dislrubution (4.5) Male pleopods (copulatory appendages) (4. 5. 9) Posterior (5"') pair of leeth of anterio-lateral region of carapace (3. 5. 7, 13) Posterior-lateral margin of carapace (13) Carapace width to length ratio (CW/CL) for crabs ==20 mm (.5. 6. 7, 9. 12. 13) Carapace width to depth ratio (CW/CD) (6. 12) Appearance of dorsal surface of carapace (I. 3) Margin of frontal area between the eyes) (3) Shape of three lobes in frontal area (between the eye sockets) (5. 7. 13) Segmentation of first antennae (3 1 Second antennae (3) Orientation of teeth at the base of the second antennae (3) Outer margin of claws (3) Tooth on inside of carpal (3) Presence of a fringe of long hair on the carpal ("wrist" segment) of claw (3) Length of legs Abdomen of same sized immature females (2) Carapace width of female in last prepuberty state (2) Carapace width of females after puberty molt (2. estimate based on iSVc growth increment with molt) Maximum carapace width (5. 8. la II) Western Baltic. North Sea. Atlantic from Norway and Iceland to Morocco and Mauritania Crescent-shaped, touch at the bend Appear to point forward Straight or convex Wider carapace 1.29-1.36 Thinner carapace 2.32-2.50 Smooth and uniform Fe\s or no bristles 3 distinct "bumps", margin scalloped, frontal area does not protrude 5-(i segments Fewer segments On \entral side Smooth Blunt No Shorter Narrower 22.5 mm 28-30 mm >9() mm Mediterranean. Black Sea, Asow Sea Parallel and straight, don't touch Appear to slant outward, tooth and central nb are elevated Concave Narrower carapace 1.22-1.27 Deeper carapace 2.19-2.26 Cardiac, hepatic and brachial regions di\ ided b> deep furrows; rough witli hairy projections Three size classes of bristles line the border of the frontal area Flatter, "bumps" not as distinct. frontal area protrudes beyond eyes 4 segments Longer, more segments On dorsal side Hairy with bumps Sharp Yes Longer and more slender Wider (mature at a sinaller si/e) 14 mm 16-19 mm 65 mm Yes, for specimens within nati\e range, (.some mixing near Strait of Gibraltar) Yes, Figure 1 (Fig. 1 15 in 5) Mostly. Figure 1. Teeth of some Atlantic specimens also slants outward (4) Mostly. Figure 1. But margin was concave in Atlantic specimen from Rabat. Morocco Y'es. Figure 1. Tables 2 and 3 Shows great promise as a discriminator Not noticeable for preserved specimens (13) No (4. 7) Not noticeable for preserved specimens (13) Yes. Figure I (Fig. 1 15 in 5. Plate 1 in 7) Toshio Furota. unpublished data No. feature varies with size of animal (4) No (4) No (4) No (13) No, (7, 13) No, some Atlantic specimens also possess this feature (4) Possibly. Figure I. (Fig. 115 in 5) No (4) No. would be difficult to quantify No, great variability (d'Lidekem d'Acoz. unpublished data) No. in Portugal ovigerous female C. maenas can be 15 mm and smaller (Kuris. unpublished data) No. maximum si/e is not reached in manv locations. these species difference in CW/CL. The ratios for two female C. maenas from Plymouth, U.K. are 1.29 and l..'^2, while that for a female C. aestuarii from Tunis. Tunisia is 1.25 (Table 3). Clark et al. (20(!)1 ) carried out the most statistically sound and geographically extensive study of carapace shape in Caiciniis. They sampled 17 populations from the Atlantic, one from Cali- fornia, and eight from the Mediterranean and measured CW. CL. and CD for a total of 1.737 specimens. While the carapaces of the Atlantic and California populations were significantly wider and thinner than those of from the Mediterranean, there was also great variation betv\een and within populations of a region. For example, the mean CW/CD for Atlantic populations varied from 1 .27 to 1 .32 and that for Mediterranean populations from 1 .23 to 1 .27. Standard deviations around the population means were over 0.02 (Table 3). 908 Yamada and Hauck Rabat Figure 2. Location of collecting sites for Atlantic and Mediterranean populations of Carcinus. The study sites of Demeusy and Veillet (1953) and Demeusy (1953, 19581 are indicated by open squares and include Sete in the Mediterranean and I.a Manche (English Channel) and Rostcoff in the Atlantic. Alnia^a's (I960) study sites are indicated by open circles. Sites for museum specimens are indicated by dark stars and include Marsala and Tunis in the Mediterranean and Rabat. Chausey Islands and Vlissingen in the Atlantic. The authors point out that it may not always be possible to cor- rectly assign any one specimen to a species. They found less overlap in the mean CW/CD ratio: Atlantic populations varied tiom 2.32 to 2.50 and Mediterranean populations from 2. 19 to 2.26 (Table 3). Our study of a limited number of museum specimens from Europe and North Africa and fresh specimens from Oregon sup- ports many of the previous findings. The shape of the male pleo- pods appears to be a good diagnostic field characteristic for sepa- rating adult males into the Atlantic and Mediterranean forms {Fig. 1; Table 2). The pleopods of C. iiiaenas curve outward in a cres- cent shape with the centers of the descents touching each other. Those of C aestuarii form straight parallel lines and do not touch. All Oregon specimens clearly display the C. maenas pattern (Table 2). Another diagnostic feature is CW/CL. For our limited sample, we found that if the ratio is > 1 .29, then the specimen is C. maenas: if it is between 1.25 and 1.29. it is C. aestuarii (Table 2). Unfor- tunately, we failed to measure CW/CD ratio before returning our museum specimens. After reviewing a previous version of this manuscript. Toshio Furota kindly supplied us with Table 5. This data set of carapace dimensions of C. aestuarii was collected by him and his coworkers while they were studying the life history of this species in Tokyo Bay (Furota et al. 1999). Furota states: "Results in morphological characteristics, including male pleopods and frontal area, clearly support that the samples are C. aestuarii." Both females exhibit a CW/CL ratio of under 1.27 and thus would be classified as C. aestuarii. One of the males exhibits a CW/CL of 1.31. While this value is not typical for C. aestuarii. it is within the range observed by Clark et al. (2001 ) (Table 3). The CW/CD ratios of 2.33 to 2.45 for three of the males, however, is unusually high for C. aestuarii. Almaga (1972) and Clark et al. (2001 ) found that CW/CD ratios of over 2.32 are chaiacteristic of C. maenas (Table 3). The reasons for this discrepancy are unknown. Since CD is more difficult to standardize than CW and CL. Furota may have measured the crabs in a different part of the body than Clark et al. (2001 ). If, however, the difference remains after a larger sample size of crabs is exam- ined using standard criteria, then a hybrid population may be present in Tokyo Bay. We found that the shape of the lobes in the frontal area between the eye sockets tends to vary between the species. The three lobes of C. maenas possess distinct "bumps" and form a scalloped mar- Field Identification of C. maenas and C. aestuarii 909 TABLE 2. Murphulo)>ical tea(uits iif European and North African museum specimens and li\e crabs collected from V'aquina Bay, Oregon. CW/CW = carapace width divided by carapace length. As asterisk indicates that frontal area for the Moroccan specimen yvas difficult to categorize. Collection Site Smithsonian Catalogue No. Carapace Width I mm) CW/CL Male Pleopod Orientation Frontal Area Species Marsala. Sicily Tunis, Tunisia Tunisia Siiutli 111 Rabat. Morocco Chau.sey Islands, English Channel Vlissingen, Netherlands '^'aquina Bay, Oregon 205783 .vS.77 1.253 .Suaighl Protrudes C aesluurii 27.54 1.263 Straight Protrudes C. aestuarii 25796."^ 42.5.5 1.279 Siraiaht Protrudes C. aesluurii 2579.'i9 5(1. 1 1 1.286 Straight Protrudes C. aesluurii 47.85 1.263 Straight Protrudes C. aestuarii 2583S0 5 1 .60 1.306 Curved C. maenas 2S3069 .36.(18 1 .304 Curved Scalloped C. maenas 1 1940S 28.93 1.317 Curved Scalloped C. nuienas 32.57 1.372 Curved Scalloped C. mueiias .35.66 1.331 Curved Scalloped C. maenas 40.32 1.370 Curved Scalloped C. mcwnas 44.53 1.350 Curved Scalloped C. maenas 45.94 1.382 Curved Scalloped C. maenas 75.17 1.348 Cur\ed Scalloped C. maenas gin between the eyes. The lobes of C. aestiicirii lack distinct "bumps" and do not form a scalloped margin. The frontal area in C. aestuarii is flat and protrudes beyond the eyes (Fig. 1). Our specimen from Morocco was the only one that we found difficult to characterize. The three lobes did not have distinct bumps nor did the frontal area protrude beyond the eyes. The frontal areas of the introduced green crabs in Tokyo Bay are flat and protruding, sup- porting the view that the specimens are C. aestuarii (Table 5). While male pleopods, carapace ratios, and the shape of the frontal area are good field characteristics for discriminating the two Carcinus species, other features can provide supportive evi- dence (Table 1 ). The fifth anterio-lateral teeth in C. aestuarii pos- sess a higher central rib and appear inore elevated than in C. maenas. The posterior-lateral margin of the carapace leading up to the 5th tooth tends to vary between the species. Typically, the margin is convex or straight in C. maenas and concave in C. aestuarii (Fig. 1 ). Thus, the fifth teeth appear to point forward in C. maenas and outward in C. aestuarii. While these features gen- erally hold true, there can be exceptions. Our specimen from Rabat exhibits the typical concave pattern of C. aestuarii. but the straight pleopods of C maenas. TABLE 3. Summary of carapace ratios of C. maenas and C. aestuarii observed by various researchers. .Asterisks drayvs attention to the unusually high ratios for C. aestuarii from Tokyo Bay. C. maenas C. aestuarii The walking legs appear longer and more slender in C. aestu- arii than in C. maenas (Fig. 1 ). While the carapace drawings are based on a composite of two specimens per species, great care was taken to correctly depict the relative proportions of the lengths and w idths of all the limb segments. We feel that further investigation into a leg length to CL index might yield an additional diagnostic feature. The last three distinguishing characteristics are related to life history features and are not useful in identifying crabs in the field. Sexual maturity in female Carcinus can be easily determined by simply attempting to lift the abdomen with a probe. If the abdomen is locked to the thorax, the female is immature, but if it lifts, she is mature. Males will mate with females prior to their puberty molt. Demeusy ( 195.3) found that the smallest mated female C. maenas from the Atlantic coast of France was 22.5 mm in CW, while the smallest C. aestuarii from Sete in the Mediterranean was 14 turn. After molting, their CWs would be 28 to 30 and 16 to 19 mm. respectively (Table 1 ). Sexual maturity in C. maenas from North- ern Europe, Maine, and Oregon typically occurs at a carapace width of 25 to 35 mm (Berrill 1982, Mohamedeen & HartnoU 1989, d"Udekem d'Acoz 1993, Yamada 2001). Great variation in this feature, however, exists with females maturing at a smaller size in the fall than in the spring (d'Udekem d"Acoz 1993). In TABLE 4. Mean carapace yvidth to length ratios (CW/CL) for male and female C. maenas and C. aestuarii of various size categories. Numbers in brackets give sample size. Data taken from Zariquiey .Alvarez 1968. Study CW/CL CW/CD CW/CL CW/CD Male Female .Male Female Zariquiey Alvarez 1968 AIma(;a 1972 1..W-1.35 1.27-1.35 2.27-2.57 1,2.3-1 27 1.24-1.27 2.25-2.31 Carapace Hidth C. maenas C. maenas C. aestuarii C aestuarii Rice and Ingle 1975 1.29-1.32 1.25 0-10 mm 1.200(1) I.IS7(5l 1,176(5) Clark et al. 2001 (range of 10-20 mm 1.272(1) 1,224(2) 1.281 (7) population means) 1.27-1.32 2.32-2.50 1.23-1,27 2.19-2.26 20-30 mm 1.354(2) 1.317(5) 1.252(11) 1.252(9) Clark et al. 2001 (range of 30-10 mm 1.323(10) 1.304(2) 1,266(16) 1.259(12) standard deviations 40-50 mm 1.348(9) 1.307(1) 1.247(7) 1.259(1) of population means) 1.25-1.36 2.22-2.60 1.21-1.29 2.12-2.32 50-60 mm 1.327(2) 1.298(1) 1.264(1) 1.227(1) Furotu unpublished data 1.25-1.31* 2.13-2.45* Mean ICW > 20 mm) 1.336 1.305 1.256 1.249 This study 1.30-1.38 1.25-1.29 Range (CW > 20 nimi 1.323-1,354 1.298-1.317 1.247-1,266 1.227-1,259 910 Yamada and Hauck TABLE 5. Carapace dimensions and characteristics of male pleopods and frontal area of green crabs collected from Shinhama lagoon in Tokyo Bay on May 5, 1999. This data set was provided by Toshio Furota. An asterisk indicates an unusually high ratio for C. aesluarii when compared with Mediterranean populations studied by Clark cl al. (2001). Carapace Carapace Carapace Male Pleopod Frontal Sex Width (mm) Length (mm) Depth (mm) CW/CL CW/CD Orientation .\rea Species Male 51.7 39.5 22.1 1.31* 2.34* Straight Protrudes C. cieskiiini 54.5 24.3 24.3 1.28 2.24 Straight Protrudes C. aesltiarii 53.6 42.3 23.9 1.27 2.24 Straight Protrudes C. aesluarii 49.0 39.0 23.0 1.26 2,13 Straight Protrudes C. aesUuirii 46.5 36.3 19.0 1.2S 2.45* Straight Protrudes C. aesuiarii 39.4 31.6 16.9 1.25 2.33* Straight Protrudes C. aesniarii Female 29.9 23.9 1.25 Protrudes C. aesluarii 36.3 29.2 1.24 Protrudes C. aesluarii Belgium, the smallest tnatuie female was only 22.1 mm. while the largest, in its pre-puberty molt, was 38.8 mm (d'Udekem d'Acoz. unpubl. data). In Mil Pontes. Portugal. C. inaenas females 15 mm and smaller have been observed to carry eggs (Kuris. pers. comm.). Mori et al. 1990 report that size of sexual maturity for C. aestuarii from Sardinia occurs around 29 mm. Thus, size at sexual maturity of the female cantiot be used to di-stinguish the two spe- cies of Carciinis. Maximum size attained by C. maenas is over 90 mm and 65 mtn for C. aestuarii (Table 1 ). Since maximum size is not reached in many locations, this feature would be of limited usefulness in distinguishing the species. DISCUSSION While Carcimis populations exhibit variation from site to site, the greatest discontinuity in morphological and genetic features occurs near the Strait of Gibraltar. Morphological and genetic studies support the \ iew that the Atlantic C. inaenas and the Medi- teiTanean C. aestuarii diverged from a cotnmon ancestor, but may still be subspecies rather than true species (Almaga 1961. Bulhn- heim& Bahns 1996. Clark et al. 2001). The Strait of Gibraltar with its high sill prevents water masses of the Atlantic and Mediteira- nean from freely intermixing, thus creating two basins with distinct water chemistry. Some surface water, however, does flow from the Atlantic into the Mediterranean and some deeper, high-saline wa- ter does flow from the Mediterranean into the Atlantic (Hopkins 1985). Thus, the Strait of Gibraltar acts only as a partial barrier to larval and gene exchange (Almaij'a 1989, d'Udekem d'Acoz 1999). For example. Almaga (1961) reports C. aesluarii morphs from the Atlantic Canary Islands and C. maenas morphs from Ceuta, just inside the Mediterranean (Fig. 2). Hybridization may explain why our specimen from Rabat. Morocco exhibited features of both species. Furthermore, evidence is accumulating that a hybrid zone may exist on the southeastern coast of the Iberian Peninsula. Fur- ther studies are needed to clarify the extent of any gene exchange between the two forms near this region of contact. From studying the literature and from examining specimens from Europe, North Africa, and Oregon, we conclude that in most cases, adult C. maenas and C. aestuarii can be distinguished in the field by the shape of the male pleopods. the frontal area, and the CW/CL. The pleopods of C. inaenas curve outward in a crescent shape with the centers of the crescents touching each other. Those of C. aestuarii forin straight, parallel lines and do not touch. The Oregon specimens exhibit C. /()(/(';/(/.v-shaped tnale pleopods. While male pleopods are diagnostic in species identification, one needs to be aware of some special circumstances when they are not. For example, when male C. aestuarii are parasitized by the castrating barnacle, Sacciilina carcini. the pleopods are often curved (Noel, pers. comm.). Alma^a (1961 ) cautions that pleopods may not be diagnostic in immature males before the pleopods harden. He states: "The thin pleopods of immature Atlantic speci- mens can be as straight as those from the Mediterranean." We found that the shape of the three lobes in the frontal area between the eye sockets tends to vary between the species. The three lobes of C. inaenas possess distinct "bumps" with a scalloped margin betv\ een the eyes. The lobes of C. aestuarii are flatter, less distinct, and protrude beyond the eyes. The frontal areas of the introduced green crabs in Tokyo Bay are flat and protruding, con- firming that the specimens are C. aestuarii (Furota. pers. comm.). All of our specimens, with the exception for the one from Moroc- co, were correctly categorized to species using this feature. This observation supports the view that hybridization may occur near the Strait of Gibraltar. Another distinguishing feature is that the carapace of adult C. inaenas is wider than that of C. aesluarii. This feature can be quantified as the CW/CL ratio. Zariquiey Alvarez (1968) finds that for each size and sex category. C. inaenas has a larger CW/CL ratio than C. aestuarii and that male C. inaenas typically have larger CW/CL ratios than females. This sex difference in adult C. inaenas was also observed by Shen (1935) who reports a CW/CL of 1.31 for males and 1.29 for females. Studies by Almai,"a ( 1972). Rice and Ingle (1975). Clark et al. (2001 ). and this study confirm the observation that C. maenas has a larger CW/CL than C. aes- luarii (Table 3). Clark et al. (2001) point out that because of natural variability within a population, it is not always possible to accurately identify any one specimen. However, if an adult Carci- mis specitnen exhibits a CW/CL of over 1 .29, it most likely is C. maenas: if it exhibits a CW/CL of under 1.27. it most likely is C. aesluarii. Two of our specimens frotii Tunis with CW/CL just under 1.29 would have presented a classification probletn had it not been for their distinct straight pleopods and flat, protruding frontal areas. Female Carcimis with borderline CW/CL ratios, however, would be difficult to classify to species. The great variability observed in the CW/CL of the introduced C. aesluarii in Tokyo Bay could support an hypothesis proposed by Geller et al. ( 1 997) that this population is composed of hybrids. Molecular genetic analysis revealed mitochondrial haplotypes FlHLD iDliNTIIICATlON OF C. MAENAS AND C. AESTUARIl 911 characteristic ot both species. The authors suggest either that the population originated from the hybrid /one near the Strait of Gibraltar or that more than one species has been introduced into Tokyo Bay. A subsequent analysis using microsatellite DNA did not support the multiple-invasion hypothesis for this population (Bagley & Geller 2000). Clearly, more research is needed to clarify the origin of this introduced population of Carciinis. While the two sibling species of Caniniis may look alike, it is possible to distinguish them in the field using the above morpho- logical criteria. By incorporating the three defining characterkstics. those of the shape of the male pleopods. the shape of the frontal area between the eye sockets, and the CW/CL ratio, successful field differentiation between these two species can be achieved in most cases. Other features, such as the orientation of the 5th aii- terio-lateral spine and the curvature of the posterior-lateral cara- pace margin are not diagnostic, but can provide supporting evi- dence for species identification. We believe that two additional features hold great promise as good species discriminators. We propose that the CW/CD ratio and the relationship of the length and width of the walking legs to CL be in\estigated. Furthermore, we suggest that greater discrimina- tion between the species could be achieved by taking the sexual difference in the carapace shape into account. ACKNOWLEDGMENTS Many people helped us m this study. The staff of the Oregon State University Valley Library and Susan Gilmont of the Guinn Library at Hatfield Marine Science Center provided us with per- tinent articles. Paul Clark. Michael Neale, and Philip Rainbow of the Natural History Museum in London. Pierre Noel of the Natural History Museum in Paris, and Cedric d'Udekem d'Acoz. research associate at the Royal Institute of Natural Sciences in Brussels, kindly provided us with copies of their manuscripts. Toshio Furota of Toho University generously sent us carapace dimensions from the introduced population of Caniiws aesniarii in Tokyo Bay. Anthony Pires translated a critical passage from the original Por- tuguese, David Reinert provided the map, and Cheryl Bright and her staff of the Smithsonian Museum of Natural History sent us the European and North African green crab specimens. Comments by Hans-Peter Bulnheim, James Carlton, Paul Clark, Cedric d'Udekem d'Aco/, Toshio Furota. Jonathan Geller. Armand Kuris, Andrew Lohrer, and William Walton greatly improved this manu- script. We thank them all. This research was funded, in part, by the National Sea Grant College program of the U.S. Depanment of Commerce's National Oceanic and Atmospheric Administration under NOAA grant number NA76RG()476. project number R/NTS-2-PD. LITERATI! Almacj-a. C. 1961. Variabilidade de alguns caracteres usados na taxononiia do gen. Curcimis Leach. Revista da Faculdade de Ciencias da Univer- sidade de Lisboa, Serie 2. C, Ciencias Natural 8:137-154. Alma(ja, C. 1962. Sur la distribution geographique du genre Caniiuis Leach (Crust. Dec. Brach.). Revista da Faculdade de Ciencias da Uni- versidade de Lisboa. Serie 2, C, Ciencias Natural 10:109-1 13. Alniai;a. C. 1972. Note preliminaire sur les rapports entre quelques dimen- sions de la carapace de Carcinus maenas (L) et C. iiieililerraiu'us Czemiavsky. Thalassia Jugoslavica 8:49-57. Alma^a. C. 1989. Biogeographic notes on the Mediterranean Brachyura. Cruslacecma 56(1): 1-1— 17. Bagley. M. J. & J. B. Geller. 2000. Microsatellite DNA analysis of native and invading populations of European green crabs. In: J. Peterson, editor. Marine Bioinvasions: Proceedings of the First National Confer- ence. January 24-27. 1999. Boston: Massachusetts Institute of Tech- nology Sea Grant, pp. 241-143. Barsotti, G. & C. Meluzzi. 1968. Osservazioni su Mytilus cdulis L e M. gallnprovincialis Lmk. Conchiglie 4:50-58. Behrens Yamada, S. & C. Hunt. 2000. The arrival of the European Green Crab. Carcinus means in the Pacific Northwest. Dreissena 11(1 ):l-7. Behrens Yamada. S. 2001. A global Invader: the European green crab (Corvallis, OR. Oregon Sea Grant and Washington Sea Grant). Berrill. M. 1982. The life history of the green crab Carcinus maenas at the northern end of its range. J. Crustacean Biol. 2:31-39. Bulnheim. H.-P. & S. Bahns. 1996. Genetic variation and divergence in the genus Carcinus (Crustacea. Decapoda). Int. Rev. Gcs. Hxdraliiol. 81: 611-619. Clark. P. F.. M. Neale & P. S. Rainbow. 2001 . A morphometric analysis of regional variation in Carcinus (Leach. 1814) (Brachyura: Portunidae: Carcininael with particular reference to the status of the two species C. maenas (Linnaeus, 1758) and C. aesluarii Nardo, 1847. J. Crustacean Bull. 21(l):288-303.. Cohen. A. N.. J. T. Carlton & M. C. Fountain. 1995. hitroduction, di.spersal and potential impacts of the green crab Carcinus maenas in San Fran- cisco Bay. California. Mar. Biol. 122:225-237. Crothers. J. & M. Crothers. 1988. A key to the crabs and crab-like animals of British inshore waters. Field Studies 5:753-806. RE CITED Czemiavsky. V. 1884. Crustacea Decapoda Pontica littoralia. Materiali and Zoographiam Ponticarn comparatam. Trudu Ohschch. Isput, Prir. imp. Khar' Kov. 13:1-268. Demeusy. N. & M. A. Veillet. 1953. Sur I'existence de deux populations de Carcinus maenas Pennant et sur les characteres morphologiques qui les distinguent. Comptes Rendus Hehdomadalres Des Seances Acad. Sci. 236:1088-1090. Demeusy. N. 1953. Comparaison biologique de deux populations de Carci- nus maenas Pennant: mue de puberte. Comptes Rendus Hehdomadalres Des Seances Acad. Sci. 236:1203-1205. Demeusy. N. 1958. Recherches sur la mue de puberte du Decapode Brachyoure Carclitus maenas Linne. .Arch. Zool. E.xp. Gen. 95:253- 492. D'Udekem d'Acoz. C. 1993. Activites reproductrices saksonnieres des dif- t'erentes classes de tallies d'une population de crabes verts Carcinus maenas (Linnaeus. 1758) dans le sud de la iner du Nord. Cah. Biol. Mar. 35:1-13. D'Udekem d'Acoz, C. 1999. Inventaire et distribution des crustaces deca- podes de I'Atlantique nord-oriental, de la Mediterranee et de eaux continentales adjacentes au nord de 25°N. Patrim Nat.. Museum na- tional d'Histoire naturelle, Paris 40, 383 pp. Furota, T.. S. Watanahe. T. Watanabe. S. Akiyama & K. Kinoshita. 1999. Life history of the Mediterranean green crab. Carcinus aestuaril Nardo, in Tokyo Bay, Japan. Crustacean Res, 28:5-15. Geller, J. B., E. D. Walton, E. D. Grosholz & G, M. Ruiz. 1997. Cryptic invasions of the crab Carcinus detected by molecular phylogeography. Mol. Ecol. 6:901-906. Grosholz. E. D. & G. M. Ruiz. 1995. Spread and potential impact of the recently introduced European green crab, Carcinus maenas. in central California. Mar. Biol. 122:239-247. Holthuis. L. B. & E. Gottlieb. 1958. An annotated list of the decapod Crustacea of the Mediterranean coast of Israel, with an appendix listing the decapoda of the eastern Mediterranean. Bull. Res. Council Israel 7B. Hopkins. T. S. 1985. Physica of the sea. In: R. Margalef. editor. Key Enxironments Western Mediterranean. New York: Pergamon Press. Le Roux. P. J.. B. M. Branch & P. A. P. Joska. 1990. On the distribution. 912 Yamada and Hauck diet and po>sible impact of the invasive European shore crab, Carciinis maeiias (L. ) along the South African coast. South African J. Mar. Sci. 9:85-93. Mohariiedeen. H. & R. G. Hartnoll. 1989. Larval and postlarval growth of individually reared specimens of the common shore crab Carciniis maenas (L). J. Exp. Mar. Biol. Ecol. 134:1-24. Mori. M.. R. Manconi & G. Fanciulli. 1990. Notes on the reproductive biology of Carcinus aestiiarii Nardo (Crustacea, Decapoda) from the Lagoon of San Teodora (Island of Sardinia, Italy). Rivista ili Idrohio- logia 29(3):763-774. Nardo, G. D. 1847. Sinomnimia modema delle specie registrate nell'opera intitolata: Descrizione de Crostacei. de Tesyacei e de Pesci che abitano le lagune e golfo veneto rappresentati dall" Abate Stefano Chiereghini. 1-127. Venice. Noel, P. Y. 1992. Cle preliminaire d'identification des Crustacea decapoda de France et des principales autres especes d'Europe. Colletion Patri- moines Naturels, Volume 9. Serie Patrimoine Scientifique. Quesada, H, C. M. Beynon & D. O. F. Skibinski. 1995. A mitochondrial DNA discontinuity in the mussel Mytihis galloprinincialis Link: Pleis- tocene vicariance biogeography and secondary intergradation. Mol Biol Evol. 12:521-524. Rasmussen, E. 1973. Systematics and ecology of the Isefjord marine fauna (Denmark) with a survey of the eelgrass iZostera) vegetation and its communities. Ophelia 11:1-507. Rice, A. L. & R. W. Ingle. 1975. The larval development of Carcinus uuiena.s (1.) and C mediterraneus Czemiavsky (Crustacea, Brachyura, Portunidae) reared in the laboratory. Bull. Br. Museum (Natural His- toiy) 28:103-119. Saavedra, C, C. Zapata & G. Alvarez. 1995. Geographical patterns of variability at allozyine loci in the European oyster Ostrea eclulis. Mar. Biol. 122:95-104. Shen, C. J. 1935. An investigation of the post-larval development of the shore-crab Carcinus maenas. with special reference to the exlernal secondary sexual characters. Proc Zool. Soc, London 1:1-33. Zariquiey Alvarez. R. 1968. Crustaceos decapodos ibericos. Investigacion pesq. 32:i-xi.. Journal ,>l Shcllli^h Rcseanli. Vol. 20. No. 3, yLVyii!). :0()1. REDUCING PREDATION ON MANILA CLAMS BY NONINDIGENOUS EUROPEAN GREEN CRABS EDWIN GROSHOLZ,' PAUL OLIN,' BRIAR WILLIAMS,' AND RICO TINSMAN^ 'Department of Environwental Science and Policy. University of California. Davis. California 95616: 'University of California, Sea Grant Extension Program. 2604 Ventura Avenue. Santa Rosa. California 9540.1: ^Bodega Marine Laboratory. P.O. Box 247. Bodega Bay. California 94923 ABSTRACT The introduced European green crab [Cciirinus imieiias) pcses a potential risk for commercial production of Manila clams (Venenipis philippiiuinim). a growing fishery in western North America. We investigated methods for modifying commercial production of Manila clams in order to reduce losses to green crab predation. In both 1997 and 1998, the timing of the outplant of seed clams was varied such that one portion of the total clam production was outplanted early in the year (March) and another portion outplanted later in the year (August/September). In July and August 2000. we examined these bags outplanted in 1997 and 1998 and quantified the clam mortality and the abundance of crabs including green crabs. We tested the effects of both year of transfer (1997 vs. 1998) and the timing of transfer (early season vs. late season) on clam and crab abundance and found significantly (p < 0.05) fewer green crabs and less predation on Manila clams in bags that were outplanted in the late season. Year of outplant also influenced clam survival and green crab abundance, such that delayed outplant significantly (p < 0.05) increased clam survival only when green crab abundances were high (1997), Our data suggest that the increased Manila clam survival in delayed outplant bags was likely due to reduced green crab recruitment and consequently lower green crab predation. We conducted additional experiments on relative predation rates by green crabs on Manila clams. We found that green crab predation is strongly size dependent, and that while juvenile green crabs were not effective predators of m;irket size Manila clams (>36 mm in width), these clams were easily consumed by adult green crabs (>50 mm carapace width). We conclude that in years of high green crab abundance, delayed outplant is an effective means of reducing losses of commercially produced Manila clams. KEY WORDS: Manila clam, Venenipis pliilqipiiumiiit. European green crab, Caixinus iihii-iHi.s. predation, recruitment, nonindig- enous INTRODUCTION More than 400 non-native species have become established in the coastal waters of the U.S., and many have resulted in substan- tial economic and ecological impacts (Ruiz et al. 1997, Cohen & Carlton I99S. Ruiz et al. 2000). In western North America, the recently introduced European green crab {Carcinus imieiuis. Lin- naeus. 17.58) no\\ poses a potential threat to the commercial pro- duction of Manila clams ( Venenipis (Ruditapes) philippinanim. .\. Adams and Reeve. 1830) (see Carlton 1992 and Coan et al. 2000 regarding subspecies) and other invertebrate fisheries in western North America (Lafferty & Kuris 1996. Jamieson et al. 1998), Like green crabs. Manila clams are also non-native, having been acci- dentally introduced with Pacific oysters (Crassostrea gii^as. Thun- berg. 1795) in San Francisco Bay in the 1940s (Cohen & Carlton 1995). Manila clams now support a growing commercial fishery with landing revenues of approximately $10.5 million in 1999 for Washington (Pacific States Marine Fisheries Commission 1999) and increasing production in Oregon. California and British Co- lumbia. Canada. European green crabs have the potential to be- come an important source of mortality for Manila clam producers throughout most of this range, where they have been reported from Morro Bay. CA to Vancouver Island. British Columbia since 1998. and have established populations between Monterey Bay. CA and Gray's Harbor. WA (Grosholz et al. 2000, G. S. Jamieson. unpub. data). In central California, where they first became established in San Francisco Bay in 1989, green crab predation has already re- sulted in losses of nearly 40% of the annual production of 5.500 kg for one producer in 1996 (John Finger. Hog Island Oyster Co., Marshall, CA. pers. comm.). European green crabs are widely acknowledged as a pest spe- cies in other areas including Australia (Thresher 1997). South Africa (LeRoux et al. 1990. Griffiths et al. 1992). and the north- eastern U. S., where they likely contributed to the demise of the soft-shell clam {Mya areiiariu Linnaeus. 1758) fishery in New England in the early 1900s (Glude 1955. MacPhail et al. 1955. Ropes 1968). In its native range along the North Atlantic coast of Europe, green crabs are prodigious consumers of invertebrates in general, and bivalves in particular (Reise 1977. Scheier & Reise 1981. Jensen & Jensen 1985. Sanchez-Salazar et al. 1987. Rafaelli et al. 1989) and are one of the primary predators of Manila clams (Cigan-i'a & Fernandez 1998. Cigan-fa & Fernandez 2000). The method of commercial production is likely to influence rates of green crab predation on Manila clams. Several culture methods are used in this fishery depending on available space, substrate type, and local predators, include bag culture, and direct outplanting (Spencer et al. 1992). During the two to four years required to grow Manila clams to market size, they can be vulner- able to various threats including crabs, waterfowl, fouling organ- isms, and parasites (Laing 1993, Paillard & Maes 1994, Smith & Langdon 1998). After the losses to green crabs experienced by central California growers in 1996. there was a need to modify the bag culture methods commonly used for Manila clam production in central California, in order to reduce green crab predation. The specific goal of this study was to investigate the effective- ness of delaying the timing of outplant of Manila clams for reduc- ing predation by green crabs. Typically, hatchery reared seed clams are initially placed in small mesh ""seed bags," then, after approximately one year, are transferred from the seed bags to larger mesh "growout bags." The original date of outplant into seed bags largely determines the timing of the transfer of clams from seed bags to growout bags, so that both outplant and transfer happen either early in the season (winter/spring) or later in the year (summer/fall) approximately a year apart. After transfer, clams remained in growout bags for typically two more years without any further tnanagetnent. We hypothesized that the timing of the 913 914 Grosholz et al. dates of oulplant and transfer (hereafter refeired to as luitplant date) from early in the year (March) to later in the year (August/ September) would reduce the number of green crabs that recruited into the Manila clam bags as megalopae or early instar juveniles. Since predation is cumulative during the time Manila clams spend in both seed and grovvout bags, our assumption was that reducing green crab recruitment into clam bags would reduce the abundance of predatory adult green crabs, thus resulting in higher survival of Manila clams at the time of harvest. MATERIALS AND METHODS Manila clams used in these experiments were grown in com- mercial culture by the Hog Island Oyster Company (HIOC). Mar- shall. CA. USA. The site for the delayed outpiant experiment was Tom"s Point in Tomales Bay (Marin Co.). CA (38'13'()2"N. I23°56'49"W). The site has a muddy sand substratum and an annual temperature range of 10-20°C and a salinity range of 20-33 ppt. Salinity retnains high during drier months from late spring through early fall (Cole et al. 1990 for additional site information on Tomales Bay). To investigate size dependent predation of Manila clams by green crabs, we conducted an additional short-term experiment on the marine reserve of the Bodega Marine Laboratory. Bodega Bay (Sonoma Co.). CA (38°20'00"N, 123°2'30"W). This si'te has sandy-muddy sand substratum with a temperature range similar to Tomales Bay. but with salinity generally near full seawater. Additional site information can be found in Grosholz et al. 2000. Delayed (hitplaiil l:\peiiiiniil The experiment consisted of two experimental treatments: the first treatment involved manipulating the timing of otitplant (and transfer) and the second treatment was the year that the clams were first outplanted (either 1997 or 1998). Early outpiant occuired in March (1997 and 1998) and late outpiant occurred in September (1997) and August ( 1998). This resulted in bags being in the field exposed to crab predation for approximately the following number of days: 1.215 days (March 1997). 1.035 (September 1997). 850 (March 1998). 700 (August 1998). Outplanted seed clams (ap- proximately 6 mm width) were held in seed bags (3 mm mesh bags, 1 m L X 0.6 m W x 0.3 m H) that were placed in rows on the substrate at a tidal height between approximately +0.3 m and + 1 .0 m. After approximately one year, clams (approximately 12 mm width) were transferred from seed bags to growout bags (6 mm mesh. I m L X 0.6 m W X 0.3 m H) and typically remained in growout bags for an additional two years until they reached com- mercial size. During this transfer from seed bags to growout bags, larger crabs and excess shell debris were removed, although smaller shore crabs such as Hemigrapsus oregonensis (Dana. 1851) were frequently missed and not removed during transfer. Between July 2000 and September 2000. we examined bags that had been outplanted both early and late in 1997 and 1998. For each of the four different outpiant periods. March 1997. September 1997, March 1998, and August 1998, we examined at least 20 Manila clam bags. For each bag, we recorded the number, size (carapace width to nearest mm), species, and gender of all crabs including European green crabs (Carcinus mcienas). grapsid shore crabs including Hemigrapsus orcgoiieiisis and Pachygnipus criis- sipes (Randall. 1839). and cancrid crabs Cancer prodiicliis (Ran- dall. 1839). Canct'r magister [V)-MVA. 1852). Cancer antcuuariiis (Stimson. 1856). Also, we recorded the number of crushed clams by counting intact hinges and measuring a subset of the remaining live clams at the widest point on the shell to nearest mm. Shells crushed by green crabs were generally identified by characteristic shell damage at the growing edge of the shell. Since clams out- planted in 1997 were already market size, we examined these at the HIOC facility as they were being harvested for sale. However, since clams outplanted in 1998 were slightly smaller and not ready for harvest, we examined these in the field using the same data collection niethods. and returned them to the field. To address the possibility that differences in the relationship between shell thickness and shell width for clams outplanted early vs. late might contribute to differences in predation rates, we mea- sured the shell width along the anterior-posterior axis and shell thickness at the posterior margin and at the pallial sinus for 100 intact shells from March 1998 and August 1998. Size Dependent Predation Experiment We conducted a second experiment in July 1999 to examine how the relationship between green crab size and Manila clam size influence predation rates under commercial culture conditions. Us- ing conimercial growout bags from HIOC. we placed 15 clams chosen from one of three size class treatments (<23 mm. 25-33 mm, or >36 inm) in commercial clam bags with one green crab chosen from one of three size class treatments (30—15 inm. 50-65 mm, or >70 mm). All experimental crabs for the two largest size classes were intermolt males, while for the smallest size class, because of the limited availability of intermolt males in this size range, we used half males and half females split evenly among clam size treatments. We used a fully factorial design with all treatment conibinations yielding nine treatments (3 clam x 3 crab treatments) with each treatment replicated five times for a total of 45 experimental units { = clam bags). We placed the 45 growout bags with experimental clanis and crabs in five blocks placed at approximately 0.2 MLLW in Bodega Harbor anchored to the substrate following standard commercial outpiant procedures. We anchored bags placed side by side and separated blocks by approximately 5 meters. After two weeks, we recovered all clani bags and counted and measured all clams and crabs. Because there was no shell breakage or other physical evidence of crabs in the smallest size class preying on clams in the largest size class, we used this treatment as a conservative ""control" treat- ment, for which we assigned mortality in this treatment as "non- predator mortality." We used this value for ""non-predator" mor- tality (4'!f ) to recalculate mortality for other treatments prior to analysis. Therefore, this correction of 49f reduces the between- treatment differences in mortality between bags with large, me- dium, and small crabs prior to statistical analysis. This resulted in more conservative between-treatment tests by making it more dif- ficult to determine significant differences between treatment, thus reducing the probability of Type II error. Statistical Analysis For the outpiant delay experiment, we conducted a two-factor analysis of variance (ANOVA) with year ( 1997 or 1998) crossed with outpiant date (early or late) as main effects and the number of green crabs per bag and the number of preyed clams per bag as dependent variables. Separate post-hoc tests of treatment means were estimated with Tukev's studenlized ranee test for each vear. Manila Clam Prki:)Ation b>' Introduced Green Crabs 915 To unaly/c the ivlalionship between the luimber ol yieeii crabs per bag and the number of preyed clams across all outplant periods, we used linear regression with the numbers of green crabs and the numbers of preyed clams as paired variates for all outplant dates. We used linear regression to analyze the relationship between green crab abundance and several other variables including the rnmiber of preyed clams per bag. the number of Heniifirapsiis on't^iiiicnsis per bag. and the number of Pachyi;ra!7siis ciassipes per bag. We also used linear regression to analyze the relationship between clamshell thicknesses on clam width for clams outplanted earlv vs. late in 1998. Lastly, to test for differences in size-specific shell thickness between treatments, we conducted independent lin- ear regressions of shell thickness on shell width for both March and August 1998 clams (100 each) using ordinary least squares regression. For the size-dependent predation experiment, we used a two- factor ANOVA with crab and clam size as main effects and clam survival as the dependent variable. As a limited test the effect of crab gender on clam mortality, we also analyzed data from this experiment in treatments with the smallest crab size class where we used equal numbers of male and female crabs. We used a t-test to compare differences in mortality for males vs. female crabs pooled across clam size treatments (n = 15). The data for all ANOVAs were tested for homogeneity of variances and data were either square root transformed (counts), log-transformed (sizes), or arc-sin square root transformed (per- centages) as needed to meet test assumptions. Ail statistical analy- ses were conducted with SAS 8.0 (Statistical Analysis Systems. Carey, NC). RESULTS Delayed OiilplanI Experiment 0\erall green crabs were most abundant in bags outplanted in 1997 and were more abundant in bags outplanted early (March 1997) than late (September 1997) (Fig. 1). Both the outplant date (F = 4.72. p < 0.05) and year (F = 16.3. p < 0.0001 ) significantly affected the average number of green crabs per bag. and there was a significant interaction between outplant date and year (F = 5.63. /' < 0.05). Post-hoc tests showed there was a strong affect of outplant date in 1997 with fewer crabs in clam bags on the later outplant date (F = 6.8, /; < 0.015). but there was no significant effect of outplant date in 1998 (F = 0.16, /> > 0.60) when overall green crab abundances for that year were lower (Fig. 1 ). Although the mean abundance of green crabs was low (mean 0.8 ± 0.66 crabs per bag in March 1997). only one green crab per bag may be necessary to produce large losses of clams (see later). A strong treatment effect was evident with the percentage of crushed clams per clam bag being greatest in the March 1997 outplant group (mean 45.5% ± 21.4) (Fig. 2). Overall, there was a highly significant effect of both year (F = 66.74. p < 0.0001 ) and outplant date (F = 31.08, /> < 0.0001 ) and there was a significant interaction as well (F = 17.31. /; < .0005). Post-hoc tests also showed a strong affect of outplani date in 1997 (/; < 0.0001 ). but not in 1998 (p > 0.40). We also found a strong positive relationship between the num- ber of green crabs per bag and the number of crushed clams, when data were pooled across outplant dates. Figure 3 shows that as the number of green crabs per bag increases, there was a significant increase in the number of preyed clams observed per bag (n = 77. r = 0.427. p < 0.001). We also found that the abundance of Hemigrapsiis was negatively associated with the abundance of green crabs. As Figure 4 shows, there was a significant negative relationship between the number of Hcmigrapsus per bag and the number of green crabs per bag (n = 8 1 . r" = 0.20. p < 0.001 ). Our analysis of the effects of year and outplant period on the number of Hemigrapsiis showed a significant affect of year (F = 7.39. /> < 0.01), no effect of outplant date (F = 0.07, p > 0.701. and a significant year by outplant date interaction (F = 11.21. /> < 0.005). The mean size of the clams in the growout bags varied as a function of outplant date (36.3 ± 4.9 mm March 1997, 34.1 ± 3.1 mm September 1997. 33.5 ± 2.9 mm March 1998. 32.4 ± 2.4 mm .August 1998) (Fig. 5). Clams from early outplant date groups (March 1997 and March 1998) were significantly larger (F = (0 V) n O c 0) 0) O 80 Mar-97 Sept-97 Mar-98 Aug-98 Outplant Date Figure L The number of green crabs <(bser\e O.dO for March. /) > 0.20 for August). Lastly, green crab size and gender were similar between treat- luents (Table 1). There were with no significant differences (/; > 0.80) among treatments in crab size and an increasing trend in crab sex ratio (male:female) with treatment date, although the low num- bers of crabs in 1 998 make these ratios a bit suspect. Size Dependent Predation Experiment The relative sizes of both green crabs and Manila clams intlu- enced predation rates of Manila clams (Fig. 7). There was a sig- 0 12 3 Green Crabs (# / bag) Figure 4. Number of green crabs observed per clam bag versus the number of native shore crabs {Hemif;rapsiis oref;iinensi>,) in outplants delay experiment. Each point represents the total number of crabs and clams for each of twenty bags sampled from all four outplant periods (March 1997, September 1997, March 1998. August 19981. Regression line is an ordinary least squares tit through all points with an uncon- strained intercept. E £, W) 0) c O 1.5 0.5 - CO March 98: y = 0.0029x -f 0.7685: R^ = 0.0028 4^1 ll ^Bl ■■- August 98: y = 0.0056x + 0.6441 ; R^ = 0.01 27 20 25 30 35 Shell Width (mm) 40 —I 45 Figure ft. The relationship between shell thickness and shell width for Manila clams from two outplant dates (March 1998 and August 1998). Each point represents single measurements of thickness and width of a single shell from either March 1998 (black trianglesi or August 1998 (gray squares). Regression lines for either March 1998 (black) or Au- gust 1998 (gray) are ordinary least squares fit through all points. Manila Clam Prldation by Introduced Green Crabs 917 TABLE 1. Abundance and size of male and female green crabs (Carciniis maeiiasi (hat recruited Into Manila clam bags during the I)ela>ed Outplant K\perimenl. For each outplant period in = 20 bags I. the mean and (me standard de\iation are sho«n for green crab abundance per bag (Mean # and S.I). #) and green crab carapace width (Mean C\\ and S.D. CVV) for both males and females as well as male to female ration (M : F ratio). Mar-97 Sep-97 Mar-98 Aug-98 Males Mean # 0.30 0.14 0.05 (1,10 S.D- # 0.47 0.36 0.00 11.31 Mean CW 54.17 59.33 36.00 4 1 .00 S.D. CW 19.23 4.93 0.00 15.56 Females Mean# 0.50 0.19 0.05 0.05 S.D. # 0.61 0.51 0.00 0.00 Mean CW 57.10 56.00 70.00 4S.O0 S.D. CW 11.17 9.13 0.00 0.0(1 M : F Ratio 0.6 0.74 1 T nificant effect of green crab size (F = 12.6. /> < 0.0001) and Manila clam size (F = 6.37. /; < 0.005) and no significant [p > 0.301 interaction. We found that both medium and large green crabs preyed upon substantial numbers of large Manila clams. However, small green crabs (30—15 mm) were not able to prey on the large Manila clams (>36 mm) (Fig. 5) and no crushed shells were seen in any replicates in this treatment. Wc found no signifi- cant effects of crab gender with similar survival of clams in bags with male (mean 84% ± 13) vs. female crabs (mean 86% ± 15) (t = -0.28. df = 1 3, /; > 0.70) using bags with the smallest crab size class (to compare males or females) pooled across clam treatments. DISCUSSION The main result was the strong effect of outplant liming on the number of crushed clams and the number of green crabs in clam bags. Clams outplanted early (March 1997) had higher numbers of green crabs and experienced greater losses than clatns outplanted later (August 1997) (Fig. 1). These data suggest that in growout bags outplanted in March 1997, adult green crabs were capable of consuming an average of 45% of Manila clams, and in the most extreme cases, more than 70% (Fig. 2). The regression analysis showed a significant (/) < O.OOl ) posi- tive relationship between increasing numbers of green crabs in growout bags and increasing numbers of crushed clams (Fig. 3). However, this relationship explained only about 40% of the varia- tion in clam mortality, due largely to bags with "zero" crabs and substantial numbers of crushed clams. It is possible that the num- ber of green crabs in these bags was initially higher, but unknow n mortality may have resulted in fewer crabs per bag. Because the results show modest numbers of green crabs per bag (on average less than 1 green crab per bag) even for clams outplanted in March 1997 (Table 1), this suggests that the high mortality in the early- outplanted bags was produced by small number of crabs. In any case, the substantial numbers of crushed clams in these experi- mental bags underscores the potential impact that green crabs can have on clams in commercial culture. Although other factors can also produce substantial clam mortality, we are unaware of any that would cause the conspicuous shell damage obser\ed along the margin of the shell (typically posterior marginl inside clam bags other than crab predation. .Another result from our study is that the effect of the timing of outplant was much smaller and not significant (p > 0.60) in 1998 in comparison with 1997, when overall green crab abundance were much lower. This suggests that the impact of green crabs may be \ ariable between years, and if green crab recruitment is poor, as in 1998, there may not be sufficient green crab predation to justify the delay in outplant (and transfer) given some cost to delayed out- plant in terms of growth (see Fig. 5). The decreased predation in clam bags outplanted later may be partly the result of less total time spent in the field. In Figure 1, there is a positive relationship between the number of days the bags were in the field and the luimber of crushed clams. Given that the March 1997 treatment had been in the field for three recruit- ment seasons and the March 1 998 treatment had been in the field for only two, it might be expected that the March 1997 treatments might have more crabs or the crabs may have attained a larger size. The delay in outplant timing had a stronger effect on reducing predation than reducing clam size. As shown in Figure 6, we found a 70% reduction in mortality between clams outplanted in March 1997 (mean 45.5% ±21.3) vs. September 1997 (mean 13.5% ± S.Ol (M1-M2 in Fig. 6). but only a 6% reduction in size between clams outplanted in March 1997 (mean 36.3 mm ± 5.0) vs. Sep- tember 1997 (mean 34.1 ±3.1)(SI-S2 in Fig. 6). Thus, the delayed outplant resulted in a substantial increase in clam survival, and therefore, commercial production, with a relatively minor reduc- tion in clam size. This differential effect on the number of days in the field on mortality vs. size likely reflects the possibility that green crab recruitment may not be as continuous throughout the year as pre- dation. The number of spring/early summer recruitment seasons that a bag may experience may largely detemiine the nuinber of crabs per clam bag. Green crabs are able to enter the growout bags only while they are smaller than the mesh, such as megalopae and early instar juveniles (generally CW <10 mm. thus carapace height <6 mm). In central California, most recruitment of early instar green crabs occurs during the spring and early summer (Grosholz et al. 2000. Grosholz. unpubl. data) and by late summer, most 100 m (/) m O 20 I SMALL CRABS 30-45 mm I MEDIUM CRABS 50-65 mm I LARGE CRABS >70 mm Small Medium 23 mm 25-33 mm Clam Size Large >36 mm Figure 7. Percentage survival of Manila clams in the size dependent predation experiment. Bar heights represent the mean survival of clams for each of the nine treatment combinations (n = five bags per treatment). Error bars represent one standard error of the mean. 918 Grosholz et al. green erahs are too large to enter the growont bags. Therefore, clams hags outplanted in the fall would not generally be subject to recruitment by young-of-the-year green crabs, although they would be receive recruits in the following years. The size dependent predation experiment provides additional support for green crabs being responsible for most of the clam mortality observed in the outplant delay experiment. Although green crabs in the smallest size class (30—43 mm CW) were not able to consume market-size clams (>36 mm I. we found medium size crabs (50-65 mm CW) could consume Manila clams of any size in commercial production, so there was no size refuge from green crab predation (Fig. 7). Also, mortality of the largest clams was the same for both medium and large green crabs suggesting that green crabs in central California, which typically reach sizes exceeding 50 mm CW in their first year (Grosholz & Ruiz 1995. Yainada. in press), can consume the full range of clam sizes in commercial bags within a one year of recruitment. In addition, within the limited size range and small samples sizes involved in this experiment, we found no significant differences in predation levels between equivalently sized male and female green crabs (/> > 0.70). Although the outplant timing appears to have influenced clam mortality, many other characteristics of the crabs and clams in these expeiiments can influence the observed patterns of mortality. Previous work by EIner and Hughes (1978) EIner (1980. 1981). and EIner and Rafaelli ( 1980) have demonstrated that characteris- tics of crabs such as chelae size. sex. molt stage, etc. may interact with previous diet experience, temperature and other environmen- tal factors to influence the rate and nature of predation. In the delayed outplant experiment, because we could not control the identity of the crabs in the clam bags (crabs recruited freely into bags), we were unable to control for differences among crabs. However, data for crab characteristics shown in Table 1 suggest that the differences in mortality among treatments were unlikely to have been influenced by differences in crab size or crab gender. The trends for crab size and gender among treatments opposed what might be expected if these factors were to have explained treatment differences in clam mortality. The relative shell thickness of Manila clams might also have influenced the differences in clam mortality between treatments. Howev er. as Figure 6 shows, there were no significant differences in the relationship between shell thickness and shell size (width) for eaily ( p > 0.60) of late (p > 0.20) outplants treatments in 1 997 with shell size explaining 1 % or less of the variation in thickness overall (R- = 0.003 for early. R^ = 0.01 for late). Therefore, difference in the timing of outplant did not produce changes in relative shell thickness. In addition to the green crabs, there were other species of crabs recruiting into the clam bags including native shore crabs Hemi- grapsus oregonensis. Puchygrapsus crassipes. and native Cancer spp. The Cancer spp. abundances were at very low abundance (<0.05 per bag) and Pacliygrapsits were equivalently rare and unlikely to have had a measurable impact on the patterns. By contrast. Hemigrapsits were commonly found in the growout bags along with green crabs (overall 1 1.3 ± 6.9). although they rarely exceeded 20 mm in carapace width, and were also unlikely to have had a substantial impact on the number of crushed shells observed in the bags (Grosholz. unpubl. data). For example, the lowest Hemigrapsiis abundance (mean per bag 7.30 ± 6.5) occurred in the outplant treatment with the highest clam mortality (March 1997) and the highest Hemigrapsiis abundance (mean per bag 15.35 ± 6.34) occuned in the treatment with the lowest clam mortality in (March 1998. see Fig. I). We found that there were fewer Hemigrapiis in 1997 when green crabs were more abundant, although the outplant date had opposing effects on Hemigrapsiis abundance in different years. We also found a significant (/) < 0.001 ) negative relationship overall between the abundance of Hemigrapsiis and the abundance of green crabs (Fig. 4). Although adult Hemigrapiis may prey on newly recruiting green crabs, the predatory relationship between these crabs may switch as green crabs approach adult size and prey on Hemigrapsiis (Grosholz et al. 2000). The strong dependence of size as a determinant of intraguild predation among crabs has been amply demonstrated in western U.S. estuaries (Dumbauld et al. 1993. Iribarne et al. 1994). In conclusion, our work demonstrated that decreased predation by green crabs on commercially produced Manila clams resulted from delaying the outplant timing of clams. We suggest that this low cost management tool could be used by Manila clam produc- ers throughout western North America to reduce predation by green crabs in regions where green crabs are abundant. However, the specific timing of the delay needed to produce this result will likely vary with latitude, since the recruitment of green crabs is delayed with increasing latitude (Yamada. in press). Regardless, the Manila clam fishery would benefit by instittiting low cost culture practices such as outplant delay, in order to avoid future losses to European green crabs. ACKNOWLEDGMENTS We thank the University of California Division of Agriculture and Natural Resources (DANR) for their support of Briar Williams thixuigh the RREA program and the Department of Environmental Sciences and Policy. We thank the Editor for her patience with revisions and two anonymous reviewers for their comments, which significantly improved the manuscript. In particular, we thank John Finger for implementing the outplant delay schedule into his Manila clam culture procedures to allow us to test our ideas about crab predation. We especially thank John and the staff of Hog Island Oyster Company for access to their properties, culture fa- cilities, and Manila clam stocks and to the director and staff of Bodega Marine Laboratory for ongoing assistance and support. LITERATURE CITED Carlton. J. T. l9'-)2. IntRiduced niarini.' and eMuannc niollusks of North America: ;m cnd-iil-lhc-2()th-ccntury perspective. J. Shcllfisli Rex. 1 1: 4^-505. Cigarria. J. & J. Fernandez. 1998. Manila clam {Riuliiapes pliilippiiuinim) culture in oyster bags: influence on survival, growth and liiomelric relationships. / Mar. Biol. Assoc. U.K. 78:551-561). Ciaarria. .1. & J. Fernandez. 2()()(). Manasjement of Manda clam beds f Influence of seed size, type of substralimi and protection on initial mortality. Aqiiuculture 182:173-182. Coan. E. V.. P. V. Scott & F. R. Bernard. 2000. Bivalve seashells of western North America: marine bivalve mollusks from Arctic Alaska to Baja California, 1st ed. Santa Barbara. CA: Santa Barbara Museum of Natural History. 746 pp. Cole, B. E., S. W. Hager & J. T. Hollihaugh. 1990. Hydrographic, hio- Manila Clam Predation by Introduced Green Crabs t)i') logical, and nulriom pruperties of Tomales Bay. California. March 1985 to May 1986. Depl. of the Interior. U..S. Geological Survey. Denver. CO. Cohen. A. N. & J. T. Carlton. 1995. Nonindigenous species in a United States estuary: A ca.se history of the ecological and economic effects of biological invasions in the San Francisco and Delta region. Report to LF. S. Fish and Wildlife Service. USA. Cohen. A.N. & J. T. Carlton. 1998. Accelerating invasion rate in a highly in\aded estuary. Science 279:555-558. Dumbauld. B. R.. D. A. Arni.strong & T. L. McDonald. 1993. Use of oyster shell to enhance intertidal habitat and mitigate loss of Dungeness crab {Cancer mii)>i.srcr) caused by dredging. Can. J. Fish. Ai/iiat. .Sci. 50: 381-390. Elner. R.W. 1978. The mechanics of predation by the shore crab Cuniniis muena.s on the edible mussel Mylilus edulis. Oecologia 36:333-344, Elner. R.W. 1980. The influence of temperature, sex. and chela size in the foraging strategy of the shore crab. Carcinus nuienus (L.). Mar. Beliav. PIn.siol. 7:15-24. Elner. R.W. 1981. Diet of the green crab Carcinus nuicna.s (L.) from Port Hebert. southwestern Nova Scotia. J. Slicllfi.sli Re.s. 1:189-194. Elner, R.W. & R. N. Hughes. 1978. Energy maximization shore crab in the diet of the shorecrab Carcinus maenas J. Anim. Ecol. 47:103-1 16. Elner, R.W. & D. H. Raffaelli. 1980, Interactions between two marine snails. LiiKtrina ruJis Maton and Litlorina nigrolineata Gray, a preda- tor. Carcinus maenas (L.) and a parasite, MicntpluiUus similis Jager- skiold. J. Exp. Mar. Biol. Ecol. 43:151-160. Glude. J. B. 1955. The effects of temperature and predators on the abun- dance of the soft-shell clam. Mya arenaria in New England. Trans. Am. Fish. Soc. 84:12-26. Griffiths C. L., P. A. R. Hockey. C. V. Schurink & P. J. Leroux. 1992. Marine aliens on South African shores: implications for community structure and trophic functioning. S. Afr. J. Mar. Sci. 12:713-722. Grosholz. E. D. & G. M. Ruiz. 1995. The spread and potential impact of the recently introduced European green crab. Carcinus maenas. in cen- tral California. Mar. Biol. 122:239-247. Grosholz, E. D., G. M. Ruiz. C. A. Dean. K. A. Shirley. J. L. Maron & P. G. Connors. 2000. The impacts of a nonindigenous marine predator in a California bay. Ecologx 81:1206-1224. Iribarne. O.. M. Fernandez & D. A. Armstrong. 1994. Does space compe- tition regulate density for juvenile Dungeness crab Cancer magisrer Dana in sheltered habitats. / E.\p. Mar. Biol. Ecol. 183:259-271. Jamieson. G. S.. D. A. Armstrong, E. D. Grosholz & R. W. Elner. 1998. Potential ecological implications from the introduction of the European green crab. Carcinus maenas. to British Columbia, Canada, and Wash- ington, USA. ./. Nal. His!. 32:1587-1598. Jensen. K. T. & J. N. .Jensen. I9S5. The importance of some epibenthic predators on the density of juvenile benthic macrofauna in the Danish Wadden Sea. J. £v/>. Mar. Biol. Ecol. 89:157-174. Laffeny. K. D. & A. M. Kuris. 1996. Biological control of marine pests. Ecology 77:1989-2000. Laing, L 1993, The response of Manila clam, Tapc\ philippinarum, juve- niles to nutritive stress. J. E.\p. Mar. Biol. Ecol. 173:1 1 1-121. Leroux. P. J., G. M. Branch & M. A. P. Joska. 1990. On the distribution diet and possible impact of the invasive European shore crab Carcinus maenas (L.) along the South African coast. S. Afr. J. Mar. Sci. 9:85-93. MacPhail, J. S.. E. L Lord & L. M. Dickie. 1955, The green crab— a new clam enemy. Prog. Rep. Allanl. Cst. Sins. 63:3-12. Pacific States Marine Fisheries Commission. 1999. Annual Fisheries Re- port Website. http://www.psmfc.org/ann_rpt/inde\.html. Paillard. C. & P. Maes. 1994. Brown ring disease in the Ruditapes: estab- lishment of a classification system. Dis. Aqua!. Organ. 19:137-146. Raffaelli, D., A. Conacher, H. McLachlan & C. Ernes. 1989. The role of epibenthic crustacean predators in an estuarine food web. Es!uar. Coas!. Shelf Sci. 28:149-160. Reise. K. 1977. Predator exclusion experiments in an intertidal mud Hat. Hetgol. Wissenschaft. Mceresuniers 30:263-27 1 . Ropes, J. W. 1968. Feeding habits of the green crab. Carcinus maenus (L.). Fish. Bull. 67:183-203. Ruiz. G. M.. J. T. Cariton, E. D. Grosholz & A. H. Hines. 1997. Global invasion of marine and estuarine habitats by non-indigenous species: mechanisms, extent, and consequences. .4m. Zoot. 37:619-630. Sanchez-Salazar, M. E.. C. L. Griffiths & R. Seed. 1987. The interactive roles of predation and tidal elevation in structuring populations of the edible cockle. Cerasloclerma edule. Esluar. Coas!. Shelf Sci. 25:245- 260. Scherer, B. & K. Reise. 1981. Significant predation on micro- and mac- robenthos by the crab Carcinus maeiuis L. in the Wadden Sea. Kiel. Meeresforsch. Sonderh. 5:490-500. Smith. M. D. & C. J. Langdon. 1998. Manila clam aquaculture on shrimp- infested mudflats. J. Shellfish Res. 17:223-229. Spencer. B. E., D. B. Edwards & P. F. Millican. 1992. Protecting Manila clam {Tapes philippinarum) beds with plastic netting. .Aquaculture 105: 251-268. Thresher. R. 1997. Proceedings of the first international workshop on the demography, impacts and inanagement of introduced populations of the European crab. Carcinus maenas. CR.LM.P. Technical Report Num- ber 1 1. CSIRO. Hobart. Tasmania. Australia, Yamada. S. A global invader: the European green crab. Oregon Sea Grant Publications. Oregon State University. Corvalis. OR. in press. .loiinial „t Shellfish Rcst-anh. Vol. 20, No. 3. 92 1 -924. 2001. RECOVERY OF THE WESTERN ROCK LOBSTER, PANULIRUS CYGNUS, FROM EMERSION AND HANDLING STRESS: THE EFFECT OF OXYGEN CONCENTRATION DURING RE-IMMERSION B. J. CREAR' - * AND G. N. R. FORTEATH' 'Tasnuiiiliin Acjiuiciilliire unci Fislierics Institute. School of AquaciiUure. University of Tasmania, PO Box 1214, Launceston. Tas. 7250, Australia: 'Tasinanian Aquaculture and Fisheries Institute, Marine Researcli Lahoratoiy, University of Tasiuauia. GPO Box I92B Hoharl. Tas. 7001, Australia. ABSTHiCT The effect of dissolved o.\ygen concentrations 1 1 10-120. 90-100. 70-80. 50-60, 30-40 and 10-20% saturation) on recovery of Panidiriis cygnus from a period of activity/emersion was investigated. Biochemical disturbances induced in P. cygniis by emersion and handling included a large decrease in the hemolymph pH (0.7 units) and an increase in the hemolymph lactate concentration (2.44 mmol L~'). Re-immersion resulted in very high rates of oxygen consumption (up to 2.5 times the pre-emersion rate), elevated hemolymph glucose (1.1^.5 mmol L"') and further increases in hemolymph lactate (2.7-1 1.1 mmol L"'). with actual concentrations dependent on the dissolved oxygen concentration. Water oxygen concentrations of 50-60% saturation or less slowed the rate of recovery; all lobsters recovering in water with 10-20% oxygen saturation died within 12 hours. Based on the rate of recovery of physiological parameters, lobsters should be allowed to recover for a minimum of eight hours between stress episodes, whilst maintaining an oxygen concentration of 70-80% saturation or above. KEY WORDS: activity, live holding, emersion, oxygen consumption, oxygen concentration, Pciiuilinis cygiiu.':. recovery, stress INTRODUCTION During the live handling of western rock lobsters (Panulinis cygnus) after capture, they are subjected to many conciitions that are Ukely to cause stress. They are often exposed to air during which time they may be handled. Their response is to attempt escape by strong beats of the tail (tail flicking). This resultant stress can be expressed quantitatively by biochemical changes (blood pH, glucose, lactate) (Spanoghe 1997) and by the develop- ment of an oxygen debt (Crear & Forleath 2001). It is presumed that when lobsters are re-immersed, they are able to recover from any physiological disturbances caused by air exposure (Taylor et al. 1997). A variety of processes take place during re-immersion — energy pools (phosphagen and ATP) are recharged, anaerobic end products are cleared from the tissues and pH disturbances are corrected. An organism-level manifestation of all of the above, essentially cellular processes of recovery, is a period of supranor- mal oxygen consumption (the oxygen debt) (Ellington 198.3). Gen- erally, physiological disturbances are fully reversed within 24 hours of re-immersion. An adequate period of recovery in water is necessary after each stage of handling or live transport to avoid compounding the stress syndrome (Taylor et al. 1997). European lobsters, Homarus gum- marus. deprived of the opportunity to recover after an episode of stress were in a significantly worse state after a subsequent period of air travel than lobsters that had been allowed to recover (White- ley & Taylor 1992). Since time between episodes of stress is often limited during post-capture handling of P. cygnus. optimising the speed of recovery becomes important. Inadequately designed holding systems, overstocking and natu- ral variability in the oxygen concentration of intake water can lead to low oxygen concentrations in live holding tanks. Concentrations as low as 35'^ saturation have been measured in tanks on board boats (Crear & Forteath 1997). As stressed lobsters have a high demand for oxygen, its availability is likely to influence both the duration and extent of recovery. Minimum dissolved oxygen con- *E-mail: hradley.crearCs^utas. edu.au centrations recommended for lobster holding systems vary from 40 to 80<7f saturation (Beard & McGregor 1991, Forteath et al. 1993, Boothroyd 1994), however there appears to be no biological ra- tionale for those recommendations. Additionally, oxygen super- saturation has been suggested as a possible tool to aid the recovery of lobsters (Forteath 1993). This paper examines the effect of dissolved oxygen concentration on the recovery of P. cygnus from a stress episode. MATERIALS AND METHODS Lobster holding conditions and respirometry are outlined in Crear and Forteath (2000). All experiments were conducted at the acclimation temperature of 23°C, which is close to the mean an- nual temperature of inshore waters inhabited by juvenile P. cygnus (Chittleborough 1975). In view of the relationship between oxygen consumption (VO, - mgO^/g/h) and P. cygnus wet body weight (Crear & Forteath 2001). a restricted weight range (367-515 g) was used in the experiments. The following specific methods were used for this series of experiments. The study was performed in two experimental series. The first series studied the oxygen consumption of lobsters recovering from a period of stress. Lobsters were removed from the holding tank and emersed for 30 minutes at an air temperature of 23°C. Con- tinual disturbance (handling) for the first 3 minutes was followed by disturbance every 5 minutes (lobsters were maintained in an open foam box when not being handled). Lobsters showed strong escape behaviors (tail flicking) during the initial period of distur- bance. The response diminished as the emersion time increased and after 30 minutes emersion the lobsters were normally unre- sponsive to disturbance. Six to twelve lobsters were tested at each of six oxygen concentrations (1 10-120%. 90-100%, 70-80%. 50- 60%. 30^0%^. 10-20%), The dissolved oxygen concentration was controlled as in Crear and Forteath (2000). As P. cygnus has a nocturnal oxygen consumption rhythm (Crear & Forteath 2001). all experiments were commenced before 0900 hours to ensure that none of the measurement periods fell during the night. Rates of V(,, were calculated immediately after placing the disturbed lob- sters in the respirometers (Oh), and after re-immersion for I, 2, 4, 922 Crear and Forteath 6. 8 and 24 hours. Ten lobsters were used to determine standard VO, (eontrol oxygen consumption rate). Tlie second series of experiments consisted of measuring the hemolyniph parameters of lobsters over the same time-period and under the same dissolved oxygen concentrations as above. Lob- sters were removed from the holding tank and disturbed for 30 minutes before being placed in water of known oxygen concen- tration. Hemolymph samples were obtained immediately after the disturbance period (0 h). and after re-immersion for 1, 2, 4, 8 and 24 hours. Dissolved oxygen was maintained within 5% of the designated concentration (Crear & Forteath 2000). Lobsters were sampled only once during each experimental run. They were re- placed in the holding tank for a muiimum of 48 hours before being used again (Waldron 1991). Control hemolymph parameters were determined for 12 lobsters: four taken directly from the holding tank at 0900 hours, and two further groups of four tested 8 and 24 hours later. Hemolymph Sampling Prebranchial hemolymph was sampled (1 niL) from the in- frabranchial sinus via the arthrodial membrane at the base of a walking leg (usually the 3rd or 4th pair). The sample was with- drawn with an ice-chilled 1 mL syringe (Luer -Tuberculin) using a 21 gauge (Luer -21G*1.5) needle. Hemolymph samples were obtained anaerobically to ensure minimum mixing with air, since changes in CO, equilibrium can alter pH values (Vermeer 1987). The hemolymph sample was taken within 20 seconds of lobster capture, and immediately placed into an ice-chilled 1-mL Eppen- dorf tube. A 250-|jiL aliquot was deproteinated by adding it to 500 |xL of ice-chilled 0.6 mol L'' perchloric acid (PCA). The deproteinated sample was snap-frozen in liquid nitrogen and stored at -86°C for later measurement of lactate and glucose. The hemolymph remaining in the original tube was used for mea- suring pH. Henmlyinph Analyses Hemolymph pH was measured with a calomel electrode (Ac- tivon Semi-Micro AEP336) connected to a pH meter (WTW pH 323). The probe was calibrated in buffer solutions chilled to the same temperature as the hemolymph samples (0°C). pH at 0°C varies from pH //( vivo at ambient temperatures (0.49 units higher than at 23"C. Crear 1498). but this was an essential concession to retard clot formation (Vermeer 1987): clots formed within .S min- utes at ambient temperature. The deproteinated hemolymph samples were centrifuged at 8000 g for 3 min. The supernatant (600 (jlL was generally ob- tained) was neutralized with 3 mol L^' KOH (6.4 p.L per 600 (xL). The samples were stored on ice for 15 minutes before being cen- trifuged at 8000 g for 3 min so that the perchlorate precipitate could be removed. The supernatant (-350 (jlL) was either frozen (-86°C) for later analysis or analyzed immediately for lactate and glucose. Lactate concentrations were determined enzymatically with the Boehringer-Mannhein analysis kit (Cat. No. 139084). The absorp- tion was measured at 340 nm on a GBC UV/VIS 916 spectropho- tometer. Glucose concentrations were deteimined with a Sigma glucose test kit (No. 510), which is based on the glucose-oxidase method. The absorption was measured at 450 nm on a GBC UV/ VIS 916 spectrophotometer. New calibration curves were made up for each sainple run and all assays were run in duplicate. Statistical Analyses Student's /-test was used to test for significant differences (P<0.05) between control and recovery values. Where appropriate. Student's r-test for samples with unequal variances was used. The one-way ANOVA was used to test for differences between treat- ments at each time-period. The Levene test was used to test for homogeneity of variance, and where necessary an appropriate transformation was pertbrmed before further analysis. Following ANOVA. means were compared by Tukey-HSD test. Correlation analysis was used to analyse the relationship between dissolved oxygen concentrations and lactate changes. The critical oxygen tension (P^.) was calculated (Cochran & Burnett 1996. Crear & Forteath 2000) from the oxygen consumption data immediately after re-immersion. All analyses were performed on the SPSS statistical package with a set at 0.05. All values are expressed as mean ± S.E. RESULTS There was no significant difference in the parameters measured for the control lobsters at 0. 8 and 24 hours, therefore the data were pooled. Survival was 100% in all treatments except for the 10- 20'7f oxygen saturation, where no animals survived for more than 12 hours. Oxygen consumption In all treatments, apart from the 10-20% treatment. VO, was significantly higher after re-immersion than the control VO,. From those high values. VO, reduced slowly over time (Fig. 1 ). with the pattern of recovery varying with the oxygen concentration. At high concentrations (90-100% and 1 10-120% oxygen saturation). VO, was not significantly different to the control VO, after eight hours re-immersion. In all other treatments recovery of VO, to control levels took longer, and in the 30-40% treatment. VO, remained significantly higher than the control VO, over the 24-hour re- immersion period. Lobsters in the 10-20% oxygen saturation treat- ment had significantly lower VO, than the control VO, at each measurement period. Upon re-immersion. VO, rates varied with the oxygen concen- tration, being highest for the 110-120% and 90-100% oxygen saturation treatments (Table 1 ). At lower oxygen saturations, VO, decreased significantly with the dissolved oxygen concentration. From this data. P^ was calculated to be 63.1%. The amount of excess oxygen consumed over the control rate during recovery also varied with oxygen concentration (Table 2). Lobsters recovered in water containing 70-80% oxygen saturation or higher consumed the least amount of oxygen during the recov- ery period. In comparison, lobsters in the 50-60% treatment con- sumed 1.3 times as much oxygen during the initial eight-hour recovery period and approximately twice as much oxygen in achieving full recovery (see Note b in Table 2). Lobsters in the 30-40% treatment consumed 0.75 times as much oxygen during the initial eight-hour recovery period, but had consumed 1.6 times as much oxygen after 24 hours although they had still not achieved full recovery. Hemocymph parameters Lobster hemolymph pH decreased significantly during the 30- minute disturbance period, from control concentrations of 8.36±0.01 to 7.66±0.03. In general, the pH returned to, or close to. the control concentrations after four hours of being re-immersed. However, during the first hour of re-immersion two distinct re- covery patterns were noted: (a) at oxygen concentrations of 70- Emersion/Re-emersion Effects on P. cygnus 923 0.16 0.14 0.12 0.10 0.08 0.06 0.04 c 0.02 o ■4-> a. ^ 0.16 F JC 0.14 3 T- 0-12 (fl r U) 0-10 o > X O 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 Time (hours) FijJuru. I. Oxygen consumption (mg O, g"' h ' )(nu'aii ± SKl of tht western rock lob.ster, Panulinis cygnus. over a 24-h period during "recovery" from disturbance (n = 6 to n = 12). The lobsters were disturbed lemersed and handled) for .3(1 minutes prior to re-immersion in water containing different concentrations of dissolved oxygen. Dotted line: control oxygen consumption rate (n = 10).*, not significantly different to the control. The lines are drawn for ease of viewing. SO^f and higher the pH remained at the low level measured after the emersion period; whilst (b) at lower o.xygen concentrations the pH increased markedly (Fig. 2. Table 3). In some treatments (70-80. 50-60 and 3()^0'.h the pH in- creased to a significantly higher level than that of the control during the recovery period, however in all those treatments the pH was not significantly different to the control after 24 hours. The pH of lobsters in the 1 10-120% treatment was significantly lower than the control, and all other treatments, after 24 hours. Hemolymph lactate increased significantly during the 30- TABLE 1. The results of the ANOVAs comparing oxygen consumption of lobsters {Panulinis cygnus) in each oxygen concentration treatment at each measurement time during the 24 h recovery period. DitTercnt letters denote significantly different results. Oxygen Concentration ( '7r ) Recoverv Time (h) 110-120 90-100 70-80 50-60 30-iO 10-20 0 1 2 4 6 8 24 a a a ab b be h a a a ab b c b b a a ab b he b c a a a a a a d b b b ab ab a e c c c c d N/,.\ minute disturbance period, from a resting concentration of 0.05 ± 0.02 mmol L~' to 2.44 ± 0.37 mmol L~' (Fig. 3). Lactate increased further during the first hour of re-immersion with the largest in- creases occurring at the lower oxygen concentrations (Table 4; Fig. TABLE 2. The amount of excess oxygen consumed (mg Oj g~') over the control rate during recovery lobsters (Panulinis cygnus) from emersion and handling. Oxygen consumed during 8-h and 24-h of recovery are shown. Dissol ved Oxygen Concentration ( ^ ) Time 10-20 30-40 50-60 70-80 90-100 110-120 8 h 24 h -0.163 N/A 0.184 0.425-' 0.334 0.593'' 0.251' N/A 0.253 N/A 0.266 N/A ^''^ - the Uibsters in this treatment did not survive for 24 h. N/.A not applicable to this time period at that particular dissolved oxygen concentration -" Oxygen consumption was still significantly higher than the control after 24 h so the total oxygen consumed during recovery would be slightly higher than this value. '' Oxygen consumption may have returned to the control prior to the 24 h period so this value may be an overestimation. "■ Although the oxygen consumption was still significantly higher than the control after 8 h it was not significantly different to either the 1 10-120% or 90-100% rate. Therefore, the total amount of oxygen consumed was only calculated up to the 8 h mark. 924 Crear and Forteath X Q. Q. E o E X Time (hours) Figure. 2. Hemoljmph pH (mean ± SE) of the western rock lobster, Panulirus cygnus. over a 24-h period during "recovery" from disturbance (n = 6 to n = 12). The lobsters were disturbed (cniersed and handled) for 30 minutes prior to re-immersion in water containing different concentrations of dissolved oxygen. Dotted line: control heniul^niph pH value (n = 12).*, not significantly different to the control (C). The lines are drawn for ease of viewing. 3). The increase was correlated (P = 0.005) with oxygen concen- tration (Fig. 4). From those high values, lactate reduced slowly over time (Fig. 3), with the pattern of recovery varying with the oxygen concen- tration. At high oxygen concentrations (70-80%. 90-100% and 1 10-120%' saturation) lactate was not significantly different to the TABLE 3. The results of the ANOVAs comparing the hemolvmph pH of lobsters {Panulirus cygnus) in each oxygen concentration treatment at each measurement time during the 24-h recovery, period. Different letters denote significantly different results. Oxygen Concentration ( % ) Recovery Time (h) 0 1 2 4 8 24 110-120 a he ahe a e b 90-100 11 c be a c a 70-80 a he e b a a 50-60 a a ab a b a 30-40 a b ab a b a 10-20 a a a b b N/A the lobsters in this trealnient did mil survive for 24 h. control after eight hours re-immersion. In the other treatments, lactate remained high for longer, and in the 30—10% treatment lactate was still significantly higher than the controls after 24 hours re-immersion. The lactate concentration of lobsters in the 10-20% treatment remained very high. Hemolymph glucose increased during the 30-minute distur- bance period from 0.35 ± 0.06 nimol L"' to 0.44 ± 0.06 mmol L"'. but it was not a significant increase (Fig. 5). After 1-hour re- immersion, the glucose concentration in all treatments was signifi- cantly higher than the control. The largest increases were measured in the low oxygen saturation treatments (Table 5). In general, the high glucose concentrations were maintained for between two and four hours, most were not significantly different to the controls after eight hours re-immersion. Lobsters in the 110-120% and 50-60% treatments still had significantly higher concentrations than the controls after eight hours re-immersion, however the con- centrations were not significantly different to those in all other treatments except for the 10-20% treatment. DISCUSSION Biochemical disturbances induced in P. cygnus by emersion and handling included a large decrease in the hemolymph pH and an increase in the hemolymph lactate concentration. Re-immersion Emersion/Re-em[-rsion Effects on P. cygnus 925 1 rrrn r C 012 4 8 24 5.0 - I 90-100% 4.0 - k 3.0 - R 2.0 ^ r\ 1.0 - 1 ^: * 0.0 - 1 1 1 1 1 1 1 0.0 - n — rm — r C 012 4 24 5.0 - T 50-60% 4.0 - \ 3.0 - i 2.0 - /\ 1.0 - / \ 0.0 - • • • 1 1 1 1 1 1 1 24 Time (hours) Figure. 3. Hemolymph lactate concentration (niniol I."' ((mean ± SEl of the western rock lobster, /"o/iu/irH,? cvg'/H.v, over a 24-hour period during "recovery" from disturbance (n = 6 to n = 12). The lobsters were disturbed lemersed and handled! for 30 minutes prior to re-immersion in water containing different concentrations of dissolved oxygen. Dotted line: control hemolymph lactate concentration (n = 12). *, not significantly different to the control (C). The lines are drawn for ease of viewing. was followed by very high rates of oxygen consuniptioii. elevated hemolymph glucose concentrations and further increases in the hemolymph lactate concentrations. The dissolved oxygen concen- tration during re-immersion influenced both the rate and the form of recovery from the physiological disturbances, with oxygen con- TABLE 4. The results of the .VNOVAs comparing the haemolymph lactate concentration Immol L ') of lobsters iPanulinis cygniis) in each oxygen saturation treatment at each measurement time during the 24-h recovery period. Different letters denote significantly different results. Oxygen Recovery Time (hi Concentration ( Vc ) 0 1 2 4 8 24 iii)-i:() a c h d c b 90-100 a c b d c b 70-80 a c b c c b 50-60 a be b cd c h 30-40 a ab a b b a 10-2(1 a a a a a N/A 0) 10.0 -1 Y = 8 - 0.07 X b-i 8.0 - 6.0 - • ^ (r^=0.87) fi> £ 4.0 - ^\ ^ Lactal (m 2.0 - • 0.0 1 1 1 1 1 1 N^ ^ A^ Dissolved oxygen (%) ^'^ - the lobsters in this Ireatnient did not survive for 24 h. Figure. 4. Increase in hemolymph lactate (mmol I. "')(•! during the first hour of recover) after re-immersion of lobsters, Panulirus cygnus. in water at different dissolved oxygen concentrati (A o u 3 O) £ a E _>. o E o> n X "rm — r C 012 4 90-100% 00 ~i — nrn — i r C 012 4 8 24 60 10-20% 24 Time (hours) Figure. 5. Heniolymph glucose concentrations (mmol L~') (mean ± SK) of the western rock lobster, Panulirus cygniis. over a 24-hour period during "recovery" from disturbance (n = 6 to n = 12). The lobsters were disturbed (emersed and handled! for 30 minutes prior to re-immersion in water containing different concentrations of dissolved oxygen. Dotted line: control heniolymph glucose concentration (n = 12). *, not signincantly different to the control (C). The lines are drawn for ease of viewing. centrations of 70-809^^ saturation or abo\e maximizing the rate of recovery. At those concentrations stress indicators returned to nor- mal levels within eight hours of re-immersion. Oxygen concentra- tions <60% saturation, which are commonly measured in holding tanks (Crear & Forteath 1997). slowed the rate of recovery. The aerobic response of P. cygiuis during recovery follows a typical Type V pattern. That is, the oxygen consumed during re- TABLE 5. The results of the .\NOVAs comparing the haemolymph glucose concentration (mmol L"') of lobsters [Panulirus cygnus) in each oxygen saturation treatment at each measurement time during the 24-hour recovery period. Different letters denote significantly different results. Oxygen Saturation ( % ) Recovery Time (h) 0 1 2 4 8 24 110-120 a cd bed b a h 90-100 a be abc b a c 70-80 a d d b a c 50-60 a cd cd b a a 30-40 a b a b ab b 10-20 a a ah a b N/A the lobsters in this trealinem did not sur\ive tor 24 li. covery exceeds the predicted aerobic oxygen deficit (Herreid II 1980). If it is assumed that the emersed lobsters can take up -50% of the oxygen that they are able to take up in water (Whiteley & Taylor 1990. Waldron 1991), and that they were fully active over the 30 minute period of disturbance, then the maximum oxygen deficit would be -0.06 mg O, g~'. In fact, this is less than one- quarter of the oxygen debt incurred at high concentrations dis- solved oxygen. Suggested uses for the excess oxygen include: (1) metabolising anaerobic end products; (2) re-establishing resting oxygen concentrations in body tissues; (3) replenishing high- energy phosphate reserves; and (4) meeting energy costs associ- ated with increased branchial chamber ventilation and hemolymph circulation (Herreid 1980. Head & Baldwin 1986). The recovery period at high concentrations of dissolved oxygen (70-80'7f and above) was similar to that observed previously for P. cyaniis (Crear & Forteath 2001) and for other crustaceans (Mc- Mahon et al. 1979. Waldron 1991 ). However, the recovery period increased as the oxygen concentration decreased, indicating that the lobsters either (a) were not accessing sufficient oxygen to repay the debt as quickly or (b) were increasing the size of the debt as a result of the processes involved in repaying the debt or (c) both. Thus: (a) The ability of lobsters to take up oxygen upon re- immersion was dependent on the dissolved oxygen concen- tration. The critical oxygen tension (P^) in this study (63.1%) was close to the 62.8% calculated by Crear and Emersion/Re-emersion Effects on P. cygnus 927 Fortealh (2001) for active P. cygnus at 23°C. The aerobic scope for activity (the amount of oxygen available to the lobsters above normal maintenance requirements) in- creased as oxygen concentrations increased up to P^.. Lob- sters with a large aerobic scope for activity should be able to shorten the time required to repay a similar oxygen debt. Lobsters in the 30-40% treatment were very limited in the amount of oxygen they were able to extract from the water (=50"* of maximum VO,). Therefore, total oxygen con- sumption over the first eight hours of re-immersion in that treatment was =25% lower than in the higher oxygen treat- ments. Access to oxygen was a major problem with these lobsters and would explain (at least partly) the slow recov- ery rates. Similarly, in response to an injection of lactate C. maeiias increased VO,. but the response was lesser, and lasted longer, under hypoxic conditions than under nor- moxic conditions (De Wachter et al. 1997). Those authors suggested that the response was due to the larger aerobic scope at the higher oxygen concentration, (b) The total amount of oxygen consumed in the 50-60% treat- ment during the first eight hours of re-immersion was -30% higher than in the treatments with higher oxygen concentrations. Although the oxygen consumed was high, they still had not repaid the oxygen debt, which indicates that there were increased energetic costs associated with recovery at that oxygen concentration. The increased en- ergetic costs could arise from: (i) the increased cost of branchial chamber ventilation and hemolymph circulation when diffusion of oxygen from the external medium to the hemolymph via an oxygen gradient would be minimal and/ or. (ii) an increased reliance on anaerobic respiration, to- gether with production of lactate and associated energetic costs of resynthesizing the substrate or oxidizing lactate further to carbon dioxide (Gade et al. 1986). The pH response of crustaceans to the combination of stressors used in this study (emersion, exercise and handling) has not often been investigated. Spanoghe (19971 recorded a similar large de- crease (0.7 pH units) in P. cygnus after one hour of emersion and handling. A pH change of that magnitude must be considered a large physiological perturbation (Vermeer 1987). Return to control pH values after re-immersion was optimized at oxygen concentra- tions of 90-100% and above. The time taken was similar to that of other crustaceans (McDonald et al. 1979. Waldron 1991. Whiteley & Taylor 1992). Schmitt and Uglow (1997) concluded that CO, accumulation was mainly responsible for emersion-induced acidosis in Nephmps nonegicus. despite the presence of high concentrations of lactate. Recovery of acid-base status during re-immersion in this study varied with the oxygen concentration, and also appears not to be determined by lactate concentrations. Lobsters at low oxygen con- centrations (50-60%. 30-40% and 10-20%) had large increases in lactate during the first hour of re-immersion, while their hemolymph pH showed significant increases. Taylor and Wheatly (1981) noted that the potential acidosis, which the increase in lactate represents, was overridden by a respiratory alkalosis due to the washout of CO, during the period of hyperventilation. In the present study, oxygen consumption upon re-immersion was rela- tively low at lower oxygen concentrations, so less CO, would have been produced, whilst high ventilation and perfusion activities would have promoted the excretion of CO^ across the gills. These two factors may have resulted in the large pH increase after one hour of re-immersion into poorly oxygenated water. At the higher dissolved oxygen concentrations, the hemolymph pH remained low after one hour of recovery, even though the hemolymph lactate was lowest in these lobsters. The high V(,, of lobsters held at the higher oxygen concentrations may result in CO, concentrations remaining elevated in the hemolymph, thus helping to maintain a low pH during the initial stages of recovery. Elimination of accu- mulated CO, is usually rapid, however, Waldron ( 1991 ) found that hemolymph CO, partial pressure remained significantly elevated for two hours during re-immersion after a period of emersion and exercise in / edwardsii. A hemolymph alkalosis was measured during the recovery pe- riod ni lobsters subjected to 70-80%' oxygen saturation or lower. Crustaceans generally hyperventilate in response to hypoxia, lead- ing to hypocapnic alkalosis due to an increase in the rate of ex- cretion of CO, (Hagerman & Uglow 1985). The pH of A. lepto- ductyhis increased by 0.16 units with exposure to 30% oxygen saturation (Sinha and Dejours 1980). In other studies in which the pH of crustaceans undergoing recovery has risen (Truchot 1975; Whiteley and Taylor 1992; Spanoghe 1997). the oxygen concen- trations in the recovery tanks may have been lower than optimal. Resting concentrations of hemolymph lactate (0.05 mmol L"') were similar to those measured in other decapod crustaceans (0.14 mmol L"' for J. edwardsii. Waldron 1991; 0.14 mmol L~' for Rcininii ninuui, Paterson et al. 1994; 0.09 mmol L"' for Carcinus maenas. De Wachter et al. 1997). There have been few reports on the hemolymph lactate concentrations of crustaceans that have been eniersed and exercised. However, in / edwardsii after a short period of exercise followed by one hour of emersion lactate in- creased by -1.0 mmol L"' (Waldron 1991) and in P. cygnus lactate increased by 2 mmol L~' after 40 minutes of emersion and dis- turbance (Spanoghe 1997); increases which were similar to this study. The slow rate of lactate removal from the hemolymph in- dicates that P. cygnus. like many other crustaceans, lack the means for rapid removal of lactate (McDonald et al. 1979, Ellington 1983, Waldron 1 99 1 ). Return to normal concentrations was optimized at oxygen concentrations of 50-60% and above. The increase in hemolymph lactate concentration of P. cygnus indicates that it was unable to maintain an adequate supply of oxygen to the tissues during the period of disturbance and needed to rely, at least partially, on anaerobic metabolism to supply its energy requirements (Spicer et al. 1990). The rise in lactate con- centration after re-immersion has also been noted in other crusta- ceans subjected to periods of exercise and/or emersion (McDonald et al. 1979, Taylor & Wheatly 1981, Whiteley and Taylor 1992). Increased hemolymph lactate concentrations after re-immersion may be due to the release of lactate previously stored in the tissues during the disturbance period, as suggested by Taylor and Wheatly (1981) and Waldron (1991). In C. destructor it appears that a steady state between tail muscle and hemolymph lactate pools is reached quite rapidly (Head & Baldwin 1986). hence lactate re- lease may not fully explain the increased concentrations. Another possible explanation is that lactate production may have increased on re-immersion due to a high energy demand requiring a contri- bution from both aerobic and anaerobic metabolism (Head & Bald- win 1986. Gruschczyk & Kamp 1990. Whiteley & Taylor 1992). Onnen and Zebe (1983) suggested that the use of anaerobic me- tabolism during the recovery process might ensure that the muscle function is restored as soon as possible. In this study, the relative increase in the lactate concentration during the first hour was de- pendent on the dissolved oxygen concentration in the water. When oxygen could not fully fuel the aerobic portion of the energy requirements of recovery, the shortfall was made up via anaerobic 928 Crear and Forteath metabolism. This suggests that the observed increase in hemolymph lactate is probably due to the continued use of anaero- bic energy sources after re-immersion, rather than the release of sequestered lactate. During recovery in the 110-120% treatment, metabolism appears to be mainly aerobic as shown by the absence of further accumulation of lactate. Hemolymph glucose concentrations of control lobsters are similar to those measured in other studies; 0.2-0.3 mmol L"' for Nephrops norvegicus (Spicer et al. 1990. Schmitt & Uglow 1997); 0.2-0.4 mmol L"' for P. cygniis (Tod & Spanoghe 1997). The maximum concentrations of hemolymph glucose measured in this study (1.0 to 4.5 mmol L"') also covered the range of maximum concentrations measured by those researchers. As has been noted in other studies (Onnen & Zebe 1983, Gruschczyk & Kamp 1990. Tod & Spanoghe 1997). there was a marked hyperglycemia in the hemolymph of P. cygnus one hour after re-immersion. In this study, the hyperglycaemia was more severe in lobsters subjected to low dissolved oxygen concentrations, suggesting that more energy substrate was required because the lobsters were relying more on anaerobic metabolism. When aerobic mechanisms of energy pro- duction are impaired, in order to provide a given amount of energy. more glucose must undergo anaerobic glycolysis (Storey & Storey 1990). as anaerobic glycolysis produces only about l/20th of the energy produced via aerobic glycolysis (Eckert et al. 1988). Dur- ing recovery in this study, anaerobic glycolysis (as indicated by lactate concentration), increases as oxygen saturation decreases. Therefore, the increases in hemolymph glucose concentrations would be expected. The time period of recovery from hyperglycaemia has not been well studied, but it was similar to that for P. cyg)ius in Spanoghe's ( 1 997 ) study, and to that recorded for Palaemon senatus and P. elegaiis after a period of emersion (Taylor & Spicer 1987). Speed of recovery to normal glucose concentrations was optimized at oxygen concentrations of 70-80% and above. In conclusion, oxygen has a considerable effect on the recovery response of P. cygnus. The duration and the effectiveness of the recovery process are of great functional importance. Recovery from anaerobic metabolism should be sufficiently rapid and com- plete for the organism to cope with further periods of stress. For example, in the case of muscles powering escape responses, this process of recovery must be sufficient to allow the organism to evade predators (Ellington 1983). Using speed of recovery as the criteria for evaluating the effectiveness of oxygen concentrations, the results show that a minimum of 70-80% saturation is required, with some indication that higher concentrations (including slight supersaturalion) may offer some benefits. ACKNOWLEDGMENTS This research was supported by the Fisheries Research and Development Corporation (Project 94/134.03) and an Australian Post-Graduate Scholarship awarded to BC. REFERENCES Beard. T. W. & D. McGregor. 1991. Storage and care of live lobsters. Laboratory Leaflet No. 66. Ministry of Agriculture, Fisheries and Food. Lowestoft. Boothroyd. F. A. 1994. Handling and maintaining live Canadian Atlantic lobsters. Infofish Int. 6:27-32. Chittleborough. R. G. 1975. Environmental factors affecting growth and survival of juvenile western rock lobsters Pamdirus longipes (Milne- Edwards). Ausl. J. Mar. Freslm: Res. 26:177-196. Cochran. R. E. & L. E. Burnett. 1996. Respiratory responses of the salt marsh animals Fundulu.s helerocUlus. Leioslomiis xanthuriis. and Palaemonetes pugio to environmental hypo.xia and hypercapnia and to the organophosphate pesticide, azinphosmethyl. J. E.xp. Mar. Biol. Ecol. 195:125-144. Crear. B. & N. Forteath. 1997. Maintaining rock lobsters in prime condi- tion on board lobster boats - Part I . Prof Fisherman. March. 2-1-26. Crear, B. J. 1998. A physiological investigation into methods of improving the post-capture survival of both the southern rock lobster. Jasiis ed- wardsii. and the western rock lobster, Panulirus cygnus. Ph.D. disser- tation. Australia: University of Tasmania. Crear, B. J. & G. N. R. Forteath. 2000. The effect of extrinsic and intrinsic factors on oxygen consumption by the southern rock lobster, Jasus edwardsii. J. Exp. Mar. Biol. Ecol. 252:129-147. Crear, B. J, & G. N, R. Forteath. 2001. Determining flow rate requirements for captive western rock lobsters (Panulirus cygnus\: effects of body weight, temperature, activity, emersion, daily rhythm, feeding and oxy- gen tension on oxygen consumption. Mar Freshw. Res. (in press), De Wachter, B., F. J. Sartoris & N. O. Portner. 1997. The anaerobic end-product lactate has a behavioural and metabolic signalling function in the shore crab Carcinus maenas. J. Exp. Biol. 200:1015-1024. Eckert, R., D. Randall & G. Augustine, 1988, Animal Physiology — mechanisms and adaptations. New York: W. H. Freeman and Com- pany. Ellington, W. R. 1983. The recovery from anaerobic metabolism in inver- tebrates. J. Exp. Biol. 228:43 1-+44. Forteath, N., P. Hart & O. Bunlcr. I99.V Approaches to recirculating sys- tems - crayfish and marine fish larvae. In: P. Hart & D. O'Sullivan, editors. Recirculation systems: Design, construction and management. Tasmania: Aquaculture Sourcebook. Forteath, N. 1995. Research at the Key Centre. .Ausuisia Aquacull. 9:49- Gade, G., R. A. Graham & W. R. Ellington. 1986. Metabolic disposition of lactate in the horseshoe crab Limidus polyphemus and the stone crab Menippe mercenaria. Mar. Biol. 91:473—479. Gruschczyk, B. & G. Kamp. 1990. The shift from glycogenolysis to gly- cogen resynthesis after escape swimming: studies on the abdominal muscle of the shrimp. Crangon crangon. J. Comp. Physiol. 159:753- 760. Hagerman, L., & R. F. Uglow. 1985. Effects of hypoxia on the respiratory and circulatory regulation of Nephrops nonvgicus. Mar. Biol. %l:2Ti- 278. Head, G„ & J. Baldwin. 1986. Energy metabolism and the late of lactate during recovery from exercise in the Australian freshwater crayfish Cherax destructor. Ausl. J. Mar. Freshw. Res. 37:641-646. Herreid 11.. C. F. 1980. Hypoxia in invertebrates. Comp. Biochem. Physiol. 67A:31 1-320 McDonald. D. G.. B. R. McMahon & C. M. Wood. 1979. An analysis of acid-base disturbances in the hemolymph following strenuous activity in the Dungeness crab. Cancer magister. J. Exp. Biol. 79:47-58. McMahon, B, R„ D, G, McDonald, & C. M, Wood. 1979. Ventilation, oxygen uptake and hemolymph oxygen transport, following enforced exhausting activity in the Dungeness crab Cancer magister. J. E.xp. Biol. 80:271-285. Onnen, T. & E. Zebe. 1983. Energy metabolism in the tail muscles of the shrimp Crangon crangon during work and subsequent recovery. Comp. Biochem. Physiol. 74A:833-838. Paterson, B. D., P. S. Exiey, & R. A. Smith. 1994. A lactic acid "ceiling" arrests profound acidosis in the blood of an ion-conforming crab, Raninu ranina. stored out of water at 25^. In: B. D. Paterson, P. S. Exley & R. Smith, editors. Live transport of crustaceans in air — Emersion/Re-emersion Effects on P. cygnus 929 prolonging the survival of crabs. Fisheries Research and Development Corporation Final Report of Project 92/7 1 . Schmitt. A. S. C. & R. F. Uglow. 1997. Hemolymph constituent concen- trations and ammonia efflux rates of Nepluops non'egicus during em- ersion. Mar. Biol. 127:403-410. Sinha. N. P. & P. Dejours. 1980. Ventilation and blood acid-base balance of the crayfish a,s functions of water oxygenation (4()-l.'il)() Torrl. Cump. Biochem. Physiol. 65A:427^32. Spanoghe, P. 1997. An investigation of the physiological and biochemical respon.ses elicited by Pamitinis cygnus to harvesting, holding and live transport. Ph.D. dissertation. Australia: Curtin University of Technol- ogy, Spicer, J. I., A, D. Hill. A. C. Taylor & R. H. C Strang. 1990. Effect of aerial exposure on concentrations of selected metabolite in blood of the Norwegian lobster Nephrops non-egiciis (Crustacea: Nephropidae). Men: Biol. 105:129-135. Storey, K. B. & J. M. Storey. 1990. Metabolic rate depression and bio- chemical adaptation in anaerobiosis. hibernation and estivation. Qiiuri. Rev. Biol. 65:145-174. Taylor, A. C. & J. I. Spicer. 1987, Metabolic respon.ses of the prawns Palaemon elegans and P. serratus (Crustacea: Decapodal to acute hyp- oxia and anoxia. Mar. Biol. 95:521-530. Taylor, E. W. & M. G, Wheatly. 1981. The effect of long-term aerial exposure on heart rate, ventilation, respiratory gas exchange and acid- base status in the craytlsh .Aiistropoiamtthiiis pallipes. J. E.xp. Biol. 92:109-124. Taylor, E. W. & N, M. Whiteley. 1989. Oxygen transport and acid-base balance in the hemolymph of the lobster, Hoinanis gammanis, during aerial exposure and resubmersion. J. Exp. Biol. 144:417—436. Taylor, H. H., B. D. Paterson, R. J. Wong & R. M. G. Wells. 1997. Physiology and live transport of lobsters: a report Irom a workshop. Mar. Freslnv. Res. 48:817-822. Tod, P. & P. T. Spanoghe. 1997. Development of improved onshore stor- age and transportation protocols for the western rock lobster, Panulirus cygnus. Canberra. Australia: Fisheries Research and Development Cor- poration— Report 94/134.06, Tiuchot, J. P. 1975. Blood acid-base changes during experimental emer- sion and reimmersion of the intertidal crab, Carcinus nuienas (L.). Re.sp. Physiol. 23:351-360. Vermeer. G. K. 1987. Effects of air exposure on desiccation rate, hemolymph chemistry, and escape behavior of the spiny lobster, Panu- lirus argus. Fish. Bull. 85:45-51. Waldron, F. M. 1991. Respiratory and acid-base physiology of the New Zealand rock lobster, Jasus edwardsii (Hutton). Ph.D. dissertation. University of Canterbury, New Zealand. Whiteley, N. M. & E. W. Taylor. 1990. The acid-base consequences of aerial exposure in the lobster, Homarus gammarus (L.) at 10 and 20°C, J. Thermal Biol. 15:47-56. Whiteley, N. M. & E. W. Taylor. 1992. Oxygen and acid-base disturbances in the hemolymph of the lobster Homarus gammarus during commer- cial transport and storage. J. Crusi. Biol. 12:19-30. JoKi-ncil of Slu-lljhh Research. Vol. 20, No. 3, 931-437, 2001, EFFECT OF DIETARY CARBOHYDRATES ON GLUCONEOGENESIS IN PREMOLT LITOPENAEUS STYUROSTRIS JUVENILES AND PRE ADULTS GERARD CUZON,' CARLOS ROSAS," * GABRIELA GAXIOLA," GABRIEL TABOADA," AND ALAIN VAN WORMHOUDT' 'cop. BP 7004, Tanivuo. Tahiti. French Polynesia: -Grupo cle Bioh>gia Marina Experimental Fac. de Ciencias UNAM. Apdo. Post. 69. IFREMER Cd. del Carmen. Canipeche. Me.xico: ^Station de Biologie Marine dii Museum National dHisioirc Naiiirelle et du College de France. BP 225. 29900. Concarneau. France ABSTRACT Because in the premolt stage L. .■.nhroMiis should produce glycogen to prepare chilin synthesis, the present paper was directed to answer the following questions: (a) Could an ab.sence of dietary carbohydrates (CHO) enhance the gluconeogenesis pathway in premolt shrimp, following the trend in species of carnivorous fish? (b) In premoulting shrimp is it necessary to force the glyconeogenic pathway through intensive feeding or is a two meal/day feeding schedule enough? (c) Is the CHO metabolism adaptation affected by the age of shrimp? and (d) are the same mechanisms working in moult stage C and D?, A comparison with shrimp fed the same diets and in stage C has been done to determme the effect of the molt stage and dietary CHO levels in glycogen accumulation of this shrimp species. Results obtained showed that in stage D shrimp the gluconeogenic pathway was enhanced, both in an intensise or in a two meal/day feeding schedule, independently of the age of the shrimp. The accumulation of glycogen to support both chitin synthesis and metabolic glucose is proposed as an explanation. Comparing the results obtained in the present paper with that of L styliroslris juveniles published earlier, it is possible to observe that both in 12 hour fasting and dunng glycemia. the glycogen concen- tration in stage C shrimp was opposite of that observed in stage D shrimp. These results show that the intensity of the gluconeogenesis pathway changes with the molting stage, being of more importance in the premolt stage (D) than during the intermoult stage (C). KEY WORDS: carbohydrate metabolism, gluconeogenesis. Digestive gland, blue shrimp Lilnpi'iiacnx .srvlimsiris. molt stage, juve- niles, glycogen INTRODUCTION Carbohydrate metabolism is complex in Crustaceans, Although metabolic aspects are identified, the whole aspect of regulation is still uncertain (Shiau & Peng 1992, Ceccaldi 1998), In recent papers L .ttYlimslris juveniles showed a limited capacity to use dietary carbohydrates. A saturation curve was measured in glyco- gen levels and a-amylase activity against dietary carbohydrate levels (CHO), Maximum values of both the metabolic indexes were obtained in shrimp fed diets containing 2l9f CHO and that was recommended as a maximum shrimp dietary CHO level tor L stylirostris culture (Rosas et al, 2001a), This limitation may be a consequence of a metabolic adaptation to use protein as a primary source of energy. Protein is the main reserve substrate in shrimp, and can be converted to carbohydrates following the gluconeogenic pathway (Campbell 1991). Recently we had observed that the gluconeogenic pathway is at constant rate in several shrimp species. In L styliro.'itris (Rosas et al. 2000), L selifenis (Rosas et al, 2001b). and L vannamci (Taboada et al, 2001 ) a relatively high glycogen level (4,3 mg/g, 4,9 mg/g, and 1 0.1 mg/g) was observed in shrimp fed during 40 days with V?( ot dietary CHO levels, indicating the importance of gluconeogenesis. The internal ammonia, ammonia excretion, glutamate dehydroge- nase activity (GDH). and osmotic pressure measurements showed that the route of protein metabolism was purposed as a donor of amino acids to obtain glycogen, Gluconeogenesis in crustaceans is a biosynthetic pathway for de luno synthesis of glucose from lactate or alanine in an inter-tissue cooperation pathway. Accord- ing to Lallier and Walsh ( 1991 ), the flux of lactate and alanine is oriented to glucose biosynthesis intensively, giving at the digestive *Corresponding author. Carlos Rosas, Apdo, Post, W. Cd, Del Carmen, Campeche, Mexico. E-mail: crv@hp.fciencias,unam,mx gland a gluconeogenic role. An elevation in phi)sphoenolpyruvate carboxykinase (PEPCK). the regulatory enzyme of gluconeogen- esis pathway, was related to the increase of gluconeogenesis path- way in L. vannamei fed with 1% dietary CHO (Rosas et al, 2001a). Similar results were seen in Chasmugiuilhiis gramdata crabs (Ol- iveira & Da Silva 1997) and in Callinectes sapidus (Lallier & Walsh 1991). Those results emphasized the gluconeogenetic role of the digestive gland. Shrimp, like other decapods, have an oiganic reserve cycle which show modifications. During the molt cycle, glycogen re- serves are actively accumulated from stage Do, preparing the ani- mal for chitin synthesis (Renaud 1949). If the shrimp can synthe- size glycogen during stage C. the mechanisms associated with clyc'ogen synthesis in stage Do to D I '"should be overstimulated. producing important changes at the tissue level. Hepatosomatic index (or Digestive gland index; DGI) has been used as an index in CHO metabolism of the digestive gland, be- cause there is a direct relation between glycogen concentration and digestive gland weight in shrimp (Gibson & Barker 1979, Jayapra- kas & Sambhu 1998, Ramos et al, 1996, Sambhu & Jayaprakas 1997), According to Renaud (1949), this DGI should be changed according to the molt stage, showing the changes in CHO metabo- lism associated with storage reserves for molting. To date. CHO metabolism has been studied in stage C shrimp but we do not know if the gluconeogenic pathway in shrimp works in the same way when it needs more CHO to be prepared for chitin synthesis. In carnivorous fish, (cod fish) fed without dietary CHO. it was observed that the hepatosomatic index was higher than that ob- served in fish fed with 25-50% dietary CHO (Hemmre et al. 1990). This was interpreted as an adaptation of cod to increase the gluconeogenic pathway to store CHO reserves. Is this type of mechanism working in the same form in shrimp? There is no information in penaeid shrimp. 931 932 CUZON ET AL. Glycemia produced by meal has been reported in several shrimp species. Shiau and Peng (1992) reported that glycemia in P. monodon was produced two to fi\e hours after feeding and de- pended on the CHO type. Similar results were reported by Abdel Rahman (1996) in P. japoniciis, by Rosas et al. (1995a) in L. setifenis adult males, by Lignot et al. (1999) in L. .mtirostris and by Cousin (1995) and Rosas et al ( 2001a) in L vannamei. Once the food is digested in the gut. chime and fine particles are digested in the lumen and absorbed by diffusion by the inner portion of the digestive gland tubules, initiating the accumulation of reserves. After absorption, glucose passed through the digestive gland (DG) wall to the blood and glycogen is stored (Al-Mohanna & Nott 1987). This mechanism should be modified when shrimp are fed without CHO. In first instance it should be fast enough to trans- form the dietary protein into glycogen during intermoult period without CHO supply. In L. sclifcnis and L. vannamei intermoult (stage C) shrimp fed without CHO, it was observed that during alycemia, glycogen is accumulated two to three hours after feed- ing, evidencing that without dietary CHO glycemia is a fast pro- cess (Rosas et al. 2001b). In premolt shrimp (stage D). this inecha- nism could be enhanced. Most experiments done v\ ith nutrients tend to test a gradient ot a nutrient (protein for example) against another (carbohydrate CHO for example). Then interpretation leads to a double sense whether one considers one nutrient in relation with another. To alleviate that problem, diets can be formulated and distributed on a pair-feeding approach, which can maintain protein constant with and without dietary CHO level. With this method it is possible to maintain the shrimp in a constant feeding schedule causing a per- manent glycemia (Cousin 1995, Rosas et al. 2000a). Because in the premolt stage, shrimp should produce glycogen to prepare for chitin synthesis, the present paper was directed to answer the following questions: (a) Could an absence of dietary CHO enhance the gluconeogenesis pathway in premolt shrimp, following the trend in species of carnivorous fish? (b) In premolt- ing shrimp is it necessary to force the glyconeogenic pathway through an intensive feeding approach or is it enough to use a two meal/day feeding schedule? and (c) Are the CHO metabolism ad- aptations affected by the age of shrimp? A comparison with shrimp fed with the same diets and in stage C (Rosas et al. 2000) was done to determine the effect of the molt stage and dietary CHO levels in glycogen accumulation of this shrimp species. Digestive gland index, digestive gland glycogen concentration, and blood glucose levels of juveniles and adults of L. styiirostris were evaluated. Two experiments were done at the "Centre Ocean- ologique du Pacifique" (Ifremer) located in Tahiti. In the first experiment the effect of dietary CHO levels on digestive gland glycogen. DGI. and blood glucose after glycemia produced by feeding in premolt shrimp were evaluated. The second experiment determined the effect of age of premolt shrimp on DGI and diges- tive gland glycogen after intensive feeding and in the absence ot dietary CHO. MATERIALS AND METHODS Animals First Experiment One hundred shrimp (9.45 ± 0.15 g wet weight) were reared for 15 to 18 days at a density of 8 shrimp/72-L (8 shrimp/nr) tank in a flow through sea water system (1 L/min; salinity of 7 O) o u >. O 16 T 14 12 10 8 ] 6 4 2 0 i- (£i Pre-prandial rif^ Glycaemia D I Post-prandial Figure 2. Digestive gland glycogen (mg/g (issue) of I.ilopenneus styl- irostris juveniles after feeding ^^ith different carbohydrates (CHO) levels. Mean ± S.E. First Experiment. DGI, %10t 9 8 7 6 - 5 - 4 - 3 - 2 - 1 - 0 Juveniles ill 0 40 Dietary CHO level, % 4 -, DGI, 3.8 3.6- 3.4- 3.2 - Pre adu ts 0 40 Dietary CHO level, % Figure 4. Wet digestive gland index ( % ) of Lilopenaens stylimstris pre adults fed with (I and 40% dietary carbohydrates (CHO) levels. Mean ± S.E. Second Experiment. Effect of Dietary Carbohydrates on Premolt Shrimp 935 1.75 1 1.7 - 1 65 "|> 1.6 c 1.55 _ • _■;•;■;■!•!■ g. 1.5 - i.1.45^ 1.4 ^ 1.35 - 1.3 - -:■:■;•:':• 40 0 Carbohydrate level, % Figure 5. Digestive gland glycogen (mg/g wet tLssue) of Lilopenaeus stvliroslris pre adults fed with 0 and 40'7 dletar> carbohydrates (CHOI levels. Mean ± S.E. Second Experlmenl. hours after a meal, a peak in DGI and glycogen concentration was measured in L vannamei juveniles (Cousin 1995) and in L. sc- tiferns adult males (Rosas et al. 1995b). Both experiments in the present study, juveniles and pre adults showed the highest DGI and glycogen concentration were obtained when shrimp were fed with low dietary CHO level or without dietary CHO. showing that without dietary CHO the gluconeo- genic pathway was induced (Fig. 1. Fig. 4). This induction means that L. xnlirostris is well adapted to use protein as a source of metabolic energy and to make glycogen. According to Cuzon et al. (1998). L. snliwstris is omnivorous-carnivorous, justifying its ad- aptation to process protein-rich diets. During the molting process (stages Do to Dl'"). crustaceans not only accumulate glycogen but also triglycerides and proteins in the DG in preparation for chitin synthesis (Santos & Keller 1993). During this time, the hexokinase activity is increased to accumu- late glycogen, reducing the pentose pathway in preparation of chitin synthesis and the demands for energy ( 1 .4 kJ representing about 259c of the energy accumulated along the intermoult cycle). which is provided through the glycolysis pathway. The highest levels of DGI and DG glycogen observed in shrimp fed without CHO reflect the amount of glycogen that the shrimp needs to support both the chitin synthesis and the production of metabolic glucose. This has been observed in other shrimp species in stage C of molt (Rosas et al. 2000, Rosas et al. 2001a. Rosas et al. 2001b). In contrast, in shrimp fed with CHO rich diets a high gluconeo- genesis is not necessary because shrimp have enough CHO from the diet to satisfy the chitin synthesis and glucose for energetic demands producing, in consequence, the lowest levels of the DGI and DG glycogen. In a recent paper (Rosas et al. 2001a). we demonstrated that L. vcmnamei can convert protein to glycogen by a gluconeogenic pathway when fed with 1% dietary CHO. In that study we mea- sured a high phosphoenolpyruvate carboxykinase (PEPCK) activ- ity indicating that the gluconeogenic pathway is induced to pro- duce glycogen and glucose. This activation allowed maintenance of a relatively high circulating glucose of 0.34 mg/niL in hemolymph in comparison to that observed in shrimp fed with 30% dietary CHO (0.45 mg/mL). A similar pattern has been ob- served in Chasmagnatus gramdata crabs (Oliveira & Da Silva 1997) in which the production of glucose from '"* C alanine was demonstrated. This type of strategy is not exclusive to shrimp. In a carnivo- rous fish such as cod. the hepatosomatic index (HIS) followed a similar pattern as with L. styUrostris with a HSI of 6.5% when Fish were fed without CHO and 5.3% when fed with 40% starch. These results suggest that cod are well adapted to obtain CHO from protein, knowing this Fish in the wild does not have dietary CHO (Hemre et al. 1990). The present results can be related to the adaptation of shrimp to respond to the protein and carbohydrate variations in wild also. Donaldson ( 1976) demonstrated that, depending on the season of the year: proteins are the most abundant molecules in benthic ecosystems ranging between 46% to 72% in comparison to 1% to 2.5% of CHO. That observation suggest that wild shrimp are ob- ligated to adapt to synthesi/e CHO according to en\ironmental protein fluctuations, adjusting their metabolism to be dependent on the abundance of proteins more than CHO. During glycemia. the highest DGI and glycogen levels were measured in L. srylirosiris fed without dietary CHO. both in an intensive feeding schedule and after a single meal. Between juve- niles (Fig. 1. Fig. 4). the highest DGI was observed in shrimp fed mtensively. indicating that in this type of feeding regimen the gluconeogenesis pathway can be forced to produce more glycogen -C fasting - D fasting 12 10 8 - 6 -- 4 2 - 0 + 10 21 33 3 20 n tf> (A — ♦— C glycaemia — ■ — D glycaemia \ ^"^\ ^--*-— E 10 - ^ ■ ■~~* Glycogen O Ol 1 10 21 33 Dietary CHO level, % Figure 6. Digestive gland glycogen (mg/g wet tissue) of Litopenaeus styliroslris fed with different CHO levels and with shrimp 12 h fasting and during glycaemia. Data for stage C from Rosas et al., (20()()). Mean ± S.E. 936 CUZON ET AL. than that observed only after a meal. In the nitensi\e feeding trial, shrimp were fed every hour starting from 0800 to 1630, which represented around seven meals a day. This rhythm of feeding procured a constant high blood glucose level and the highest gly- cogen level in the DG. Feeding has been shown to decrease hemolymph hyperglycemic hormone (CHH) levels activating the glycogen synthesis, resulting in increase of glycogen content in different tissues (Santos & Keller 1993). If high glucose remains for a long time in the blood after feeding, the control of CHH could be reduced maintaining the glycogen synthesis for a long time and evidencing the poor regulation of CHO metabolism by shrimp. Similar results were shown by Cousin (1995) in a previous study in which he observed the postprandial effect of different starches in L. r(j;iy(i/;)(f/ juveniles. All these results suggest that the gluco- neogenesis process is fast and independent of the molt stage. In previous work, we observed in L. stylirnstiis fed without CHO that the peak of glycogen after a single meal is between two to three hours (Rosas et al. 2000). In this present study made with shrimp in stage D, glycemia was reached two hours after feeding when shrimp were fed after 1 2 hours fasting. These results could indicate that the molting stage changes the intensity of the process but not its rate. The use of labeled CHO in diets could help confirm this hypothesis. The age of shrimp does not affect the intensity of gluconeo- genesis in shrimp fed without CHO. As in juveniles, pre adults had the highest DGI and DG glycogen levels when they were fed without CHO. The shrimp culture industry had demonstrated that shrimp broodstock are more carnivorous than juveniles, because at that age the shrimp needs more protein and lipids for sexual matu- ration (Bray & Lawrence 1992). For this reason the gluconeogen- esis in the adult phase could be more important than in juveniles. Comparing the results obtained in the present paper with those for L. styUrostns juveniles and published in Rosas et al. (2000), it is possible to observe that both in 12 hour fasting and during glycemia. the glycogen concentration in stage C shrimp was the inverse of that measured in stage D shrimp (Fig. 6). These results show that the intensity of gluconeogenesis pathway changes with the molting stage, being more important in the pre molt stage (D) than during the inter molt stage (C). In the same paper, a signifi- cant induction of a-amylase and a-glucosidase activities related to an increase in dietary CHO levels were measured in L. snliroslris juveniles indicating that in stage C the dietary CHO affected di- rectly the CHO metabolism. During growth, the molting process requires the shrimp to make dramatic physiological, behavioral, and metabolic changes to satisfy energetic and structural demands (Bauer 1996, Carvalho & Phan 1998, Charmantier et al. 1994). To satisfy energetic de- mands, L snliroslris, as other shrimp species, is well adapted to use protein independently of the dietary CHO. The DG changes observed in the present study confirm that assumption and the limited ability of this shrimp species to use dietary CHO, mainly when the glycogen demands are enhanced by the molting stage. ACKNOWLEDGMENTS This study was partially financed by the CONACyT-Sisierra grant to Carlos Rosas. We also acknowledge financial support by ANUIES-Me.xico and ECOS-Nord-France programs. Thanks to Dr. Ellis Glazier for editing this manuscript. LITERATURE CITED Ahdel Rahman, S. H. 1996. Evaluation of various diets for the optimum growth and survival of larvae of the penaeid prawn Penaeiis japoniciis Bate. Aqiwciilt. Niilr. 2:151-155. Al-Mohanna, S. Y. & J. A. Nott. 1987. R cells and the digestive cycle m Penaeus semisulcatus (Crustacea Decapoda). Mar.Biol. 95:129-137. Aquacop. 1973. Determination des stades dintermue chez Macrtihrciciiiiiiii rossembergii (caridae) et Penaeus mergiiiensis (Peneaidea). Rapport Interne, CNEXO/COP 1-37. Bauer, R. T. 1496. Role of the petasma and appendices masculinae during copulation and insemination in the penaeoid shrimp. Sicyonia ilorsalis (Crustacea: Decapoda: Dendrobranchiata). Imerwhr. Reprocl. Dev. 29: 17.V1S4. Bray. W.A. & A. L. Lawrence. 1992. Reproduction of penaeus species in captivity. In A. W. Fast & I. J. Lester, editors. Marine Shrimp Culture: Principles and Practices. El.sevier Science Pub. 93- 1 70 pp Carvalho, P. S. M. & V. N. Phan. 1998. Oxygen Consumption and Am- monia Excretion During the Moulting Cycle in the Shrimp Xiphope- naeus kroyeri. Comp. Biochem. Physiol. 1 19A:839-8-t4. Ceccaldi, H. J. 1998. A synopsis of morphology and physiology of diges- tive system of some crustacean species studied in France. Rev in Fisli- eries Science 6: 1 3- 1 4. Charmantier, G., C. Soyes. Aquacop. 1994. Effect of moult stage and hypoxia on osmoregulatory capacity in the peneid shrimp Penaeus vannamei. J. Exp. Mar. Biol. Ecol. 178:223-246. Cousin. M., 1995. Contribution a I'etude de I'utilisation des glucides et du rapport proteine/energie chez P. vannamei et P.srylirosrris. PhD thesis. InstituI National Agronomique Paris-Grignon. 181 pp. Cuzon. G. & Aquacop. 1998. Nutritional review of Penaeus shiirostris. Rev. in Fish. Scien. 6:129-141. Donaldson, H. A. 1976. Chemical composition of sergestid shnmp (Deca- poda: Natantia) collected near Bermuda. Mar. Biol. 38:51-58. Drach, P. & C. Tchemigovtzeff. 1967. Sur le method de detrmination des stades dintermue et son application generale aux crustaces. Biologie Marine 8:595-610. Gibson, R., & P. L. Barker. 1979. The decapod hepatopancreas. Oceanogr. Mar Biol. Ann. Rev. 17:285-346. Hemmre, G. I., O. Lie, G. D. Lumberstsen, A. Sundby. 1990. Dietary carbohydrtaes utilization in cod (G. moroua). Hormonal response of insulin, glucagon and glucagon-like peptide to diet and starvation. Comp. Biochem. Physiol. 97A:41-44. Jayaprakas, V. & C. Sambhu. 1998. Growth characteristics of white prawn Penaeus inilicus (Decapoda-Crustacea) under dietary administration of protein hormones. Indian J. Mar. Sci. 27:389-395. Lallier. F. H. & P. J. Walsh. 1991. Metabolic potential in tissue of the blue crab. Callinecles sapidus. Bull. Mar. Sci. 48:665-669. Lignot. J. H.. J, C. Cochard, C. Soyez, P. Lemaire & G. Charmantier. 1999. Osmoregulatory capacity according to nutritional status, molt stage and body weight in Penaeus sryliroslris. Aquacuhure 170:79-92. Oliveira, G. T. & R. S. Da Silva. 1997. Gluconeogenesis in hepatopancreas of Chasmagnalhus granulata crabs maintained on high-protein or car- bohydrate-rich diets. Comp. Biocliem. Physiol. 1 18A: 1429-1435. Ramos, L., C. Vazquez Boucard, M. Oliva, E. Garcia & 1. Fernandez. 1996. Variations in the total lipids content, lipids and fatty acids class during the ovarian maturation in the Penaeus schmini and P. notialis. Rev. Invest. Mar 17:157-165. Renaud, L. 1949. Le cycle des reserves organiques che/ les crustaces decapodes. Ann. Inst. Oceanolog. 24:249-357. Rosas, C, A. Bolongaro-Crevenna. A. Sanchez. G. Gaxiola, L. Soto, & E. Escobar. 1995a. Role of the digestive gland in the energetic metabolism oi Penaeus setiferus. Biol. Bull. 189:168-174. Rosas, C. A. Sanchez. E. Diaz. L. Soto. G. Gaxiola. R. Brito, M. I. Baes & R. Pedroza. 1995b. Oxygen consumption and ammonia excretion of Effect of Dietary Carbohydrates on Premolt Shrimp 937 Penatnis selifcnis, P. schinilti, P. iltttucirum and P. noticdi^ postiarxae fed purified test diets: effects of protein level on suhstratc metabolism. Aqiial. Liv. Ri'Miiin: X: 161-169. Rosas. C. G. Cu/on. G. Gaxiola. L. Arena. P. Leniaire. C. Soyez. & A. Van Wormhoudt. 2000. Intluence of dietary carbohydrate on the metabo- lism of ju\enile Liiopciuieiis sryliioslris. J. E\p. Mar. Bio. Ecol. 249: 181-198. Rosas, C, G. Cuzon. G. Ga.\iola. Y. Le Pnol. C. Pasciial. J. Rossignyol. F. Contreras. .\. Sanchez & A. Van Wormhoudt. 2001a. Metabolism and growth of juveniles of Liropeiuieiis vunnamei: effect of salinity and dietary carbohydrate level. J. E.xp. Mar. Biol. Ecol. 259:1-22 Rosas. C. G. Cuzon. G. Taboada, C. Pascual. G. Gaxiola & A. Van Wormhoudt. 2001b. Effect of dietary protein and energy levels (P/E) on growth, oxygen consumption, hemolymph and digestive gland car- bohydrates, nitrogen excretion and osmotic pressure of Lilopciuieiis vannamei and L. «'r//i'//(i juveniles (Crustacea, Decapoda; Penaeidae). Aijuaciil!. Rc'i. 32:1-20. Sambhu. C. & V. Jayaprakas. 1997. Impact of human chorionic gonado- tropin on digestive enzymes activity and nucleic acid contents of pearl- spot. Etropliis surutensis (Bloch) and whiteprawn, Peiiaeus imliciis {M. Edwards). Proc. Indian Null. Sn. Acatl.lB BioiSci.) 63:.5l5-522. Santos, E.A. & R. Keller. 1993. Crustacean hyperglicemic hormone (CHH) and the regulation of carbohydrate metabolism: current perspectives. Comp. Biochem. Physiol. 106A:405-4II. Shiau, S.-Y.. S.Y. Peng. 1992. Utilization of different carbohydrates at different dietary protein levels in grass prawn. Penaens monodon, reared in seawater. Aquaculture 101:241-250. Taboada. G., G. Cuzon. G. Gaxiola. C. Pa.scual, L. Arena, A. Van Worm- houdt & C. Rosas. 2001. Influence of dietary carbohydrate levels on growth, survival and glucid metabolism of the white shrimp Litope- naeiis vannamei. .Aipiaciillure In press. Zar, J. H. 1974. Bioestatistical Analysis. Englewood Cliffs; Prentice Hall. .lounuil i>f Shellfish Kcscunh. Vol. 20. No. 3, 439-943. 2001. THE CONTENT OE ASCORBIC ACID AND TOCOPHEROL IN THE TISSUES AND EGGS OF WILD MACROBRACHIUM ROSENBERGIl DURING MATURATION RONALDO O. CAVALLI,' - MONTAKAN TAMTIN/ PATRICK LA YENS,' PATRICK SORGELOOS.' HANS J. NELIS/ AND ANDRE P. DE LEENHEER' Lahonitory af Aqiiucultiirc and Anemia Reference Center. Ghent University. Rozier 44, 9000 Ghent. Belgiunt: ^Lahoratorio de Maricitltura, Departaniento de Oceaiuiiirafia. Fitnda^do Universidade Federal do Rio Grande. CP 474. 96201-900, Rio Grande. Brazil: ''Petchahuri Coastal Aquaculture Station, Haad Chao Saniran. Muang. 76100 Petchahiiri. Thailand: ^Laboratory of Pharmaceutical Microhiologx. Ghent University. Harelbekestraat 72. B-9000. Ghent. Beli-ium: ^Laboratory of Medical Biochemistry and Clinical Analysis. Ghent University. Harelbekestraat 72. B-9000. Ghent. Belgiiun ABSTRACT Vanalicins in the cuncentratiiins of ascorbic acid (AA) and tocopherols in association with the gonadal developmenl of the freshwater prawn Macrobraihiiim rosenhergii were investigated in females captured in the IWae Klong River. Thailand. Mean ovarian AA levels ranged from 210 to 540 |xg/g dry weight (dw) and were at least 11-fold higher than midgut gland (MG) levels. Variations in ovarian AA levels are believed to be related to the biosynthesis of steroid hormones, the formation of collagen, and the deposition of egg yolk compounds. a-Tocopherol (a-T) was the predominant form of vitamin E in prawn tissues and eggs. The level of a-T in the MG was constant, whereas in the ovaries, it ranged from 143 to 425 (.ig/g dw. The incorporation of a-T into the ovary was highly correlated (r- = 0.87) to ovarian lipid levels, which probably reflects the role of this vitamin as a maior antioxidant agent. The present results provide further evidence of the essentiality of these vitamins in crustacean reproduction. KEY WORDS: ascorbic acid, tocopherols. Mucrobruchium rosenhergii. wild, reproduction, nutrition INTRODUCTION Although in the last two decades much progress has been achieved in the understanding of vitamin metabolism in crusta- ceans (Conklin 1997), knowledge concerning the role of vitamins in crustacean reproduction is still limited (Harrison 1990, Harrison 1997). As a result, most information on vitamin functions and requirements are adopted from literature on fish and other verte- brates, rather than being deri\'ed from studies with crustaceans (Harrison 1990). Vitamin E (tocopherol) and vitamin C (ascorbic acid [AA]) are considered essential dietary components for crustaceans (Conklin 1997). The biological activity of vitamin E is widely accepted to be at least partially related to its antioxidant properties, as it reacts rapidly with organic free radicals that may damage membrane- bound polyunsaturated fatty acids (PUFA) (Burton & Trabor 1990). Vitamin E and C are known to act synergistically, with vitamin E reacting with lipid peroxy radicals donating a hydrogen atom and forming a vitamin E radical, which is then regenerated by AA (Packer et al. 1979). The imponance of vitamin E for fish reproduction has long been recognised (Watanabe & Takashima 1977, Watanabe et al. 1985), but has only been recently demon- strated in crustaceans (Cahu et al. 1991, Alava et al. 1993b. Cahu et al. 1995). Aside from its role in the recycling of vitamin E, AA also participates in the enzymatic processes involved in the formation of collagen (Barnes 1975; Hunter et al. 1979) and in the biosyn- thesis of steroid hormones (Hilton et al. 1979: Seymour 1981). Although the need for AA in diets for fish broodstock has been well established (Watanabe & Takashima 1977, Sandnes et al, 1984, Soli man et al. 1986, Waagbo et al. 1989, Dabrowski 1991, Blom & Dabrowski 1995), the essentiality of AA in crustacean reproduction was initially inferred from a study evidencing its variation in the ovary of Palaeiiuiii serratiis Pennant (Guary et al. 1975). More recent studies on penaeid shrimps have confirmed its importance in crustacean reproduction (Alava et al. 1993a, Alava et al. 1993b, Cahu et al. 1995). Under rearing conditions, feed regimes for the freshwater prawn Maerohrachium rosenhergii (de Man) range from fresh food to foimulated feed, and thus vitamin rations may vary con- siderably. As no information is available on the status of vitamin E and C in adult prawn tissues during maturation, it is not possible yet to establish criteria for evaluating the ovarian status of these vitamins or to recommend modifications in the dietary vitamin content so as to optimize broodstock performance and offspring quality. With this in mind, this paper aims to present baseline data on the concentrations of AA and tocopherols in the midgut gland, ovary, and eggs of wild M. rosenhergii throughout sexual matu- ration, and it discusses the possible roles that these vitamins may have in the reproduction of this species, MATERIAL AND METHODS Live mature M. rosenhergii females were obtained from fish- ermen in the Mae Klong River, Amphur Muang, Province of Samut Songkhram, Thailand. Captures were made in single col- lections on July and September 1998. After capture, female prawns were grouped in five stages of gonadal development, according to the si/e, colour, and aspect of the ovary (Chang & Shih 1995), i.e., (1) no ovarian tissue is visible, which is characteristic of both nondeveloped and spent females; (II) the ovary has a small yellow- colored spot near the posterior part of the carapace; (III) the ovar- ian tissue turns orange and is visible from the posterior part of the carapace to the area just in front of the epigastric tooth; (IV) the ovarian tissues have grown and extended to the area of the epi- gastric tooth; and (V) the ovarian tissues have extended to the anterior part of the carapace. Females in all five stages of gonadal development were sampled on July and September, Females were then blotted dry and were individually measured (total length from the tip of the rostrum to the end of the telson) and weighed (to the nearest 0,1 g). The ovary and midgut gland were quickly dissected, weighed, and immediately frozen at -20"C. Grayish, eyed-eggs were also sampled and conserved in a 939 940 Cavalli et al. TABLE 1. Weight, total lensth, ovary weight, and midgut gland weight of wild ,V/. rosenhergii females at different stages of gimadal development. Each value is the mean of four separate prawn samples analyzed individually, except for stages 1 and 11 where tissues of three prawns were pooled. Within rows, values with different superscript letters indicate significant differences (/*< 0.051. 1 Stage of Gonadal Development 11 III IV Pmwn weight (gl Pmwn length (cm) Midgut gland weight (g) Ovary weight (g) 33.4+ 12.4 14.4 ± 1.1 1.37 ±0.50 0. 1 7 + 0.06" 34.3 ± 10.1 14.7 ± 1.4 1.38 ±0.39 0.53 ±0.15' 41.4+ 12.5 15.2 ± 1.9 1.79 ±0.54 1.16 ±0.35'"' 36.7 ±11.6 15.4± 1.7 1.53 ±0.48 1.62 ±0.51'''" 38.8 ± 10.2 15.5 ± 1.4 1 .73 ± 0.45 2.33 ± 0.60" similar manner. The gonado-somatic index (GSI) and midgut siv nialic index (MSI) were calculated as the percentage of gonad and midgut gland to total body weight, respectively. All samples were conditioned in a Styrofoam box with dry ice and were transported by air to Belgium. Samples for AA and tocopherol analysis were then maintained at -80°C and -20°C until analysis, respectively. The content of AA of the samples was determined according to Nelis et al. (1997). while a-tocopherol (a-T), 7-tocopherol (7-T). and 8-tocopheroi (8-T) levels were estimated following Huo et al. (1999). Tissue samples from the same wild M. rosenhergii females were utilized in a parallel study describing the variation of total lipids, lipid classes, and fatty acids (Cavalli et al. 2001). Statistical analysis of the data was undertaken with one-way analysis of variance (ANOVA) followed by Tukey's honest sig- nificant difference (HSD) test. An alpha level of 0.05 was used to identify significant differences. Correlations were determined us- ing linear regression analysis. A minimum of three replicates for each tissue and eggs were analyzed. When tissues of one indi- vidual were insufficient for analysis, tissues were pooled from 2 to 4 individuals. This was especially true for the early stages of gonadal development. Results are presented as means ± SD. RESULTS Prawn weight, total length, and midgut gland weight presented no significant differences between the various stages of matura- tion, but ovarian weight tended to increase, especially after stage III (Table 1). Changes in MSI and GSI during maturation are shown in Figure 1. MSI was constant irrespective of maturation stage, whereas GSI presented a significant increase from stages I to V. Figure 2 summarizes the data on the concentrations of AA in the midgut gland (MG) and ovary throughout maturation. The content of AA in the MG was constant in the earlv stages of 1^^ MSI , r r 1 [ > ■ I i^^ "(381 I II III IV V Stage of development Figure 1. Changes in GSI and MSI of wild M. rosenhergii females at different stages of gonadal development. maturation, but decreased significantly from stage 111 (mean of 24.9 (xg/g dw) to stage V (12,8 jjLg/g dw). In the ovary, the AA concentrations were stable at around .540 p-g/g dw at stages I and II. but declined significantly between stages II to IV. At stage V, ovarian AA levels increased significantly to 370 |xg/g dw. Within the same stage of gonadal development, mean AA levels were 1 1- to 29-fold higher in the ovary than in the MG. Eggs contained 128.2 ± 28.3 j.Lg AA/g dw (Table 3). The main form of tocopherol present in the MG. ovary, and eggs was a-T (Tables 2 and 3). In the MG, there were no signifi- cant differences in tocopherol content, as the variations were large. The content of a-T in the ovary increased significantly during the initial stages of gonadal development (from stages I to III), and decreased afterwards. The levels of 7-T were not significantly different at any given stage of maturation, and 8-T levels were below the detection limit during the initial stages of maturation. The mean content of a- and 7-T in the eggs was 324.7 and 22.8 (jLg/g dw. respectively (Table 3). No 8-T was detected in the eggs. Linear regression analysis of the total lipid content in the ovary (data from Cavalli et al. 2001) against the concentration of a-T produced a correlation coefficient (r") equal to 0.87 (Fig. 3). Cor- relations between ovarian total lipids and 7- and 8-T levels were not significant. DISCUSSION Although only a limited number of studies report AA levels in crustaceans, it is well documented that tissue levels vary with season, ontogenetic development, dietary intake, moulting, and reproductive cycles (Guary et al. 197.5. Magarelli & Colvin 1978. Coglianese & Neff 1981. Merchie et al. 1995). In fish, several 600 1 g 500 ■ 400 300 200 ^ 100 0 120 5 inn u 80 i 3. 60 c 40 O) a 20 •0 ? 0 I II III IV V stage of development Figure 2. Concentration of ascorbic acid (micrograms per dw ) in the midgut gland and ovary of wild M. rosenhergii according to the stage of gonadal development. gram of females Ascorbic Acid .-^nd Tocopherol in Wild Prawn Bkoodstock 941 TABLE 2. Concentration of tocopherols (niicrogram.s per yrani of dr> Height) in the midgut gland and o>ar.\ according to the stage of gonadal development of w'M M. rosenbergii females. Kach value is the mean of four separate pra«n samples analv^ed individuallv. except for stages I and II where tissues of three prawns were pooled. \\ ithin rows, values with different superscript letters indicate significant differences (/• < 0.05). Stage of Gonadal Development I II Ill IV V Midgut gland u-Tocophcrol -y-Tocopherol 6-Tocopher(il Ovary a-Tocopherol 7-Tocopherol 8-Tocopher(il 153.7 ±72.2 21.1 ± 18.1 0.5 + 0.7 142.6+ 13.0" 11. (1± 1.9 n.d. 174.9 ±70.3 13.0± 11.3 0.3 ± 0.3 334.0 + 79.9-''' 22.1 ±2.1 n.d. 136.3 + 92.8 9.8+1.6 n.d. 425.0 ± 67.7-' 13.1 ±0.8 n.d. 31.6+ 12.3 2.7 ± 0.9 n.d. 260.3 ± 33.5"'' 16.3 ± 1.5 n.d. 102.3 ±26.5 9.7+ 1.6 1.0 ±0.5 279.1 ±37.1"'- 17.9±7.1 0.6 ± 0.4 n.d. = not detec ted aulhors (Se.Miiour 1981. Sundries &. Braekinari 1981. Dabrowski 19911 have demonstrated that the levels of AA in the ovaries change during the reproductive cycle. Sandnes and Braekman (1981) showed a rise in ovarian AA concentration during ovarian growth followed by a decrease in the final stages prior to spawn- ing, and they discussed whether this variation could be related w ith sex steroid synthesis. Guary et al. (1973) also postulated that the decrease in AA levels in the maturing ovary of P. serratus could be connected to steroidogenesis. In the present study, the decline of AA levels in the ovary between stages II to IV coincides with an active phase of ecdysteroid hormone accumulation in the ma- turing ovaries of M. rosenbergii (Wilder et al. 1991 ). Furthermore, significant levels of cholesterol, the chief precursor of steroid hor- mones (Kanazawa & Teshima 1971 ). were present in the ovary of M. rosenbergii throughout maturation (Cavalli et al. 2001). These findings reflect a possible demand for AA by the hydroxylating reactions needed for steroidogenesis in the ovarian follicle cells, and they agree with results revealing the possibility of endogenous production of steroid hormones in crustaceans ( Kana/.aw a & Teshima 1971. Shih & Liao 1998). The decrease in the ovarian AA content between stages II and IV could also be linked to the biosynthesis of collagen, as the hydroxylation reaction necessary for the synthesis of this fibrous protein requires the presence of AA at adequate levels (Hunter et al. 1979). In this respect, the concentration of total ovarian protein has been shown to increase lineariy along with GSI in M. rosen- bergii {Lee & Chang 1997). Guary et al. (1975) suggested that the formation of egg yolk compounds, such as polysaccharides and glycogen, also require TABLE 3. Concentration of ascorbic acid and tocopherols in the eggs of wild M. rosenbergii females. Mean ± SD (ng/g dry weight) Asccirbic acid a-Tocopherol 7-Tocopherol 6-Tocopherol 128.2 + 28.3 324.7 ± 77.3 22.8 ± 10.4 n.d. n.d. = not detected considerable amounts of AA and therefore could be an additional cause for the decrease in ovarian AA levels, particulariy at the final stages of gonadal development. Although the possibility that these metabolic processes consume some AA cannot be ruled out, the observation that the levels of AA in the ovary of M. rosenbergii increased from stage IV to V suggests that the deposition of AA into the ovary at the final stages of maturation occurs at a much higher rate than its catabolism. The raise in ovarian AA content in the final stages of matura- tion may be related to an increased requirement in the egg at a later stage of life (Hilton et al. 1979). Indeed, it was found in various fishes (Hilton et al. 1979. Dabrowski 1991. Blom & Dabrowski 1995) and crustaceans (Guary et al. 1975. Coglianese & Neff 1981. Alava et al. 1993a) that AA levels in the ovaries are usually higher than in other tissues. Similarly, in the present study, the ovary of M. rosenbergii contained significantly more AA than the MG. Sandnes et al. (1984) and Soliman et al. (1986) confirmed that an important share of the broodstock dietary AA intake is trans- ferred to the oocytes where it is stored for use during embryogen- esis and larval development. This clearly indicates a preferential transfer of reserve AA to the embryos, which is particulariy im- portant in M. rosenbergii since the embryos and eariy larvae are totally dependent on the yolk reserves for normal organogenesis and physiological functioning (Harrison 1990). Several authors have shown that the viability of fish eggs (Sandnes et al. 1984. Soliman et al. 1986. Waagbo et al. 1989) and shrimp eggs (Cahu et al. 1995) was directly related to their AA content. The miprove- ment in egg viability with increased AA levels was attributed to the protection of membrane-bound lipids against oxidation and by the action of this vitamin in the synthesis of stable forms of col- lagen (Cahu et al. 1995). as suggested for fish (Waagbo et al. 1989). In an eariier study. De Caluwe et al. ( 1995) collected M. rosen- Ijeniii eggs 2 days after fertilization and found that mean AA levels in the eggs varied between 210 and 382 jjig/g dw. The upper limit of this range agrees well with the ovarian AA content at the final stages of maturation in the present study (around 370 jig/g dw), but is relatively higher than the 1 28 (Ag AA/g dw found in the e22s. However, as the eggs sampled here were "eyed" and gray colored, and were thus at the final stages of embryonic develop- ment (New & Singholka 1982). this suggests that the developing 942 Cavalli et al. TL = 31.229 + 0.08176 x a-tocopherol (r* = 0.87) 30' — 100 i 150 300 350 400 450 500 550 200 250 a-tocopherol (ng/gdw) Figure 3. Linear regression analysis between the total lipids (TL) and the contents of a-T in the ovary of wild M. rosenbergii females. embryos po.ssibly consumed AA. Moreover, the fact that newly hatched, nonfeeding larvae of M. rosenbergii contained from 149 to 265 (JLg AA/g dw (Cavalli et al.. 2000) further indicates that AA was indeed consumed by the developing embryo. This possibility is also supported by the results of Sato et al. ( 1987) who demon- strated a continuous decrease in AA content of rainbow trout eggs during embryonic development. Conversely, Guary and Guary (1975) reported that the eggs of P. serratiis and Cancer pagurus (L.) seemed able to synthesize AA during the early stages of embryonic development, and hence. AA contents after spawning were found to be similar to those just before hatching. However, from a metabolic standpoint, it seems unlikely that a female shrimp would accumulate considerable amounts of AA into its gonad (Guary et al.. 1975) if the eggs were able to biosynthesize it during the early stages of embryonic development. Therefore, it remains to be confirmed whether the biosynthesis of AA occurs during the embryonic development of crustaceans. Watanabe et al. (1985) reported that vitamin E. together with lipids, was easily incorporated into red sea bream eggs. According to the present results, the incorporation of a-T into the ovary of M. rosenbergii was highly correlated to ovarian lipid levels. This finding is in agreement with the antioxidative role of this fat- soluble vitamin, which requires its close association with lipids, particulariy membrane-bound PUFA. Therefore, it is possible that to fulfill its vitamin E requirements, reproductive M. rosenbergii females would depend more on the dietary intake than on body reserves, as was hypothesized for lipids (Cavalli et al. 2001 ). Nev- ertheless, the contribution of MG and muscle reserves inay also be of some importance. Data from the studies of Castillo et al. ( 1989) and Alava et al. ( 1993b) indicate that «-T might have been trans- ferred from these tissues to the eggs of P. indicus and P. japonicns. respectively. It is still unclear whether this is also true for M. roscnliergii. De Caluwe et al. (1995) found that M. rosenbergii eggs had from 711 to 1,287 |xg a-T/g dw. These concentrations are much higher than those found in the present study, and they suggest a comparatively lower dietary intake of vitamin E under natural conditions. In fact, the rate of incorporation of vitamin E into the eggs of P. indicus and M. rosenbergii was shown to increase with dietary levels of a-TA (Cahu et al. IWl, De Caluwe et al. 1995). M. rosenbergii females fed a diet containing 223 |jig a-TA/g dw produced eggs with an average of 7ll|xg a-T/g dw, while the content of a-T in the eggs almost doubled to 1,287 p.g/g dw when dietary a-TA levels were increased to 2,025 pg/g dw. In summary, the present study provides further evidence of the importance of AA and tocopherols in the reproduction of crusta- ceans, and consequently suggests that feeding a diet deficient in either vitamin C or E could virtually impair broodstock perfor- mance and offspring viability. More research is necessary to de- termine optimal dietary levels for crustacean broodstock. ACKNOWLEDGMENTS We wish to thank Mr. Nopadol Phuwapanish (Department of Fisheries, Thailand) for his help in the collection and sampHng of the prawns, and Dr. Greet Merchie (INVE Aquaculture, Thailand) for the transportation of the samples. We also thank Mathieu Wille for critically reading the manuscript. The assistance of Petra Rigole in the analytical work is also greatly appreciated. This research was funded in part under grant No. 200796/96-8 from the Brazilian Council for Science and Technology (CNPq). Alava, V. R., A. Kunazawa. S. Teshima & S. Koshio. 1993a. Effect of dietary L-ascorbyl-2-phosphate magnesium on gonadal maturation of Penaeus jiiponicus. Nippon Suisun Gakknishi 59:691-696. Alava. V. R.. A. Kanazawa. S. Teshima & S. Koshio. 1993b. Effects of LITERATURE CITED diclary vitamins A, E. and C on the ovarian development of Pciniciis japdiiicii.s. Nippon Siiisan Gakkiii.slii ,''9:1235-1241. Barnes. M. J. 1975. Function of ascorbic acid in collagen metabolism. Ann. N. Y. Acad. Set. 258:264-277. Ascorbic Acid and Tocopherol in Wild Prawn Broodstock 943 Blom. J. H. & K. Dubruwski. 1995. Reproductive success ol teni;ile rain- bow trout (Oncorhynchiis mykiss) in response to graded dietary ascor- byl monophosphate levels. Biol. Reprod. 52:1073-1080. Burton. G. W. & M. G. Trabor. (1990). Vitamin E: antioxidant activity, biokinelics and bioavailability. Anim. Rev. Niitr. 10:357-382. Cahu. C, M. Fakhfakh & P. Quazuguel. 1991. Effect of dietary a-tocoph- erol level on reproduction of Penueus imlicus. In: P. Lavens. P. Sorgeloos, E. Jaspers & F. Ollevier. editors. Larvi '91: Fish & Crus- tacean Larviculture Symposium. European Aquaculturc Society 15: 242-244. Cahu, C. L., Cuzon G. & P. Qua/uguel. 1995. Effect of highly unsaturated fatty acids, a-tocopherol and ascorbic acid in broodstock diet on egg composition and development of Pciuifiis iihIIcils. Cmiip. Binchciii. Physiol. 112A:41 7-424. Castillo, M. .\.. C. Cahu & G. Stephan. 1989. Teneur en a-tocophero! des oeufs et des organes des reproducteurs de Pciuieiis indicus. Oieiinis 15:553-560. Cavalh. R. O., E. Vanden Berghe, P. Lavens. N. T. T. Thuy. M. Wille & P. Sorgeloos. 2000. Ammonia to.\icity as a criterion for the evaluation of larval quality in the prawn Macnibnichiitni rosenbcrgii. Coiiip. Biu- chein. Physiol. 125C;333-343. Cavalli. R. O.. M. Tamtin. P. Lavens & P. Sorgeloos. 2001. Variations in lipid classes and fatty acid contents in the tissues of wild Mucro- hrachium rosenheri^ii (de Man) females during maturation. Aqiuicul- hire 193:311-324. Chang. C. F. & T. W. Shih. 1995. Reproductive cycle of ovarian devel- opment and vitellogenin profiles in the freshwater prawn. Mucro- brachiiim ro.'ienheri^ii. Invert. Reprod. Develop. 27:11-20. Coglianese. M. & J. M. NetT. 1981. Evaluation of the ascorbic acid status of two estuarine crustaceans: the blue crab. Ciillinecte.s .sapidu.s and the grass .shrimp, PaUuinonetes piigio. Comp. Bioihcm. Physiol. 68.^:451- 455. Conkim, D. E. 1997. Vitamins. In: L. R. D'Abramo. D. E. Conklin & D. M. Akiyama. editors. Crustacean Nutrition. Baton Rouge. LA: World Aquaculture Society, pp. 123-149. Dabrowski. K. 1991. A.scorbic acid status in high-mountain charr. Sulveli- iiH.s alpimis. in relation to reproductive cycle. Environ. Biol. Fish. 31:213-217. De Caluwe. J.. P. Lavens & P. Sorgeloos. 1995. The intluence oi Macro- hrachium rosenbergii broodstock diet on egg and larval characteristics. In: P. Lavens. P. Sorgeloos. E. Jaspers & F. Ollevier. editors. Larvi '91: Fish & Crustacean Larviculture Symposium. European Aquaculture Society 24:79-82. Guary. M. M. & J. C. Guary. 1975. Variations de la teneur en acide ascorbique dans les oeufs de Cancer paguriis et Pohicmon serniuis (Crustacea: Decapodal au cours de I'embryogenese. Mar. Biol. 32:357- 363. Guary. M. M., H. J. Ceccaldi & A. Kanazawa. 1975. Variations du tau.\ d'acide ascorbique au cours du developpemenl de I'ovarie et du cycle d'interniue chez Palaemon serratus (Crustacea: Decapoda). Mar. Biol. 32:349-355. Harrison. K. E. 1990. The role of nutrition in maturation, reproduction and embryonic development of decapod crustaceans: a review. J. Shellfish Res. 9:1-28. Harrison. K. E. 1997. Broodstock nutrition and maturation diets. In: L. R. D'Abramo. D. E. Conklin & D. M. Akiyama. editors. Crustacean Nu- trition. Baton Rouge. LA: Worid Aquaculture Society, pp. 390-408. Hilton. J. W.. C. Y. Cho. R. C. Brown & S. S. Slinger. 1979. The synthesis, half-life and distribution of ascorbic acid in rainbow trout. Comp. Bio- chem. Physiol. 63A:447-453. Hunter. B.. P. C. Magarelli, D. V. Lightner & L. B. Colvin. 1979. Ascorbic acid-dependent collagen formation in penaeid shrimp. Comp. Bioeliem. Physiol. 646:381-385. Huo. J. Z.. H. J. Nelis. P. Lavens, P. Sorgeloos & A. P. De Leenheer. 1999. Simultaneous determination of a-tocopheryl acetate and tocopherols in aquatic organisms and fish feed. J. Chromalogr. B 724:249-255. Kanazawa. A. & S. Teshima. 1971. In vivo conversion of cholesterol to steroid hormones in the spiny lobster. Ptmulini.s japoniiiis. Bull. Jap. Soc. Sci. F«/i. 37:891-897. Lee. F. Y. & C. F. Chang. 1997. The concentrations of vitellogenin (vi- tellin) and protein in hemolymph, ovary and hepatopancreas in differ- ent ovarian stages of the freshwater prawn, Macrohrachium rosen- bergii. Comp. Biochem. Physiol. lI7A:433-439. Magarelli, P. C. & L. B. Colvin. 1978. Depletion/repletion of ascorbic acid in two species of penaeid: Penaeus califoniiensis and Penaeiis styl- iroslris. In: J. M. Avault. editor. Proceedings of the Annual Meeting of the World Aquaculture Society, January 3-6. 1978. Atlanta. GA. Merchie. G.. P. Lavens. J. Radull. H. Nelis. A. De Leenheer & P. Sorgeloos. 1995. Evaluation of vitamin C -enriched Anemia nauplii for larvae of the giant freshwater prawn. Atiiiacult. Int. 3:355-363. Nelis. H. J.. A. De Leenheer. G. Merchie. P. Lavens & P. Sorgeloos. 1997. Liquid chromatographic determination of vitamin C in aquatic organ- isms. J. Chromatogr. Sci. 35:337-341. New. M. B. & S. Singholka. 1982. Freshwater prawn farming: a manual for the culture of Macrohrachium rosenbergii. FAO Fisheries Technical Paper No. 225, Rome: FAO. Packer. J. E., T. F. Slater & R. L. Willson. 1979. Direct observation of a free radical interaction between vitamin E and vitamin C. Nature 278: 737-738. Sandnes. K. & O. R. Braekkan. 1981, Ascorbic acid and the reproductice cycle of ovaries in cod {Gadus mon-hua). Comp. Biochem. Physiol. 70A:545-546. Sandnes. K.. U. UIgenes. O. L. Braekkan & F. Utne. 1984. The effect of ascorbic acid supplementation in broodstock feed for reproduction of rainbow trout (Salmo gairdneri). Aquaculture 43:167-177. Sato, M., R. Yoshinaka, R. Kuroshima, H. Morimoto & S. Ikeda. 1987. Changes in water soluble vitamin contents and trans-aminase activity of rainbow trout egg during development. Bull. Jap. Soc. Sci. Fish. 53:795-799. Seymour, E, A, 1981. Gonadal ascorbic acid and changes in level with ovarian development in the crucian carp, Carassius carassius (L.). Ctmip. Biochem. Physiol. 70A:45 1-453. Shih J. T. & C. F. Liao. 1998, Conversion of cholesterol to sex steroid-like substances by tissues of Mictyris hrevidacnlus in vitro. Zool. Stud. 37:102-110. Soliman, A. K„ K, Jauncey & R. J. Roberts. 19X6. The effect of dietary ascorbic acid supplementation on hatchabiliiy, survival rate and fry performance in Oreochromis mossambicus (Peters), .\q]iaculture 59: 197-208, Waagbo, R„ T, Thorsen & K, Sandnes, 1989, Role of dietary ascorbic acid in vitellogenesis in rainbow trout (Salmo gairdneri). Aquaculture 80: 301-314. Watanabe, T, & F, Takashima, 1977, Effect of a-tocopherol deficiency on carp. VI. Deficiency symptoms and change of fatty acid and triglyc- eride distribution in adult carp. Bull. Jap. Soc. Sci. Fish. 43:819-830. Watanabe. T.. A. Itoh, S, Satoh, T. Kitajima & S. Fujita. 1985. Effect of dietary protein levels and feeding period before spawning on chemical components of eggs produced by red sea bream broodstock. Bull. Jap. Soc. Sci. Fish. 51:1,501-1509, Wilder. M, N,. T, Okumura & K, Aida, 1991. Accumulation of ovarian ecdysteroids in synchronization with gonadal development in the giant freshwater prawn. Macrohrachium rosenbergii. Zool. Sci. 8:919-927. Joiirihil of Sbi'lllhli Rcsccuvh. Vol :(). No, 3. 44?-y49. 2001. GROWTH AND GAMETOGENIC CYCLE OE THE CRESTED OYSTER, OSTREA EQUESTRIS (SAY, 1834), IN COASTAL GEORGIA RANDAL L. WALKER AND ALAN J. POWER Marine Extension Service. Slwllfisli Ai/iuu iilfure Lalniratory. University of Georgia. 20 Ocean Science Circle. Savaiuuili. Georiiia 3141 1 -101 1 ABSTRACT Crested oysters. Ostrea equestris (Say. 1834). were collected from the mouth of the Savannah River on December 2, 1998 and transferred to pearl nets suspended from the main dock at the Skidaway Institute of Oceanography on the Skidaway River. To determine the reproductive cycle and growth of oysters, 50 oysters were collected monthly, measured for shell length, and a gonadal sample was taken for histological analysis. Oysters grew from a mean size of 18.8 mm in January 1999 to 28 mm by November 1999 — a rate of 0.9 mm per month. Of 574 oysters .sectioned for histological examination, males dominated the population (49.59^^ ). with only yS'i being females. 4.5'7r were indeterminate and 42.5% being hermaphrodites. Crested oysters were reprodiictively active year around in Georgia with spawning and spent oysters occurring in most months. In males, spawning and spent stages ranged from 23.7% of the oysters in January to 86.4% in August, while females in these stages ranged from 10.7% in February to 77.8% in August. Brooding larvae were found in the histological samples between April and October. Spawning of crested oysters in Georgia occurs year round with major spawning starting in February and continuing into fall. KEY WORDS: crested oyster, Oxtreii eqiicstris. growth, gametogenesis. hermaphrdditism INTRODUCTION The crested oyster. Ostrea equestris (Say 1834), occurs from North Carolina to both coasts of Florida, throughout the Gulf of Mexico and tt) the West Indies (Abbott 1074). Externally, the crested oyster appears similar to the eastern oyster, Crussostrea virginiea (Gnielin 1791), but can be distinguished from it by a single row of denticles along the inner lateral margin of the upper valve with corresponding depressions on the lower valve (Galtsoff & Men'ill 1962). The crested oyster is a high saline organism occurring in estuarine areas greater than 20 ppt to oceanic condi- tions (Galtsoff & Merrill 1962). It grows to 72 mm in length in the Gulf of Mexico (Gunter 1951) and to 82 tTiin on the Atlantic coast (Galtsoff & Merrill 1962). Crested oysters do not form beds like C. virginiea (Galtsoff & Merrill 1962). O. equestris will replace C. virfiiiiica on beds when salinity increases to 30 ppt (Parker 1960). Hoese (1960) observed in Texas at the end of a drought, an O. equestris - Braeliidontes exiistiis community was replaced by a C. virginiea - B. reciiiTiis community. Although there have been several reports published on the bi- ology of O. equestris from the Gulf of Mexico region (Gunter 19.S1. Menzel 19.S5, Parker 1960, Hoese 1960) and one distribu- tional study on the Atlantic coast (Galtsoff & Merrill 1962), stud- ies on life history traits for Atlantic coast populations are absent. This study describes the growth and gametogenic cycle of the crested oyster in the coastal waters of Georgia. MATERIALS AND METHODS Crested oysters were collected by the R/V Georgia Bulldog while trawling with a 3 m conch try net at the mouth of the Savannah River (32°02'42" X 80°50'21") 1998 December 2. The sampling depth was 6 m and the sediment was dominated by shelly material. Surface salinity was 24 ppt. Oysters attached to relic C. virginiea shells were returned to the laboratory where 50 non- randonily selected oysters were measured for shell length (i.e.. hinge to lip] lo the nearest 0.5 mm with Vernier calipers. This initial sample was non-random because large-sized oysters were selected for ease of identification. After tneasuring each oyster, the internal lateral margin of each oyster shell was examined to ensure identification and a 1cm" gonadal sample was removed and pre- served in Davidson's solution. Remaining oysters were placed into 12-mm mesh pearl nets and suspended from the main dock at the Skidaway Institute of Oceanography on the Skidaway River. Each month 50 oysters were randomly gathered, measured for shell length, and a gonadal sainple (ca. 1 cnr) was dissected from each oyster. These oysters were selected by gathering two C. vir- giniea shells from the pearl nets with attached crested oysters. Tissues were processed according to procedures outlined in How- ard and Smith ( 1983). Prepared slides of gonadal tissue were ex- amined with a Zeiss Standard 20 microscope (x20). sexed and assigned to a developmental stage as described by Pouvreau et al. (2000). A staging criteria of 0 to 6 was employed for early active (I), activity developing (2). near ripe (3). ripe (4), partially spawned (5). spent (6). and inactive (0). Water temperatures and river salinity from the Skidaway River were taken from the Marine Extension Service's dock which is adjacent to the Skidaway Institute of Oceanography's dock at 0800 h Monday through Friday. RESULTS Monthly mean water temperature and salinity values for the Skidaway River are given in Figure 1. Temperatures ranged from 12.2'C in Januai-y 1999 to 30.9 C in August 1999. while river salinity ranged from 23.4 ppt in July 1999 to 29.9 ppt in Decetiiber 1999. River salinity values were abnormally high especially during spring due to prolonged drought conditions in the southeast and Georgia. Crested oyster mean size from December 1998 to November 1999 is given in Figure 2. The mean size of oysters in December lOOS (22 mm) was artificially high because larger-sized animals were selected to ensure the correct identification of the animals at the start of the experiment. The January 1999 size (18.8 mm) represents a random mean size. Thus, oysters grew from 18.8 mm in January 1999 to a mean size of 27.8 mm by November 1999 — a rate of 0.9 inm per month. A total of 574 oysters were sexed from December 1998 to November 1999. Indeterminate oysters occurred in every month except March and April and accounted for 26 indi\ iduals (4.5%). 945 946 Walker and Power p— -^, j-^^~~s — y r\^ < -•-Salinity I -o— Temperature Nov-98 Dec-98 Jan-99 Feb-99 Mar-99 Apr-99 May-99 Jun-99 Jul-99 Au9-9 Monlh Se[v99 Ocl-99 Nov-99 Dec-99 Figure I. The mtiin iiiiinthl> water temperature and salinity of the Skidaway River. (Jeorgia from November I'WS to November IW^ (vertical bars indicate ± 1 standard error from the mean). Ocl 99 No. -99 Figure 3. The reproductive phases (early active = 1, activity develop- ing = 2, near ripe = 3. ripe = 4, partially spawned = 5. and spent = 6) of female and male crested oysters, Ostrea eqitestris. from the coastal waters of Georgia between December 1998 and November 1999. DISCUSSION Male indi\ iduals occurred every moiuh and totaled 284 indi\ idiials (49.5%), while only 20 female oysters (3.5<7r) occurred. There were no female oysters identified in the December 1998, March 1999, April 1999 and September 1999 samples. Two hundred forty four oysters (42.5%) were hermaphroditic. Hermaphrodites oc- curred in each month sampled. Brooding larvae in the samples were observed from April to October. Crested oysters in Georgia were reproductively active year around (Fig. 3). Near ripe and spawning ripe female and male oysters were found in all months. Female and male spawning and spent oysters occurred year-round with these stages dominating from March through December. Male spawning and spent stages ranged from 23.7% of the oysters in January to 86.4% in August, while in females these stages ranged from 10.7% in February to 77.8% in August. Brooding larvae were found in 3.1% of the histological samples occurring in each month from April to Octo- ber. Spawning appears to be a prolonged event in Georgia starting in Februarv and continuing into fall. Dec-98 Jan-99 Feb-99 Mar-99 Apr-99 May-99 Jun-99 Jul- Lug-99 Sep-99 Oct-99 Nov-99 Figure 2. The mean shell length (mm ± standard error) of crested oysters from pearl nets suspended in the Skidaway River, Georgia from December 1998 to No\ ember 1999. Within the Ostreidae family, oysters of the genus Ostrea are brooding species, whereas, members of the Cnissoslreci genus re- lease their gametes into the water column (Galtsoff 1964). Brooded larvae have been documented in Ostrea equcstris by Gut- sell ( 1926) and Menzel ( 1955). and were noted in the present study from histological samples taken between April and October. Water temperature is a controlling factor in regulating spawn- ing and gametogenesis in marine bivalves (Giese 1959. Sastry 1975). In general spawning periods and gametogensis in marine bivalves start earlier and last longer in southern geographical areas than in northern ones (Eversole 1989. Thompson et al. 1996). Ostrea equestris in the subtropical waters of Georgia are repro- ductively active year around (Fig. 3). Although Menzel ( 1955) did not follow gametogensis of O. equestris on an annual basis, he showed that they continually recruited from the first of May uinil the first of November indicating prolonged spawning seasoning in Louisiana. Other warm water Ostrea species also have long spawning periods and periods of no reproductive inactivity. In Costa Rica. O. irulescens was observed spawning 9 out of 14 months, with no reproductive resting stages observed. In New Zealand. Tiostrea {= Ostrea) chilensis spawns year around with peak spawning in late summer-autumn and brood larvae found from July to January (Jeffs 1 998. Jeffs & Hickman 2000). In colder waters. O. chilensis has one spawning period per year in spring to early summer with brooding lasting only 7 to 8 weeks (Chaparro & Thompson 1998). Likewise, in Argentina O. puelchana spawns from December to February; however, no resting stages of repro- ductive activity were observed with rapid proliferation of ovocytes occurring in March-April (Castro &. Mattio 1987). In California O. litrida is reproductively active year around, but spawns only during .seven months with peak spawning in June and July (Coe 1932a). Spawning in O. Iiirida ceases when water temperatures fall below 16°C (Coe 1932b). In Ireland. O. edulis spawns from May to September with minimum reproductive activity from October to February (Wilson & Simons 1985). Ostrea equestris is a protandric species as are other species of Ostrea. Menzel ( 1955) followed the growth and reproductive cycle of crested oysters fioni set and observed that 74.4% of the oysters were males and 10% females. Of the remaining oysters. 13% were GRDWTH and G.AMETOGENIC CYCLE OF CRESTED OYSTERS 947 indeterminate and only 2.6% were hermaphroditic. In our study, males al.so dominated (49.5%); however, the percentage of her- maphroditic indi\iduals (42.5%) was much greater than that ob- served in crested oysters from Louisiana. The crested oysters from Georgia were well beyond the set stage when collected from the Savannah River. Since the crested oyster is protandric as suggested by Menzel (1955), then this could explain the difference in the percent hermaphroditism between populations. The oysters from Louisiana were younger and developing first as males, thus the 747f males, before changing to females or hermaphrodites at a later age or size. Oysters gathered in December 1998 in Georgia already ranged in size from 8.7 mm to 34 mm in shell length, and had probably already past through the initial se,\ change stage from males to females. In oyster populations, the closer the pro.ximity of one oyster to another, the higher the proportion of males in that population (Burkenroad 1931, Smith 1949, Menzel 1951, Buroker 1983). In the present study, crested oysters were held close to- gether in suspended pearl nets, which may explain the much skewed male to female ratio observed (1:0.07). Oysters of the genus Ostrea have greater occurrences of her- maphrodites than species of Crassostrea (Table I ). Galtsoff (1964) stated that oviparous species of oysters (Crassostrea) are not usu- ally hermaphroditic, while larviparous species of oysters (Ostrea) as a rule change in sexuality (i.e., from an initial male phase followed by alternating female and male phases). Ostrea ediilis has long been recognized as a protandric species: spat sexually mature as males, spawn, then switch to a female phase. At age one year, these oysters enter a male phase followed again by a female phase. TABLE L .Sex ratios and percent hermaphroditism in oyster species. Family Genus species Number Sex ratio (F:M» Hermaphrodite Source Ostreiiiae Criiwostrca aiii^itUtUi 1.00:0.79 Pelseneer 1926 Crci.\siistrea coinmeniiilis 0.05:1.00 1.00:0.37 Roughley 1933 Crassostrea cucidlata 333 1,00:1.(10 0.9 Lasiak 1986 Crassostrea gigas 120 0.8 Aniemiya 1929 77(1 1 .00:0.8 1 Buroker 1983 976 1.00:0.90 115 1.00:0.37 Mori 1982 in: 1.00:1.17 100 1.00:0.96 1377 1.00:1.00 0.23 Steele & Mulcahy 1999 Crassostrea glomerata 121 2.00: 1 .00 0 Ansari & Ahmed 1972 Crassostrea gryphoicles 925 1.00:0.92 Durve 1965 Crassostrea luadrasensis 1.00:1.16 1.00:1.11 1.50:1.00 Stephen 1980 Crassostrea paraihaiieii.sis 1.00:1.08 1.00:1.09 Singarajah 1980 Crassostrea rhizophorae 226 1.00:2.32 1.00:3.00 Veiez 1976 Urpi et al. 1983 1833 l.0():().26 Littlewood & Gordon 1988 5.36 1.00:1.00 1 .00: 1 .00 Gruzetal. 1989 Velez 1991 Crassostrea tutipa 1.00:1.00 Yankson 1996 Crassostrea virginiea 1.00:1.00 Pelseneer 1926 744 1 .00: 1 .00 Burkenroad 1931 1408 1 .00: 1 .26 1.5 Menzel 1951 1.00:1.00 <1 Kennedy 1982 6.50:1.00 3.00:1.00 Heffernan et al. 1989 Ostrea comtnereiaVis 2.70:1.00 Roughley 1933 Oslrea ecjiieslris 469 1 .00:7.43 2.6 Menzel 1955 574 1.00:14.2 42.5 This Study Ostrea forkskali 43 2.60: 1 .00 Hulings 1986 Ostrea irklescens 470 1.00:3.09 2 Fournier 1992 Ostrea luriilu 238 90.0 Coe 1932b Ostrea inadrasensis 1.00:1.17 Stephen 1980 Ostrea pitekhana 109 84 1 .00:0.79 1.00:0.12 Morriconi & Calvo 1989 Saecoslrea ciuiilhita 497 1.78:1.00 1.00:1.38 Braley 1982 Morton 1990 Tiostrea chilensis 1542 1.00:0.15 96.1 Jeffs 1998 1S16 1.00:7.38 71.9 Jeffs & Hickman 2000 948 Walker and Power Most two-year-old oysters are feniales. In addition, female stages for this species have been observed developing within the gonads as the older gonadal material was releasing sperm (Cole 1942). Mann ( 1979) observed 29% hermaphroditic individuals in O. edii- lis while finding no hermaphrodites among C. f;igas. For O. luriJa. Coe (1932a) found that of 238 oysters 90% was hermaphrodites. Foamier ( 1992) also found a higher ratio of males to females (3:1) in O. iiidescens with 2% hermaphroditism. Castro and Mattio (1987) described O. puekhana as a rhythmic consecutive her- maphroditic species. In Tiostrea chilensis populations, hermaph- roditic individuals accounted for 72% and 96% of the oysters sampled in New Zealand (Table 1: Jeffs 1998, Jeffs & Hickman 2000. respectively). The growth of crested oysters in Georgia was slow — a rate of 0.9 mm per month. The largest oyster observed was 36. 1 mm in the September sample, which is half that of the maximum reported size (82 mm in shell length) of the crested oyster (Galtsoff & Merrill 1962). In the Gulf of Mexico region, Menzel (1955) showed that 12 oysters grew from a mean size of 13.1 mm to 29 mm in 9 months — a rate of 1 .76 mm per month. The observed slow growth of the Georgia crested oyster is maybe a result of compe- tition for both space and food. Crested oysters were gathered from a natural set in the Savannah River. When oysters were brought to the surface, relic C. viii;inica shells were already densely covered by crested oysters. No attempt to reduce the densities of the shells was made prior to or during the time that the crested oysters were held in pearl nets at the main dock on the Skidaway River. Thus, oysters that were collected in an overcrowded state remained in dense numbers throughout the sampling period. High stocking densities of bivalves result in slower growth rates (Sheldon 1968. Walker 1984. Holliday et al. 1993. Jensen 1993. Adams et al. 1994). LITERATURE CITED Abbott. R. T. 1974. The American Seasliell. Second Edition. New York: Van Nostrand Reinhold Co. Adams. M. P.. R. L. Walker & P. B. Heffernan. 1994. The effects of stocking density, bag mesh size, and bottom sediment on the growth and survival of the eastern oyster. Ciassostrea virginica. with special emphasis on controlling oyster spat fouling. J. Applied Acjiuiculturc 4:25^t4. Amemiya, I. 1929. On the sex-change of the Japanese common oyster Ostrea gigus Thunberg. Proc. Imp. Acad. (Japan). 5:284-286. Ansari. F. & M. Ahmed. 1972. Seasonal gonadal changes in the oyster Crassostrea ghmerata Gould. Paki.slan J. Zool. 4:35^3. Braley, R. D. 1982. Reproductive periodicity in the indigenous oyster. Saccostrea cuctdkila in Sasa Bay. Apra Harbor. Guam. .Mar. Biol. 69:165-173. Buroker, N. E. 1983. Sexuality with respect to shell length and group size in the Japanese oyster, Cras.<:o. 5 individuals per station for each of the three sites) were used in these analyses. For each species, the number of fish caught per gill net deployment were compared with an ANOVA using site, day of the year, and time of day as factors. Data for bluefish, striped bass, and weakfish met both the assumptions of homogeneity of vari- ance and normality after transformation with the reciprocal trans- formation (Zar 1996). Data for croaker and spot satisfied the as- sumptions of homogeneity of variance and normality after loga- rithmic transformation. Counts for Atlantic menhaden satisfied the assumption of homogeneity of variance with the reciprocal trans- formation but not normality regardless of the transformation (log + 1. In -I- I. sqrt + 1. reciprocal). Fish Assemblage-Habitat Relationships Transient fish species abundance associations were compared across sites using detrended correspondence analysis (DCA). DCA was used as a descriptive tool to characterize the fish assemblages observed at each site on the basis of abundance. DCA ordinations spatially aggregate similar samples and separate dissimilar ones on the basis of species abundances within a sample. All DCA analyses (CANOCO for Windows version 4.0 1998) were detrended with second order polynomials (per ter Braak 1995) to avoid potential loss of gradient information during the detrending procedure (Minchin 1987). Species-samples biplots were made using CANO- DRAW software (version 3.1. Similauer 1998). Species-specific Length Data Total lengths (mm) for the six numerically dominant species were compared with species-specific one-way ANOVAs using site as a factor. RESULTS Analyses Piankatank River Temperature and Salinity Data Neither water temperatures nor salinity values were signifi- cantly different among sampling sites in 1997 (ANOVA, p < 0.05). Water temperature and salinity conditions observed in the Pianka- tank River during 1997 were similar to those observed during 1993-96 (Fig. 2. R. Mann, unpublished data). Species-specific Abundance Data Fourteen different transient fish species were observed in gill net collections from Palace Bar reef (Table 1 ). Ten of these four- teen species were observed at Ginney Point (oyster shell bar) and nine were observed at Roane Point (sand bar). Atlantic croaker, Atlantic menhaden, bluefish, spot, striped bass, and weakfish were the most abundant fish species at all three sites (Table I ). O-iSTHR Reefs as Fish Habitat 953 Figure 1. Map of the Piankutank River in relation to the Chesapeake Bay showins; saniplin)" locations after Harding and Mann (19991. Palace Bar reef (C), Ginney Point (an oyster shell bar. A) and Roane Point (a sand bar, Bl were sampled to provide data for reef vs. non-reef habitat comparisons. Abundances of Atlantic croaker, Atlantic menhaiien, ancj striped bass were significantly greater at sites with oyster shell substrate (Palace Bar reef and Ginney Point) than at the sand bar site (Roane Point) but there was no significant difference in abun- dance of these three species between the oyster reef and the oyster bar (Table 2; ANOVA. Tukey test, p < 0.05; Figs. 3. 4, and 7). Bluefish were significantly more abundant at the oyster reef than at any other site (Table 2; ANOVA. Tukey test, p < 0.05). Spot were significantly more abundant at the oyster bar than at either the oyster reef or the sand bar (Table 2; ANOVA. Tukey test, p < 0.05). Weakfish abundance was low relative to the other species and similar across all three sites (Table 2; ANOVA. Tukey's test, p < 0.05). In general, fish abundance increased at night across all sites. Atlantic croaker, bluefish. and spot were significantly more abun- dant from dusk to dawn (2000-0800) than during the day (Figs. 3. 5. 6; ANOVA. Tukey's test, p < 0.05). Striped bass were signifi- cantly more abundant from dusk to dawn than at mid-day ( 1 200- 1600; Fig. 7, ANOVA. Tukey test, p < 0.05). Atlantic menhaden and weakfish were significantly more abundant during darkness (2000-0800); abundances observed between midnight and 0400 were higher than at any other time for both menhaden and weak- fish (Figs. 4 and 8; ANOVA. Tukey's test, p < 0.05). Fish abundances varied seasonally. Bluefish were significantly more abundant in May and September than from June to August (Fig. 5. ANOVA, Tukey's test, p < 0.05). Striped bass and Atlantic menhaden were significantly more abundant in May than at any other time during the year and more abundant in late June than during late July and August (Figs. 4 and 7; ANOVA. Tukey's test, p < 0.05). Weakfish were significantly more abundant in late July (Fig. 8, ANOVA, Tukey's test, p < 0.05). Atlantic croaker abun- dance was significantly greater during July and early August while 22 20 H. 18 3 2:- 16 14 12 10 30 u 28 a 26 3 f. 24 i) a. 22 (1) 1-. 20 cd ^ 18 16 _L May Jun Jul Aug 1997 Sep Oct Figure 2. a-Mean salinity (ppt) and b-water temperature (°C) values (±standard error) for Ginney Point and Palace Bar reef. Piankatank River, Virginia from May through September 1997 after Harding and Mann (2001 ). Data from these two sites were averaged since there was no significant difference in temperature or salinity between sites (ANOVA. p < 0.05). Reference mean values for temperature and sa- linity data from 1993-1996 are plotted with a solid line (± standard error), 1997 data are indicated by lines with symbols (± standard error). 954 Harding and Mann TABLE 1. Total number of transient fish species collected with yill nets at Palace Bar oyster reef. Ginney Point (oyster ban, and Roane Point (sand bar), Piankatank River. \ irginia during 8 thirty-six-hour stations conducted from May 22 to September 3. 19^7. Common Name Scientific Name Palace Bar Reef Ginney Point Roane Point Atlantic croaker Micropogonias itndiilcitiis 121 120 70 Atlantic menhaden Brevoortia tyrannus 480 455 195 Bluefish Pomulomus sahatrix 65 35 20 Spot Leiostomus xanthunis 221 258 150 Striped bass Moroiie saxatilis 62 98 10 Weaicfish Cynoscion regalis 14 11 7 Blueback herring Alosa aestivalis 3 5 4 Buttert'ish Peprihis triaianthus 1 0 0 Cownose ray Rhinoptera bonasus 1 0 0 Gizzard shad Dorosoma cepedianiim 1 0 0 Hog choker Trinecles iiuuiilatiis 3 0 0 Silver perch Bairdiflla chxrsotira 38 3 2 Spotted seatrout Cvnoscion netnilosu!, 4 8 4 Summer flounder Paralichthyes dfnuinis 0 2 0 spot were significantly less abundant in Aug(.ist (Figs. 3 and 6: ANOVA. Tukey's test, p < 0.03). Fish Assemblage-Habitat Relationships A detrended correspondence analysis (DCA) using all samples and all species (Fig. 9) aggregated all but one of fourteen species (summer flounder) and all but two of 231 samples (the two samples containing flounder) from all three sites along a single axis virtually on top of each other. This cohesive spatial grouping indicates strong similarity of most species and samples across all sites. Axis I describes a gradient in diurnal light levels moving from left (dark) to right (light). Axis II represents a seasonal gra- dient in water temperatures moving from bottom (lower water temperatures) to top (warmest water temperatures). The variance explained by the axes was 21.6'7r (Axis I) and 38.2"^ (Axis 11). If rare species or species where the total number of fish ob- served across all three sites was less than fifteen are removed from the analysis, eight species remain (Table 1 ). A second DCA using only these eight species in the gill net samples (Fig. 10) shows a lack of spatial aggregation of samples by site in ordination space as would be expected by site-specific fish assemblages. Thus, the samples from all three sites show a ubiquitous distribution. Axis I represents a gradient in diurnal light levels moving from left (dark) to right (light). Axis 11 represents a seasonal gradient in water temperatures moving from bottom (lower water temperatures) to top (warmest water temperatures). Fishes that were more abundant from dusk to dawn during late May, June, and early September (spot, bluefish) are grouped toward the middle of the plot to the left of fishes that were more abundant from dusk to dawn in July (silver perch, weakfish; Fig. 10). Primarily nocturnal species (Atlantic menhaden and spotted seatrout) are grouped to the left (dark) side of Axis I. Striped bass were most abundant in May and early June during daylight hours as indicated by their position in the lower right corner of the plot (Fig. 10). Atlantic croaker were frequently caught between dawn and dusk during the warmer months as indicated by their position in the upper right comer of the plot (Fig. 10). The variance explained by the axes was 28.5% (Axis I) and 48.8% (Axis II). Species-specific l^ength Data Atlantic croaker. Atlantic menhaden, and striped bass observed at Palace Bar reef are significantly larger than fishes of these species observed from either the oyster bar or the sand bar (Table 3: ANOVA, Tukey"s test, p < 0.05). Spot from the oyster bar are larger than spot from any other site (Table 3; ANOVA, Tukey's test, p < 0.()5). Bluefish from the reef are slightly but not signifi- cantly larger than fish from other sites and weakfish from all sites are of similar length (Table 3: ANOVA, Tukey's test, p > 0.05). DISCUSSION There was no clear delineation of habitat use by transient fishes along a gradient of estuarine habitat types (oyster reef to sand bar). Atlantic croaker, Atlantic menhaden, bluefish, silver perch, spot, spotted seatrout, striped bass, and weakfish were found in all habi- tat types examined. The ubiquitous distribution of these common species indicates a lack of site-specific fish assemblages in these habitats. It is unreasonable to expect site-specific groupings of TABLE 2, Summary of ANOVA results (p-values) for species-specific abundance (number of a species collected per gill net deployment) of the six most abundant transient fish species observed in the Piankatank River in relation to site, day of the year, and time of day. Asterisks indicate results that were significant at the p < 0.1)5 level. Factor df Atlantic Croaker Atlantic Menhaden Bluefish Spot Striped Bass Weakfish Site Day of the year Time of day 0.01* 0.01* <0.01* 0.01* 0.02* <0.01* <0.01* 0.12 <0.01* 0.02* <0.01* <0.01* <0.01 <0.01* <0.01* <0.0I* 0.01* <0.01 Oystkr Reefs as Fish Habitat 955 □ Roane Point ^ Ginney Point ■ Palace Bar reef r- A. "n^ir^l 'Mh I n-^M : r-^« i r^if-^ G. rv-^ r^Mlrr-^B iJ, i.l .^1^ I Qs^Mi JBLL^EsBB. '■ ^^^ I mi^Ji I .... I I .... I r ^ .— , r^-^ I I . ^ I 0000-0400 O40O-O8OO 0800-1200 1200-1600 1600-2000 2000-2400 Time of day Figure 3. Species-specific abundance for Atlantic croalier in relation to time of day and day of the year for A.) May 22-23, B.) June 5-6. C.) June 19-20, D.) July 2-3, K.) July 17-18, F.) .\ugust 4-5, G.l August 18-19, and H.) September 2-3. 1997. □ Roane Point 100 „ A. Ginney Point Palace Bar reef 187 222 100 ^ C. 80 60 40 20 0 ^_K: r-^ i .... I .... I 100 ^ D 80 60 ;_ 40 20 [ n 1.^-6??^ 1 .... I ' 100 E. 80 - 60 40 20 0 I .... I .... 1 .... I .... I 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 r- F. r- G. I irniiM I ii 1^ i . Ks?a . i . . . , i 1 .... I I .... I _^L 0000-0400 0400-0800 0800-1200 1200-1600 1600-2000 2000-2400 Time of day Figure 4. Species-specific abundance for Atlantic menhaden in rela- tion to time of day and day of the year for A.) May 22-23. B.I June 5-6. CM June 19-2(1. D.I July 2-3. K.l July 17-18. F.I August 4-5. G.l August 18-19, and H.I September 2-3, 1997. generalist species such as these that are opportunistically using available habitat. It is more likely that habitat use by these eight t1sh species relates to variations in habitat quality indicated by habitat-specific productivity. In general, the smallest fish are found on the sand bar. the site with the least habitat heterogeneity. As habitat complexity in- creases along the gradient from oyster shell bar through oyster reef, transient fish size and abundance increases. The oyster reef may ha\e relatively higher food availability, a wider diversity of food types because of increased habitat heterogeneity, or greater abundance of high quality food relative to other habitat types. Dietary analyses on bluefish (Harding & Mann 2000) and striped bass (Harding & Mann, unpublished data) from these sites cor- roborate these functional relationships between reef habitats and transient fishes. Bluefish from sites with oyster shell substrate consume more teleosts than bluefish from the sand bar (Harding & Mann 2001 ). Bluefish from Palace Bar reef consume a wider di- versity of prey items than fish from other sites (Harding & Mann 2(J0I) while reef striped bass consumed more teleosts in general and naked gobies in particular than tlsh from other sites (Harding & Mann, unpublished data). In other words, the observed differ- ences in fish abundance and size across habitat types may relate to habitat producti\itv as enhanced by ecological and structural com- plexity. Presence/absence and abundance data from this study demon- strate that these transient finfish employ generalist lifestyle strat- egies (Sale 1980) and are opportunistically using the range of available habitat on a local scale. The habitats of interest herein 956 Harding and Mann □ Roane Point ^ Gimey Point ■ Palace Bar reef 10 8 6 4 2 r A. n i^ i . , . i , , , , 1 1 ^B i m, . i i rrn ^" 10 8 6 F^ . r i i i 4 2 0 - 1 R?3 1 1 I^T^sH i 1 iH ^ . . .^ rr-i. .B i I , i.\\.\ , I m, , . I £iz^a_j_i_ I rri. . . I . g^ . I 10 - G. 8 E. 6 L 4 ■_ 2 0 ■ . . . SSS3 . \ . SSSi 10 ^H. I rrksya «iUJ=jj □ Roane Point ^ Ginney Point ■ Palace Bar reef A. .rJ, 1^^ ll. K^^^l .... lrv-i-„ .1.^.1 ES. 60 C. 50 40 30 20 lU i n m™ r-^B I .... I ....'.... 1 1 60 50 40 30 20 10 0 i .... I .... i .... i — R^ H J K^:^ I I I ESS— 1 \ .i^M,.l 60 F. 50 i 40 30 20 10 n -..1,-^,1 ,,,,I,.^,I.L_. G 60 50 40 30 20 10 n '' , . 1 , ^^ , 1 , . , , r^-^^,^ ,^.™_ 60 , H. 50 40 30 20 10 : n r-r-W?^ 0000-0400 0400-0800 0800-1200 1200-1600 1600-2000 2000-2400 LJ_EZ:tssHL_ I ■ ... I r^Mi 0000-0400 0400-0800 0800-1200 1200-1600 1600-2000 2000-2400 Time of day Figure 5. Species-specific abundance for bluefish in relation (o time of day and day of the year for \.) May 22-23, B.) June 5-6, C.) June 19-20, D.) July 2-3, K.) July 17-18, F.) August 4-5, G.) August 18-19. and H.) September 2-3, 1997. represent a gradient or dine of habitat complexity eomnionly ob- served in temperate estuaries; namely a cline moving from simple, unstructured hard sand bottom habitats through hard bottom shell habitats with little vertical relief culminating in complex, three- dimensional reef structures created and maintained by oysters. These biogenic reef structures naturally ranged in size from acres to hectares and historically were dominant habitat types in Chesa- peake Bay. This gradient of habitat types is a temperate analog to tropical coral reef systems ranging in scale from patch reefs through much larger reef systems (e.g., the Great Barrier Reef). The transient fish communities associated with temperate and tropical reef habitats are composed primarily of generalists that will opportunistically use available habitat (Sale 1980, Ebiing & Hixon 1993, Roberts Time of day Figure 6. Species-specific abundance for spot in relation to time of day and day of the year for A.) May 22-23. B.l June 5-h, C.) June 19-20, D.» July 2-3, E.) July 17-18. F.) August 4-5, G.I August 18-19, and H.) .September 2-3, 1997. 199,^). The structural and ecological complexity of reef habitats makes them attractive foraging habitat for transient finfish as well as aggregation sites. Historically, shallow portions of Chesapeake Bay were characterized by a mosaic of habitat types including biogenic structure ranging from seagrass beds to oyster reefs ex- tending across spatial scales ranging from kilometers to 10s of kilometers. The development of large biogenic reef structures was facilitated by the evolution of the Chesapeake Bay estuary (Hargis 1999). The parallel development of the Bay's fish fauna favored transient fishes with broad habitat and dietary requirements (gen- eralists) that were able to opportunistically use the dynamic estuarine habitat. These fishes successfully use the modem Chesapeake habitat in spile of relatively recent habitat alterations, namely the decline of both seagrass beds and oyster reefs during the late 20th century. Oyster Rhefs as Fish Habitat 957 n Roane Point 20 ^ A. 15 10 5 0 Ginney Point J_ .^ Palace Bar reef _^aKix^^^Hj 20 ^ B. 15 10 5 0 ' ■ l^^^^ I , , ,^ I , , , , I . lit^ . I , 20 D. 15 L 10 L 0 L_g ■ I I I ■ ■ I I ■ tysi ■ I 20 E. 15 10 ^ 5 0 , l>^Nl , I 20 F ^■ 15 L j 10 L i 5 L n : .... 1 ... . r^. .. 1 .... 1 .... 1 .... 1 20 15 10 5 0 20 15 10 5 0 G pH. 1 1 1^ 1 1 1 1 i j 1 — 1 0000-0400 0400-0800 0800-1200 1200-1600 1600-2000 2000-2400 Time of day Figure 7. Species-specific abundance for striped bass in relation to time of day and day of the year for A.) May 22-23, B.i June 5-6, C.l June iy-2(), I).) July 2-3. K.l July 17-18. F.) August 4-5. G.) August 18-19. and H.l September 2-3. 1W7. Previous discussions of oyster reef habitats as essential fish habitat for transient finfish (Breitburg & Miller 1998. Coen el al. 1999) have examined fish species richness data from a geographic range of oyster reef habitats including both natural and restored reefs of varying ages. Coen et al. (1999) suggest that the use of oyster reef habitats by transient fish species "portends the reef habitats" importance as essential fish habitat, but many functional relationships remain to be evaluated". This study presents a unique comparison of transient fish use of oyster reefs in relation to other locally available habitat types and is the first to provide data to describe fish habitat use at Level 1 (presence/absence). Level 2 (abundance) and Level 3 (size) levels of EFH designation. These data clearly show that these transient generalist fishes do not rely exclusively on oyster reef habitats. From a local historical per- □ Roane Point ^ Ginney Point ■ Palace Bar reef r ^- ^ i 1 1 1 1 1 . . . i . . . . .... 1 r^-,. . , 1 20 B. 15 L 1 j 10 - ! 1 5 L 0 :, ... i .... i .... i .... i .... i .... i I I . I I I . . ■ ■ I . ■ . ■ I . . . . I I .... I .... I ■ ... I .■■. I 20 15 10 5 0 20 15 10 5 0 20 15 10 5 0 r F- ■: . I I I ■ ■ ■ . I ^H. 1 m, iBJ 1 1 1 j .,1.1,1,1 i i 1 ■ . — i 0O00-O4O0 0400-0800 0800-1200 1200-1600 1600-2000 2000-2400 Tline of day Figure 8. Species-specific abundance for weakfish in relation to lime of day and day of the year for A.) May 22-23. B.I June 5-6. C.) June 19-2(1. D.) July 2-3. E.) July 17-18, F.( August 4-5, C;.l August 18-19. and H.) September 2-3. 1997. spective. the continued presence of these species in the lower Chesapeake in the absence of natural oyster reefs for the past 20-(- years (Hargis 19991 is an obvious indicator that oyster reef habitat is not essential for these opportunistic fishes. The habitat value of oyster reefs to transient fishes is much more complicated than a binary distinction (essential or not essen- tial). Evaluations of oyster reefs as fish habitat must consider reefs in the context of locally available habitat types (per Minello 1999; this study) if accurate descriptions of habitat importance are to be made, particularly for transient finfish species. Continued exami- nation of the functional ecological relationships between oyster reefs and the trophic communities that they support will provide data on which to base habitat distinctions at all four levels of EFH description and related resource management decisions. Gradients 958 Harding and Mann +4,0 X < "tST Atjant_ic_ menhaden Airantic croaker Silver perch Bluefish. cownose ray, hog choker, spot, weakfish Blueback hemng -3.0 Butterfish Seatrout Gizzard shad Summer flounder Striped bass 1.1.1.1,1.1,1,1,1,1.1.1.1.1,1,1,1,1,1.1.1,1.1.1,1,1.1.1,1,1.1.1.1,1,1,1,1,1.1,1, 1,1,1,1,1,1,1,1,1,1 25,2'2"2^2■2■2^?,2;2^^2■2,?,2■2,■2■2?2^y,2;2^?,2^2.Y,0,2■,2^2,Y.^2,?.2■l■2■,2? -12,2,2,2.2,2,2,2 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3.3,3,3,3,3,3,3,3,3,3,3,3,3,3, 3,3,3,3,3,3,3 ® -5.0 +40.0 Axis I Figure 9. Specie.s-saniple biplot for detrended correspondt'iicc anal>sts iDCA) describing transient finfish assemblages and species abundances across a gradient of habitat types ranging from sand bar through three dimensional oyster reef. Fourteen species from two hundred and thirty-one samples collected at Palace Bar reef ( I), Roane Point (2 1, and Ginney Point {3) with gill nets are presented. Axis I represents a gradient in diurnal light levels moving from left (dark) to right (light). Axis II represents a seasonal gradient In water temperatures moving from bottom (lower water temperatures) to top (warmest water temperatures). + 2.5 -3.0 ® ® ^®'® ® ©1.1.1,2 3,3,3.3 ' ® @ ® ® GL Striped bass -1.0 Axis I ■2.0 Figure 10. Species-sample biplot for detrended correspondence analy- ses (DCA) of common fish species across a gradient of habitat types ranging from sand bar through three dimensional oyster reef. Eight species from 201 samples collected at Palace Bar reef (1), Roane Point (2), and Ginney Point (3) with gill nets are presented. Axis I represents a gradient In diurnal light levels moving from left (dark) to right (light). Axis II represents a seasonal gradient in water temperatures moving from bottom (lower water temperatures) to top (warmest wa- ter temperatures). TABLE 3. Average total length (mm) of the most common transient fish species (standard error) collected with gill nets at Palace Bar oyster reef, Ginney Point, and Roane Point, Plankatank River, Virginia. Site-specific species total lengths were compared with species-specific ANOVAs. Horizontal lines under site-specific species average lengths values indicate sites where statistically similar sizes of a particular species were observed (ANOVA, Fisher's test; p<0.05). Fish Palace Bar Reef Ginney Point Roane Point Atlantic coaker Atlantic nienliaden Bluefish Spot Striped bass Weakfish 311.6 (3.3) 262.1 (I.M) 307,1 (.5.7) 29.'i.l (2.,S) 240.2 (3.9) 246.8(1.6) 298.3 (5.8) 239.7(1.0) 297.1 (6.1) 199.1 (1.6) 294.5 (5.1) 2864 (8.9) 205.1 (1.4) 261.7(3.3) 198.5(1.7) 278.6(10.7) 3124(20.8) .302(20.7) in physical habitat complexity relate to gradients in habitat pro- ductivity and thus habitat value or importance. A gradient of tenns to describe habitat value that reflects the ecological value of a habitat would be a more realistic tool for habitat distinction. Given their physical and trophic complexity, oyster reefs are important habitat for transient estuarine finfish. however, on the basis of these data, we question the use of term "essential" with regard to oyster reef habitats given the generalist nature of the transient fish species that use these habitats. We suggest that oyster reef habitats are not essential for these fishes but that oyster reef habitats are of higher quality than other locally available estuarine habitat types and thus are better or perhaps even optimal for these fish in terms of growth, reproductive success, and survival. ACKNOWLEDGMENTS Support for this project was provided by the U.S. Environmen- tal Protection Agency Chesapeake Bay Program (CB993267-02-1, CB993267-03. and CB993777-0I) and the Virginia Institute of Marine Science Department of Fisheries Science. The following Oyster Reefs as Fish Habitat 939 individuals graciously donated llicir time to assist with field sam- pling: F. Arazyus, I. Bartol. R. Boger. S. Brooke. J. Brust. N. Clark. V. Clark. C. Conrath. S. Drake. J. Duggan. K. Farn- swortli. C. Fennessy. J. Coins. S. Goodbred, A. Hayes. E. Hinchey. K. Hovel. D. Kerstetter. T. Lewis. K. Murray. J. Nestlerode. S. Neubauer. J. Newton. J. Oliver. C. Robertson. V. Rose, H. Simpkins. M. Soutliuorth. C. Squyars. M. Thompson. B. Trainum, M. Wagner. K. Walker, J. Walter. G. White, and H. Yarnall. J. Duggan provided valuable assistance designing and building the gill net reel used to store and deploy nets. This is Contribution No. 2429 from the Virginia Institute of Marine Sci- ence. College of William and Mary. Gloucester Point. VA. LITERATURE CITED Ahk', K. 1999. Measures of juvenile fish habitat quality: examples from a National Estuarine Research Reserve. In: L. Benaka. editor. Fish habi- tat: es.sential fish habitat and rehabilitation. Bethesda. Maryland: American Fisheries Society. Symposium 22. pp. 1,14-147. Bahr. L. M. & W. P. Lanier. 1981. The ecology of intertidal oyster reefs of the South .Atlantic coast: A community profile. U.S. Fish and Wildlife Service. Washington D.C.: Office of Biological Services. FWS/OSB/- 81/15. 105 pp. Benaka, L. 1999. Fish habitat: essential fish habitat and rehabilitation. Bethesda. Maryland: American Fisheries Society. Symposium 22. 45*} pp, Breitburg, D. 1999. Are three dimensional structure and healthy oyster populations the keys to an ecologically interesting and important fish community. In: M. W. Luckenbach. R. Mann & J. A. Wesson, editors. Oyster reef habitat restoration. A synopsis and synthesis of approaches. Gloucester Point. Virginia: Virginia Institute of Marine Science Press, pp. 2.^9-250. Breitburg. D. & T. Miller. 1998. Are oyster reefs essential fish habitat ' Use of oyster reefs by ecologically and commercially important species. J. of Shellfish Res. 17:1293. Burchmore. J. J.. D. A. Pollard, J. D. Bell. M. J. Middleton, B. C. Pease & J. Matthews. 1985. An ecological comparison of artificial and natural rock reef fish communities in Botany Bay. New South Wales. Austra- lia. Bull. Mar. Sci. 37:70-85. CANOCO for Windows 4.0 1998. Center for Biometry. CPRO-DLO. Wageningen. The Netherlands. Chen. H.. P. Hyer. A. Kuo & C. Fang. 1977. Hydrography and hydrody- namics of Virginia estuaries XI. Mathematical model studies of water quality of the Piankatank estuary. Special report No. 1 24 in Applied Science and Ocean Engineering. Gloucester Point. Virginia: Virginia Institute of Marine Science. 80 pp. Coen. L., M. Luckenbach & D. Breitburg. 1999. The role of oyster reefs as essential fish habitat: a review of current knowledge and some new perspectives. In: L. R. Benaka, editor. Fish habitat: Es.sential fish habi- tat and rehabilitation. Bethesda. Maryland: American Fisheries Society. Symposium 22. pp. 438-454. Coen. L. & M. Luckenbach. 2000. Developing success criteria and goals for evaluating shellfish habitat restoration: Ecological function or re- source exploitation? Ecological Engineering. 15:323-343. Ebeling, A. W. & M. A. Hixon. 1991. Tropical and temperate reef fishes: a comparison of community structures. In: P. F. Sale, editor. The Ecol- ogy of Fishes on Coral Reefs. San Diego. California: Academic Press, pp. 509-563. Harding. J. M. & R. Mann. 1999. Fish species nchness in relation to restored oyster reefs, Piankatank River. Virginia. Bull. Mar. Sci. 61: 289-300. Harding. J. M. & R. Mann. 2000. Naked goby [Gohiosonia h clionid sponges. For more information on approaches to tapho- nomic analysis, we refer the reader to Davies et al. (1990). Swept Area and Catch per I'liit Effort (CPl'E) Analysis Swept area w as calculated lor each dredge tow from the 5-sec position logs and the dredge width. Site coverage was estimated by dividing the site into 10 x 10-m squares, 100 nr in area, and calculating the total swept area in each of these squares. The choice of a 10 x 10-ni square was made primarily for ease of data analysis and presentation; however, the dimensions were consid- ered to be the smallest areal dimension that was distinctly larger than the accuracy for which dredge position was known. Dredge position was judged to be accurate within 5 to 10 m. Accuracy was Swept Area: Grid 36, April 30, 2000 -1 — I — I — I — I — I — \ — I — I — I r 10 12 14 16 18 20 22 24 26 28 30 3 2 4 6 Position (10-m increments) Fiyurt- 2. t'o>erage map of grid 36 (one of the two experimental sites) for an 8-h dredging event that occurred on No\eml)er 21, 19W. Posi- tion is identified on the .v and v axes as the number of lO-m intervals with the (O.O) location in the northwest corner. Contours are the total swept areas dredged in each 10 x l()-m section. Contours were gener- ated from the total estimated coverage for each 10 x 10-m square. or abundant. The longest dimension of each oyster and box >2() mm was measured. The number of live and dead spat <2() mm was counted. Wet meat weights were taken for each live oyster. At the same time, a random sample of 10 oysters per bushel was selected for P. marinus analysis. Mantle and rectal tissue were taken for analysis of infection intensity following standard procedures (Powell & Ellis 1998). A taphonomic assessment was made for each box and live oyster. The following types of damage were recorded as present or absent on each individual; minor chipping on the shell margin, major chipping on the shell margin, presence of perforations through which tissue could be observed or. for boxes, through which the inside of the shell could be observed, cracking without breakage, and presence of abrasions. Minor and major chipping were distinguished by assigning to the major category cases where small chips were distributed around most of the shell inargin or cases where larger pieces of shell were missing. Abrasions were T — r 4 26 28 30 32 Position (10-m increments) Figure 3. Coverage map of grid 36 (one of the two experimental sites) for an 8-h dredging event that occurred on April 30, 2000. Position is identified on the .v and v axes as the number of 10-m intervals with the (0.0) location in the northwest corner, ((mtours are the total swept areas dredged in each 10 x 10-ni section. Contours were generated from the total estimated coverage for each 10 x 10-m square. 964 Powell et al. greater for the two-dredge boat because the dredge was released port and starboard forward of amidships, whereas the single- dredge boat released the dredge from the stern. Coverage was assumed to be 100% if the total swept area in a 10 x 10-m square during a single daily event reached 100 m'. Samples for analysis were obtained from typical dredge tows. A typical dredge tow covered about X)0 to 600 m" and intersected 40 to 4,'i 10 X 10-m squares. Individual 10 x 10-m squares were not sampled for this report. CPUE was calculated as the number t)f bushel sacks per square meter of dredged area. The number of market-size oysters per sack averaged ,'^18. Statistical Analysis Most statistical analyses used ANOVA with all interaction terms. Block (the site pairs), treatment (experimentals and con- trols), and in some cases, time were class variables. For time trends in CPUE. a Spearman's rank correlation was used. For ANOVA analyses, data on number and volume were normalized to the total volume of shell per bushel (the sum of the volumes of live oysters, boxes, and shell). This was done because the bushel sample did not come from a known area — the dredge catches much more than one bushel — and the volume of debris including biont overgrowth was not a constant over the study, whereas to a first approximation, the volume of shell was. To further assess the influence of using differing sample sizes, we also conducted analyses on ratios of variables, such as the ratio of live oysters to boxes. Such ratios are not influenced by the volume of sample collected. With the ex- ception of presence-absence data that were recorded as a 0 (absent) or a 1 (present), all variables were ranked prior to analyses. Con- sequently, all statistical analyses were effectively nonparametric. Finally, analyses using variables that potentially were influenced by size, such as Dermo infection intensity, condition index, etc., included size class as a class-variable covariate (Underwood 1997). Size classes were juvenile (20-6.^.4 mm), submarket (63.5- 76.2 mm), and market (>76.2 mm). RESULTS Swept Area Coverage The one-dredge boat addressed an average of 37.039 m" of bottom duriiit! a sinele 8-h dav. The two-dredae boat addressed 13- c 17H (U S 4 u §21-1 225-1 c o •§29 o D- 33-1 37- 41 45- Swept Area: Grid 10, November 20, 1999 -« ; \ 1 e- 1 -T-<^ 13 I 21 45 1 — I — I — I — I — I — I — r»-{ — I — r 17 21 25 29 33 37 41 Position (10-m increments) Fljjure 4. Coverage map of grid 10 dint' of tho two experimental sites) for an 8-h dredging event that occurred on November 2(1, 1999. Position is identified 1 00 m- per 10 x lO-m square, meaning that 50% to >I00% of a 10 X lO-m square near the epicenter was covered by the dredge during that day. Exainples of individual trips are shown in Figures 2 through 5. Total coverage for the study is depicted in Figures 6 and 7. In the heavily dredged central areas of each experimental site, coverage in a 10 x lO-m square varied from 300 to >600 ni" by the end of the study. That is, an area equivalent to the entire area of a 10 x lO-m square was impacted in that 10 x lO-m square by the dredge three to six times during the study. Although we cannot unequivocally show that each square meter in the 10 X 10-m square was addressed by the dredge three to six times because the position of the dredge on the bottom is not known precisely, certainly we can assume that the entire 10 x 10-m area of such a square was impacted multiple times. Daily and Cumulative CPVE CPUE was calculated each time a bushel sack was filled during the day to determine whether the rate of capture declined (Fig. 8). CPUE did not change significantly over the course of the day for any of the 10 dredging e\'ents (five 8-h events, two experimental sites; Spearman's rank correlation, a = 0.05). Site Status at Study Inccptiiin All four sites, the two experimental sites and the two control sites, were sampled prior to the initiation of dredging. ANOVA analyses revealed that the experimental sites diverged from the control sites significantly for a few variables, including the amount of fouling (P = 0.02), taphonomic status (boxes only, P = 0.0001), and condition index (P = 0.0001). The significant dif- ference in taphonomic status came from significant differences in Swept Area: Grid 10, July 4, 2000 "I I I — I — I — I — I — I — I — I — I — I — r-»i — 1 — 1 — I — I — I — I — r 3 5 7 9 II 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 Position (10-m increments) Figure 5. Coverage map of grid 10 (one of the two experimental sites) for an 8-h dredging event that occurred on July 4, 2000. Position is identified on the .v and y axes as the numher of 10-m intervals with the (0,0) location in the northwest corner. Contours are the total swept areas dredged in each 10 x 10-m section. Contours were generated from the total estimated coverage lor each 10 x 10-m square. 966 Powell et al. Total Swept Area for Experimental Grid 36 1 3 5 7 9 11 \^ 15 17 19 21 23 25 27 29 Position ( 10-m increments) Figure 6. Cumulative coverage map of grid 36 (one of tlie l«o experi- mental sites) for the entire study (five 8-h dredging events = 40 h of dredging). Position is identified on tlie -»• and v axes as the number of 10-ni intervals with the ((1,0) location in the northwest corner. Con- tours are the total sviept areas dredged in each 10 x 10-m section. Contours were generated from the total estimated cov erage for each 10 X 10-m square. iiiidor chipping and abrasion. Because of the iniperf'ect nature of the controls (e.g., Lindegarth et al. 2000), the evaluation of the effects of dredging will focus on the significance of interaction terms between treatment and time in further analyses. Immediate Impuct of a Dredging Event The question of the immediate inipact of dredging was assessed at the time of the first two dredging events, which took place within 10 days of each other in late October/eaily November 1999. In this case, the experimental sites were sampled before the first event and after the second. In only one case was a uniform effect on the two experimental sites found. Minor chipping of live oysters increased after di'edging on both experimental sites (Table 1 ). In general, the incidence of dredge damage increased in live oysters and boxes after dredging (Table 1). In general, the increase in dredge damage was similar for boxes and live oysters. However, 16 h of dredging did not result in significant changes in oyster health as measured by mor- tality, P. maiiniis infection intensity, or condition index. Such a result might be expected, however, because the samples were taken immediately after the second 8 h of dredging had been completed. Influence of Coverage Coverage, defined as the total swept area for the dredge, in- creased at each experimental site with each dredging event in the time series (Fig. 9). This increase was not uniform over the ex- perimental site, however. Dredging tended to be heavier in some portions of the site than in others (Figs. 2-5). As a result, when a sample was taken, the coverage for the specific area sampled dif- fered, sometimes considerably, between the three replicate samples. This local variation in coverage within an experimental site might diffeienlially influence each of the three samples taken each time period. Because the market-size oysters are sorted from the total dredge haul onboard the boat while the boat is underway, and the remaining material subsequently redistributed in a quasi- random manner over the site, one might not expect significant relationships to exist between the coverage of any specific set of 10 X lO-ni squares and the status of the oysters, boxes, and shell located there after dredging had ceased. This is particularly true in our case where any one sampling tow intersected 40 to 45 10 x lO-m squares over the 300 to 600 m distance of the tow. Never- theless, if such a relationship did exist, coverage of necessity would have to be included in the statistical analysis as a covariate. Accordingly, we examined the degree to which the three samples at a given site and sampling period differed as a function of the coverage recorded for the specific area of the site from which each sample was taken. Instances where differences in coverage between the three samples significantly impacted a variable within each site and sampluig time did not occur more often than expected by chance (binomial test, a = 0.05), as anticipated from the dredging and onboard culling process and the length of the sampling tows. Ac- cordingly, coverage was not included as a covariate in further analyses. Cumulative Impact of Dredging As time passed and more dredging events occurred on the experimental sites, the difference between the control sites and experimental sites might be expected to diverge. Such variance would be evidence of long-term impacts of dredging. Accordingly, we examined the lime series for significant interaction terms be- tween time and treatment. Because we anticipated that the simi- larity of the control and experimental sites should decrease over time as dredging impact accumulated on the experimental sites, we also examined the level of significance achieved in comparisons of control and experimental sites at each individual time period when a significant ■'time*treatment" inteiaction occuned. We expected the difference between control and experimental sites to reach increasingly higher levels of significance as time passed and the cumulative impact of the dredge increased. Changes in taphonomic status can be expected to occur as Repeated Dredging on Delaware Ba^ Oyster Reef 967 Total Swept Area for Experimental Grid 10 1 — r 29 31 15 17 19 21 23 25 27 Position (10-m increments) T 37 T 43 45 Figure 7. Cumulative coverage map of grid 10 (one of the two experimental sites) for the entire study (five 8-h dredging events = 40 h of dredging!. Position is identifled on the .v and v axes as the numher of lO-m intervals with the (0,0) location in the northwest corner. Contours are the total swept areas dredged in each 10 x 10-m section. Contours were generated from the total estimated coverage for each 10 x 10-m square. dredging occurs. Indeed, the time*treatment interaction term was significant for total taphonomic impact and for many of the indi- vidual taphonomic indicators for live oysters and for boxes (Table 2 1. However, a consistent pattern over time did not occur. Oysters from experimental sites were more heavily damaged in November and July and less heavily damaged in April, for example (Table 2. Fig. 10). Minor chipping rose slightly over the winter and declined into the summer, presumably as the oysters were able to add new shell, thereby removing the evidence of old chips (Fig. 1 1 ). Oys- ters from the control sites and experimental sites followed very similar patterns in that the incidence of minor chipping declined simultaneously in both. Similarly, abrasive wear rose throughout the study for oysters from both the experimental and control sites (Fig. 12). Oysters from the experimental and control sites differed significantly in the incidence of abrasion in late November. April, and July, but the relationship was inconsistent over time. Abrasion was more common in oysters from control sites than from experi- mental sites in April and was less common at the other two times. More severe forms of shell damage such as breakage, cracking, and shell perforation occurred larely (Fig. 13), and were not sta- tistically related to the presence or absence of dredging. Boxes followed a similar pattern. Boxes from the control sites diverged from the experimentals at the beginning of the experi- ment, but subsequently did not differ much (Fig. 10). The instance of minor chipping declined in the summer (Fig. 1 1 ). The frequency of abrasion rose over the course of the experiment, though less regularly than for live oysters, and boxes from experimental sites diverged from the control boxes only at the inception of the study (Fig. 12). Not surprisingly, considering the similarity of the boxes and live oysters throughout the study, the ratio between the two for the various taphonomic indicators varied little (Table 2, Fig. 14). Oysters from the control and experimental sites differed little beyond the taphonomic indicators of diedge damage. Time*treat- ment interaction terms were not significant for the ratio of live oysters to boxes, indicating that the pattern of mortality did not change between control and experimental sites (Fig. 15). The ratio 96S Powell et al. Grid 10 Catch Per Unit Effort November 11, 1999 0.004- 0.000- November20, 1999 0.007- I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 5 10 15 20 25 30 # Time Segments 0.000- III I I I I I I I I I I M II I I II I I M M I I I I I I I I I I I I I I 2 10 18 26 34 42 50 58 66 74 82 # Time Segments Grid 36 Catcii Per Unit Effort November 21, 1999 0.004 0.000 1 M I I I M I I I I I I M I I I I M I M 2 10 18 26 34 42 50 # Time Segments 0.005 0.004 0.003 0.002 0.00 April 30, 2000 II I II 2 6 I I I I M I I I I I II M I M I 10 14 18 22 # Time Segments Figure 8. Examples of time trends in CPUE for daily dredging events. CPUE is measured in bushels caught per square meter of swept area. The X a.xis records consecutive time segments when one or more bushel sacks were filled. Each CPUE estimate is based on the swept area during that time segment. of juveniles to submarkets and submarkets to market-size oysters also did not change, indicating that the processes of growth and mortality changed in similar ways in oysters from control and experimental sites. The time*lreatnient interaction was significant for condition index, but this result was principally due to the divergence of the experimental and control oysters at the begin- ning of the study (Table 2). Subsequently, condition index de- clined slightly over winter in both treatments, rose somewhat into the summer, and then declined again, as is typically the case in Delaware Bay (Fig. 16). Thus, the experimental and control sites yielded very similar values of condition index throughout the study. DISCUSSION Coverage and Short- Term Impaets Total dredge coverage for the study was about 240,000 ni" on each experimental site (Fig. 9). The most heavily dredged areas were completely covered by the dredge four to six times during the study. This degree of impact is somewhat greater than is typical for the Delaware Bay seed beds under the direct-market harvesting scenario that exists today. Estimates of swept-area coverage in 2000 by industry vessels during the fishing season range from about 50% to 800% of bed area (the latter indicating that the bed Repeated Dredging on Delaware Bay Oyster Reef 969 TABLH 1. Results of ANOV'A tests iiii time for the llrst t"() dredf-inj; events that took plaee within 10 da>s of each other in late Ottoher/earlj November. 1999, in which the condition of the site before and immediatclv after 16 h of dredging was compared. P values under the time and time*site columns record the results of ANOVA analysis for the variable time (before vs. after dredging) and the interaction term with site (the tv\o experimental sites!. The column "'.'Site" records which of the two sites. El or E2, were signincantl\ different. The directional arrows indicate whether dredging resulted in an increase (upl or decrease (down) in the variable. Only taphonomic indicators showing significant results are shown. N.S. not significant (a = (1.05). Variable Time Time*,Site ?Site Direction Total live oysters Total boxes Live/Dead ratio Number drills Fouling status Number li\e spat Number dead spat Condition index Pcrkiusus marinits infection intensity Taphonomic status, live Minor chipping Abrasion Shell perforation Taphonoinic status, dead Minor chipping Major chipping Abrasion Taphonomic status, ratio hve/dead NS NS 0.03 NS NS NS NS NS NS 0.0001 0.0001 0.005 0.02 0.0001 0.004 0.02 0.0001 NS NS NS NS NS NS NS 0.04 0.000.^ NS 0.02 NS NS NS 0.0004 NS NS 0.0007 NS E2 E2 E2 E2 E1,E2 E2 E2 E2 El El E2 was completely dredged eight times in 2000). depending on the seed bed, but only one bed exceeded 150% of bed area ( 1.5 times; Banta et al. tinpubl. data). Average coverage for the fished oyster beds in 2000 was 849f . Thus, our study produced dredge swept- area coverages of approximately double to triple the oyster fish- ery's impact on an average heavily fished oyster bed in Delaware Bay. It follows that any impacts recorded in this study are likely, on the average, to be more severe than typically observed under normal fishing activities in Delaware Bay. Sixteen hours of dredging, two 8-h events within a 10-day period at the beginning of the study, produced barely detectable changes in the oyster population. Given the rigorous nature of dredging, one might anticipate that shell cracking and breakage would occur often enough to be readily detected. In fact, these 250000 -^200000- 0) Q. CD g CO > E o 150000 100000- 50000- I Experimental 1 ^1 Experimental 2 ^E. im^ 12 3 4 5 Time Figure 9. Cumulative swept-area coverage for the two experimental sites. Time segments are defined in Figure 10. severe forms of impact occurred rarely. The incidence of cracking and breakage did not normally exceed about 2% of the live oysters, and major chipping was restricted to about 10'7r of the population. These more serious forms of shell damage did not increase in frequency after 16 h of dredging. Minor chipping and abrasion were more common, about 70% of the oyster population had minor chipping and a more variable number. 10% to 50%. showed evi- dence of abrasion. These latter two less serious forms of dredge damage did increase in frequency after a 16-h period of dredging. However, no other discernible impacts were found. Long- Term Impact: Dredge Damage Dredge coverage accumulated at an average of about 47.500 m~ per 8-h event on each experimental site (Fig. 9). Although the impact of a single event was barely measurable, dredge datnage might be expected to accutnulate over the study. No such trend was observed. Many of the taphonomic indicators of dredge damage showed time-dependent trends that differed between control and experi- mental sites. However, for the most part, these effects were limited to minor chipping and indications of abrasive wear, rather than the more serious aspects of shell damage defined as major chipping, breakage, cracking, and shell perforation. Furthermore, the antic- ipated divergence between control and experimental sites was most noticeable on the sampling dates thai immediately preceded and followed the winter months. By summer, most differences had disappeared. Two possible reasons exist for the failure of dredge damage to accumulate over the study as might have been anticipated. ( 1 ) The taphonomic indicators rnight represent damage that occurs from sources other than dredging. Certainly, chipping and breakage can be predator induced, for example (McDemiott 1960, LaBarbera 970 Powell et al. n ^ * - -5. = ^ =S! s 2 S a- ■? -S S- = o »; S a i n I c ; ^ "^ i^ ."' 3 I- — c a^ r •= 0- '- aj C ^^ — s n = C ■2 S: „ 1- ^ E •^ ^ :P a> 4!2 E '3.1 o 0/ c X 3 ZL _aJ ?:- ■o n 2 a. 3 ^ n = 3 ^ 4^ 0^ wl Si: Q. o 3 E 3 ■a M 3 o U 1 r^i U. n &i S u CO Q. 3 OX 1 5C' < ■c 3 C 0.* II c n 3 S 44 3 S _= ^ ■o c 'E 0£ n ■■*■ 3 ^ ^ ,^_, £ 3 1 C 3 > 3. a.< — 3 cj: ^ O' c E ^ ^ L. a 'Si: 0^ u -o X (U 3j 3 'x 5 ^ 11 X H _3 It v .O ^- 8. E X 3 — £ K CO = a n -S £ -< C5 c L. > •" =2 O z L. ■q. E £ < .£ X **- =3 X _s^ ■o c o 3 c 3 © VI 3 n 0^ X < a^ u > ■C < ^w > Z :5 o •O < it z C c o IE =3 u u ^^^^ x/ ^ 0.800- T3 ^ 0.600- n 0.400- 0.200- —A- Control n Don — •— Experimental 1 1 1 1 1 3 4 Time 1,400- ^i!^-^ 1.200- ▲ ^^""^ ^^ .^^^ 1.000- 0.800- 0.600- 0.400- 0.200- -A- Control n nnn — •— Experimental 1 1 I 1 1 3 4 Time Fijjuri' 10. Tinie-de|K'iulfnl chaiijjfN in the tapli(in(imic condition of live oy.sters and boxes during the .stud.v. Tuplionomic condition is calculated as the sum of minor chipping;, major chipping, ahrasion, hreakage, cracking, and shell perforation. For example, if two of these conditions were present, the sum recorded would take the value of 2. Time segments are: 1. October 30-31, 1999; 2, November 11-13. 1999; 3, November 20-21, 1999; 4. April 2S-3(). 2000; 5. Jul> 2-1. 2000; 6. August 31. 200(1. 1981, Elner & Luvoie 19S3). (2) Rapid repair of niiiior shell dam- age may mitigate the dredging impact. Repair should occur most rapidly during the warmer months and. in fact, the incidence of minor chipping decreased substantially during the summer. This decrease suggests that rapid shell repair during the warmer months is an important mechanism minimizing the observable impact K^'i dredging, although some effect on the animal's energy budget, in diverting energy to shell growth, might be anticipated (Bernard 1474). The incidence of abrasive wear increased throughout the study on both sites (Fig. 12). Abrasion on the shell surface cannot be repaired and so would be expected to accumulate. However, the fact that this taphonomic indicator increased in both the experi- mental and control oysters suggests that its origin is something other than dredge damage. Increased clionid activity was noted during the study (Fig. 17) and wciuld explain this observation. Juvenile, submarket, and market-size oysters might be expected to differ in their sensitiv ity to shell damage by dredging. No evi- dence could be obtained for a differential effect in this study, however. Lons-Term Impact: Population Health One assumes that oysters, like other bivalves, respond at the physiological level to the stress of capture and release (Pekkarinen & Suoranta 1995). No physiological indices of stress were evalu- ated in this study. However, a variety of population health indica- tors were assayed during the study that should integrate changes in underlying physiological state via measures of growth, disease pressure, and mortality. These variables included the ratio of live oysters to boxes, condition index. P. nuuimis infection intensity, and ratios between oyster size classes. Essentially no significant effects could be discerned for any of these measures. The ratio of live oysters to boxes remained rela- tively similar between control and experimental sites (Fig. 15), Even when the oyster and box populations were apportioned into three size classes, the ratio of live oysters to boxes was not intlu- LIVE OYSTER BOX 0.8- ▲ ^/*^^t. cn c 8:0.6- ./ ^^-, E o c o Q.0.8 (0 XJ 0.6- f Shellfish Research. Vol. 20, No. 3. 977-4X9, 2001 . A fishp:ries model for managing the oyster fishery during times OF disease JOHN M. KLINCK,' ERIC N. POWELL,' JOHN N. KRAEUTER,- SUSAN E. FORD," KATHRYN A. ASHTON-ALCOX^ ^Center for Coastal Pliysical Oceanography. Crittenton Hall. Old Dominion University. Norfolk. Virginia 23529: 'Haskin Shellfish Research Laboratory, Rutgers University, 6959 Miller Ave., Port Norris. Ne^v Jersex 0S.U9 ABSTRACT Setting the yeuily ailocatum lor a fished stock is always an uncertain endeavor. Populations suffering significant mortality from disease require particularly careful management. Disease mortality is not a standard component of fisheries models, however. Here, we develop a model for the management of fished oyster populations in which disease mortality is a controlling influence. The model requires a quantitative estimate of abundance by size class, some knowledge about growth rates to establish the size range recruiting into the fishery, and an estimate of the anticipated natural mortality rate. The latter is of considerable consequence because small changes in mortality rate effect large changes in predicted outcomes. The model permits investigation of scenarios that include a range of allocations, timing of fishing seasons, variation in fishing efforts within sea.sons to establish a preferred harvest level, variations in the distribution of fishing among beds to minimize overharvesting of disease-affected beds (area management), and rebuilding plans to increase total stock abundance after epizootic mortality or periods of overharvesting. The model is sufficiently general that it can be applied to any commercial shellfish species. Simulations show that appropriate timing of the fishing season with respect to the timing of disease mortality can more than double the yearly allocation to the fishery. Some harvested animals would otherwise have died from disease. Besides disease, the other model parameter that most affects simulation outcome is the abundance of submarket-size oysters that can be expected to recruit to the fishery in the simulated year. Population stability is strongly determined by the number of recruits available to replace the deaths that decimate the market-size population each year. The number of recruits is a function of survivorship in previous years, but also the anticipated growth rate that defines the size range of oysters at the beginning of the fishing year that can be expected to recruit to the fishery. This modeling exercise points to the critical need to understand population dynamics and survival of size classes below market size that are not often the targets of investigatory activities. A'£)' WORDS: fisheries models, management, mortalitv rate, disease, Dermo, stock assessment INTRODUCTION At otie time. Delaware Bay was one of the leading producers of American oysters (Cras.wstrea virginica) in the United States (Mackenzie Jr. 1996, Ford 1997). Like many other East coast estuaries, diseases, principally caused by the protozoan parasites Haplosporidium nelsoni and Perkinsiis marinas, have seriously reduced oyster abundance and, consequently, the oyster industry has declined (Ford 1997. Fegley et al. 1994). The diseases have also necessitated changes in the way oysters are harvested. At one time, oystermen transplanted oysters from upbay seed beds down- bay to leased grounds for grow-out. These higher-.salinity leased grounds have fallen into disuse or production on them has been seriously curtailed by the high rates of disease mortality of oysters planted there. Consequently, new approaches have been proposed (Ford & Haskin 1988. Ford 1992, Powell et al. 1997). One of these options is to concentrate all phases of production on the upestuary seed beds where lower salinity limits disease-produced mortality, at least to sorne extent. Concentrating all phases of production on the seed beds necessitates increased vigilance by oyster tiianagers (HSRL. 2000. 20011. In Delaware Bay, the highest quality oysters are produced natu- rally on the lower third of the seed beds (Fig. I ). Consequently . in recent years, oyster production has focused on two production schemes: ( I ) direct marketing of oysters produced on these lower seed beds, essentially a wild fishery; and (2) transplanting oysters downbay to these lower seed beds with subsequent recapture for sale. Upbay oysters tend to be smaller and of lower tnarket quality. Both conditions improve within a few months to a year when moved downbay (Powell et al. 1997). One necessary consequence of both of these approaches is a need to estimate the total allowable catch (TAC) from the seed beds each year, essentially the setting of a yearly allocation for transplant downbay and for direct mar- keting (HSRL. 2000, 2001). The problem of setting an allocation is an interesting one be- cause oyster diseases produce high natural mortality rates that are relatively unpredictable from one year to the next (Jordan 1995, Powell et al. 1996, Brousseau et al. 1998). This obviates the pos- sibility managing under standard fisheries guidelines that focus on biomass at maximum sustainable yield (B,„„.) and normally define S,„„ as one-half of the carrying capacity (A72) (May et al. 1978. Applegate et al. 1998. Restrepo et al. 1998). Disease mortality is not a standard component of fisheries models. Even in cases where natural mortality may be principally caused by disease, a standard fisheries modeling approach has been used (Allen. 1979). Nor is it clear that K is a concept that can be applied straightforwardly to populations whose dynamics are inherently controlled by disease. Population models addressing the issue of carrying capacity and disease, such as described by Heesterbeek and Roberts ( 1995) and Swinton and Anderson ( 1993). have not yet been applied to marine species, although the need for doing so continues to increase (Har- vell et al. 1999). What seems clear is that oyster populations are well below pre-disease levels in many Mid-Atlantic bays today (Andrews 1988. 1996. Gottlieb & Schweighofer 1996. Ford 1997) and that recovery to pre-disease levels is unlikely with the continued pres- ence of disease. The purpose of this contribution is to develop a model for the management of fished oyster populations in which disease mortality is a controlling influence and for which manage- ment options cannot be described in terms of fi„„, and A'. We use, as an example, the oyster populations and oyster fishery of Dela- ware Bay (Fig. I ). Because many of the same issues exist for a 977 978 Klinck et al. DELAWARE RIVER NEW JERSEY DELAWARE Figure 1. Lucatiun of the oyster seed beds in Delaware Bay referred to in this report. number of commercial shellfish species where disease mortality is significant (e.g.. Sousa 1991. Miller & Lawrenz-Miller 1993. Meyer et al. 1993. Park et al. 1999, Bower et al. 1999. Sunila et al. 1999). this modeling approach may ha\e a much wider applica- bility. The Model The modeling exercise begins with the familiar equation dS dt = -ZV (1) that describes a change in abundance (iV) as a function of mortality rate (Z). Equation ( 1 ) is solved under the boundary condition that N = N„ at r = 0: Mr) = N„e.- (2) where t is time in days. For a fished species, total mortality rate. Z can be decomposed into a natural mortality rate. in. and a mortality rate due to fishing. /; Z = in + f. Thus, N(t) = N.e' +fi' (3) Application of equation (3 1 to management of the Delaware Bay oyster industry faces a number of serious problems, all of which stem from the time-dependency of the two mortality rates. The first of these problems is that the natural mortality rate. in. is principally determined by Dermo disease. Dermo disease intensi- fies during the summer months, generally culminating in increased mortality in late summer and fall (Andrews 1988, O'Beim et al. 1994. Powell et al. 1996. Oliver et al. 1998. Brousseau & Baglivo 2000). Thus, m is not constant o\er the year. Secondly, the fishery does not impact the oyster population equivalently throughout the year. Typically, a spring and a fall season are separated by a summer period when harvesting is minimal and fishing normalh ceases during the winter (HSRL. 2000. 2001). Equation (3) as- sumes that the two monality processes, fishing (/) and natural mortality (in), covary such that neither varies seasonally or, if they do vary seasonally, both rise and fall proportionately during the year. This is patently not the case for the natural mortality rate of most estuarine shellfish and rarely the case for fishing mortality for most estuarine shellfisheries. Finally, some portion of fishing mor- tality is likely to be compensatory because some of the oysters harvested during the spring and early summer would ha\e died from Dermo disease in the following late summer and autumn. Whether different sources of mortality are additive or compensa- tory is a critical issue in the population dynamics of species where mortality from parasitism and disease is important (Jakobsen et al. 1988. MeiBner & Bick 1997. see also Thomas & Kunin 1999). In this case, knowing how the two sources of mortality, from disease and fishing, interact may be particularly important because one of them, fishing, can be manipulated in time to maximize the com- pensatory component and, consequently, reduce the effect of the fishery on the stock. A more complex formulation than equation (3) is needed to address issues of seasonality in mortality rates and the interaction between mortality rates. The following model is designed to re- solve the issue of time-dependent mortality rates and compensa- tion between mortality rates, while retaining the desirable charac- teristic of allowing an analytical solution. For simplicity in model construction, we have chosen to define the process of recruitment into the fishery as a rate, like mortality. This requires that the rate of recruitment into the fishery be a function of the number of market-size animals present: however, a standard stock- recruitment relationship (Hilbom & Walters 1992) is rarely, if ever, present (Loosanoff 1966, Hofstetter 1977, Hofmann et al. 1992. Whitlaich & Osman. 1994). In the Delaware Bay fishery, market-size oysters are typically >76 mm in length, although no regulatory size limit is in force. Oyster growth rates vary within and between bays along the East and Gulf coasts (O'Beim et al. 1996. Mackenzie & Wakida-Kusunoki 1997. Dittman et al. 1998). On the direct-market beds of Delaware Bay. growth rates are such that oysters of about 63.5-76 mm in January , defined as submar- kets, can be expected to recruit to the fishery in the coming year. Because no regulatory size limit exists and oysters recruit to the fishery in an uneven and unpredictable way (non-knife edge), we resolve the issue of dependency of recruitment on the abundance of market-size oysters by defining recruitment rate such that the recruits reach market size on the first day of the year. This permits the dependency to be artificially defined at the beginning of each simulated year by appropriate choice of recruitment rate (R) with respect to market-size abundance (A'^). Accordingly, equation (3) describing the change in abundance over time is expanded to dN — = [-Z[t) + R(t)]N (41 where Z represents the sum of a series of time-dependent mortality rates. For the purposes of the present application to management of the oyster fishery in Delaware Bay. Z includes/,, the fishing rate during the first (spring) season, /,, the fishing rate during the A FiSHHRiEs Model for Managing During Disease 979 second (autumn) season, and in. the non-t'ishing mortality rate (mainly Dermo-produced during late summer and early fall). Each of these processes is active over only part of the year. Hence, for market-size oysters dN -— = [-/•,(;) - f.(t) - iMt) + Rin] N. at The general solution for this model is (5) (6) where A',, is the number of market-si/e oysters present at the be- iiinnina of the vear and Ail) = J ,',[-/i(t) -./:(t) - m(T) + R{j)](h. (7) The population ks invariant over a year if A{T) = 0, where T = 1 year (or 365 days since t is in days). When A(T) = 0, the various removal processes just match recruitment mlo the fishery over the year. The effect of each term included in the right side of equation (7) can be calculated by integrating the right side of equation (5) with respect to that term. For example, the cumulative number of oysters per area removed by fishing during the spring season (0^,(7)) is ^':'^ = /v' /■.(tIMtV/t. (8) The structure of the various terms is required to proceed. First let recruitment R be an impulse or delta function at the beginning of the year, or R{t) = R„ bit - e) (9) where /?,, is the total recruitment rate and e is a small number to have recruitment occur shortly after / = 0. If the recruitment is expected to double the number of oysters at the beginning of the year, then i?„ = log, 2. This impulsive form resolves the issue of dependency of recruitment into the fishery on N. Each of the other terms is modeled with a box-like structure so that the process turns on and off at certain times. That is. each term turns on once and then turns off (Fisz. 2). This formulation allows 0 0 I I 1 5 10 I I I I I I I I I I I I 40 45 50 Figure 2. Examples of the time-dependent rate functions used for mortality rates (equation |10|). (irev line, the time-dependent fishinj; mortality rate, a = 0.5. Black line, the time-dependent Perkiiiiu\ muri- nils mortality rate, a = 10. more than one process to be on at a time, so the model is com- pletely general. Each term has the form 1 B(/:f, ./,) = - laiih ■ tanh (10) The 'box' function described by equation (10) has the value one- half at the times /, and r,. or Bit^ ) = B(r,) = 0.5 (Figure 2). The parameter <(, with units of days, controls how fast the process turn.s on or off. So. for small a (at a f,) = /, - 'i- That is. the integral of BU: r,. t^) beyond the end of the 'box' defined by /, and ?, is just the time the process is switched on, Tt - 'i I'll days). The various sources of mortality are represented as (J\.f\. m) = (./„„,/:„. m„)BU: t,. I,) (12) where ;, and /, are different for each process. So each of the fishing and mortality processes are controlled by a constant am- plitude times the 'box' function which has start and slop param- eters. Furthermore, the steady solution is constructed in parts as each of the processes can be integrated in time as f: (/,.,/,. m) e. But this is always true after the first day. Further development of the model requires establishment of a goal for the management program. One option, given the uncer- tainties in year-to-year variations in natural mortality that preclude evaluation of the stock abundance relative to a known value such as carrying capacity K. is to adopt the goal that market-size abun- dance should not decline over the year due to activities of the fishery. Stated mathematically; N,. = N, (15) The condition expressed by equation (15) necessitates that the rate processes in equation (7) sum to zero [AiT) = 0| or R„ -f\„BI{T: r,. t.) -t\„BhT: f,. /j) - m„B/(r; ?,. fj = 0. (16) where f,. t^. ',. '4. 's- and ^,, are the transition times for fishing and natural mortalitv. The form of the 'box' function means that each 980 Klinck et al. Dermo mortality Spring season Fall season I I II I I II I I! I I I I I I I I 250 300 350 Figure 3. An example time series of mortality rates showing the two independent fishing seasons and the lime of highest Perkiiisiis muniiiis induced mortality. lenii In equation ( 16) reduces to the time spLui that each process is on times a constant amplitude. An example time series is shown in Figure 3. So. equation (16) becomes A little reanansement aives hi - '"oCh - '5) = 0. <'2 - 'i!/i„ + ('4 - '31/2(1 = R.. - (',, - '>)'"(, (17) (18) Equation (18) gives a relationship between the two fishery rates compared to recruitment and natural mortality. The rest of the calculation is determined by the choice of the relationship between the two fishing mortalities. For example, sup- pose that the per day fishing rates for spring and fall are the same. This simply says that the daily effort and catch per unit effort are constant. In the case of the Delaware Bay oyster fishery, daily effort and CPUE (measured in terms of bushels d"') are very similar in the fall and spring and CPUE varies little between years because changes in stock abundance do not substantially influence CPUE (when measured in terms of bushels d~'). (Banta et al.. in press). Thus, /,y ~ /|,, and the fishing rates are easily estimated by: fw = l^„- ('(,- '5>"'o t,+t. (19) In some cases, such as after an epizootic, maintaining a stable population may not be desirable. In this case, the balance between recruitment and mortality established by equations (6) and ( 15-18) must be altered. Equation (6) becomes in (20) where t) is the desired ratio between the abundance of market-size oysters at the beginning of the year {N,^„) and at year's end {N,^^,^^y. A'j^i/W,^,^,, = T|. Equation (20) can be rearranged to Nit) = N/'""-'-"<''-^>l (21) Equation (17) then becomes /^,. -./i()('2 - 'i) -./2o('4 - '.^1 - '"(i('<, - '5) = loaM)- (--) Once again, letting /|q = Z,,, provides a relationship analogous to equation (19). /hi R-{t^-t^)m„-log,.{r\) TABLE 1. (23) The total numher of bushels of market-size and submarket-size oysters on the Delaware Bay oyster beds in late-October. IW*. listed in bushels based on an estimated 348 market-size oysters per bushel and 499 submarket-size oysters per bushel. Market-size Submarket-size Seed Bed Oysters Oysters Transplant Group Low-mortality Arnolds 3.182 1 1 ,646 Round Island 1 .y.'^4 3.847 Upper Arnolds 3,412 12.410 Medium-mortality Upper Middle 0 0 Middle 19,150 30.535 High-mortality Beadons 899 4.017 Naiituxent Point 3,047 8.691 Direct-market Group Medium-mortality Cohansey 72,936 127.380 Sea Breeze 11,226 6.246 Ship John 39,293 45.325 Shell Rock 8,258 21.620 High-mortality Bennies 7,470 16.632 Bennies Sand 951 1 .93 1 Egg Island 462 618 Hawk's Nest 9.447 16.774 Hog Shoal 1.681 1.857 New Beds 4,840 7.626 Strawberry 1.279 2.651 Vexton 1.815 I..S57 A Fisheries Model for Managing During Disease 981 60 >50 40 O ^ 30- H Z uj 20 O cr LU D. 10 UPPER - UlbUlJ 1 1 1 1 1 1 \ r 1 i 4 o CD CJ) CD CT> a5Cr)CDCT)0>cr)CT)a5a> o T_ IM CO 'J- in CO h- CD a> o m m m m fii m Cll CD m CD o O) cn CT> cr> o> en en CD CJl o CvJ 60 > F,0 1- 1 < 1- 40 CL O ^ 30 I- Z LU 20 cc Q. 10 UPPER CENTRAL T -r i 1 1 1 1 4 - m o T— CM ro en CD CD CT> c^ Ch cn CD CJ> CD CD CD CJ) CD CD CD CT) CD CJ) CJ) O O O CvJ Figure 4. Yearly estimated natural mortality as a percent iif total market-size abundance for the three bed groupinss defined in Table 1, for the decade of the IWds. Dermo became a significant source of mortality in iyy(»; as a consequence, mortality for years prior to ly^O is not included in the lime series. Uala The data come from the Delaware Bay oyster stock assessment program thai has produced a yearly autumnal survey of the seed beds since the early 1950s. A stratified random sampling method is used for the survey. Each bed (Fig. I ) is divided into a series of contiguous 25-acre grids that fall into one of three strata. The strata consist of "test" areas, typifying the highest quality areas of the bed that sampling over the course of many years has shown to have a high percentage of living oysters 75% or more of the time, "high quality" areas in which oysters were abundant 25-75% of the lime, and "low quality" areas in which oysters were abundant less than 25% of the time. The survey consists of about 100 samples cov- ering the primary and most of the minor oyster beds. Each sample represents a composite of 3 one-third bushels from three ran- domly-directed one-minute tows within each sampled grid. The survey instrument is a standard 1.27-m commercial dredge on a typical dredge boat, the /TV Howard W. Sockwell in this case. Oyster abundance is quantified in terms of numbers per m" by correcting the raw counts obtained from the three measured (by DGPS navigation) tows by a measure of dredge efficiency (Powell el al. submitted). Data from the 2000 survey are used in model simulations (Table I ). The seed beds were divided into three groups based on the level of natural mortality normally experienced during the 1990s (HSRL 2001). These groups will be referred to as the low-mortality, me- dium-mortality, and high-mortality groups (Table 1. Fig. 4). The bed groupings follow the salinity gradient, as anticipated from the importance of Dermo disease in the mortality pattern, such that the lowest natural mortality rates are on the upbay, lowest-salinity beds and the highest natural mortality rates are on the downbay. highest-salinity beds. The time series of natural mortality since 1 990, when Dermo began to become an important source of mor- tality in Delaware Bay. is characterized by two Dermo epizootics, an extended epizootic in 1991-1995 and a shorter epizootic in 1998-1999. Percentiles of natural mortality rate were calculated from the rank-order of yearly mortalities for U)9()-2000 obtained from this time series. The oyster seed beds were also divided Into two groups based on the estimated size of the smallest individual that could recruit to the fishery during one year. These sizes were 6.-i.5 mm for all high-mortality beds (listed in Table 1 ) with the exception of Shell Rock and 70.0 mm for all other beds (HSRL. 2001 ). Finally, some beds, as shown in Table 1 . are normally used in a transplant pro- gram; others are used lor direct-marketing. The distinction is un- important for this treatment, but is retained to maintain consistency with referenced sources (e.g.. HSRL, 2001 ). RESULTS Compensation Between h'ishinf; and \utural Morlality The impoitance of compensatory mortality was examined using a Dermo mortality rate characteristic of epizootic conditions on the high-mortality beds (Figure 4): 0.5 yr"'. Simulations were run for cases with two fishing seasions, the first opening and closing before 982 Klinck et al. P. murinus reaches levels capable of generating mortality in the population and the second varying in its timing with respect to the time period of highest mortality from disease (Table 2). Simula- tions were run under the provision of no net change in market-size abundance over the year. Values used for the simulations were the sums of the estimated abundances of market-size and subniarket- size oysters for the high-mortality direct-market beds (Table 1 ). In past years, these beds have supported the vast bulk of industry effort in Delaware Bay (Banta et al.. in press). Appropriate timing of the fishing season with respect to the timing of disease mortality can more than double a seasonal allo- cation to the fishery under the constant abundance scenario (Table 2). In the twenty simulations provided in Table 2. the highest yearly fishing allocation is 43.261 bushels. The lowest allocation is 33.969 bushels, about 20'7f less than the highest value. Differences in allocation often reach 50% within an individual fishing season. For example, in the first set of simulations in Table 2. the second fishing season was translated in time with respect to the timing of Dermo-induced mortality. Setting the second fishing season prior to the time of greatest disease mortality provides an allocation of 18.575 bushels. Setting the season after disease mortality ceases provides an allocation of 9.361 bushels. In each of these cases, population abundance at the end of the year is the .same. Not surprisingly, varying the timing of disease mortality relative to a designated fishing season, examples of which are shown in the second group of simulations in Table 2. has a similarly profound effect on the yearly allocation. In Delaware Bay. the oyster season normally opens in early April. Many times, the season has been split into two parts, one from April to June and one from September to November. An alternative is one long season beginning in April and extending into the summer or early autumn. To the extent that this season terminates prior to the late summer rise in disease mortality, the yearly allocation will increase over the alternative of a fall season. However, as the summer season begins to overlap the period of highest disease mortality, the yearly allocation will decrease as some oysters, otherwise harvestable, succumb to disease. An ex- treme example in which the fishing season is varied such that more and more of the season coincides with the time of greatest disease mortality is shown in the third group of simulations in Table 2. In this case, disea.se mortality occurs for a short, hut intense, time in late summer. Total lost allocation reaches about 16% when the fishing season extends into and after the season of most intense natural mortality. Historically, the fall har\est has obtained a higher market price. Hence economic issues encourage harvesting after Dermo disease has waned. However, this strategy minimizes the yearly allocation. Simulations in Table 2 suggest that the per-bushel dock-side price must at least double to compensate for the reduction in yearly allocation in epizootic years when the rate of natural mortality is high. A doubling of price is, in fact, rarely achieved, suggesting that the attraction of the higher market price in the fall is more often than not false economics in the oyster industry in these years. Harvest and. probably, dock-side value is maximized by concen- trating the fishery, to the extent possible, prior to the time that disease mortal itv increases in late summer. TABLE 2. The estimated yearly allocation for spring and autumn tlshinj; seasons during a year when 50'7f of the market-size oysters succumbed to Dermo disease. Day 1 was set at January 1, Data are presented in bushels. One bushel -37 L and contains approximately 348 market-size ovsters. Fishing Mortality Total .Season 1 Season 2 Dermo Mortality (In Bushels) (In Bushels) (In Bushels! (In Bushels) 43,261 24.686 1 8.575 28.797 41.872 24.686 17,186 30.185 38.307 24.686 13,621 33.750 35,080 24.686 10,394 36.977 34,048 24.686 9,361 38.010 33,961 24,600 9,361 38.096 35,080 24,686 10,394 36.977 38,307 24,686 13,621 33.750 41,872 24,686 17,186 30.185 43.261 24,686 1 8.574 28,795 43.261 24,686 18,575 28,797 42.176 20,287 21,888 29.881 40.044 17,214 22,829 32,013 37.927 14,948 22,978 34,1.30 36.228 13,208 23,020 35,829 42,560 24,685 17,874 29,497 40,526 24,686 15,840 31,531 38,039 24,686 13,353 34,018 35.932 24.686 11,246 36,125 34.5 1 2 24,686 4.826 .^7.-545 ^ear Days Y ear Days 90 150 150 240 330 Solid lines show the timine of the iwo fishina seasons. Dashed line shows the timing of Dermo mortality. A Fisheries Model eor Managing During Disease 983 VarUilidii in Disease Mortality Rate Timing of disease mortality varies from year to year and is often difficult to measure in a timely fashion for near-real-time re-evaluations of management decisions. Varying the duration of disease-mortality from 60 to 120 days, as shown in simulation sets 1 and 4 in Table 2. has only a minor intluence on cumulative yearly allocation, however, as long as the yearly natural mortality rate does not change. The daily rate is lower in the case of the longer time course, but the cumulative mortality integrated over the vear is about the same. Small variations in the natural mortality rate typically have profound consequences in fisheries models (Clark 1999), and this model is no exception. The yearly loss to Dermo disease is deter- mined by a number of factors, including temperature, salinity, food supply, and the history of disease within the populations, that determine late summer and fall infection intensity (Hofinann et al. 1995, Powell et al. 1996, Cook et al. 1998, Soniat et al. 1998). Variation in the rate of Dermo-induced mortality significantly af- fects the yearly allocation estimated by the model (simulation set I, Table 3). For example, the yearly allocation more than doubles with a decrease in natural mortality rate from 0.5 yr"' to 0.3 yr~ . This range of natural mortality rates is routinely observed in Dela- ware Bay (Fig. 4). Not surprisingly, the significance of compen- satory mortality in determining the preferred timing of the second fishing season declines at low natural mortality rates (simulations sets 1, 2, and 3. Table 3), The precautionary approach to fisheries management (FAO, 1995, Restrepo et al. 1998; see also Francis & Shotton 1997) would dictate the choice of a relatively high natural mortality rate in forecasting yearly allocations, perhaps 0.5 yr"' for Delaware Bay. Variation in the timing of the fall fishing season significantly affects the yearly allocation, a difference of about 8.000 bushels in the simulations presented in Table 3. In contrast, even a small drop in natural mortality rate, from 0.5 yr"' to 0.4 yr^'. increases the allocation by about 10.000 bushels. .\n increase to 0.6 yr"' essen- tially eliminates the yearly harvest. Thus, the model, and the popu- lation, are very sensitive to year-to-year variations in natural mor- tality rate. Precautionary management designed to minimize the likelihood of an underestimate in natural mortality rate is one consequence of this sensitivity. Increased predictability in the time course of Dermo disease, permitting a less conservative choice of natural mortality rate, would substantially increase the yearly al- location. Influence of Spatial Variations in Satural Mortality Spatial changes in the rate of natural mortality, as exist along the salinity gradient, may exert a significant influence on the es- timated yearly allocation. Is this variation important in the practi- cal context of management decision-making? The Delaware Bay oyster beds were partitioned into three groups according to the long-term average rate of natural mortality, the principal source of which is P. imiriniis infection. Yearly allocations to the oyster industry were evaluated for each of the three bed groups under a TABLE 3. Thf i-slimatfd yearly allocation for spring and autumn fishing seasons as a function of natural mortality rate for three different timings for the autumn fishing season. Day 1 was set at January 1, Data are presented in bushels. One bushel -37 I, and contains approximately 348 market-size oysters. Fishing Mortality Dermo MortaHtv Total Season 1 Season 2 Dermo Mortality Rate (yr'l (In Bushels) (In Bushels) (In Bushels) (In Bushels) 0.01 7 1 ..s 1 1 39.534 3 1 .976 546 0.10 66,314 37.154 29.1,59 5,734 0.20 59,854 34.084 25,769 12.204 0.30 52,469 30.424 22,044 19,589 0.40 43,845 25,949 1 7,895 28.212 0.50 33,480 20,287 13,192 38,577 0.60 20.496 12,771 7,724 51,561 0,70 3,158 2,037 1,121 68,899 0.01 71,710 46.705 25,005 347 0.10 68.275 44,065 24.209 3.783 0.20 63,662 40,624 23,037 8,395 0.30 57,901 36,470 21,430 14,156 0.40 50,482 31,317 19,165 21,575 0.50 40,526 24,686 15,840 31,531 0.60 26,371 15,703 10.668 45,686 0.70 4,395 2,540 1,8,54 67,662 0.01 71,754 46,704 25,049 304 0.10 68,718 44,065 24,653 3,339 0.20 64,567 40,624 23,943 7,490 0.30 59.266 36,470 22,796 12,791 0.40 52,260 31,316 20,943 19,797 0.50 42,560 24,685 17,874 29,497 0.60 28,227 15,703 12,523 43,830 0.70 4,832 2,540 2,292 67,225 Year Days Year Days 9(1 1511 150 240 330 Solid lines show the timing of the two fishing seasons. Dashed line shows the timing of Dermo mortality. 984 Klinck et al. range of natural mortality rates that covered the range observed in these bed groups during the 1990s. For ease of comparison, the same oyster abundances were used for each bed group. This abun- dance was the 2000 abundance for the high-natural-mortality group (Table I ). Not surprisingly, a much higher allocation was available for the low-mortality beds than for the high-mortality beds in these simu- lations (based on equivalent abundances between groups) (Figure 5). As before, the yearly allocation declined as natural mortality increased for each bed group. The range of allocations across the observed range of yearly mortality rates, a range that is assumed to document the range of anticipated mortality rates for a coming year, was much larger for the high-mortality group than for the other two groups. In practice, the fishery operates principally on the high- mortality beds, hence the use of these abundance values in the simulations shown in Figure 5. A factor of two exists between the highest and lowest allocations for these beds, depending on the natural mortality rate that might occur in any given year. Obvi- ously, errors in the prediction of natural mortality rate are most grave for this group. In fact, if stable market-size abundance is the desired outcome, taking the median rate of natural mortality for the decade of the 1990s for the low- and medium-mortality groups as the value for the coming year would introduce no more than about a 12% error in achieving this goal on these beds, even if natural mortality rate approached either the highest or lowest recorded values. In contrast, the error introduced for the high-mortality group would exceed ?i0'7c. Thus, increased levels of natural mor- tality increase the need for a precautionary approach to stock man- agement given the limited ability at present to predict the course of Dermo disease (Powell et al. 1996, Soniat & Kortright 1998). For the high mortality group, use of the 64r/; or 75tli percentile would seem appropriately conservative. Influence of Variations in Growlli Rale or Pre-reeniit Abundance Variation in the number of recruits into the fishery in a given year is also important. The premise of the model is to manage the fishery under the proviso of no net change in market-size abun- dance. Accordingly, the allocation is a function of the availability of new recruits and the natural mortality rate. The number of new recruits is controlled not only by settlement and recruitment into the population and subsequent survival of juveniles, but also by the yearly growth rate that establishes the smallest-sized oyster antic- ipated to recruit to the fishery during the season. Because the number of oysters typically declines in the larger size classes, slower growth can have a dramatic effect on yearly allocation by reducing the number of new recruits. The simulations shown in Figure 5 were based on the assump- tion that oysters 63.5-76 mm in size recruit to the fishery in a given year (market size, 2:76 mm). This is a growth rate typical of the higher-salinity oyster beds in Delaware Bay. Growth rate slows upbay. The equivalent simulations are shown for the case where oysters 10-16 mm are assumed to recruit to the fishery during the season in Figure 6. The difference in allocation is based on the observed number of oysters in these two size-classes on the high- mortality beds in 2000. About 47% of the oysters in the 63.5-76- mm size-class were s70 mm in size. 70000 Bushels allocated, Low- mortality beds Bushels allocated, Medium- mortality beds Bushels allocated. High- mortality beds Natural Mortality Rate (m) Figure 5. Yearly allocation (in bushels) for \arioiis rates of natural mortality (principally caused by Perkinsiis marinus). Simulations were run for three groups of beds characterized by lo», medium and high rates of natural mortality and for the range of natural mortality rates observed during the decade of the 1990s for these beds (Figure 4), In each case, the fishing season «as initiated on day 90 and ceased on day 288 (the 2001 fishing season for New Jersey). Perkinsiis Hiori/iH.v-niortality occurred between days 180 and .^00 at the yearly rate shown in Figure 4. Oysters 63,5-76 mm were assumed to recruit to the fishery. A Fisheries Model for Managing During Disease 985 35000 Bushels allocated, Low- mortality beds Bushels allocated, Medium- mortality beds Bushels allocated. High- mortality beds Natural Mortality Rate (m) Figure 6. Yearly allocatldii (in bushels) for various rates of natural mortalit) (principally caused by Perkiiisiis marimis). Simulations were run for three groups of beds characterized by low. medium and high rates of natural mortality and for the range of natural mortality rates observed during the decade of the I99l)s for these beds (Figure 4). In each case, the fishing season was initiated on day 90 and ceased on day 288 (the 21)01 fishing season for New Jersey). Perkiimis (naniius-mortality occurred between days 180 and 3()t) at the yearly rate shown in Figure 4. Oysters 70-76 mm were assumed to recruit to the fishery. A reduction in the number of recruits has two significant ef- fects. First, the total yearly allocation is significantly lower under the "slow-growth"' assumption (compare Figs. 5 and 6). In addi- tion, the range in allocation is increased when evaluated across the range of natural mortality rates observed during the decade of the 1990s. For the high-mortality beds that are most sensitive to this effect, the allocation estimated from the median rate of natural mortality for the decade of the 1990s is now a factor of 2 different from the extreme values rather than the -30% difference observed with the higher growth rate in Figure 5. As growth slows, the importance of an accurate estimate of anticipated natural mortality rate increases and, as a practical matter, the more precautionary must be the management evaluation prior to the season. Post-Epizootic Population Recovery Natural mortality rates during Dernio epizootics may exceed 0,7 yr"' (Mackin 1959. Andrews 1988, Powell et al. 1996). Con- sequently, oyster population abundance typically declines signifi- cantly during epizootic years. Management of the oyster resource might include rebuilding stocks after an epizootic. Management goals might include a return to the long-temi average population level or some higher level (e.g., 15th percentile). in Delaware Bay, 2000 end-of-year weighted-mean oyster abundance (calculated as # s 76 mm -f 0.5 x #63.5 - 76 mm) was near the decadal average weighted-mean oyster abundance ( 1989- 2000) and about 67% of the decadal 75r/! percentile weighted abundance (Fig. 7). Thus, a one-year return to the 15lh percentile weighted abundance would require an end-of-year ratio of /V,,,,,, Figure 8 shows the relationship between fishery allocation, natural mortality rate, and a desired degree of increase (or de- crease) in market-size abundance at year's end. The 2000 allo- cation for Delaware Bay was 40,000 bushels for the entire bay (all beds), a low level required by the ongoing Dermo epizootic at that time. Approximately the same allocation could be achieved in 2001 just on the high-mortality beds, under the manage- ment goal of no net change in oyster abundance (Fig. 8, upright triangles). This option exists because a significant increase in juvenile oyster abundance occurred in 2000 that dramatically increased the number of new recruits available to the fishery in 2001 (Fig. 9). However, because the number of bushels avail- able for harvest on the medium-mortality beds upbay was about 60,000 bushels (HSRL 2001) an opportunity would exist to rebuild population abundance on the high-mortality beds and still increase the yearly allocation. Figure 8 shows that rebuild- ing to half of the 15th percentile abundance goal (Ny^^/N^^m = 1.24) can be accomplished even at a relatively high natural mor- tality rate, if the allocation on the high-mortality beds is reduced to about 20,000 bushels. Reaching the 75r/! percentile weighted abundance goal is achievable if the yearly allocation on these beds is limited to about 8,000 bushels (Fig. 8). Even higher abundances could be achieved: however, only if the natural mor- tality rate is at or below the long-term mean. Thus, recovery from the 1998-1999 epizootic with a return to above average oyster abundances could be accomplished in one year even under the precautionary management assumption that natural mor- tality rates will approach the 15tli percentile of the decadal time series. 986 Klinck et al. w CT) r^ o "* O ,— 00 LO 05 CO C\J (T) 00 O) 05 05 o 05 05 05 05 05 05 05 o> 05 05 05 o 05 05 05 05 05 05 ^— T— T- r- f— CM T— ''" T" ▼~ ■*" ^~ Year Figure 7. Rank-order of oyster ueishted-niean abundance (calculated as # >76 min + 0.5 x # 63.5-76 mm) for the hlKli-niortallt> Delaware Bay oyster beds for 1989-21)11(1. DISCUSSION Setting the yearly allocation for a fished stock is always an uncertain endeavor. Choice of the natural mortality rate is critical in most fisheries models. Populations suffering significant mortal- ity from disease, particularly those where year-to-year changes in mortality rate cover a wide range, require particularly careful man- agement. Unfortunately, routinely-used fisheries models do not consider mortality from disease. Compensatory mortality is frequently observed in populations where disease mortality is significant. Some animals dying in other ways would otherwise have died from disease. In a com- mercial species, fishing mortality is an important additional source of mortality and compensatory processes should be important. Thus, a fisheries model developed to evaluate a diseased popula- tion must be capable of resolving compensatory mortality origi- nating from the timing of the various sources of mortality during the year. The model presented here permits the evaluation of compen- satory processes, thereby giving guidance to appropriate timing of fishing seasons for maximizing yield. In some cases, the economic attraction of a fall harvest may be offset by the decreased yield available to the fishery. An earlier season permits a higher harvest because some of the animals taken by the fishery would subse- quently be lost to the population through disease. Although a number of shellfish models exist (Kobayashi et al. 1997. Powell et al. 1997. Dowd 1997), few fisheries models tai- lored specifically to shellfish exist (e.g., NEFSC, 2n00a. 2000b). Fisheries models require a forward prediction of abundance under a chosen natural mortality rate in. However, small changes in mortality rate effect large changes in predicted outcomes. The present model is no exception. Outcomes are strongly influenced by the chosen yearly mortality rate. Accordingly, the accuracy of 100000 Rebuilding Plans Yearly Ratio (T+1/T) Natural Mortality Rate (m) Figure 8. Yearly allocation (in bushels! for \arious rales of natural niortalit> (principally caused by Perkinsus marinus). Simulations were run for the high-niortalily beds over a range of natural mortality rates obser\ed during the decade of the 199(ls (Figure 4). Kach simulation was run to produce a desired change in niarkel-si/e oyster abundance over the year, calculated as A'(7'+ll/.V(7'). In each case, the fishing season was initiated on day 9(( and ceased on day 288 (the 2tH)l fishing season for New Jersey I. Perkinsus wanKH.v-mortality occurred between days 18(1 and 30(1 at the yearly rate shown in Figure 4. Oysters 6.^.5-76 mm were assumed to recruit to the fishery. Negative numbers indicate cases where an insufficient number of new recruits was available to meet the demands of the rebuilding program. A Fisheries Model eor Managing During Disease 987 CO CD O O o CJ) en CD en ¥> c» cr> CM 00 CXI CTl 05 (J) C\J (3) CD en CO o en P Year Figure 9. Rank-order of the year-to-vear change in oyster weighted- mean abundance (calculated as # >76 mm + 0.5 x # 63.5-76 mml for the high-mortality Delaware Bay oyster beds for 1989-2000. forward prediction is controlled, in part, by the accuracy at which the disease process can be predicted in oyster populations. Besides disease, the other model parameter that most affects the outcome of simulations is the abundance of submarket-size oysters that can be expected to recruit to the fishery in the simulated year. A small drop in growth rate can dramatically alter the estimate of the yearly allocation because market-size is on the tail end of the size-frequency distribution. Thus, the definition of submarket size is critical. In addition, the abundance of submarket-si/e individuals establishes the resilience of the population to transient increases in disease-induced mortality and to errors in management that lead to o\ erfishing. As the number of submarketsize oysters declines, sus- tainability of the population becomes less certain and the accuracy of model simulations becomes more critical. In essence, as the ratio of submarket to market-size abundance approaches 1.0. reli- ance on more conservative management approaches should in- crease. Although the maximum life span of Crassostrea virginica is at least 15 yr (Lavoie & Bryan 1981). the average life span in most East and Gulf coast bays probably does not exceed four due to fishing and disease. Thus, population stability is strongly deter- mined by the number of recruits available to replace the deaths that decimate the market-size population each year. In Delaware Bay in 2000, the number of recruits was insufficient to permit rapid re- covery from the most recent P. mariiuis epizootic without conser- vative management measures because the epizootic resulted in lower survival of the submarket-size class, a frequent characteristic of epizootic conditions (Powell et al. 1996). and because spat settlement rates were low prior to this time interval (HSRL 2001 ). In 2001. conditions changed. An increase in recruits to the fishery provided an opportunity to rebuild population abundances to levels abo\e the decadal mean, while still permitting an increase in al- location over 2000. This modehng exercise points to the critical need to understand population dynamics and survival of size classes below market size that many times are not the targets of investigatory activities (Hofmann et al. 1995). The objective of management of the oyster fishery should be the maximization of yield under conditions that stabilize popula- tion abundance at some preferred level. What that preferred level is cannot be ascertained by the model described here. However, once that preferred level is chosen, the model presented here per- mits the evaluation of management scenarios targeting this objec- tive. The model requires a quantitative estimate of abundance by size class, some knowledge about growth rates to establish the size range recruiting into the fishery in any given year, and a decision about the natural mortality rate anticipated. The latter is of con- siderable consequence and. so. probably should be chosen conser- vatively, perhaps a value at or above the long-term average. Once these data are available, the model permits investigation of sce- narios that include a range of allocations, the timing of fishing seasons, and the variation of fishing efforts within seasons in order to establish a preferred harvest level for the oyster fishery. The model also supports investigation of recovery scenarios after an epizootic and the evaluation of population resiliency to future epi- zootics or overestimates of fishery yield. The model is sufficiently general that it likely can be applied to any commercial shellfish species, providing that the stock assessment program provides sur- vey data adequate for the modeling exercise. ACKNOWLEDGMENTS Special thanks to Larry Hickman. Captain of the FA/ Howciid W. Sockwelt and to Bivalve Packing for providing the F/V Howard W. Sinkuflt and logistical support for the 1999 stock survey. We thank the HSRL survey team. R.D. Barber. J.A. Gandy. and B.M. Brewster, for analysis of the 2000 survey samples and S.E. Banta, M. Cummings and A.J. Bonner for help in data analysis. This study was funded by an appropriation from the State of New Jersey to Rutgers University in support of the stock survey and associated oyster research program and by funding from the Sea Grant Oyster Disease Research Program, contract number 4-25238. LITERATURE CITED Allen, R. L. 1979. A yield model for the Foveaux Strait oyster iOslreii hitaria) fishery. Rapp. P. v. Reun. Cons. Int. Explor. Mer. 175:70-79. Andrews. J. D. 1988. Epizootiology of the disease caused by the oyster pathogen Perkinsus marinus and its effects on the oyster industry. Am. Fish. Soc. Spec. Publ. 18:47-6.^ ■Andrews. J. D. 1996. History of Perkinsus maiimis. a pathogen of oysters in Chesapeake Bay 1950-1984. J. Shellfish Res. 15:1.V16. .Applegate. A.. S. Cadrin. J. Hoenig. C. Moore. S. Murawski & E. Pikilch. 1998. Evaluation of existing overfishing definitions and recommenda- tions for new overt'ishing definitions to comply with the sustainable fisheries act. Final Report. Overfishing Definition Review Panel, NOAA, NMFS. 179 pp. Banta. S. E.. E. N. Powell & K. A. Ashton-Alcox. in press. Evaluation of dredging behavior by the Delaware Bay oyster fishery. N. Am. J. Fish. Manage. Bower. S. M., J. Blackbourn. G. R. Meyer & D. W. Welch. 1999. Effect of Perkinsus qugwadi on \arious species and strains of scallops. Dis. Aqiiat. Org. 36:143-151. Brousseau. D. J.. J. A. Baglivo. 2000. Modeling seasonal proliferation of the parasite. Perkinsus marinus (Dermo) in field populations of the oyster. Crassaslrea virginica. J. Shellfish Res. 19:13,3-138. Brousseau. D. J.. J. C. Guedes. C. A. Lakatos, G. R. LeCleir & R. L. Pin- sonneault. 1998. A comprehensive survey of Long Island Sound oys- ters for the presence of the parasite. Perkin.sus marinus. J. Shellfi.'ih Res. 17:255-258. Clark. W. G. 1999. Effects of an erroneous natural mortalily rate on a 988 Klinck et al. simple age-structured stock assessment. Can. J. Fish. Aquat. Sci. 56: 1721-1731. Cook. T.. M. Folli, J. Klinck. S. Ford & J. Miller. 1998. The relationship between increasing sea-surface temperature and the northward spread of Perkinsus marinus (Dermo) epizootics in oysters. E.sniarine Coastal Shelf Sci. 46:587-597. Dittman, D. E.. S. E. Ford & H. H. Haskin. 1998. Growth patterns in oys- ters. Cmsso.'Hrea virginica. from different estuaries. Mar. Biol. {Bcil.\ 132:461-469. Dowd. M. 1997. On predicting the growth of cultured hixalves. Ecol. Model. 104:113-131. FAO. 1995. Precautionary approach to fisheries. Part 1: Guidelines on the precautionary approach to capture fisheries and species introductions. FAO Fisheries Technical Paper 350/1, 52 pp. Fegley. S. R., S. E. Ford, J. N. Kraeuter & D. R. Jones. 1994. Relative effects of harvest pressure and disease mortality on the population dynamics of the Eastern oyster {Crassoslrea virginica) in Delaware Bay. Final Report #NA26FL0588 to the National Oceanic and Atmo- spheric Administration. Rutgers University, Haskin Shellfish Research Laboratory. Ford, S. E. 1992. Avoiding the transmission of disease in commercial culture of molluscs, with special reference to Perkin.sii.s nmriniis (Dermo) and Haplosporidiwn nelsoni (MSX). / Shellfish Res. 1 1 :539- 546. Ford. S. E. 1997. History and present status of molluscan shelllisheries from Barnegat Bay to Delaware Bay. NO.\A Tech. Rpt. NMFS 127: 119-140. Ford, S. E. & H. H. Haskin. 1988. Management strategies for MSX (Hap- losporidium nelsoni) disease in Eastern oysters. Am. Fish. Soc. Spec. Piibl. 18:249-256. Francis, R. I. C. C. & R. Shotton. 1997. "Risk" in fisheries management: A review. Can. J. Fish. Aqual. Sci. 54:1699-1715. Gottlieb, S.J. & M. E. Schweighofer. 1996. Oysters and the Chesapeake Bay ecosystem: a case for exotic species introduction to improve en- vironmental quality? Estuaries 19:639-650. Harvell, C. D., K. Kim, J. M. Burkholder, R. R. Colwell, P. R. Epstein, D. J. Grimes, E. E. Hofmann, E. K. Lipp. A. D. M. E. Osterhaus. R. M. Overstreet, J. W. Porter, G. W. Smith & G. R. Vasta. 1999. Emerging marine diseases — climate links and anthropogenic factors. Science [Wash. D.C.) 285:150.5-1510. Heesterbeek, J. A. P. & M. G. Roberts. 1995. Mathematical models for microparasites of wildlife. In: B. T. Grenfell & A. P. Dobson. editors Ecology of infectious diseases in natural populations. Cambridge, UK: Cambridge Univ, Press, pp. 90-122. Hilbom, R. & C. Walters. 1992. Quantitative Fisheries Stock Assessment. New York: Chapman & Hall. 570 pp. Hofmann. E. E.. E. N. Powell. J. M. Klinck & G. Saunders. 1995. Model- ing diseased oyster populations I. Modelling Perkinsus marinus infec- tions in oysters. / Shellfish Res. 14:121-151. Hofmann. E. E., E. N. Powell. J. M. Klinck & E. A. Wilson. 1992. Mod- eling oyster populations. III. Critical feeding periods, growth and re- production. J. Shellfish Res. 11:399-416. Hofstetter. R. P. 1977. Trends in population levels of the American oyster. Crassostrea virginica (Gmelin^. on public reefs in Galveston Bay. Texas. Texas Parks Wildl. Dept. Tech. Ser. 24:1-90. HSRL. 2000. Report of the 2000 Stock Assessment Workshop for the New Jersey oyster seed beds. Haskin Shellfish Research Laboratory. Rutgers University. 7 pp. HSRL. 2001. Report of the 2001 Stock Assessment Workshop for the New Jersey oyster seed beds. Haskin Shellfish Research Laboratory. Rutgers University. 23 pp. Jakobsen. P. J.. G. H. Johnsen & P. Larsson. 1988. Effects of predation risk and parasitism on the feeding ecology, habitat use. and abundance of lacustrine threespine stickleback (Gasterostens aculcatus). Can. J. Fish. Aquat. Sci. 45:426-431 Jordan. S. J. 1995. Population and disease dynamics of Maryland oyster bars: a multivariate classification analysis. J. Shellfish Res. 14:459- 468. Kobayashi. M.. E. E. Hofmann, E. N. Powell, J. M. Klinck & K. Kusaka. 1997. A population dynamics model for the Japanese oyster, Crassos- trea gigas. Ac/uacultiire. 149:285-321. Lavoie, R. E. & C. F. Bryan. 1981. Oyster bed development in the Dunk River, P.E.I., 1975-1980. Can. Tech. Rpt. Fish. Aquat. Sci. 1015:1-14. Loosanoff V. L. 1966. Time and intensity of setting of the oyster Cras- sostrea virginica. in Long Island Sound. Biol. Bidl. (Woods Hole) 130:211-227. Mackin. J. G. 1959. A method of estimation of mortality rates m oysters. Proc. Nad. Shellfrsh. Assoc. 50:41-51. Mackenzie Jr.. C. L. 1996. History of oystering in the United States and Canada, featuring the eight greatest oyster estuaries. U.S. Fish Wildl. Seiv. Mar. Fish. Rev. 58:1-78. Mackenzie Jr.. C. L. & A. T. Wakida-Kusunoki. 1997. The oyster industry of Eastern Mexico. U.S. Fish Wildl. Sen: Mar. Fish. Rev. 59:1-14. May, R. M., J. R. Beddington. J. W. Horwood & J. P. Shepherd. 1978. Exploiting natural populations in an uncertain world. Math. Bio- sciences ■i2:2\9-252. MeiBner. K. & A. Bick. 1997, Population dynamics and ecoparasitological surveys of Cornphium volulator in coastal waters in the Bay of Meck- lenburg (southern Baltic Sea). Dis. Aquat. Org. 29:169-179. Miller. A. C. & S.E. Lawrenz-Miller. 1993. Long-term trends in black abalone, Haliotis cracherodii Leach, 1814, populations along the Palos Verdes Peninsula, California. J. Shellfish Res. 12:195-200. Moyer, M. A.. N. J. Blake & W. S. Arnold. 1993. An ascetosporan disease causing mass mortality in the Atlantic calico scallop. Argopeclen gih- Ims (Linnaeus. 1758). J. Shellfish Res. 12:305-310. NEFSC. 2000a. 30"' northeast regional stock assessment workshop (30''' SAW): Stock assessment review committee (SARC) consensus sum- mary of assessments. Northeast Fish. Sci. Cent. Ref. Doc. 00-03. 477 pp. NEFSC. 2000b. 31" northeast regional stock assessment workshop (31" SAW): Stock assessment review committee (SARC) consensus sum- mary of assessments. Northeast Fish. Sci. Cent. Ref Doc. 00-15. 400 PP- O'Beirn, F. X., C. C. Dean & R. L. Walker, 1994. Prevalence ot Perkinsus marinus in the Eastern oyster Crassostrea virginica in relation to tidal placement in a Georgia tidal creek. Northeast Gulf Sci. 13:79-87. O'Beirn, F. X„ P. B. Heffernan. R. L. Walker & M. L. Jansen. 1996. Young-of-the-year oyster, Crassostrea virginica. reproduction in coastal Georgia. Estuaries 19:651-658. Oliver, L. M., W. S. Fisher, S. E. Ford. L. M. Ragone Calvo, E. M. Bur- reson. E. B. Sutton & J. Gandy. 1998. Perkinsus marimts tissue distri- bution and seasonal variation in oysters Crassostrea virginica from Florida. Virginia and New York. Dis. Aquat. Org. 34:51-61. Park. K-I.. K-S. Choi & J-W. Choi. 1999. Epizootiology of Perkinsus sp. found in the Manila clam, Ruditapes philippinarum in Komsoe Bay, Korea. / Korean Fish. Soc. 32:303-309. Powell, E. N., J. M. Klinck & E. E. Hofmann. 1996. Modeling diseased oyster populations. II. Triggering mechanisms for Perkinsus marinus epizootics. J. Shellfish Res. 15:141-165. Powell. E. N., J. M. Klinck, E. E. Hofmann & S. Ford. 1997. Varying the timing of oyster transplant: implications for management from simu- lation studies. Fish. Oceanogr. 6:213-237. Restrepo. V. R., G. G. Thompson. P. M. Mace, W. K. Gabriel, L. L. Low, A. D. MacCall, R. D. Methot, J. E. Powers, B. L. Taylor. P. R. Wade & J. F. Witzig. 1998. Technical guidance on the u.se of precautionary approaches to implementing National Standard 1 of the Magnuson/ Stevens Fishery Conservation and Management Act. NOA,\ Tech. Mem. NMFS-F/SPO-31 54 pp. Soniat. T. M. & E. V. Kortright. 1998. Estimating time to critical levels of Perkinsus marinus in Eastern oysters, Crassostrea virginica. J. Shell- fish Res. 17:1071-1080. Soniat, T. M., E. N. Powell, E. E. Hofmann & J. M. Klinck. 1998. Under- standing the success and failure of oyster populations: the importance A Fisheries Model for Managing During Disease 989 of sampled variables and sample timing. J. Shellfish Res. 17:1 149- 1165. Sousa. W. P. lyyi. Can models of soft-sediment community structure be complete without parasites? Am. ZihiI. 31:821-830. Sunila. I.. J. Karolus & J. Volk. 1999. A new epizootic of Haplosporidiiim iielsoni (MSX). a haplosporidian oyster parasite, in Long Island Sound. Connecticut. / Shellfish Res. 18:169-174. Swinton. J. & R. M. Anderson. 1995. Model frameworks for plant- pathogen interactions. In: B. T. Grenfell & A. P. Dobson. editors. Ecol- ogy of infectious diseases in natural populations. Cambridge, UK: Cambridge Univ. Press, pp. 280-294. Thomas. C. D. & W. E. Kunin. 1999. The spatial structure of populations. J. Anim. Ecol. 68:647-657. Whitlatch. R. B. & R. W. Osman. 1994. A qualitative approach to man- aging shellfish populations: assessing the relative importance of trophic relationships between species. .]. Shellfish Res. 13:229-242. Jminml of Shcllfi.Ji Hc.u;iirli. Vol. 20. Nii. .\ 991-W.S. 2()()1. A DNA PROBP: for TRANSCRIPTION ANALYSIS OF THE PROTEOLYTIC ENZYME CATHEPSIN B IN THE PACIFIC OYSTER, CRASSOSTREA GIGAS (THUNBERG, 1793) K. M. DONALD, A. J. S. HAWKINS,* AND G. R. SMKRDON Phiuouth Marine Lahoratory, Cciilre for Coastal ami Marine Science, Prospect Place. Plyniouili. Devon. PLI ^DH. United Kiiijidon] ABSTRACT Molecular lechniques ha\c been employed lo develop a prohe lor the proleolylie enzyme catliepsiii B In the Pacific oyster. C. gigas. Degenerate primers were used to amplify a 450 tiase pair (bp) fragment of the cathepsin B gene. Deduced amino acid sequence indicates 60. 59. and 59% identity with cathepsin B from human, rat and cow. respectively. Expression of cathepsin B was detected, by RT-PCR. in both l.irval and adult tissues. Northern hlot analysis demonstrated expression of a single transcript of approximately 2.4 kb. KEY WORDS: proteolytic enzy ines. gene expression, cathepsin B, Crcissosrn'{i gigas INTRODUCTION The continuous bieakciown. replacement and/or renewal of pro- teins, known as protein turnover, is an essential component of nietabolistn. enabling development, adaptation, repair and regula- tion (Hawkins 1991. Ciechanover & Schwartz 1994), Neverthe- less, the intensity of protein turnover varies greatly according to genetic and environmental influences, with significant conse- quences for growth and survival atnong animals generally (Hawk- ins & Day 1999), Comparisons between individual shellfish have shown that whole-body specific activity of the key proteolytic enzyme cathepsin B correlated with the intensity of whole-body protein turnover, and was associated with both higher energy e.\- penditure and slower growth rate (Hawkins & Day 1996). Cathep- sin B is a proteolytic enzyme belonging to the papain cysteine protease superfamily (Hasnain et al. 1993), and exhibits both endo- and exopeptidase activities (Koga et al. 1991 ). Mammalian cathep- sins have been extensively studied, due to the correlation between cathepsin activity and cancers and brain disorders, such as Alz- heimer's disease and multiple sclerosis (Bernstein 1994), In com- parison, cathepsin activity in invertebrates has received compara- tively little attention (Zeef & Dennison 1988). Here, to enable future work elucidating tnechanistns that regu- late cellular levels of cathepsin B. we describe the development of a molecular probe to monitor expression of the cathepsin B gene in the commercially important Pacific oyster Crassasirea gificis at different stages of development and in different adult tissues. Dur- ing development, there are "critical windows" of increased protein turnover, when major structural changes occur as C. gigas indi- viduals develop into different larval forms and then metamorphose from a free-swimming planktonic larva into a sessile calcified adult. Within five to 10 hours of fertilization, a non-feeding tro- chophore larva is formed, which within 24 to 48 hours develops into a planktotrophic. straight-hinge veliger larva (Dekshenieks et al. 1993). Between two and three weeks PF, the larva develops a foot and a pair of eyespots. This is the pediveliger form. Once the larva reaches -300 |j.in. usually after appioxitnately 25 days PF when grown at 23°C. metamorphosis occurs. This involves a major anatomical reorganization; the velum and foot are lost, the organ systems are rearranged and the gill and adductor muscle become enlarged (Galtsaff 1964). Following tnetamorphosis. C. gigas re- main benlhic and undergo no futlher re-organization of their body plan. Proteolysis continues to maintain normal functions, such as sanitation and breakdown of degraded proteins (Ciechanover & Schwartz 1994). However we do not expect levels to be as high as in the larvae, due to substantial changes in body plan and a weight specific decrease in metabolic rate. The only exception to the relatively low levels of protein turnover expected in adult tissues tnay be in the gonads immediately prior and post spawning, as high energy expenditure and protein turnover is necessary for spawning to take place (Peek & Gabbott 1990). In summary, we have used molecular techniques to develop a probe to detect transcripts of the proteolytic en/yme cathepsin B gene in C. gigas. We have used the cathepsin B probe to quantify cathepsin B gene transcripts at times of expected high and low energy expen- diture and protein turnover, which included key points in larval development and different tissues from adult oysters. MATERIALS AND METHODS DNA Extraction *Corresponding author. Crassostrea gigas individuals from Whitstable oyster hatchery were acclimated for a tninimutii of .seven days in a system of recirculated sea water at I5'C. Following this acclimation adduc- tor muscles were dissected from individual oysters and ground under liquid nitrogen. Fifty tnilligrams of the ground tissue was transfened to 500 |jil extraction buffer (50 niM Tris-HCl pH8. 100 mM NaCl. 50 mM EDTA, 1% SDS). Proteinase K (50 (xg) was added and the solution incubated at 55°C for 1 hour. Phenolxhlo- roform extraction was performed on the solution and DNA pre- cipitated by incubation with 0.1 volumes 3M NaOAc and 2.5 volumes 1009(- ethanol at -20''C overnight. The precipitated DNA was harvested by centrifugation, washed in 709^ ethanol and dried in a vacuum desiccator. The dried pellet was re-suspended in 500 |j.l TE and incubated with 5 p.1 RNase ( 10 mg.ml" ' ) at 37°C for 30 minutes and then with 10 |jl1 proteinase K ( lOmg.ml"' ) at 55°C for 30 minutes. A second phenol:chlt>roform extraction and ethanol precipitation were then performed and the DNA harvested by cen- trifugation. The pellet was washed in 10% ethanol. dried by vacuutii desiccation and re-suspended in 500 ptl TE. The purity and quantity of DNA was measured spectrophotometrically and by agarose gel electrophoresis. PC R- Amplification Amplification of genornic DNA was performed using 100 ng extracted DNA template per reaction. Degenerate primers were 991 992 Donald et al. designed from conserved regions of tfie cathepsln B gene, using the sequence entries in the EMBL database and software at the Human Genome Mapping Project Resource Centre (HGMP-RC). Cambridge UK (see Fig. 1). Cathlf: 5'-gg(a/c/t) tg(c/t) (a/c/g)a(ayg/t) ggl (ayg)g(a/c) tat-3'; Cathlr: 5'-ga(a/g) ttg gc(a/c/g/t) a(c/t)(a/c) age cag-3' The initial cycling parameters included a denaturation step of 92°C (5 min), followed by 35 cycles of 48°C ( 1 min). 72°C (1 min). and 92°C (1 min). A final annealing phase at 48°C (2 min) was fol- lowed by an extension phase at 72°C (5 min) and storage at 4°C. Due to the limited quantities of initial PCR product, a second amplification was performed with identical primers using 5 jj.1 of the initial PCR product as template and with the annealing tem- perature increased to 50°C and the number of cycles reduced to 30. Aliquots of the second amplification reaction were visualized b\ agarose gel electrophoresis (1.2%). Sequencing To verify that the correct gene had been amplified, the PCR product was cloned into the Siiuil site of pBluescript SK^. to produce plasmid pCCcathI, and sequencing was performed fol- lowing the dideo.xy chain termination method (Sanger et al. 1977) and using the Sequenase II sequencing kit (Amersham). Transcript Uclection Two methods were used to detect CathB transcripts, reverse transcription (RT)-PCR and northern blotting. There are merits and drawbacks to both techniques. During RT-PCR, RNA is reverse transcribed to make cDNA, which is then amplified by PCR using gene specific primers. RT-PCR is a very sensitive technique. How- ever it is non-quantitative and can only be used to look for the presence/absence of any expressed gene. Also, transcript size can- not be determined using RT-PCR as only part of the transcript is amplified. Northern blotting involves the hybridization of a radio- labelled probe to size separated transcripts. This technique can be used to identify transcript size and for quantitation, but is less sensitive than RT-PCR. RNA Extraction and RT-PCR RNA was extracted from adults and from larvae at "critical windows" in development when protein turnover was expected to be high; at one day post-fertilization the larvae is developing from the trochophore into the veliger form and around settlement the larvae changes from a free-swimming larvae to a calcified adult. At both these stages energy expenditure and metamorphosis are high. Adult tissue, dissected from six individual oysters several days post spawning, and larvae, sampled at one day PF and around settlement metamorphosis, were obtained from the Whitstable oys- ter hatchery. The oysters were prepared for RNA extraction by grinding the adult tissue under liquid nitrogen and homogenizing the larvae in lysis buffer. RNA was then extracted using the SV total RNA isolation system (Promega). The quantity and purity of RNA were measured spectrophotometrically and RNA quality was determined by electrophoresis. RT-PCR was performed using 100 ng RNA as the template with the Access RT-PCR system (Promega) in conjunction with non-degenerate primers designed directly from C. gigas sequence data. Cgigas. cathlf: ?'-ggt tgc aat ggt aga tat cc-3'; Cgigas.cathlr: 5'-gaa ttg gcc ate age cag ta-3' The Promega Access RT-PCR system uses avian myeloblastosis virus (AMV) reverse transcriptase for first strand cDNA synthesis, and the thermostable Tfl DNA polymerase for second strand cDNA synthesis and amplification. First strand cDNA synthesis was performed with an initial 45-minute incubation at 48°C. Fol- lowing inactivation of AMV-RT by a two minute incubation at 94"C, second strand synthesis, consisting of 40 cycles of 53"C ( 1 min), 68°C (2 min) and 94"C (30 sec) was perfomied. A final extension phase at 68°C (7 min) was followed by storage at 4°C. Following the initial RT-PCR, a nested PCR reaction was per- formed on 2 p,l product using an annealing temperature of 56°C, 30 cycles and Taq polymerase (Promega) and a new forward primer again designed from specific C. gigas sequence. Cgigas.cathnf 5'-ctg aag gtg cct ggt ccc-3': Cgigas.cathlr: 5'-gaa ttg gcc ate age cag ta-3' 301 850 a b c d e f forward primer TGGGGACGGC TGTAATGGTG GCTATCCTGC TGAAGCTTGG AACTTCTGGA TGGGGACGGC TGTAATGGTG TGGTGATGGT TGTCAAGGTG TGGAGATGGA TGCAAGGGTG TGGCGATGGT TGTGATGGAG GGC TGTAATGGTG T CC A A A G G 1301 a CTACTGGCTG GTTGCCAACT b CTACTGGCTG GTTGCCAACT c CTACTGGCTG GTTGCCAACT d TTACTGGTTG ATTGCCAATT e GTACTGGCTG GTGGCCAACT f TTACTGGCTT CTTGCAAATC reverse CTGGCTG GTNGCCAACT primer T A T Figure 1. Alignment of cathepsin H sequences from (a) Homo sapien aegypti; and (I) Triticnm uesliyum. Conserved regions are l)oxed; pr GCTATCCCTC TGGAGCATGG AACTTCTGGA GATTTCCTGG TGTAGCATGG GACTATTGGG GATATCTGGG TCCGGCTTGG CAGTTCTGGG GATATCCTAT CAGTGCGTGG CAATACTTCG GCTAT (216 degeneracies) CCTGGAACAC CCTGGAACAC CTTGGAACCT CATGGAATGA CCTGGGGTGA AGTGGAACAG C TGACTGGGGT TGATTGGGGT TGACTGGGGT AGATTGGGGT CGATTGGGGA AGGCTGGGGC (32 degene 1350 GACAATGGCT GACAATGGCT GATAATGGCT GAGAAGGGAC GACAATGGTT GATGACGGGT racies) : (bl Bos tauriis: (c) Mus musculus; (d) Schistosoma japoniciini: fe) Aedes niers designed from such regions are shown in bold. Gene Expression of Cathepsin B in C. gigas 993 Aliqiiots of the amplification reaction were analysed by agarose gel electrophoresis (1.59f) to check amplification efficiency. Northern Blotting For the probe, an Xbal and EcoRI (Promega) fragment con- taimng the cathepsin B sequence was excised from the pCGcathI plasmid by restriction digest with Xba\ and EcoRI (Promega). The cathepsin B fragment was purified by agarose gel electrophoresis and extraction from the gel was performed using the Qiaex ex- traction kit (Qiagen). 25 ng of this extracted DNA was used for priibe construction using a Prime-It II random primer labelling kit (Stratagene) and '"PdATP (Amersham). RNA totalling up to 30-100(j.g from larvae one day PF and around settlement and from both adult digestive gland, adductor muscle and gonads and larvae at different stages of development were analyzed using Reliant RNA gels in MOPS buffer (PMC BioProducts). RNA was electrophoresed for two hours at 3.5 volts per cm. RNA markers (Promega) were run alongside the RNA samples so the size of the CathB transcript could ultimately be determined. Following electrophoresis the markers were stained with ethidiuni bromide. The RNA samples were then blotted onto Hybond N-i- membrane and the position of the RN.^ markers re- corded on the membrane which was then placed at 8()"C for two hours to fix the RNA to the membrane. The Hybond N-i- membrane was pre-hybridized for one hour at 42°C in 10 ml ULTRAhybT^-' ultrasensitive hybridisation buffer (Anibion). '-P-labelled probe was added to the membrane which was left shaking at 42''C. Following hybridization for 16 hours, the probe was removed and the membrane was washed for 2x5 min in 2 X SSC, 0.1% SDS at 42°C and then for 1x10 mm in 0.1 x SSC, 0.1% SDS at 42°C. The membrane was then sealed in Saran wrap and exposed to X-ray film for three days at -80°C. Figure 2. Amplincation of cathepsin B fragmint from adull C. i;igiis DN.\ using primers cathlf and cathlr. (PCR conditions are described In section 2.2). A total of 5 ^1 PCR products were analyzed on a 1.2'7f agarose gel. Lane I = <))xl74/HaelII size markers: lane 2 = -ve control lacking template DN.A; lane 3 = template I)N,\ extracted from adduc- tor muscle. The sizes of the molecular weight markers are indicated in base pairs. GGT TGC XAT GGT XGA TAT CCT GAA GGT GCC TGG TCC CAA TTT AAG ' GCNGRYPEGAWSQFK 'aGC AAG GGA CTC AGC TCA GGT TGG CTA TAT GGT GAT AAA AAG TAc" SKGLSSGWLYGDKKY 'tGC AAA GCT TAT AGT TTA CCT CCT TGC AGT CAT CAT GTA AAT GGT CKAYSLPPCSHHVNG 'tCT CAC CCT GCT TGC ACA GGC AGC ACT AGA ACT CCT GTC TGC ACa" H V AAG GAA TGT GCC GCT GAA TCA GGC AGG ACT TAC AGT TCA GAC ATC ' CGC CAC GGA AAA TCA GTC TAC TCC GTC AGA GGT GAG GCC AAC ATC RHGKSVYSVRGEANL 'aTG CAT GAG ATT ATG ACC CGC GGA CCT ATG GAA GCT TCA TTC ACT M M GTT TAC CAA GAC TTT TTG GCT TAC AAA TCA GGA GTT TAC CAT CAT VYQDFLAYKSGVYHH ^GTG ACC GGA AGT GCT CTT GGA GGT CAT GCT GTC AAA ATC ATG GGA V V TGG GGA GTA GAA AAT GGC ACT CCT TAC TGG CTG ATG GCC AAT TC W N WGVENGTP Figure 3. Nucleotide sequence and deduced amino-acid sequence of the cloned 450 b.p. cathepsin B fragment from C. gigas. Primer se- quences are sho«n in bold. RESULTS Isolation of a C. gigas Cathepsin I! Clone Amplification of genomic DNA with degenerate primers de- signed from conserved regions of 3 vertebrate. 2 invertebrate and 1 plant cathepsin B nucleotide sequences (Fig. 1 ). resulted in am- plification of a 450 bp fragment (Fig. 2). Sequence Analysis of the C. gigas Cathepsin B Fragment The nucleotide and deduced amino acid sequence of the ca- thepsin B clone are shown in Figure 3. As predicted from the PCR amplification, the cloned cathepsin B fragment was 450 bp long. The deduced amino acid sequence was compared with cathepsin B amino acid sequences for Homo sapiens (human), Rattiis norvegi- ciis (rat) and Bos laiinis (cow) (Fig. 4). The C. gigas cathepsin B fragment showed 60% identity to H. sapiens. 59% identity to R. non'egiais and 58% identity to B. lunnis at the amino acid level. Expression Analysis of the Cathepsin B Gene in C. gigas A cathepsin B transcript was detected in both adult and larval tissues after reverse transcription and two rounds of PCR (Fig. 5), indicating cathepsin B expression in both adult and larval tissues. For a quantitative but less sensitive technique, northern blot analy- sis was performed on a selection of adult tissue and larvae at different stages of development (Fig. 6). The probe identified a C. gigas GCNGRYPEGA WSQFKSKGIS SGWLVGDKKV CFAYSLPPCS HHVMGSHPAC H. sapiens GCNGGYPAEA WNFWTRKGLV SGGLYESHVG CRPYSIPPCE HHVNGSPPPC F. Norvegicus GCNGGYPSGA WKFWTPKGLV SGGVYHSHIG CLPYTIPPCE HHVNGSPPPC B. taurus GCNGGFPSGA WKFWTKKGLV SGGLYNSHVG CRPYSIPPCE HHVNGSPPPC C gigas TG3TRTP/CT KECAAESGRT YSSDLRHGKS '.■YSVRGEANL -MHELHTRGP H. sapiens TGEGDTPKCS KICEPGVSPT YKQDKHYGYN SYSVSNSEKC IMAEI YKNGP R Worvegicus TGEGDTPKCN KMCEAGYSTS YKEDKHYGYT SVSVSDSEKE IMAEIYKNGP B. taurus TGEGDTPKCS KTCEPGYSPS YKEDKHFGCS SYSVANNEKE IMAEIYKNGP C gigas MEASFTVYvD FLAYKSGVYH HVTGSALGGH AVKIMGWGVE NGTPYWU-yUJ H sapiens VEGAF£VY:D FLIYKSGVY. HVTGLMKGGH A:?I1GWGVE NGTPYML.AN F Worvegicus VEGAFTVFSD FLrYKSGVYF HEAGl VKGCH AlPILGWGrE NG.PYWLVAN B. taurus VEGAFSVYSD FLLYKSGVYO HVJGEIMGGH AIPILGWGVE NGTPYWLVGN Figure 4. .Amino acid sequence comparison of cathepsin B from C. gigas with sequences from Homo sapiens. Rattits norvegicns. and Bos taurus. Amino acid residues that are identical to C. gigas cathepsin B are boxed. 994 Donald et al. 1 2 3 4 5 6 7 cathepsin B "amplicon Figure 5. RT-PCR analysis of cathepsin B gene expression in C. gigas. Total RNA was reverse transcribed and submitted to two rounds of PCR as described pre> iousl>. A total of 5 pi PC'R products were ana- lyzed on a 1.2% agarose gel. Lane 1 = kHindlll markers: lane 2 = -ve control; lane 3 = 100 ng RNA from adult digestive gland; lane 4 = 100 ng RNA from larvae 2 days before settlement; lane 5 = 100 ng RNA from larvae 2 days post settlement. cathepsin B transcript of approximately 2.4 kb in larval tissue and in adult gonad tissue. No expression was detectable in RNA ex- tracted from adult digestive gland or adductor muscle. DISCUSSION To our knowledge, this is the first study to adopt molecular techniques to clone and partially sequence a gene encoding a the proteolytic enzyme cathepsin B in any shellfish. The amino acid sequence of cathepsin B in Crassostrea gigas showed 60. 59 and 58% identities to cathepsin B amino acid sequences in the human, cow and rat. respectively. Despite the obvious evolutionary dis- tance between C. gigas and the three mammalian species, there is still a high degree of similarity between the oyster and mammalian cathepsin B sequences. It has been previously reported that bivalve cathepsins do share certain biochemical properties with mamma- lian cathepsins (Zeef & Dennison 1988). For example, the surf clam [Spisiila solidissima) has been shown to possess a cathepsin with similar specificity and inhibition properties to mammalian cathepsin B (Chen & Zall 1986). Figure 6. Northern blot analysis of C. gigas total RNA from adult and larval tissue. \n [a-'-P|.\TP-labelled 450-bp fragment of C. gigas ca- thepsin B DN.\ was hybridized with total RN.\ from adult and larval tissue. Lane 1 = 30 ng RNA from adult digestive gland; lane 2 = 30 pg RN.\ from adult adductor muscle; lane 3 = .Ml (ig RN.\ from adult gonads; lane 4 = 30 pg RN.\ from larvae 1 day post fertilization: lane 5 = 100 pg RNA from larvae 1 day post fertilization; lane 6 = 30 pg RN.4 from larvae 2 days before settlement: lane 7 = 30 pg RNA from lar\ae 2 days post settlement 2 // 1 RN.\ markers (Proniegal were run alongside RN.\ samples, the molecular weights were indicated. We detected cathepsin B gene expression m all developmental stages that we tested, reflecting the essential nature of proteolysis in repair, sanitation and metabolic regulation (Hawkins 1991, Ciechanover & Schwartz 1994). Northern blot analysis indicated that the level of cathepsin B expression was higher in larvae than in adult tissue; so much so that the level of cathepsin B expression was too low to be detected by northern blot analysis in adult digestive gland or adult adductor muscle tissue. This was consis- tent with our expectation of fast protein turnover during periods of rapid larval development (Fujii et al. 1991. Hawkins 1991). In adults, cathepsin B gene expression was only sufficient to be de- tected at the northern blot level in gonadal tissue. It is likely that this relatively high level of transcription occurred because the oysters had recently spawned, after which there is significant breakdown and reorganization of the remaining cells (Peek & Gab- bott 1990). In summary, we have developed a molecular probe to measure expression of the cathepsin B gene in C. gigas. We have detected expression of that cathepsin B gene throughout development, using RT-PCR. Findings indicate that the level of cathepsin B expression varied according to life stage, with higher levels of expression in larval tissue. Using this probe, our future work will relate expres- sion to both the quantity (by western blotting) and activity (by bio- chemical assay) of cathepsin B, thus helping to resolve the mecha- nisms regulating cellular levels of this key proteolytic enzyme. LITERATURE CITED Bernstein. H. G. \994. The many faces of lysosomal proteinases (cathep- sins) ni human neuropathology. A histochemical perspective. Eur. J. Histoclwm. 38:189-192. Chen. H. L. & R. R. Zall. 1986. Partial panlicalion and characterization of cathepsin D-like and B-like proteases from surf clam \ iscera. J. Food Sci. 51:71-75. Ciechanover, A. & A. L. Schwartz. 1994. The iibiqimm-niediated proteo- lytic pathway: mechanisms of recognition of the proteolytic substrate and involvement in the degradation of native cellular proteins. FASEB J. 8:182-191. Dek.shenieks. M. M., E. E. Hofinan & E. N. Powell. 1993. Environmental effects on the growth and development of eastern oyster, Crassostrea virginica (Gmelin. 1791), larvae: A modeling study. / Shellfisli Res. 12:241-254. Fujii. Y., M. Taguchi. K. I. Kobayashi & S. Horiuchi. 1991. Immu- nochemical studies on cathepsin D-like enzyme in the tadpole tail of Riiihi Lutesbeianu. Zoo. Sci. 8:511-520. Galtsaff, P. S. 1964. The American oyster Crassostrea virginica Gmelin. Fisliery Bull. 64:355-380. Hasnain. S.. T. Hirama. C. P. Huber. P. Mason & J. S. Mort. 1993. Char- acterization of cathepsin B speciticity by site-directed mutagenesis. J. Biol. Chem. 268:235-240. Gene Expression of Cathepsin B in C. gigas 995 Hawkins, A. J. S. 1991. Protein turnover: a functional appraisal. Fiiiicr. Ecol. 5:222-233. Hawkins. A. J. S. & A. J. Day. 1996. The metabolic basis of genetic differences in growth efficiency among marine animals. / Ex/j. Mm: Biol. Ecol. 203:93-11?. Hawkins. A. J. S. & A. J. Day. 1999. Metabolic interrelations iinderljmg the physiological and evolutionary advantages of genetic diversity. Amcr. Zoo. 39:401-411. Koga. H., H. Yamada. Y. Nishimura. K. Kalo & T. Imoto. 14^)1. Multiple proteolvtic action of rat-liver cathepsin-B - specificities and pH- dependences of the endo- and exopeptidase activities. 1. Biocheiii. (To- kyo) 110:179-188. Peek. K. & P. A. Gabbotl. 1990. Seasonal cycle of lysosomal enzyme activities in the mantle tissue and isolated cells from the mussel Mytilus edulis. Mar. Biol. 104:403-412. Sanger F.. S. Nicklen & A. R. Coulson. 1977. DNA sequencing with chain- temiinating inhibitors. Proc. Nam. Acad. Sci. USA 75:5463-3467. Zeef L. A. H. & C. Dennison. 1988. A novel cathepsin from the mussel [Pcnui periHi) (Linne; Coiiip. Biochem. Phwioi 906:201-204. JnuiiHil of Shellfish Research. Vol. 20. No. .^, 497-1(101. 2001. IMPROVED METHOD FOR ROTAVIRUS DETECTION IN OYSTERS USING RT-PCR: SUITABILITY OF A COMMERCIAL PCR KIT C. S. SANTOS, C. RIGOTTO, C. M. O. SIMOES,' AND C. R. M. BARARDI' * Laboratorio de Virologia Aplicada. Departamento de Cienclas Fannaceutkas. Ceiilro dc Ciciicias da Saiide. Universidade Federal de Saiila Catariiui HH040-900, Floriandpolis. Santa Calarina. Brazil: 'Departamento de Microhiologia e Parasitologia. Centra de Ciencias Bioldgicas. Universidade Federal de Santa Catarina 88040-900. Floriandpolis. Santa Catarina, Brazil. ABSTRACT Commercially markeled kits are now a\ ailablc for PCR reactions. This study was conducted to determine the suitability of one of these kits (PCR supermix kit from Life Technologies Inc) for use in environmental testing for rotavirus in commercially cultured oysters. We focused on developing a rapid, efficient and inhibitor-free oyster-processing procedure which could be used for sensitive viral genome amplification by reverse transcripts PCR (RT-PCR) in raw oysters using a commercial kit for genome amplification. Rotavirus SAl 1 strain was used to evaluate the efficiency of virus recovery. Oyster tissues were seeded with 3 x 10' ffu of rotavirus. Oyster-processing included elution in TPB (Tryptone phosphate broth )/glycine (pH 9.0). virus precipitation using polyethylene glycol, sonication, and RNA extraction. RNA extraction methods evaluated included CTAB/SDS followed by phenol/ chloroform extraction (standard method), or the commercial reagent TRIZOL LS'^'. Both were equally effective for removal of PCR inhibitors. Detection limit of the method described in this study was 0.03 ffu rotavirus SAl 1 recovered from the oyster tissues. We conclude that the Supermix - commercial PCR kit can provide a more rapid and sensitive alternative for virus detection in oysters when compared to traditional PCR protocols. This is especially beneficial when large numbers of environmental samples require analysis. KEY WORDS: oysters, RT-PCR, PCR supermix" kit. rotavirus INTRODUCTION Enteric virus transmission through consuiiiplion of fecaliy- contaminated shellfish is a significant public health concern (Ri- chards 1983. Jaykus el al. 1994). Fecal coliforms have long been considered the best, if imperfect, indicators of fecal pollution. However, these indicators have come under increasing scrutiny as acceptable indicators of viral health hazard in commercial shell- fish, and waters used for their cultivation (CDC, 1991 ). The pri- mary arguments against use of fecal coliforms as indicators of fecal pollution are: 1. Shellfish and waters with acceptable levels of fecal coliforms may still be contaminated with enteric viruses. 2. Depuration methods for removal of bacterial pathogens from shellfish have not consistently removed viruses. 3. Depurated shellfish have been proven a source of outbreaks of viral-induced gastroenteritis (Sobsey et al. 1987, Caul et al. 1993). In addition, Abad et al. (1997) demonstrated that pathogenic viruses could be detected in mussels from areas regarded as un- polluted, safe for swimming, and suitable for harvesting shellfish using standard bacterial indicators. In the absence of effective indicators, investigators have attempted the direct detection of vi- ruses from shellfish and their cultivation waters. Shellfish are readily contaminated with viruses present in sewage contaminated waters due to the concentration effect of filter feeding (Larkin & Hunt l982.Gerba 1988, Cromeanset al. 1997, Barardi etal. 2001 ). Since viruses do not replicate in shellfish tissues, the vector po- tential of shellfish is thought to be due to the stability of viruses in their tissues. The risk of acquiring a shellfish-borne viral disease is substantial as they are often eaten raw, including the intestinal tract (Heller et al. 1986, Wanke & Guerrant 1987, Speirs et al. 1987. Bouchriti & Goyal 1993. Kohn et al. 199.3). In addition, cookins '^Corresponding author. Tel.: -(■55-48-331-5207: fax: 1-55-48-331-9258; E- mail: ccblcrb@ccb.ufsc.br experiments performed with artificially contaminated mussels re- vealed that five minutes after the opening of the mussel valves, rotaviruses and hepatitis A virus could still be recovered in steamed shellfish. Under commercial depuration conditions, health-significant enteric viruses, such rotavirus and hepatitis A virus, could be recovered from bivalves after 96 hours of immer- sion in a continuous flow of ozonated marine water (Abad et al. 1997). Rotaviruses represent 80';^ of recognized viral etiologies, and 141) million cases of diarrhea per year worldwide. Structurally, rotaviruses belong to the "naked"" class of viruses. These are gen- erally more environmenlally resistant than enveloped viruses (Gerba et al. 1996). They are commonly found in wastewater and in addition can be concentrated by shellfish. Thus, the environment constitutes a significant reservoir for the virus (Bajolet & Chip- paux-Hyppolite 1998). Examination of shellfish for viral contami- nation has generally relied on cell culture (Berg et al. 1984, Cromeans et al. 1997, Sufien & Sobsey 1999). However, wild-type enteric viruses do not readily gi'ow in cell culture. They also gen- erally require long periods of adaptation and specific antigen as- says for virus detection, since no visible cytopathic effect is pro- duced in vitro. Polymerase chain reaction (PCR) can be used to enzymatically amplify nucleic acid sequences present in low copy numbers in environment samples to detectable levels. The speed, specificity, low cost, and ease of use of PCR have led to its in- creasing use in environmental science (Metcalf et al. 1995. Ab- baszadegan et al. 1999). Despite the sensitivity of PCR reactions, interfering substances in the environmental samples can severely limit the usefulness of this approach (Schwab et al. 1993. Atmar et al. 1993, Shieh et al. 199,5). The overall objective of this study was to evaluate and establish a new, molecular-based protocol to routinely monitor shellfish for infectious viruses, using rotavirus SAl 1 and the oyster Crassos- irea gigas as a virus-shellfish model system. We evaluated a simple protocol for oyster extract preparation, and a cost-effective method for viral RNA isolation from these extracts. After cDNA 997 998 Santos et al. synthesis, the strategy was to use a PCR kit together with specific primers capable of rapidly identifying the presence of infectious viral pathogens in environmental samples through genome ampli- fication with the accuracy and speed required for routine shellfish monitoring by laboratories. Rotaviruses were chosen as a model to develop this assay as they are responsible for severe gastroenteritis in humans and ani- mals. They have been implicated in outbreaks of waterborne gas- troenteritis in many countries (Ansari et al. 1991 ). After replicating in the gastrointestinal tract, they are excreted and may be dispersed in environmental waters (Kapikian & Chanock 1996, Hart & Cun- liffe 1997), The stability of human rotaviruses in environmental waters, and their resistance to physico-chemical treatment pro- cesses in sewage treatment plants may facilitate their transmission (Hurst & Gerba 1980). In addition, their presence in drinking water, (Deetzet al. 1984,Strappe 1991) sea water, (Goyal & Gerba 1983) and shellfish (Lewis & Metcalf 1988, Hatliger et al, 1997) has been described. MATERIALS AND METHODS \ iriis and Cells Simian rotavirus SAll (group A, simian serotype G3) was propagated and assayed by indirect immunofluorescence in MA104 cells, a continuous line of fetal Rhesus monkey kidney. Cells were cultivated in Eagle's minimal medium (MEM) supple- mented with IX non-essential aminoacids. 15 niM Hepes buffer, 2 niM glutamine and lO^^f (v/v) fetal calf serum (Birch et al. 1983). Infected fluid was titrated by indirect immunofluorescence assay as described by Barardi et al. ( 1998) using 10 |jLg/ml of trypsin in the maintenance medium (as described above without Fetal Calf Serum). Expcriimiilal Design fur Viral Recovery Studies Oyster Inoculation Oysters (Crassosrreii gigas) were obtained from a cultivated system oyster farm in Florianopolis City. Santa Catarina State. Brazil. Oyster shells were scrubbed with a stiff brush in running potable water (ca. 0.7 ppm free chlorine). Shell surfaces were then coated with a 70% ethanol solution and left to air dry for 30 minutes at room temperature in a biological safety cabinet. Shells were then opened at the hinge with an autoclaved oyster knife. Each oyster was seeded with 300 |j.l of rotavirus-infected tissue culture supernatant containing 3 x 10^ focus forming units (ffu) using a micropipette tip inserted at three points of the visceral area as described previously (Cromeans et al. 1997. Barardi et al, 1999). After adsorption for 30 minutes at room temperature, the oyster extract was prepared (see later in text). Negative controls consisted of washed and disinfected shells from non-seeded oys- ters from the same source. Tissues were then homogenized with a shaft blender (Ultra-Turrax T-25 Ika®) at 24,000 rpm for 30s. The resulting homogenate was placed in a 50 ml centrifuge tube, shaken at 250 rpm for 30 niin. at 22°C. and centrifuged at 10.000 X.i; for 30 min at 4°C, The pellet was discarded and supernatant pH adjusted to 7,5 using 2M HCl, Polyethylene glycol solution (PEG, MW = 8000) (50%, w/v) prepared in 10% TPB was added to a final concentration of 8% (w/v). The mixture was stirred for 2 h at 4°C, then centrifuged at 1 0.000 xg for 20 min a 4"C, The supernatant was discarded and the pellet resuspended in 5,0 ml of 0,15 M Na.HPOj. pH 9,0, The resuspended pellet was then sonicated at high power twice for 30s and transferred to a fresh centrifuge plastic tube and the pH read- justed to 7,4 using 2M HCl, This preparation was designated oys- ter extract B (Fig, 1), To study the effectiveness of this virus extraction method usuig both PCR and RT-PCR. unseeded oysters were extracted using the same procedure. These oyster extracts were then seeded with the same amount of virus used for preparation of spiked-whole-oyster extracts. Seeded extracts were designated as positive controls-A (Fig. 1). Viral RNA Extraction from O.vsters Extracts Two techniques of viral RNA extraction were compared for their effectiveness in detecting Rotavirus genome from oyster ex- tracts by RT-PCR. Technique I (Barardi et al. 1999): An aliquot (0.5 ml) of each oyster extract: A (positive control) and B (experi- mental extract) (equivalent to 2.5 x 10^ ffu) were used for viral RNA isolation. An equal volume of trichlorotrifluoroethane (Aid- rich) was added to remove lipids from each aqueous oyster extract, and the suspensions centrifuged at 6.(!)()0 rpm in a microcentrifuge (Eppendorf-) for t"ive minutes at room temperature. The aqueous phase was transferred to another sterile Eppendorf® tube. Tris-HCl (pH 7.5), EDTA, SDS and proteinase K were added at final con- centrations of 10 mM, 5 niM. 0.5% (w/v) and 400 jjig/ml, respec- tively. Samples were then incubated at 37"C for 30 min. Cetyltri- A unseeded ovsler B oj'sier seeded with rotavirus Tissues homogenized in TPB1U%. glycine 0 U5M pH 9,0 Spin at 10 OOQXg for .lOmin Adjusted the ptt of supernatant to 75 Polyethylene glycol solution for virus precipitation Stirred for 2h at 4°C Spin al 10 OOOXg for 20min at 4"C Discarded supenutant Oyster Extract Preparation The following method for preparation of virus extracts from oysters was examined for effectiveness of viral recovery for the subsequent RT-PCR assay (Barardi et al. 1999 with minor modi- fications): 10-20g of oysters flesh were transfen^ed to a sterile Schott!!' bottle containing 100 ml of pre-chilled (ice bucket) 10% (v/v) tryptose phosphate broth (TPB) (100% TPB contains 20g tryptose, 2.0g glucose, 5.0g NaCl and 2.5g Na,HP04, pH 9.0) prepared in 0.05M glycine (pH adjusted to 9.0 using 2N NaOH). Pellet suspended in Na.HP04(pH 9 0) Soni cation ( Positive contrul ) Oyster extract "X" seeded with the same amount of virus used tor exper. extract Experimental oyster extract "B* Figure L Outline of techniques used to prepare oysters extracts A. and B prior to detection using RT-I'CR. Rotavirus Detection in Oysters Using a PCR Kit 999 iiiethylammonium bromide (CTAB) and NaCI were added tn final concentrations of 1.3% (w/v) and 0.52 M. respectively, and samples incubated at 56°C for 30 min. Samples were subsequently extracted twice with an equal volume of phenol-chloroform- isoamyl alcohol (25:24:1). The aqueous phase was transferred to another microfuge tube and an equal volume of chloroform added. The aqueous phase was then precipitated in 3 vol of chilled 100*^ ethanol at -20°C. Resulting pellets were washed with chilled 70% (v/v) ethanol. suspended in 50 p.1 of Milli-Q water, and stored at -20°C for RT-PCR assays. Technique II: An aliquot (0.5 ml) of oyster extracts (A and B) was used for viral RNA isolation. Ac- cording to manufacturer specifications, each sample was added to 1.5 ml of TRIZOL LS'^' reagent and residual cells in the sample lysed by passing the suspension several times through a blue tip form P-1000 micropipette. Samples were incubated for five min- utes at room temperature, 0.4 ml of chloroform added, and the sealed tube shaken vigorously by hand for 15 sect)nds. Samples were then incubated at room temperature for 10 minutes followed by centrifugation at 12.000 xg for 15 minutes at 28"C. The aque- ous phase was transferred to a fresh tube and 0.5 ml of isopropyl alcohol added, followed by incubation at room temperature for 10 minutes and centrifugation (12.000 x?. 10 min. 4°C). The super- natant was removed and pellet washed once with 1 ml. 75% etha- nol. At the conclusion of the procedure the RNA pellet was air dried, suspended in 50 |j.l of Milli-Q water, and stored at -SO^C for RT assays. RT-PCR Assays The level of rotavirus detection by RT-PCR was determined for the extracts A (positive control) and B (experimental extract) using 2-fold serial dilutions of viral RNA ranging from 1:5 to 1 : 5,000,00 and corresponding, respectively, to 2 x 10^ to 0.015 ffu. Primers: the oligonucleotide primers BEG9 (5'-GGC TTT AAA AGA GAG AAT TTC CGT CTG G-3') and END9 (5'-GGT CAC ATC ATA CAA TTC TAA TCT AAG-3') (Gouvea et al. 1990), were used for reverse transcription and first PCR. These primers pro- duce full-length copies of gene 9 ( 1062 base pairs) from any group A rotavirus strain. The primers ET3 (5'-CGT TTG AAG AAG TTG CAA CAG-3') and END9 were used for the serotype-specific semi-nested PCR (second amplification-374 base pairs) (Gouvea et al. 1990), as the SAl 1 simian rotavirus, which belongs to serotype G3, has been used as a model for group A rotaviruses in this study. Two RT-PCR protocols were compared for their sensitivity to detection of the lowest quantity of viral RNA. Protocol 1: (Barardi et al. 1999 with minor modifications). 5.0 p,l aliquots of diluted RNA isolated either by technique I or II (in separate reactions) and 200 pmol BEG9 and END9 primers were heated at 99"C for 5 min, followed by chilling on ice for 1 min. The denatured RNA was added to the reaction mix consisting of 20 niM Tris-Cl, pH 8.4, 50 mM KCl, 0.4 niM each dATP. dCTP, dTTP and dGTP. 1.5 mM MgCK. 10 niM DTT and 40 units M-MLV reverse transcriptase (GibcoBRL). Reverse transcription of viral genomic RNA was carried out at 37''C for 90 min After this step, 5 units Tcuj DNA polymerase were added, and tubes overlayed with mineral oil. PCR amplification was carried-out for 25 cycles of 94°C for 1 min, 55°C for 2 min and 72=C for 3 min, and a final 10 min incubation at 72°C. For semi-nested PCR, 2.0 jxl of the first PCR product was added to the reaction mix of 50 p-l containing 20 mM Tris-HCl, pH 8.4, 50 mM KCl, 0.2 mM each dATP, dCTP. dTTP and dGTP, 3.0 mM MgCU, 100 pmol ET3 and END9 primers, and 5 units Tucj 1 2 345 67 89 500 374bp 500 B 1 23456789 41 If 100 374bp 374bp 374bp Figure 2. Semi-nested PCR products (A and B-tradltional PCR and C and D-Supernii\" kit) from simian rotavirus RN.\ extracted from oysters extracts A and B. (A) A extract (positive control). Lane 1: molecular weight marker: lanes 2 to 8: serial dilutions of viral RNA corresponding from 2 x 10^ to 312 tTu respectively: lane 9: negative control. (B) B extract (c\p. extract). Lanes 1 and 9: molecular weight marker: lanes 2 to 8: serial dilutions of v iral RN .\ corresponding from 1.6 X U)' to 625 ffu respeclivelj. (C) B extract (exp. extract). Lanes 1 and 15, molecular weight marker, lanes 2 to 1.^: serial dilutions of viral RNA corresponding from 1 x 1()~ to (1.(17 ffu. lane 14: negati\e control. )D) .\ extract (positive control). Lanes 1 and 17: molecular weight marker: lanes 2 to 15: serial dilutions of viral RN,\ corresponding from 1 X W to 0.015 ffu. lane 16: negative control. DNA polymerase. Tubes were then overlayed with mineral oil and the same PCR piogram used. Protocol 2: The PCR Supermix^ kit (GibcoBRL) was used in this protocol. The manufacturer's proto- col was followed, except for the concentration of MgCK (see later in text) Eive p,l aliquots of diluted RNA isolated either by tech- nique I or II (in separate reactions) were heated at 99°C for 5 min. then chilled on ice for 1 min. The denatured RNA was added to the reaction mix consisting of 50mM Tris-HCl. pH 8.3. 75 irM KCl. 3.0 mM MgCK. 0.5 mM each dATP. dCTP. dTTP and dGTP. 5 mM DTT. 100 pmol BEG9 and END9 and 40 units M-MLV reverse transcriptase. Reverse transcription of viral genomic RNA was carried out at 37°C for 90 min. After this step. 2.0 p.1 of reverse transcriptase product and 100 pmol BEG9 and END9 were added to 45 p.1 of the commercial reaction mix containing 22 mM Tris-HCl. pH 8.4. 55 mM KCl. 1.65 mM MgCU, 0.22 mM each dNTP and 22 units/ml Tac/ DNA polymerase. Additional MgCU was added to the reaction mix to achieve a final concentration of 3.0 niM. Tubes were overlayed with mineral oil and heated at 80°C 1000 Santos et al. TABLE 1. Sensitivity of RT-PCR for rotaviru.s detection using both PCR protocols H ith tlie different oyster extracts. Limit of Viral Detection (ffu) Oyster Extracts PCR Traditional PCR Supermix Kit A (positive control) B (experimental test) 312 0.03 625 0.15 tor 1 min. The same PCR piogram described above was then used. PCR products were analyzed by electrophoresis in 1.5% (w/v) agarose, or 10% (w/v) polyucrylamide gels. Both were stained with ethidium bromide at 10 |j.g/ml, and visualized under UV light. Either nucleic acids isolated from non-seeded oyster extracts or just distilled water were used as negative controls for the RT-PCR reactions. RESULTS Sensitivity of Rotavirus Detection Using RT-PCR Assav.v The sensitivity of two RT-PCR assays for virus detection was evaluated using total RNA isolated from oyster extracts A (posi- tive control) and B (experimental extract). The RNA from both extracts were isolated according to Techniques 1 and II. Both tech- niques were equally effective for RNA isolation for the RT-PCR reactions (not shown). The minimum amount of virus (ffu) that could be detected in the second PCR (semi-nested) using the tra- ditional PCR protocol (protocol 1 ) was 3 1 2 ffu for extract A (posi- tive control) and 625 ffu for extract B (experimental extract). PCR Supermix'-' kit (protocol 2) was also used for oyster extracts A and B In this case, the minimal amount of virus (ffu) that could be detected in the second PCR (semi-nested) was 0.03 ffu for extract A (positive control) and 0.15 ffu for extract B(exp. extract). These results can be seen in Figure 2. Table 1 summarizes the sensitivity of RT-PCR using both PCR protocols. DISCUSSION Nested or semi-nested RT-PCR systems are among the most sensitive molecular biological detection methods available today, with the ability for detection of as little as a single copy of a viral genome (Hafliger et al. 1997). This paper describes the evaluation and establishment of a rapid and reliable protocol for virus detec- tion in artificially-inoculated oysters using a commercial PCR kit for genome amplification. Rotavirus SAI 1 and the oyster Cnis- sostrea gigas were used as a model system to evaluate the efficacy of two viral RNA isolation techniques, and the sensitivity of sub- sequent virus detection using two different RT-PCR methods. Two procedures were compared for purification of viral RNA from artificially infected oyster extracts. There was no relevant differ- ence in rotaviius detection by RT-PCR between the two methods. TRIZOL LS'*' reagent, a mono-phasic solution of phenol and guanidine isothiocyanate, followed by chloroform proved to be effective for viral RNA purification. So, only results obtained with the less expensive method for nucleic acid extraction (phenol- chlorophorm) were presented, using two different RT-PCR proto- cols. Comparison of results obtained by RT-PCR. followed by semi-nested PCR detection, had confirmed that when the PCR Supermix" (a ready-to-use mixture of recombinant Taq DNA polymerase, salts, magnesium, and deoxyribonucleotide triphos- phates) is used, the limit of virus detection was 0.03 ffu for the positive control (A), and 0.15 ffu for the experimental extract (B). In both cases, this was 4.000 to 10.000 times more sensitive when compared with the vims detection limit using the traditional PCR protocol where all reagents were added separately. This study demonstrates improvements and simplification of a highly sensitive and simple method to prepare oyster tissue ex- tracts, and application of a specific and sensitive RT-PCR based system for the detection of rotavirus. We anticipate that the meth- ods described here will be further improved and optimized in a single multiplex PCR reaction for detection of a greater variety of human enteric viruses commonly found contaminating shellfish and aquatic environments. ACKNOWLEDGMENTS C. S. Santos was a Masters Student fellow from CAPES (Co- ordenadoria de Apoio a Pesquisa e Ensino Superior). C. Rigotto is a UFSC Master Student also supported by CAPES. This work was supported by the BMLP (Bra/ilian Mariculture Linkage Program) financed by CIDA (Canadian International Cooperation Agency). The support of CAPES. BMLP and CIDA support is gratefully acknowledged. We thank Dr. James J. Smith for his review of, and comments on this manuscript. LITERATURE CITED Abad. F. .\.. R. M. Pinto. R. Gajardo & A. Bosh. 1997. Viruses in mussels: public health implications and depuration. J. Food Protect 60:677-68 1 . Abbaszadegan. M.. P. Stewart & M. Le Chevallier. 1999. A strategy for detection of viruses in groundwater hy PCR. Appl Environ. Microbiol. 65:444-449. Ansari. S. A.. V. S. Springthorpe & S. A. Saltar. 1991. Survival and vehicular spread of human rotaviruses: possible relation to seasonality of outbreaks. Rev. Inf. Dis. 13:448-461. Atmar. R. L.. T. G. Metcalf. F. H. Neill & M. K. Estes. 1993. Detection of enteric virus in oysters by using the polymerase chain reaction. Appl. Environ. Microbiol. 59:631-635. Bajolel. O. & C. Cippaux-Hyppolite. 1998. Rotaviruses and other diarrheal viruses. Bull Soc. Pathol. Exot. 91:432^37. Barardi. C. R. M.. K. R. Emslie. G. Vesey & K. L. Williams. 1998. Development of a rapid and sensitive quantitative assay for rotavirus based on flow cytometry. J. Virol. Methods 74:31-38. Barardi. C. R. M.. H. Yip. K. R. Emslie, G. Vesey. S. R. Shanker & K, L. Williams. 1999. Flow cytometry and RT-PCR for rotavirus detection in artificially seeded oyster meat. Int. J. Food Microbiol. 49:9-18. Barardi. C. R. M.. C. S. Santos & C. M. O. Simoes. 2001. Ostras de qualidade em Santa Catarina. Ciencia Hoje 29:70-73. Berg. G.. R. S. Safferman. D. R. Dahling, D. Berman & C. J. Hurst. 1984. U.S. EPA Manual of Methods for Virology. EPA-600/4-84-013. U.S. Cincinnati: Environmental Protection Agency. Birch. C. J.. S. M. Rodger, J. A. Marshall & I. D. Gust. 1983. Replication of human rotavirus in cell culture. J. Med. Virol. 1 1:241-250. Bouchriti. N. & S. M. Goyal. 1993. Methods for the concentration and detection of human enteric virus in shellfish: a review. Microbiologica 16:10.5-114. Caul. E. O.. N. J. Sellwood. D. W. Broun, el al. 1993. Outbreaks of gastroenteritis associated with SRSVs. PHLS. Microbiol. Dig. 10:2-8. Centers for Disease Control. 1991. Gastroenteritis associated with con- sumption of raw shellfish-Hawaii. 1991. Morbid. Mortal. Weekly Rep. 40:303-305. Rotavirus Detection in Oysters Using a PCR Kit 1001 Cronieaiis. T. L., O. V. Nainan & H. S. Margolis. 1997. Detection of hepatitis A virus RNA in oyster meal. A/i/j/. Eiiviroii. Microhial. 63: 2460-2463. Deetz, T. R.. E. M. Smith. S. M. Goyal. et al. I9S4. Oeciirrence of rota and enterovirus in developing nations. Water. Res. 1 8:567-.'i7 1 . Gerba. C. P. I98S. Viral disease transmission by seafoods. FdocI TccIiiihI. 42:99-103. Gerba. C. P.. J. B. Rose. C. N. Haas & K. D. Crabtree. 1996. Waterbome rotavirus: a risk assessment. Wen. Res. 30:29-30. Gouvea. V., R. I. Glass. P. Woods, et al. 1990. Polymerase chain reaction and typing of rotavirus nucleic acid from stool specimens. J. Clin. Microbiol. 28:276-282. Goyal. S. M. & C. P. Gerba. 1983. Viradel method for detection of ro- tavirus in seawater. J. Virol. Methods 7:95-106. Hatliger. D.. M. Gilger. J. Luthy & P. Hubner. 1997. .Seminested RT-PCR systems for small round structured viruses and detection of enteric viruses in seafood. Int. J. Food Microbiol. 37:27-36. Hart, C.A. & N. A. Cunliffe. 1997. Viral gastroenteritis. Ciin. Opin. Infect. Pis. 10:408-113. Heller. D.. O. N. Gill. E. Raynham. T. Kirkland. P. M. Zadick & R. Stanwell Smith. 1986. An outbreak of gastromtestina! illness associated with consumption of raw depurated oysters. Br. Med. .1. 292:1726- 1727. Hurst. C. J. & C. P. Gerba. 1980. Stability of simian rola\ irus in fresh and estuarine water. Appl. Environ. Microbiol. 39:1-5. Jaykus. L. A., M. T. Hemard & M. D. Sobsey. 1994. Human enteric pathogenic viruses. In: C. R. Hackney & M. D. Pierson, editors. En- vironmental indicators and shellfish safety. New York: Chapman & Hall. pp. 92-153. Kapikian. A. Z. & R. M. Chanock. 1996. Rotaviruses. In: B. N. Fields. D. N. Kmpe. R. M. Howley. R. M. Chanock. J. L. Melnick. T. P. Monath. B. Roi/.man & S. E. Satraus. editors. Fields Virology. 3.ed. Philadel- phia: Lippincott. pp. 1657-1708. Kohn. M. A.. T. A. Farlev. T. Ando. et al. 1995. An outbreak of Norwalk virus gastroenteritis associated with eating raw oysters: implications for maintaining safe oyster beds. J.AMA. 273:466-471. Larkin. E. P. & D. A. Hunt. 1982. Bivalve moUusks: control of microbio- logical contaminants. Bioscience 32:193-197. Lewis, G. D. & T. G. Metcalf. 1988. Polyethylene glycol precipitation for recovery of pathogenic viruses, including hepatitis A virus and human rotavirus, from oyster, water, and sediment samples. Appl. Environ. Microbiol. 54:198,^-1988. Metcalf. T. G.. 1, L. Melnick & M. K. Estes. 1995. Environmental virol- ogy: from detection of virus in sewage and water by i.solation to iden- tification by molecular biology — a trip of over 50 years. Annn. Rev. Microbiol. 49:461-487. Richards, G. P. 1985. Outbreaks of shellfish-associated enteric virus illness in the United States: reqiusitc for de\elopnienl of viral guidelines. J. Food P rot. 48:815-823. Schwab. K. J., R. De Leon & M. D. Sobsey. 1993. Development of PCR methods lor enteric virus detection in water. Water Sci. Tech. 27:21 1- 218. Shieh. Y. S. C. D. Wait, L. Tai & M. D. Sobsey. 1995. Methods to remove inhibitors in sewage and other fecal wastes for enterovirus detection by the polymerase chain reaction. / Virol. Methods 54:51-66. Sobsey. M. D.. A. L. Davis & V. A. Rullman. 1987. Persistence of hepatitis A virus and other viruses in depurated eastern oysters. In: Proceedings Oceans 87, Halifax, Nova Scotia, vol 5. Washington. DC: Marine Technology Society, pp. 1740-1745. Speirs. J. I., R. D. Pontefract & J. Harwig. 1987. Methods for recovering poliovirus and rotavirus from oysters. Appl. Environ. Microbiol. 53: 2666-2670. Strappe. P. 1991. Rotavirus detection: a problem that needs concentration. Water Sci. Tech. 24:221-223. Sufien, E, & M. Sobsey. 1999. Recovery and detection of enterovirus, hepatitis A virus and Norwalk virus in hardshell clams {Mercenaria tnercenaria) by RT-PCR methods. / Virol. Methods 77:179-187. Wanke. C, A, & R. L. Guerrant, 1987. Viral hepatitis and gastroenteritis transmitted bv shellfish and water. Infect. Dis. Clin. N. Am. 1:649-664. Mmrmil nj Slwlljisl, Research. Vol. 20. No. .^. lOO.VIOIO. 2(101. POPULATION STRUCTURE AND DEMOGRAPHY OF THE PUELCHE OYSTER (OSTREA PUELCHANA, D'ORBIGNY, 1841) GROUNDS IN NORTHERN PATAGONIA, ARGENTINA MARCELA S. PASCUAL,'* EDUARDO A. ZAMPATTl,' AND OSCAR O. IRIBARNE"** ' Institiito de Biologi'a. Marina y Pesqueiu. "Alte. Stonii". CC 104, (8520) San Antonio Oeste. Rio Negro. Argentina: -Faciiltad de Ciencias E.xacta.s y Naturales. Universidad Nacional de Mar del Plata. CC 573. (7600) Mar del Plata. Argentina AliSTR.\CT The three main grounds of ihe flat oysler ( Oxtrea puclclmmi. D'Orb. ) are located at the northwestern portion of the San Malias Gull, at Northern Patagonia. Argentina. The population structure and demography of the grounds at Banco Reparo. Las Grutas and Bajo Oliveira were .studied through several diving and dredge surveys performed between \'^%5 and 1991. The complete population of the grounds located at Banco Reparo and Bajo Oliveira, and the population of the central part of Las Grutas ground, were assessed, measuring oyster density and spatial distribution, size structure, proportion of clustered and free oysters, proportion of oysters carrying epibiotic males, and oyster's life habit. The puelche oyster is distributed almost continuously along the northern coastal area of the San Matfas Gulf, at depths ranging from two to 20 minutes. The grounds are likely to be interconnected by larval dispersion. The three studied grounds differed in their demographic characteristics, and it was difficult to compare biomass between the grounds due to different methods of assessment. However, the results enabled us to conclude that high oyster density areas were always limited on any of the grounds, with maximal densities only on relatively small patches. Patches of recruits were never found in the surveys, suagesting that a clear spatial segregation of settlement does not occur in this species. Oyster size structure at Banco Reparo and Las Grutas was similar, while individuals at the Bajo Oliveira population were very large, with mean population size 20 mm larger than on the other two grounds. Size classes below 45 mm were always poorly represented or absent on the three grounds. The portion of the population comprising of individuals <30mm was represented by the epibiotic males who represented a "hidden" mode. The three populations differ in their dynamics, mainly in the turnover rates of individuals. At Bajo Oliveira the medium size individuals are absent (there is a hiatus between the two modes) and the mean size of adult oysters is the largest (89 mm). In contrast, at Las Grutas. the intermediate sizes between the two modes are present (there is no hiatus between the two modes) and the mean size of adults is the smallest (66 mm). Banco Reparo occupies an intermediate position. Based on this information we suggest that turnover rate is then maximal at Las Grutas. intermediate at Banco Reparo. and minimal at Bajo Oliveira. The puelche oyster grounds at the San Matfas Gulf offered the rare opportunity of studying the structure of one of the few natural and untouched flat oyster populations in the world. Work helped support the closure of the fishery as a policy mainly based on the vulnerability of the stocks. KEY WORDS: oysters. O-srici; piicUhaiia. Patagonia, natural grounds, demography, density distribution INTRODUCTION Oslrca puclcluiiui (d'Orbigny 1841). ciuiimonly named "ptielche" or "Patagonian". is a flat oyster belonging to the Family Ostreidae (Subfamily Ostreinae) (Stenzel 1971). This species is distributed from Rio Grande do Sul (30°S, 50°W. Brasil) to Bahia Camarones. Northern Patagonia (44°S. 66°W. Castellanos 1957) (Fig. 1 1. Dense aggregations ("grounds'") have been reported only at the southern limit of their distribution (Valette 1929. Castellanos 1957. Fernandez Castro 1987. Pascual & Bocca 1988. Pascual 1993). However, there has been recent evidence of the presence of the species in deep waters (30-35 m depth) off the State of Santa Catarina, Brasil (Carlos Polis & Jaime Ferreira. pers. comm.). The distributional area of O. puelchana should then be extended north- wards. Isolated individuals regularly appear along the northern Argentinean coast as by-catch in the mussel dredge fishery off Buenos Aires shores (M. Lasta. pers. comm.). and at intertidal sites of the same area. The occasional appearance of Hat oysters has been reported for several localities along the coast of Buenos Aires Province (35-40°S: 57-62°W). The presence of this species in Argentina was first reported by D'Orbigny (1841) who. as indicated by Castellanos (1957) misi- dentified this species as Ostrea spiela. a non commercial Hat oyster that cohabits with O. puelchana at shallow coastal waters of the San Matfas Gulf (40°40'S; 63°30'W) (Fig. 1). O. spreta is a small oyster, pink to purple colored, that attains a maximal size of 50 mm and forms beds inside San Antonio Bay (Fig. 1 1, in envi- ronments of strong tidal currents. Extreme types of both species *E-niail; ostralff'canaldig.com.ar **E-mail: osiriba(S'mdp.edu.ar may be easily recognized but a wide gradient of morphological characteristics makes the identification of young indi\ iduals dif- ficult. Ostrea puelchana has solid shells, lamellated and subnacreous, circular or subsquare. The upper or right shell is flat and lamel- lated. The lower or left shell is bigger, lamellated and convex (Castellanos 1957). The individuals have a diverse coloration, the juveniles are homogeneously cinnamon colored but frequently showing shells with longitudinal purple estriae. The adult speci- mens loose their cinnamon color and show yellow greenish or gray shells. The puelche oyster lives isolated or in clusters. A cluster is formed when the larvae settle and grow on the shells of a "found- er" oyster. The maximal size registered in this species is 140 mm of total height (TH: largest diameter from the umbo to the opposite margin). This size is rare, however, with size generally up to 120 mm in deep populations. Age is difficult to assess in this species; attempts at age revelation from readings on shell cuts have been unsuccessful. Nevertheless, external characters such as the thick- ness of the ligament and the rate of infection by perforating sponges, Clioiui cclatu or the Mitylid Litophaga patagonica. sug- gest that it is a long-lived species. The structure of some of the puelche oyster populations of the San Matfas Gulf has been studied for the last 20 years. Most of this information is contained in unpublished technical reports (Vacas 1978. Vacas 1979). a thesis (Pascual 1993). and some has been published (Morriconi & Calvo 1989). A great amount of research effort was dedicated to the sexuality of this species (Calvo & Morriconi 1978. Morriconi & Calvo 1979. Fernandez Castro 1987, Fernandez Castro & Lucas 1987. Pascual et al. 1989. Pascual & 1003 1004 Pascual et al. Zampatti 1995, Pascual 2000). O. puelchanci. while fitting the general reprodiieti\e pattern of broeiding Ostreas, shows an alter- native breeding system that makes it unique among Ostreids. In natural grounds, oysters change their sex to the female phase when they reach a shell size of 50-55 mm. At this point, the oyster acquires the ability of "carrying" small conspecific individuals settled on an expansion ("platform") of the anterior margin of the concave shell (Calvo & Morriconi 1978). These epibiotic oysters are always males during the reproductive season and their gonads mature synchronously with the gonads of its female "carrier" (Calvo & Mon'iconi 1978). The epibiotic males have their growth rate severely reduced by the influence of the female carrier (Pas- cual et al. 1989). This "dwarf' condition assures the persistence of each couple (carrier oyster + epibiotic male) for long periods of the carrier's lifetime. The populations of the puelche oyster of the San Mati'as Gulf have been protected by a legal fishing closure established in 1971 . This rare legal interdiction has been based for many years on the assumption of the vulnerability of the stocks. However, it was not before 1985 that a thorough sampling of the oyster grounds was initiated through the San Mati'as Gulf Oyster Project. This project was intended to gain knowledge on the demographic structure and spatial extent of the puelche oyster stocks, and on the reproductive ecology of the species. Simultaneously, an Oyster Culture Project has been developed. The present paper presents information obtained from exten- sive surveys performed between 1985 and 1991 on the population structure and dimensions of the three main oyster grounds of the NW San Mati'as Gulf: Banco Reparo (BR). Las Grutas (LG). and Bajo Oliveira (BO) (Fig. 1). MATERIALS AND METHODS Study Site and Background Information The Banco Reparo oyster ground i; Mati'as Gulf (40°40'S; 63°30"W) (Fig. located in the NW San 1 ). The bottom is sandv. Fisjiire 1. Lociition of the three main ovsler jjniunds alon^ the San Mati'as Gulf coast. Northern Patagonia ( 1 = Banco Reparo; 2 = Las Grutas, and 3 = Bajo Oliveira). partially covered by pebbles and mollusk shells. Currents are strong (26-52 cm s"' ) and depth ranges from 2-3 m at low tide. Tidal range is 7.62 m in spring, and 5 m during mean tides (Ser- vicio de Hidrografi'a Naval. 1969). Seawater temperature ranges from 7°C (August) to 23.5°C (January); and salinity ranges from 34-36'^f (Fernandez 1989). The highest values of chlorophyll a are reached in March (23 mg 1"'). while the maximal values of phosphates and nitrites/nitrates are reported in spring (0.13 jjig 1"'. and 9.25 (j,g P' respectively) (Fernandez 1989). Dissolved oxygen fluctuates seasonally reaching its maximal value in August (7.69 mg r') (Fernandez 1989). The Las Grutas oyster ground is located on open waters on the NW coast of the gulf 15 km south from Banco Reparo (40°48'S; 65°05'W. Fig. 1 and Fig. 3). The bottom is composed of coarse sand and shell, occasionally interrupted by limestone platforms crossed by channels filled with sand. Tidal currents are weaker than in Banco Reparo (20-30 cm s"') and depth ranges from 2.5 to 6 111 at low tide. Oysters live on sandy flats, on limestone platforms, and among dense stands of the macroalgae Codhmi sp. Water temperature ranges from 8°C (August) to 2rC (January), and salinity between 34 and 35%(! (Fernandez 1989). The highest values of chlorophyll a were reported for March (42 mg 1"'. in 1987). Phosphates peak in April ( 10.5 p-g r' ). and nitrates/nitrites peak in March (13.5 jjig 1"'). Dissolved oxygen shows a marked seasonality, reaching its winter peak on August (6.8 ml P') (Fernandez 1989). The Bajo Oliveira oyster ground is located south from Las Grutas ground, also in the NW of the gulf (40"50'S; Figs. 1 and 3). The bottom is predominantly sandy and flat, covered by pebbles, small shell pieces and mollusk shells. Depth ranges from 4-22 m at low tide. Available oceanographic information is scarce and was gathered during a two survey conducted during 1971 and 1991 over the entire gulf, both performed in autumn, during which only one or two stations were located at Bajo Oliveira (Carreto & Ve- rona 1971. Esteves et al. 1991). During those surveys salinity ranged from 33.9-34. 9%f, chlorophyll a from 0.7-1.8 mg 1"'. ni- trates were 0.78 |a,g I"', phosphates were 1.33 |j.g 1"' and water temperature was 17°C. Sampling Design At Banco Reparo. an area 1853 m long and 400 in wide, cor- responding to the surface actually occupied by oysters, was posi- tioned, on the basis of land references, and marked by buoys. A rectangular grid, comprising of 18. 2-m wide transects, each sepa- rated by a 100 m distance, was designed. Once each transect was positioned and marked, all the oysters found in each linear segment ( 10 X I m) at one side of the first sub-transect were collected by divers. The procedure was then repeated in the opposite direction. Thus. 20 sampling units were obtained on each sub-transect, and 80. as a maximum, over the entire transect. The eastern limit of each transect was determined by the absence of oysters along a distance of 100 m. The described procedure was performed 18 times and took 10 working days, from 23 October to 12 November 1985. The number of free oysters and clustered oysters was reg- istered in each sample. All the oysters visible to the naked eye and one out of ten samples were collected and sized (total height in mm). Density data (oysters 10 m"~) were used to produce a map of the population density distribution at the ground, this was used to estimate the area with different oyster densities. The area of the sectors included between isopleths was estimated by graphic com- puter software. PuELCur; O'lSTiiR Grounds in Northern Patagonia 1005 e , Figurt 2. Distribution and ahundance of oysters in Banco Kcparo. Maximal densities are concentrated in the black stratum. Densities are number of oysters Id m '. Complementarily, sampling was performed to assess the pre- dominant life position of individuals of different size ranges. Two transects, located in the center of the ground, were selected, and the divers registered, for each individual present along the two transects, its size and life position (oysters lying on its flat or on its concave shell). Oysters were allocated into three size groups: small (<3 cm); medium (3-7 cm) and large (>7 cm). Sampling of oyster population structure at Banco Reparo was repeated October I9S6 and November 1987 in the areas identified as high oyster density spots during the 1985 survey. A sample of all the oysters present in a 100 m transect was collected. Oysters from the samples were measured (total height: TH). The number and size of epibiotic (dwarf) oysters per carrier oyster were reg- istered. At Las Grutas only the central part of the ground, which was determined by previous sampling to have the highest density (marked by a star in Fig. 3). was prospected to describe the popu- lation structure. A station was fi.xed by a marking buoy and acted as a starting point from which seven radially spreading transects were laid out. Five to seven sampling units ( I m") were collected along each transect, at 1 5-20 m intervals. All the oysters within the frame were collected. The survey was completed between 16 to 28 November 1987. The sampled oysters were separated as free and clustered oys- ters; they were measured (total height in mm) and counted. The epibiotic oysters fixed on the platform of carrier oysters (previ- ously sacrificed) were also counted and measured. Following the same procedure as in Banco Reparo. life position of oysters was also evaluated by diver collection of all oysters along a 200 m transect. Oysters lying on their flat shell or concave shell were collected in separate bags. The Bajo Oliveira oyster ground, due to its great extent, had to be surveyed with a fishing research vessel equipped with a com- mercial dredge (mouth width: 2.5 m). Two surveys were con- ducted in June 1987 and Mav 1988. The first was the more detailed OYSTERS 10 m-' K / ] I 0.06-0 006 ) / / 0 n 0.6-0.061 I f if/ Sr< in 0.61-6 )) I / IV >6 / // \/ III SAN MATIAS GULF Figure 3. Oyster grounds at Las Grutas (the star indicates the central area of the ground where the survey was performed), and Bajo Oliveira with a map showing distributi9 0,24 12.9 3.09 Biiju Oliveira I ().()6-().()06 11.19 0.03 0.01 11 0.6-0.061 27..^4 0.57 0.78 III 0.61-6 46.60 2.51 1.39 IV >6 .^.2-S 7.06 - density for the complete prospected area was 8 oysters per ni" (s = 5.3; N = 41). The oyster ground of Bajo Oliveira occupies an area of 88 km", forming a ground elongated from north to south within the isobats of 10 and 20 m (Fig. 3). The northern portion of the ground extends out from the 10 m isopleth. extending onshore to a depth of 4 m (Fig. 3). The higher density area (Fig. 3; Table 1) (3.25 km") was located in the NW portion of the ground. The mid density areas occupy a central fringe and represent a 52.60% of the total area. The western and eastern sides of the ground are defined by low-density patches (Fig. 3). The maximal oyster density re- ported at this ground was eight individuals per ni". The population is concentrated in a narrow depth range (5-13 m). Fishing hauls were also concentrated in this depth range and mainly in the 10 m isobat (46.6%). Population Structure The population of Banco Reparo is homogeneous, lacking spa- tial segregation of adults and juveniles. Free oysters (non clus- tered) had a mean size of 68.69 mm (s = 17.34; N = 1327) (Fig. 4). .Ml samples had a very low number of individuals smaller than 40 mm (<59f). The percentage of oysters foi'ming clusters in the 1985 samples ranged from 0-34%. Over the total stock of oysters examined in 1985, 1986 and 1987. the percentage of clustering was 19% (N = 1574); 88% of the clusters composed of 2 oysters, 9% by 3 oysters, and 3% by 4 oysters (N = 1 14). No clusters comprising of more than four oysters were reported at this ground. Mean size of free oysters (x = 77.93 mm; s = 13.3; N = 122) was only slightly higher than the mean size of clustered oysters (X = 71.13 mm; s = 16.87; N = 122) (Fig. 4). The 1986 and 1987 samples enabled the completion of the demographic study of the ground including the component corresponding to the epibiotic oysters settled on the platform of carrier oysters. The size fre- quency distributions were clearly bimodal (Fig. 4). In 1986, the first mode (epibiotic males) was 10.69 mm (s = 7.06; N = 111), and the second (free oysters), was 70.62 mm (s = 15.09; N = 209). In 1987. the modes were 12.69 mm (s = 6.82; N = 152), and 77.56 mm (s = 12.15; N = 388). The percentage of "carriage" (oysters earning dwarf males), in oysters bigger than 50 mm. estimated from samples of the 1986 and 1987 surveys (N = 946). was 32%. The carrier oysters of PuELCHE Oyster Grounds in Northern Patagonia 1007 BANCO REPARO LAS GRUTAS 35 30 25 20 15 10 5 > o z UJ 3 o 111 D free oysters ■ clustered oysters ■ I iririr m I su 40- free oysters 30 epibiotic males A 20 A 10 0 -L \LaJ l_ 5 25 35 45 55 65 75 85 95 105 115 total height (mm) Figure 4. Banco Reparo. Size fre(|ucnc) distribution (total height in mm) ol free and clustered oysters (abovel, and size frei|ucncv distri- hulion of the complete population composed of free and epibiotic (dwarf males! oysters (below I. Banco Reparo held a maximum of seven epibiotic males on the platform per oyster; 619^ of the oysters carried one epibiotic male (mean = 1.69; s = 1. 1 2; Table 2). The total population of oysters at Las Grutas had a mean size of 66.67 mm (s = 18.67; N = 722). Free (non clustered) oysters had a mean size (x = 69.05 mm; s = 18.18; N = 332) 10 mm higher than the group composed by clustered oysters (x = 59.33 mm; s = 23.99; N = 317) (Fig. 5). The whole oyster population (including epibiotic males) presented a clear biinodal structure. The first mode, representing epibiotic males fixed on carrier oys- ters, was 8.2 mm (s = 6.3; N = 420), the second, representing the rest of the population, was 66 mm (s = 18.6; N = 722) (Fig. 5). The clusters were composed of a maximum of 1 1 oysters of TABLE 2. Mean number of epibiotic males per carrier oyster at the three prospected grounds (Banco Reparo, Las Grutas and Bajo Oliveira). Oyster Ground N" of Epibiotic Males/Carrier Oyster SD N Banco Reparo Las Grutas Bajo Oliveira I h<) I.4N i.y? I 12 ().K6 0.95 365 260 151 >- O z UJ O UJ 20& ISO- 100 5 15 25 35 45 55 65 75 85 95 105 115 TOTAL HEIGHT (mm) Figure 5. Las Grutas. Size frequency distribution (total height in nun) of free and clustered oysters (above), and size frequency distribution of the complete population composed of free and epibiotic (dwarf males) oysters (below). different sizes, those composed of only two oysters represented 479f of the total (mean = 3.26; s = 1.63). The proportion of clustered oysters was high in this ground; 449^ of the sampled oysters (N = 627) had this life habit. The total estimated carriage (number of oysters >50 mm carrying epibiotic males) in the ground was 349c (N = 490). A maximum of six epibiotic males per carrier was registered, but 687r of the carriers (N = 259) were oysters carrying only one epibiotic male (Table 2). The population of Bajo Oliveira is composed of large individu- als. The mean size of oysters per haul in 1987 ranged from 77.(5- 94.2 mm (Fig. 6). Free oysters (isolated or clustered) presented a TABLE 3. Percentages of clustering and carriage in oysters of the three grounds. Ground % Clustering N % Carriage N Banco Reparo 19 1574 32 946 Las Grutas 44 627 34 490 Bajo Oliveira 3.9 862 88.28 862 BR vs LG ts = 3.68* ts = 0.77 n.s. LG vs BO ts = 20.31*** ts = 21.83*** Br vs BO ts = 11.64*** ts = 21.51*** SD = standard dcMatuin. Banco Reparo (BR). Las Grutas (LG). and Bajo Oliveira (BO). Peicentages (arcsin transformed) were compared between grounds by a test of equality of percentages (P = 0.001) (Sokal & Rohlf 1969). 100« Pascual et al. BAJO OLIVEIRA BANCO REPARO LAS GRUTAS >- o z UJ o epibiotic males III. ..ill free ^ oysters 0 % clustering 45 55 65 75 85 TOTAL HEIGHT (mm) 95 105 115 125 Figure 6. Bajo Oliveira. Complete size structure of the oyster popu- lation, including the component represented by dwarf males carried on the platform of carrier oysters. Samples collected on the 1987 sur\ey . UJ z LU o a. LU 0. O flat ■ concave Figure 7. Fife position of oysters at Banco Reparo and Las Grutas. Proportion of oysters living with its concave shell up ("concave"), and with its Hat shell up ("Hat"). Oysters were divided into three size groups: small (S = TH < 3 cm), medium (M = TH = 3-7 cm), and large (L = TH > 7 cm). mean total height of 89.01 mm (s = 13.39: N = 1550) in 1987. anij 90.2 mm (s = 7.52; N = 412) in 1988. The entire population, including epibiotic males, is bimodal (Fig. 6). The mode repre- senting the epibiotic males is 9.09 mm (s = 6.53; N = 532). CaiTiage levels found on this ground were very high both in 1987 (83.6-97.6%) and in 1988 (74.19-100%). Mean carriage obtained pooling all samples was 88.28% (N = 862). Carrier oysters had a mean of 1.95 epibiotic males per oyster (s = 0.95; N = 151; Table 2) with five the maximum reported. The oysters of Bajo Oliveira are predominantly non-clustered. Pooling all data from 1987. 0% of clustering (N = 302) was found (Fig. 6), and 6.8% in 1988 (N = 560). Total clustering for both surveys was 3.9% (N = 862). All clusters were of two oysters. The comparison of clustering and carriage percentages between grounds showed significant differences (P < 0.001 ) in ail pairs of comparis(Mis, with the exception of the percentages of carriage in Banco Reparo and Las Grutas where no significant differences were detected (P > 0.001 ) (Table 3). Life position A total amount of 1 .858 oysters were sampled at Banco Reparo. 89.8% were found lying on their tlat shell. Life position at this location appears related to the oyster's size with 81.8% of oysters smaller than 3 cm found lying on their concave shell and medium and large oysters (90.8% and 94.4%. respectively) found lying on their tlat shell (Fig. 7). Life position of 1.779 oysters was assessed at Las Grutas, 61% of which were lying on their tlat shell. When data were separated by size groups, it was concluded that even when the proportion of oysters lying on their tlat shell was higher, the percentages were almost equal in the case of large individuals (Fig. 7). DISCUSSION Surveys performed on the three main oyster grounds in the San Mati'as Gulf (Banco Reparo. Las Grutas and Bajo Oliveira) en- abled us to conclude that the puelche oyster is distributed almost continuously along the northern coastal area of the San Mati'as Gulf, at depths ranging from 2-20 m. The long planktonic larval life of this species (20 days. Pascual et al. 1989; Pascual and Zampatti 1995) suggests that larvae bom in any of the studied grounds are capable of being transported long distances before settling. The grounds or high density areas, are likely interconnected by larval dispersion. This hypothesis rests upon the dynamics of residual coastal currents at the northern portion of the gulf (Lanfredi & Pousa 1988). The main cun'ent has two components; one flowing N-NE. and the second, flowing S-SE. The tidal wave follows the shoreline. Banco Reparo and Puma Viilarino (Fig. 1 ) may then act as barriers, forcing the water to circulate describing a clockwise gyre. The three studied grounds differed in their demographic char- acteristics. It was difficult to compare the grounds regarding their biomass due to the different assessment methods used. However, the results show that high oyster density areas were not large on any of the grounds, with maximal densities only on relatively small patches. Oyster size structure at Banco Reparo and Las Grutas was similar. Size classes below 45 mm were better represented at Las Grutas than at Banco Reparo, even when on both grounds indi- viduals <20 mm were not present in the samples. The Bajo Oliveira population showed a different demographic structure; it was represented by very large individuals, with mean size 20 mm larger than on the other two grounds, and no individuals <55 mm. Nowhere did we find patches of recruits, as is the case of other native gulf species such as the scallop. Aequipecten telniekinis. or the mussel, Mytihts edulis platensis, suggesting that a clear spatial segregation of settlement does not occur in this species. On all three oyster grounds the portion of the population composing of indi\iduals ranging from 1-30 mm was represented by epibiotic males. The resulting scenario showed that, in the case of the puelche oyster populations, a distinction should be made between the ap- parent demographic structure, represented by adult and sub adult oysters, and the real demographic structure, that includes the small epibiotic oysters, an inconspicuous portion of the popu- lation. This "hidden" mode, added to the free portion of the popu- lation, shaped the bimodal size distributions typical of these popu- lations. The apparent structure is intriguing in the case of a protan- PUELCHE OlSTER GROUNDS IN NORTHERN PaTAGONIA looy dric, sex alternating species, because the portion of tlie population that is absent, or poorly represented, corresponds to the free males (going through their first sexual maturation). These males are sexually more acti\e than those in which the spermatogenesis oc- curs as a post-spawning event (females that become males after their first main spawning of the season) (Morriconi & Calvo 1979). In conclusion, the apparent population is primarily female. The usual life position of free oysters is that in which the Hat shell lies against the bottom. This orientation seems to be hydro- dynamically more stable, facilitating water flow and minimizing the risk of being drifted by currents. The pattern observed at Banco Reparo. where this life position is exclusive for adult oysters, seems to be adequate for shallow and strong energy environments as an adaptation to resist smothering, condition under which the animals are capable of living for long periods. Clusters are typical in this species, they are formed by the fixation and subsequent growing of recruits on a "founder" oyster (Morriconi & Calvo 1989). The incidence of this life habit differs among grounds (LG > BR > BO) (Table 3). At Bajo Oliveira. clusters are poorly represented (4%). This gradient among grounds is repeated by the number of oysters forming clusters (LG > BR > BO). Morriconi and Calvo ( 1989) compared the incidence of this lil'e habit in the three populations. The gradient that they described is similar to the one reported in this paper, even when the percentages of clustering are lower. This difference could be because their sampling lacked a specific design (Morriconi & Calvo 1989). The collection of samples in some spots of the ground may render a biased image of the overall structure. The authors describe, as well. the sex ratio of free and clustered oysters at Las Grutas, concluding that the proportion of males is higher on clustered (51.9%) than on free oysters (31.6'^f ). This result would be expected if we analyze the size structure of both groups: clusters include a higher propor- tion of young oysters, ergo, of oysters still in their male protandric sexual phase. This result, however, is difficult to interpret because even when the authors mentioned that they dealt with individuals >50 mm. did not report the size structure of the studied oysters (Morriconi & Calvo 1989). Here we propose that the different levels of clustering among grounds is a result of other regulatory mechanisms that operate on larval settlement occurring on living oysters, or on the survival of new settlers. These mechanisms, already proven, are mainly graz- ing (or bulldozing) by chitons, and smothering (Pascual 1997). Grazing accounts for 887f post-settlement mortality (Pascual 1997) of recruits settled on all hard surfaces, in the three months following recruitment. This mechanism, an ecological contin- gency, may explain the demographic differences among grounds where the mortality agents operate with varying intensity. The proportion of carrier females is similar at Banco Reparo (32%) and Las Grutas (34%). but is considerably higher at Bajo Oliveira (88%) (Table 3). This topic was previously treated by Morriconi and Calvo (1989), who described the exi.stence of a gradient in carriage, among grounds, inverse to the clustering gra- dient (LG < BR < BO). This result could again be the result of a biased sampling. The method we used for sample collection was of "total cleaning" of the sample area, thus avoiding bias in the rep- resentation of all the size classes present in the population. The similar percentages of carriage in Banco Reparo and Las Grutas are consistent with the similar size structure at both grounds. The probability of "carrying" epibiotic males increases with the oyster's size. On oysters larger than 30- .'i? mm. time when car- riage of epibiotic males begins, the platform develops, progres- sively widening. The growth of the epibiotic male produces, ad- ditionally, a hollowing on the platform that constitutes, once the epibiotic male detaches, an increasingly favorable substrate for new settlers (Pascual 2()()()). This development pattern of the plat- form with oyster size is expressed by the number of epibiotic males earned per oyster; at Banco Reparo and Las Grutas. 61% and 68% of the carriers respectively carry one epibiotic male, while at Bajo Oliveira. 63% of the carrier oysters hold two to five epibiotic males. The high percentage of carriage at Bajo Oliveira may then be explained by the larger size of the individuals that compose this population. The higher number of males per oyster at Bajo Oliveira suggests that recruitment could be heavier at this ground, even when the absence of oysters <55 mm in Bajo Oliveira suggests the opposite. Previous research focusing recruitment pat- terns on shells and platform of oysters suggest that mortality agents (mainly grazing or bulldozing) strongly operate on recruits at this ground during the three months following recruitment (Pas- cual 1997. Pascual 2000). Considering different pieces of information regarding the three studied populations, and even when the age of the oysters cannot be determined in an absolute manner, we are able to conclude that the three populations differ in their dynamics, mainly in the turn- over rates of individuals. At Bajo Oliveira. on one extreme, me- dium size individuals are absent and mean size of adult oysters is highest (89 mm). In compailson. at Las Grutas. intermediate sizes between the two modes are present and the mean size of adults is the lowest (66 mm). Banco Reparo occupies an intermediate po- sition. Turnover rate is then maximal at Las Grutas. intermediate at Banco Reparo. and minimal at Bajo Oliveira. The puelche oyster grounds at the San Matfas Gulf offered the rare opportunity of studying the structure of one of the few natural and untouched flat oyster populations in the world. The information gathered from the present work and from pre- vious published information, enables us to draw the following scenario: 1. The puelche oyster grounds are mainly composed by low density populations. 2. Turnover rates in the three grounds are low even when Las Grutas shows a relatively higher rate. 3. Adults exert an attraction over the larvae ready to settle, live oysters being the main settlement sub,strate for oyster larvae in natural grounds (Pascual & Zampatti 1995: Pascual 2000). 4. Post-settlement mortality is very high (95-98%) mainly due to grazing by chitons (Pascual 1997). 5. Growth rate of individuals seems to be very low . The prob- ability of carrying dwarf males on the shell platform in- creases as the size of the female oyster increases (Pascual 2000). 6. The male portion of the population is mainly represented by the epibiotic males fixed on the platform of female carrier oysters ("a hidden mode"). 7. In natural grounds. 33 to 88% of the individuals >55 mm are "long living couples" composed by the female and its epi- biotic dwarf male (Pascual et al. 1989). The existing information provides enough reasons to protect this species, sustaining the fishing closure as a policy mainly based on the vulnerability of the stocks. Commercial exploitation of this species is likely to be supported by aquaculture. 1010 Pascual et al. ACKNOWLEDGMENTS We thank L. Orensanz and A. Parma for the review of this work, to our dear friend. A. Bocca. who participated on the first survey, and to Cap. Migueliz for "impossible" hauls performed with the "F.R.V. Cap. Canepa" at Bajo Oliveira. Financial support was provided by the International Fiiundation for .Science (Grant A704/1-3 to M. S. Pascual). Calvo. J. & E. R. Moniconi. 197S. Epihiontie et protancJrie che/ Ostrca imekhuna. Haliotis 9:85-88. Carreto. J.J. & C. A. Verona. 1971. Fitoplancton. pigmentos y condiciones ecologicas del Golfo San Manas. I. Coinrih. N°235, Argentina: Insi. Biol. Mar.. Mar del Plata. 19 pp. Castellanos. Z. A. 1957. Contribucion al conocimiento de las ostras del literal argentino (O. ptieklmna y 0. sivela). Technical Report. Argen- tina: Min. Agric. Can. Nac. 52 pp. Esteves. J. L.. M. Soli's, M. Gil. M. Commendatore & H. Ocariz. 1991. Resultados de la campana oceanografica Austral, abril 1991. Centro Nacional Palagonico. Conicet. Unpuhl. report. Fernandez. M. O. 1989. Paranietros hidrobiologicos del Golfo San Mati'as. Inst. Biol. Mar. Pesq. "Alte. Storni". Unpubl. Tech. Rep. 10 pp. Fernandez Castro. N. 1987. Hermaphrodilisme et sexualitc de riiuilre plate Ostrea puelcluiiui id' Orhigny) en elevage experinienl:il. Hiilunis 16: 283-29.1. Fernandez Castro. N & A. Lucas. 1987. Variahility in the frequency of male neoteny in Ostrca pucIlIhiiui (Mollusca:Bivalvial. Mar. Biol. 96: 359-365. Iribarne. O. O.. M. L. Lasta. H. C. Vacas. A. M. Parma & M. S. Pascual. 1991. Assessment of abundance, gear efficiency and disturbance in a scallop dredge fishery: results of a depletion experiment. In: S. Sham- way & P. Sandifer. editors. N° I . Scallop Biology and Culture. A tribute to James Mason, pp. 242-248 Lanfredi. N. & J. Pousa. 1988. Mediciones de corrientes. San Antonio Oeste. Universidad Nacional de La Plata. Fac. Grupo de Oceanografi'a Costera. Technical Report N°88/01. 10 pp. Morriconi. E. R. & J. Calvo. 1979. Ciclo reproductivo y altemancia de sexos en Ostrea piielchana. Physis 38:1-17. Morriconi. E. R. & J. Calvo. 1989. Alternative reproductive strategies of Ostrea piielchaiia. Hiilrohioloi>ia 185:195-203 Pascual, M. S. 1993. Contingencia y adaplacion en la ecologia reproductiva de Ostrea puelclninu. Doctoral Thesis. Argentina: Universidad Nacio- nal de Mar del Plata. 1 83 pp. Pascual, M. S. 1997. Carriage of dwarf males by female puelche oysters: the role of chitons. J. Exp. Mar. Biol. Ecol. 212:17.V85. LITERATURE CITED Pascual. M. S. 2(100. Dwarf males m the Puelche oyster [Ostrea ptielcliana. D'Orb.): differential mortality or selective settlement'^ ,/. Sliellfi.sh Res. 19:815-820. Pascual. M. S. & A. H. Bocca. 1988. Cultivo experimental de la ostra puelche, Ostrea puelchaim d'Orb. en el Golfo San Matias, Argentina. In: J. 'Verreth, M. Carrillo, S. Zanuy & E. A. Huisman editors. Aqua- culture Research in L.itin America. The Netherlands: Pudoc Wagenin- gen. pp. 329-345. Pascual. M. S.. O. O. Inbarne. E. A. Zampatti & A. H. Bocca. 1989. Female-male interaction in the breeding system of the puelche oyster [Ostrea piieUhanu d'Orb.). / E.\p. Mar. Biol. Ecol. 132:209-219. Pascual. M. S.. A. G. Martin, E. A. Zampatti. D. Coatanea. J. Defossez & R. Robert. 1991. Testing argentinian oyster. Ostrea puelchaim. in .sev- eral French oyster farming sites. International Council for the Explo- ration of the Sea. CM. 1991/K:.30. Pascual, M. S. & E. Zampatti. 1995. Chemically mediated adult-larval interaction triggers settlement in Ostrea puelclmna: applications in hatchery production. Aquaciilttire 133:33—14. Servicio de Hidrografia Naval, Cdo. J. Armada. 1969. Carta h-262. Puerto San Antonio. (IV-1969). Sokal. R. R. & J. Rohlf 1969. Biometry. The principles and practice of statistics in biological research. Freeman & Co. 776 pp. Stenzel, H. B. 1971. Treatise on Invertebrate Paleontology. Part N. Mol- lusca 6. Bivalvia. Oysters 3:953-1224. Vacas. H. C. 1978. Informe sobre captacion y crecimiento de Ostrea piiel- i liana en colectores artitlciales en la Bahia de San Antonio y Banco Reparo. Direccion General de Pesca y Recursos Mari'timos de la Pro- vincia de Rfo Negro. Argentina. Unpubl. Tech. Rep. 1 1 pp. Vacas, H. C. 1979. Informe sobre captacion y crecimiento de Ostrea ptiel- cliana en colectores artificiales en la Bahia de San antonio y Bco. Reparo. Direccion General de Pesca y Recursos Mari'timos de la Pro- vincia de Rio Negro. Argentina. Unpubl. Tech. Rep. 10 pp. Valette. L. H. 1929. Resultados experimentales de ostricultura. Min. Ag- ricultura (Dir Ganaderia). .■\rgentina. Unpubl. Tech. Rep. 21 pp. Joiinnil ,if Shvllfish Research. Vol. 20, No. .^. 1(11 1-1017, :()0I. A COMPARATIVE STUDY OF ANTl-PERKINSUS MARINUS ACTIVITY IN BIVALVE SERA ROBERT S. ANDERSON AND AMY E. BEAYEN Chcsiipciikc Biological Lalnirainry. University of Maryland Center for Environmental Science, P.O. Box 3,S. Soloinoiu. Marylaiu! 20688 ABSTRACT The eastern oyster Cni.\M>\lirci viri^inica has been decimated by a protislan parasiie HciI^iumi.s iiuiiimis: however, other bivalves appear to be more resistant to this pathogen. To better understand the basis for this difference in susceptibihty. a comparative study of the activities of anti-P. mariiuis serum proteins of several bivalve species was carried out. Sera from mussels not known to develop P marinus disease. Mytiliis eduli.s and Geiiliensia demissa, contained high anti-P. marinus activity. About 25% of M. ediiHs serum samples contained <10 kDa anti-P. marinus peptides; the possibility of seasonal, geographic, or other reasons to explain this variability requires additional study. Anti-P. nuiiiiiiis peptides in G. demissa serum were apparently absent. Measurable anti-P. marinus activity was present in C. virginica and C. niftas sera, but at levels many hundred-fold lower than that of the mussels. The greater P. inariniis resistance of C. gigas vs. C. virginica could not be explained by differences in anti-P. nuiriniis activity of their sera, Hemocyte Ivsates from all the bivalves examined produced marked inhibition of the growth of P. marinas, suggesting that antiinicrobial agents may be secreted by hemocytes into the serum. These factors may also participate in intracellular destruction of P. marimis, since the kilhng ability of the hemocytes of the different species closely mirrored the anti-P. marinus activities of their sera. The data suggest that C. virginica lacks active anti-P. nnirinus serum agents typical of M. edulis and G. demissa; however. P. marinus resistance of C. gigas seems not to depend upon elevated le\els of antimicrobial serum factors. KEY WORDS: Pcrkiinus marnuis. Crassostrea virginica. Crassostrea gigas. Geukensia demissa. Mytiius edulis. antimicrobial proteins, defensins INTRODUCTION Circulating hetiiocytes form the basis for host defense in bi- valve molltisks and many other invertebrates. Typically these cells mediate the destruction of foreign microorganisms encountered by chance contact in the hemolymph. or as a result of chemotactic migration (Cheng & Howland 1979). The microbes can be killed intracellularly or extracellularly by a variety of cytotoxic factors released from the cells. Many cytotoxic molecules have been de- scribed in the hemocytes and sera of bivalves such as lysosomal hydrolases (Pipe 1990) particularly lysozyme (McDade & Tripp 1967. Hardy et al. 1976). reactive oxygen species (Adema et al. 1991, Anderson 1996), reactive nitrogen species (Conte & Otta- viani 1995), as well as antimicrobial proteins and peptides (Roch et al. 1996). These activities may be constitutive, or may be stitnu- lated or suppressed by various treatments. Furthermore, the ex- pression of particular cytotoxic agents in bivalves can ditter be- tween species (Anderson 1994), seasons of the year (Santarem et al. 1994). and geographic locations (Fisher et al. 1996). Bivalves are generally able to mount an adequate defense against a majority of microbial invaders, thereby containing or resolving infections by evoking several complementary mechanisms of innate immu- nity. However, some pathogenic species, such as Perkinsiis mari- nus, can subvert or escape these defense mechanisms in their sus- ceptible bivalve hosts. A number of small, cationic. cysteine-rich antimicrobial pep- tides have been described in bivalve mollusks. especially tnussels. Mytihis galloprovincialis serum contains MGDl. a 4.4 kDa de- fensin-like peptide with 8 cysteines in its primary structure (Hubert et al. 1996a). Another isoform (1VIGD2) has recently been de- scribed, and both peptides are abundantly expressed in hemocyte granules (Mitta et al. 1999b). Bacterial challenge triggers release of MGDl by the hemocytes into the cell-free hemolymph or serinii (Mitta et al. 1999b). Peptides resetiibling arthropod defensins, with sequence similarities and 6 cysteines (defensins A and B) exist in the hemolymph of Mrtiliis edulis (Charlet et al. 1996). Another group of 4 kDa antimicrobial peptides have been found in mussels that arc distinct from defensins; to date .S isoforms are known: mytilins A and B from M. edulis (Chariet et al. 1996). and mytilins C, D and Gl from M. galloprovincialis (Mitta et al, 2000), Ex- perimentally unchallenged M. gallopr stems at I2^C. containing 25 ppt Instant Ocean (.'\quarium Systems. Inc.. Mentor. OH). Levels of nitrate/nitrite and ammonia were moni- tored and full or half water changes made whenever these param- eters were outside the normal range. Bivalves were routinely ac- climated in the tanks for at least one week before experimentation. The animals were fed live cultured algae or reconstituted frozen algal slurry 5 days per week. The live algae was a mixture of IsDchnsis galbaiia and Thalassiosira weissflogii: the frozen slurry contained Isochrysis sp.. Chaetocenis gracilis, and Teiraselmis sp. Cultivation of Perkinsus marinus Perkinsus marinus (strain I) cultures were originally provided by Drs. M. Faisal and J. La Peyre (Virginia Institute of Marine Science). The cultures were maintained in DME/HAM F- 1 2 me- dium, with phenol red, reconstituted with 10 ppt artificial sea water and contained 1% penicillin-streptomycin solution, 2% fetal bo- vine serum, and was HEPES-buffered at pH 6.5. All the compo- nents of the medium were purchased from Sigma Chemical Co. This medium is a modification of that originally described by Gauthier and Vasta (1993). Anri-Perlvinsus marinus Activity of Sera and 6 bivalves and held at I °C prior to centrifugation (.300 g. IO°C, 10 min) to separate the hemocytes from the serum. The supernatant (senrni) was filter- sterilized by passage through a 0.2 |a.m Whatman Puradisc syringe filter and frozen (-20"C) until further use, the cell pellet was discarded. Subsequently, the protein content of the serum was measured (BCA protein assay kit. Pierce Co.) and the desired concentrations reached by dilution with filter-sterilized Instant Ocean sea salts (.Aquarium Systems, Inc.), adjusted to 25 ppt sa- linity (lO). Cultured P. marinus cells were washed and resus- pended at a known density in DME/HAM F-I2 media without phenol red. this P. marinus suspension was mixed with a dilution of serum to give a known final concentration of serum protein containing 1x10'^ P. marinus cells per ml. and incubated at 26''C. Sterile technique was used throughout all experiments. As indi- cated on the Figures. 0.5 ml aliquots were periodically removed from the culture tlasks and read at 560 nm in a spectrophotometer: these OD readings were converted to numbers of P. marinus cells/ ml from a previously constructed standard curve. In this way, the effects of various serum concentrations on the growth kinetics of this protistan were determined. This same basic method was also used to obtain a measurement of relative anti-P. marinus activity ill the sera of the bivalve species under study. P. marijuis cultures, set up as described, were in active growth phase for at least 200 h; exposure to serum during this period often markedly inhibited this growth. Therefore, the Ab^^o of 170-hour cultures exposed to vari- ous serum concentrations were compared to the 170-h Ab,,,,, of serum-free controls to determine the concentration of serum pro- teins required for 50'^!r inhibition of normal growth (EC^,,). The anti-R marinus activity of u k- oysters pattern of anti-P. inaiiniis activity seen in the sera and cell extracts was also recorded when the ability of hemocytes to kill P. luariiuis in vitro was tested (Fig. 5). C. virginica hemocytes were able to kill -25% of the parasites and C. gigas hemocytes killed -10%. but there was considerable variation and no significant differences between the oyster hemocyle killing indices. Since C. virginica has very lim- ited ability to destroy P. inarinits in vivo, it was surprising that a modest hemocyte killing index was consistently recorded. A simi- lar finding was reported by Volety and Fisher (2000). in which hemocytes from Floridian oysters averaged 57% P. marinus killing capacity. Since it is likely that P. marinus loses some of its viru- lence during laboratory culture (Bushek et al. 1997); it would be interesting to te.st the ability of hemocytes to kill P. marinus freshly harvested from naturally infected oysters. In any case, it appears that resistance of C. gigas to P. marinus cannot be ex- plained by enhanced capacity for hemocytic killing. In contrast to the oysters. G. dcmissa and M. edulis cells routinely and efficiently destroyed P. mariiuis in vitm. In conclusion, the data suggest that low-level anti-P. marinus activity can be detected in the serum and cells of C. virginica and C. gigas. but differences between these species are insignificant and cannot be used to explain reported differences in resistance to P. marinus disease. However, the serum and cells of the mussels C dcmissa and M. edulis have high levels of anti-P. marinus activity. In all bivalve species tested, the hemocytes contained anti-P. marinus activity, and it is likely that the hemocytes secrete serum antimicrobial agents. It appears that the high anti-P. mari- nus acti\ity characteristic of M. edulis and C. dcmissa sera may depend on cytotoxic molecules other than, or in addition to, lysozyme or antimicrobial peptides. ACKNOWLEDGMENTS This project was funded in part by NOAA grants NA86RG0037 and NA46RG()()91. as part of the Oyster Disease Research Program. This is Contribution No. 35 1 2 of the University of Maryland Center for Environmental Science. Chesapeake Bio- logical Laboratory. LITERATURE CITED Adema CM.. W. P. W. van der Knaap & T. Smiiiia. 1 991. Molluscan heniocyte-mediated cytotoxicity: the role of reactive oxygen interme- diates. Rev. At/iiul. Sci. 4:201-223. Anderson. R. S. 1994. Hemocyte-derived reactive oxygen intermediate production in four bivalve mollusks. Dev. Camp. Immunol. 18:89-96. Anderson, R. S. 1996. Production of reactive oxygen intermediates by invertebrate hemocytes: immunological significance. In: K. Soderhiill. G. Vasta & S. Iwanaga. editors. New Directions in Invertebrate Im- munology. Fair Haven, New Jersey: SOS Publication.s. pp. 109-129. Anderson R. S., K. T. Paynter & E. M. Burreson. 1992. Increased reactive oxygen intermediate production by hemocytes withdrawn from Cras- so.itreu virginica infected with Perkinsiis marinus. Biol. Bull. 183:476- 481. Bushek, D.. S. K. Allen Jr.. K. A. Alcox. R. G. Gustafson & S. E. Ford. 1997. Response of Crassoslreti virginica to in vitro cultured Perkinsus marinus: preliminary comparisons of three inoculation methods. / Shellfish Res. I6:479-J85. Charlet. M.. S. Chernysh. H. Philippe. C. Hetru. J. A. HolTmann & P. Bulet. 1996. Isolation of several cystein-rich anlmiicrohlal peptides from the hlood of a mollusc. Mytihis edulis. J. Biol. Cliem. 271:21808- 21813. Cheng, T. C. 1992. Selective induction of release of hydrolase from Cras- soslrea virginica hemocytes by certain bacteria. / Invertehr. Pathol. 59:197-200. Cheng. T. C. & K. H. Howland. 1979. Chemotactic attraction between hemocytes of the oyster, Crassostrea virginica. and bacteria. / Inver- tehr. Pathol. 33:204-210. Cheng. T. C, G. E. Rodrick. D. A. Foley & S. A. Koehler. I97.S. Release of lysozyme from hemolymph cells of Mercenarui nu-rcenurui during phagocytosis. / Invertehr. Pathol. 25:261-26.s. Chu. F-L. E. & J. F. La Peyre. 1989. Effect of environmental factors and parasitism on hemolymph lysozyme and protein of American oysters {Crasso.strea virginica). J. Invertehr. Pathol. 54:22-t-232. Conte, A. & E. Ottaviani. 199.5. Nitric oxide synthase activity in molluscan hemocytes. FEBS Lett. 265:120-124. Fisher. W. S.. L. M. Oliver & P. Edwards. 1996. Hematologic and sero- logic variability in Eastern oysters from Apalachicola Bay. Florida. / Shellfish Res. 15:555-564. Gauthier. J. D. 1998. Development and uldi/ation of an in vitro culture system for the characterization of gene expression in the oyster parasite Perkinsus marinus. Ph.D. Dissertation. College Park. Maryland: Uni- versity of Maryland. 193 pp. Gauthier. J. D. & G. R. Va.sta. 1993. Continuous in vitro culture of the eastern oyster parasite Perkinsus marinus. J. Invertehr. Pathol. 62:321- 323. Hardy. S.W., T. C. Fletcher & L. M. Gerrie. 1976. Factors in hemolymph of the mussel. Mytilus edulis L., of possible significance as defense mechanisms. Biochem. Soc. Trans. 4:473^75. Hubert. F.. T. Noel & P. Roch. 1996a. A member of the anthropod defensin family from edible Mediterranean mussels (Mytilus galloprovincialis). Eur J. Biochem. 240:302-306. Hubert. F.. W. van der Knapp. T. Noel & P. Roch. 1996b. Cytotoxic and antibacterial properties of Mytilus galloprovincialis. Ostrea edulis and Crassostrea gigas (bivalve molluscs) hemolymph. Aquiu. Livutg Re- sour 9:115-124. Hubert, F., E. L. Cooper & P. Roch. 1997. Structure and differential target sensitivity of the stimulable cytotoxic complex from hemolymph of the Mediterranean mussel Mytilus galloprovincialis. Biochim. Biophys. .Acta. 1361:29-41. Ito. Y.. .\. YoshikavMi, T. Hotani. S. Fukuda. K. Sugimura & T. Imoto. 1999. Amino acid sequences of lysozymes newly purified from inver- tebrates imply wide distribution of a novel class in the lysozyme fam- ily. Eur J. Biochem. 259:456-461. Jolles. J.. A. Fiala-Medioni & P. Jolles. 1996. The rununant digestion model using bacteria already employed early in evolution by symbiotic molluscs. J. Mil. Evol. 43:523-527. La Peyre, J. F., F.-L. Chu & J. M. Meyers. 1995. Haemocytic and humoral activities of eastern and Pacific oysters following challenge by the protozoan Perkinsus marinus. Fish Shellfish Immunol. 5:179-190. McDade. J. E. & M. R. Tripp. 1967. Lysozyme in the hemolymph of the oyster, Crassostrea virginica. J. Invertehr. Pathol. 9:531-535. McHenry. J. G. & T. H. Birkbeck. 1979. Lysozyme of the mussel, Mytilus edulis (I,.). Mar. Biol. Lett. 1:111-119. Meyers. J. A.. E. M. Burreson, B. J. Barber & R. Mann. 1991. Suscepti- bility of diploid and triploid Pacific oysters. Crassostrea gigas. and eastern oysters, Crassostrea virginica. to Perkinsus marinus. ,/. Shell- fish Res. 10:433-437. Mitta, G., F. Hubert, T Noel & P. Roch. 1999a. Myticin. a novel cysteine- rich antimicrobial peptide isolated from haemocytes and plasma of the mussel Mytilus galloprovincialis. Eur. J. Biochem. 265:71-78. Mitta. G.. F. Vandenbulcke. F. Hubert & P. Roch. 1999b. Mussel defensins are synthesized and processed in granulocytes then released into the plasma after bacterial challenge. J. Cell Sci. 112:4233^242. Mitta, G.. F. Vandenbulcke. F. Hubert, M. Salzet & P. Roch. 20110. In- volvement of mytilins in mussel antimicrobial defense. J. Biol. Chem. 275:12954-12962. ANT]-PeRKINSI'S MMilNVS ACT1\'IT\- IN SERA 1017 Morvan. A.. E. Bachere, P. P. Da Silva. P. Pimenta & E. Miallie. 1994. In vitro activity of the antimicrobial peptide magainin 1 against Bonamia ustrea, the intrahemocytic parasites of the flat oyster Osireii cdulis. Mol. Marine Biol. Bioiech. 3:327-333. Nilsen. I. W.. K. Overbo. E. Sandsdalen. E. .Sandaker. J. Sletten & B. Myrnes. 1999. Protein purification and gene isolation of chlamysin. a cold-active lyso/yme-like enzyme with antibacterial activity. FEBS Lett. 464:153-158. Pierce. J.C. W. L. Maloy. L. Salvador & C. F. Dungan. 1997. Recombi- nant expression of the antimicrobial peptide polyphemusin and its ac- tivity against the protozoan oyster pathogen Perkinsiis murinus. Mol. Marine Biol. Biotech. 62:248-259. Pipe. R. K. 1990. Hydrolytic enzymes associated with the granular haemocytes of the marine mussel AA(/7»,s eiliili.\. Histoehem. ,/. 22:595- 603. Roch. P. 1999, Defense mechanisms and disease prevention in farmed unertebrates. Aquaculltire 172:125-145. Roch. P.. F. Hubert. W. van der Knaap & T. Noel. 1996. Present knowl- edge on the molecular basis of cytotoxicity, antibacterial activity and stress response in marine bivalves. Ital. J. Zool. 63:31 1-316. Santarem, M. M.. J. A. F. Robledo & A. Figueras. 1994. Seasonal changes in hemocytes and serum defense factors in the blue mussel Myiiliis galloprovinciatis. Dis. Aijiiat. Ori<. 18:217-222. Volety. A. K. & W. S. Fisher. 2(100. In vitro killing of Ferkimits manmis by hemocytes of oysters Crassosirea virginica. J. Shellfish Res. 19:600. Volety, A. K.. L. M. Oliver, F. J. Genthner & W. S. Fisher 1999. A rapid tetrazolium dye reduction assay to assess the bactericidal activity of oyster (Crassostrea virginica) hemocytes against Vibrio pura- haeinolylicus. Aquacuhure 172:205-222. Joiinial o) Shellfish Research. Vol. 20. No. 3. 1(114-1(124. 2(101. HISTOPATHOLOGY OF THE INFECTION BY PERKINSUS ATIJ^NTICUS IN THREE CLAM SPECIES {RUDITAPES DECVSSATVS, R. PHILIPPINARUM AND R. PULLASTRA) FROM GALICIA (NW SPAIN) M. CAMINO ORDAS, J. GOMEZ-LEON, AND ANTONIO FIGUERAS* Institiito de Investigaciones Marinas. Conscjo Supvrior de litvesligaciones Cicutificas (CSIC). Ediiardo Cahello 6. 36208 Vigo. Spain ABSTRACT In this study, we have determined the histopathology of the infection by the protozoan parasite Pcrkinsus cilUmliciis in three exploitable clam species (carpet-shell clam Riidiuipes clecKssalK.s. Manila clam R. philippiiianim. and the species R. piiUastra) from Galicia (NW of Spain). In histological preparations from infected animals, typical mature trophozoites of Perkinsiis. containing lipid droplets and vacuole but no vacuoplast, were ob.served. The trophozoites were usually in groups surrounded by a light halo. Pre-palintomic tomonts, as well as palintomic tomonts with a variable number of daughter cells, were also present in the host tissues. Less frequently, iminature trophozoites with an undifferentiated cytoplasm were observed. The presence of P. atkmncus cells was always associated with a strong hemocytic infiltration of the surrounding tissues. Occasionally, the parasite cells were internalized by granular hemocytes. The moiphology and distribution of the P. atlwnicus were similar in the three clam species studied, this being the first tmie that R. pidlastra from Galicia has been found infected by P. athmticus. The morphology of the parasite's life cycle stages and the histopathology of the infection resemble those reported on other species of the genus Perkinsiis KEY WORIXS: Pcrkinsus utianticus. protistan. Protozoa. Riulilapc. ilcciissaius. Riuliiapcs plulippinanwi. Riuliuipes pulhistni. his- topathology INTRODUCTION Protozoan parasites of the genus Perkinsiis (Apicomplexa. Per- kinsea) have been reported from marine molluscs all around the world. P. niariiuis (Mackin et al. 1950) has been associated with heavy oyster mortalities in the East and Gulf coasts of the USA since it was first described (Andrews 1946. Burreson et al. 1994). P. olseni parasitizes the abalone {Haliotis ruber) in Australia (Lester & Davis 1981). and the recently described P. qiigwadi infects the Japanese scallop (Patinopecten yessoensis) in Canada (Blackbourn et al. 1998). Until now. the only Perkinsiis species described in Europe has been P. atlanliciis, parasite of the carpet- shell clam (Riiditapes deciissatus) in Portugal (Azevedo 1989). On the other hand, Perkinsiis sp. infections have been described in Manila clam [Riiditapes philippinanim) from Japan (Hamaguchi ct al. 1998. Maeno et al. 1999) and Korea (Choi &. Park 1997). in the pearl oyster [Pincuida maxima) and 30 other molluscan species from Australia (Goggin & Lester 1987, Norton et al. 1993). and in more than 30 marine bivalve species in different areas of the Pacific and Atlantic oceans (Perkins 1988). In Spain. Perkinsiis sp. has been found in the caipet-shell clam (Riiditapes deciissatus) and the Manila clam (R. pliilippinariim) from the South Atlantic. Mediterranean and Cantabric coasts (Cigarria et al. 1997. Montes et al. 1995a, Navas et al. 1992. Saarista et al. 1995). Two other bivalve molluscs of economical itnportance. R. pidlastra and Venenipis aureus, have been reported to be infected by this parasite in Huelva (SW Spain) (Navas et al. 1992). In Galicia (NW Spain), clam aquacuiture is an activity of high economical importance. The production of Riiditapes deciissatus. R. pidlastra and R. philippiiianim in this region in 1998 totaled 914, 2554 and 604 Tm respectively, this representing an annual income of approximately 4900 million pesetas (I US$ = 185 pesetas) (data from the Consellerfa de Marisqueo, Pesca y Acuicul- tura of the Xunta de Galicia). The presence of a Perkinsiis-Wke organism was associated with severe carpet-shell clam mortalities in a depuration plant in Meira (Galicia) thirteen years ago (Figueras et al. 1992. Gonzalez Herrero et al. 1987). Later. Per- kinsus sp. was detected in carpet-shell clams and Manila clams frotii different locations along the Galician coast (Figueras et al. 1996). A strong hemocytic infiltration was always observed in the Perkinsus-'mfecled tissues, this probably affecting the clam health and endangering its production (Figueras et al. 1992). Recently. Novoa et al. (2001 ). using molecular tools based on the 18S rRNA gene, have detiionstrated the identification of the Perkinsus species observed in carpet- shell clams from Galicia as P. atlanticus. This study was conducted to describe the histopathology of the Perkinsiis atlanticus infection in the three economically most iin- poitant clam species in Galicia; carpet-shell clam. Manila clam and R. pidlastra. MATERIALS AND METHODS Clams Market-sized carpet-shell clams [Riiditapes deciissatus). Ma- nila clams (R. philippinanim) and R. pidlastra. were collected in Cainpelo (Rfa of Ponlevedra). Meira. Vilaboa and Arcade (all Ri'a of Vigo), and Camarii'ias (Central Ria) (all Galicia. NW Spain) (Fig. 1). Approximately, one hundred individuals were examined frotn each location. The anitnals were processed immediately after beina received in the laboratory. Diagnosis of Perkinsus *Corresponding author. E-mail: patol@iiin.csic.es The presence of Perkinsus atlanticus infection in the clams was assessed using Ray's Fluid Thioglycollate Medium (RFTM). Briefly, one pair of gills from each clatn was introduced into Fluid Thioglycollate Medium (FTM) (Ray 19.54. Ray 1966). and incu- bated at room temperature in the dark for three to five days. Then. 1019 1020 Camarinas Campeio Vilaboa Arcade Figure 1. Clam sampling sites in the Ria of Pontevedra (Campeio). Ria of Vigo (Meira, \ ilaboa and Arcadel and Central Ria (Camarinas), all of them located in Galicia (NVV Spain). gills were placed on a slide, stained with Lugoi's iodine and ob- served with a light microscope. Histology The lemaining soft tissues were fixed in Davidson's fixative (Shaw & Battle 1957) for 24 hours and a transverse section ap- proximately 5 mm thick including mantle, gonad, digestive gland, gills, kidney, and foot, was excised from each clam. Tissue samples were embedded in paraffin and .'i-ij.m sections were stained with hematoxylin-eosin. The histological sections were observed under a light microscope (Nikon). RESULTS The RFTM-treated gills of several carpet-shell clams, Manila clams and R. puUastra showed dark blue spherical bodies, typical of hypnospores of Perkinsus. Only those clams diagnosed as posi- tive for Pi'i-kiiisi(\ by RFTM were examined histologically. Macioscopically. the most heavily infected individuals showed milky-white cysts in their body surface. Under the microscope, the morphology and distribution of Perkinsus atlaiuiciis life cycle stages in the host tissues were similar in the three clam species studied, with strong hemocytic infiltration of the infected tissues and encapsulation of the parasite (Fig. 2. Fig. .^. Fig. 4). The histopathology of the infection by P. atlcDiticus in Ruditapes dc- ciissatits is described in detail below. The parasite cells were mainly located in the gills and connec- tive tissue close to the digestive epithelium and digestive tubules (Fig. 4A). In heavy infections, the parasite was detected in all clam organs examined, including the gonads and the muscle tissues (Fig. 4B). Granular hemocytes containing one (Fig. 5A) to ten (Fig. 5B) P. atlanticiis cells were frequently observed. The most prevalent P. atlanticiis life cycle stage present in the clam tissues was the mature trophozoite (Fig. 6). Its large eccentric vacuole lacked a vacuoplast. and a dark nucleolus was present inside the nucleus (Fig. 6C). The trophozoites were usually grouped in clumps of 2 to 20 cells that were usually similar in size ^^mmci-'m',^-::^^. V Figure 2. Mature trophozoites of Perkinsus atlaiitieus in Manila clam {Ruditapes philippinurum) tissues. This stage has a large eccentric vacuole without a vacuoplast, and a dark nucleus. (A) Perkinsus at- laiitieus located in the mantle edge. Note the strong hemocytic reaction associated with the infection. (B) Clam's gills heavily infected hy Per- kinsus atlaiitieus. In this case, the parasite cells are grouped in small clumps (fewer than five individuals per clump). Most of the groups are surrounded h\ a light halo. )C) Accumulation of Perkinsus atlaiitieus trophozoites in the tissue surrounding digestive tubules. The halo around the large parasite clump is surrounded by numerous hemocytes. Bar =100 fini, and stage. Palintomic cells, named tomonts by Perkins (1996), were also frequently detected in the clam tissues (Fig. 7). They were larger than mature trophozoites, and were not usually found clustered. Before the onset of palintomy, mature trophozoites lost the vacuole, which mixed with the cytoplasm (Fig. 7A). This pre- palintomic stage was siinilar in size to the toniont. and showed a dense undifferentiated aspect. Once the palintomy started, two- and four-daughter-cell to- monts were clearly identified in the clam tissues (Fig. 7B. Fig. 7C, Fig. 7D). When the number of daughter cells was too high, the tomont had a granulated aspect with several dark nuclei (Fig. 7E). Sometimes, the vacuole did not disappear friim mature trophozo- ites before the palintomy and during palintomy underwent divi- sions to give the daughter cells inside the tomont (Fig. 7B, Fig. 7C, Fig. 7F). Occasionally, small iminature trophozoites without vacu- ole were observed. Their cytoplasm was undifferentiated but nucleus and nucleolus were well defined (Fig. 5B. Fig. 70, white ■> • ^ .•»., A •' ■"'■ ' 'S' '•'.•I'i '■ >■ •-: Figure 3. Mature troplKi/ditis iil I'likiiisiis dtlaiiliiiis in RiidiKipirt puUastra tissues. Note tlie similarity iil these parasite cells with those present in Manila and carpel-shell clam tissues. The trophozoites ha\e a large eccentric \acuole and a dark nucleus, and are associated with strong hemocyte infiltration. They can appear individually (A) or in large clumps (B). In the latter case, a light halo usually surrounds the grouped parasite cells. Bar = 5(1 pni. Figure 5. Mature I'erkinsiis (itlunticiis trophozoites internalized h\ carpet-shell clam [RiidiUipes deciiwciliis) hemocytes, probably granu- locytes. (A) Hemocytes containing one P. ailanticiis cells and (B) con- taining to ten P. allanlicus cells. Bar = 25 pni. ^ 1 v5;^ ^ 0 ^ ■■ » m- •ij*-.-,* .■■■ ■ r- * .4 ^' ■■- •V' «=• »s IV V^v •' ^t Figure 4. Mature trophoziiiles ol Piikiii\ii\ allaiilicii\ in carpit-shill clam {Riiditiipes decussatus) tissues. The morphology of this lil'e-cycle stage, with a large eccentric vacuole and a dark nucleus, is similar to that of the Perkinsiis atlaiiliciis present in Manila clams and K. pullus- Ira. (Al Perkinsiis allanlicus (arrows) in the gill epithelium. Note the strong hemocytic reaction and the disorganization of the tissues. (B) Perkinsiis atlanticiis in the muscle tissue. The clump, constituted by four mature trophozoites, is surnumded by a light halo. Among the numerous hemocytes, pieces of muscle tissue are seen (arrows). Bar = 50 ^im. ":-.\:--M >'.%' "r; :ni>- ?V:.' '''"'■ y* /••^i: •B- •->.•■ Figure 6. General view of carpet-shell clam {Rudilapes decussatus) tissues heavily infected with Perkinsiis allanlicus. Note the strong hemocytic infiltration (A) and the light halos surrounding the parasite clumps (B). The most prevalent stage is the mature trophozoite (detail in C), with an eccentric vacuole and a nucleus containing a dark nucleolus. Immature trophozoites are also present in the clam tissue (B), with an undifferentiated cytoplasm and no vacuole. Several dark lipid droplets are present in their cytoplasm. The immature tropho- zoites are grouped and surrounded by a light halo. Bar = 10() pm (A), 50 Jim (Bl, and 15 uni (C). 1022 Ordas et al. Figure 7. Palintomy of Perkinsiis allaiiliciis in the carpet-shell clam {Riidilapes deciissaliis) tissues. (Al Mature tnipho/oites after loosing the vacuole by its mixing with the cytoplasm. The nucleus is distin- guishable as a dark spot. Two- (B, C) and four- iD) cell daughter tomonts. Nuclei of the four- cell tomont are pointed with arrowheads. (El Tomont in advanced palintomy, with a high number of daughter cells whose nuclei are observed as dark spots. (B, C, Fl Palintomy of trophozoites whose vacuoles were not lost prior to palintomy. The vacuole is successively di\ided with the cytokinesis (B. Cl. until the tomont acquires a niultivacuolated aspect (Fl. ((il Small immature trophozoites without vacuole, with undifferentiated cytoplasm but nucleus and nucleolus well defined (white arrows). Bar = 25 fifim. arrows). Generally, the diameter of the protozoan cells infecting the host tissues varied from 3-15 |j.ni. Frequently, a light halo, which resembled an empty capsule, sunounded the clumps of parasite cells (Fig. 2C. Fig. 4B. Fig. 6). A high number of hemocytes were grouped around these capsules. The presence of every lite cycle stage of Peikinsiis culanlicus was associated with a strong host hemocytic reaction (Fig. 2. Fig. 3. Fig. 4. Fig. 6). When the infection was heavy and the P. atlaiuiciis cells were distributed in all host tissues, they looked disorganized with massive hemocyte infiltration. DISCUSSION In general, the morphology and distribution of Perkinsiis at- Uintiius in the tissues of the three clam species from Galicia were similar to those reported before for Perkinsiis infecting different host species around the world (Lester & Davis 1981. Azevedo 1989. Monies et al. 1995a. Sagrista et al. 1995. Perkins 1996. Blackbourn et al. 1998. Hamaguchi et al. 1998, McLaughlin & Faisal 1998a, Bower et al. 1999. Maeno et al 1999). Cultured P. inariniis and Perkinsiis sp. cells have been reported to develop vacuolated immature trophozoites inside the tomont (Perkins 1996. McLaughlin & Faisal 1998b). as we found for P. ciilaiiliciis in the clam's tissues. On the other hand, and in contrast to P. tnarimis (Perkins 1996), no vacuoplast was observed inside the P. arlanti- ciis vacuole. P. olseni also lacks this inclusion (Lester & Davis 198 1 ). The presence of a Perkinsiis species in R. piillastni has only been reported in the South Spanish ,'\tlantic coast (Na\as et al. 1992). Therefore, this is the first time that Perkinsiis is described in such economically valuable clam species in Galicia. Recently, a new protozoan parasite morphologically similar to Perkinsiis has been described in carpet shell clam from Galicia (Figueras et al. 2000). Moreover, this new species, named Pseiidoperkinsiis tapetis. develops hypnospores in RFTM indistin- guishable from Perkinsiis hypnospores. In spite of the coexistence and visible similarity of the two organisms (Ordas & Figueras 1998. Novoa et al. 2001), the parasitic cells observed and de- scribed in this paper should be mainly Perkinsiis since its preva- lence is much higher than that of P. tapetis (Novoa et al. 2001 ). The initial response of the host to the P. inarinns infection is the hemocytosis and the migration of hemocytes to the infection site (Lauckner 1983). This could be the origin of the strong hemocytic infiltration associated with the presence off. atlanticiis in the clam tissues. When the P. atlanticiis infection rate is very high, the accumulation of hemocytes can be macroscopically observed as white cysts as previously reported by Lauckner (1983), Azevedo (1989), and Sagrista et al. (1995). The encapsulation of the parasite is as.sociated with the secre- tion of a "specific" protein by the granulocytes that had migrated to the infection place and surround the parasite cells (Monies et al. 1995b. Monies et al. 1996. Monies et al. 1997). In this study, the capsules that surrounded the P. atlimiicus clumps seemed empty. This could be due to the histological processing that might have extracted the capsule material. However. P. nuiriniis releases pro- teases when cultured in vitro (La Peyre et al. 1995) that alter several host defense parametres (Garreis et al. 1996) and degrade certain proteins of the oyster hemolymph (Oliver et al. 1999). Therefore, it could also be possible that P. alhmticiis secretes similar proteases in vivo, these being responsible for the digestion of the capsule material and interference with the host defense response. The oyster tissues with higher prevalence of Perkinsiis inarinns are the gills and the digestive gland (Lauckner 1983. Oliver et al. 19981. The hypothesized strategy of P. inarinns for entering into the host (through feeding) would explain this distribution (Lauck- ner 1983. Andrews 1988). In the carpet-shell clam, the connective tissue of the organs shows the highest Perkinsiis infection degree (Chagot et al. 1987). These results agree with our observations of the clam tissues infected with P. atlanticiis. Cheng (1988) defined Perkinsiis inarinns as an extracellular parasite, although it can be phagocytized by the hemocytes and be destroyed intracellularly. In contrast. Perkins (1996) considers HiSTOPATHOLOG'i' OF THE INFECTION BY P. ATLANTICVS 1023 that, althiuigh the parasite cells can be found tree in the host tissues, they are most commonly located in the phagosomes of the hemocytes. In any case, we have frequently observed one to ten P. attuiuicits cells inside a clam hemocyte. If the heniocyte is not able to destroy the internalized Perkinsus cells, the parasite could grow inside the blood cell until tmally destroying it. thus diminishing the host cellular defence factors. However, a relationship between the hemocyte lysis and the capsule formation around the Perkinsus cells has been reported by Montes et al. (IQQ.'ib. 1996. 1997). constituting a host reaction to the hemocyte destruction by the parasite proliferation. The morphology of the host hemocytes that have internalized the P. atlanticus cells corresponded to granular hemocytes. In fact, it is well known the higher phagocytic activity of the granulocytes compared to the hyalinocytes in bivalve mol- luscs (Fisher 1986. Feng I9S8). ACKNOWLEDGMENTS M. C. Ordus thanks the .\unta de Galicia for her fellowship. The authors thank Dr. Iva Dykova for the revision of the manu- script, and Mrs. B. Villaverde for her help in the histological processing of the samples. This work has been partially funded by the Commission for Cultural. Educational and Scientific Exchange between the United States of America and Spain. LITERATURE CITED Andrews. J. D. 1988. Epizootiology of the disease caused by the oyster pathogen Perkinsus mariniis and its effects on the oyster industry. Am. Fisli. Soc. Spec. Publ. 18:47-63. Andrews. J. D. 1996. History of Perkinsus marinns. a pathogen of oysters in Chesapeake Bay 1950-1984. J. Sliellfisli Res. 15:13-16. Azevedo. C. 1989. Fine structure of Perkinsus uilanticus n. sp. (Apicom- plexa. Perkinsea) parasite of the clam RuJitapes decussatus from Por- tugal. J. Parasilol. 75:627-635. Blackbourn. J.. S. M. Bower & G. R. Meyer. IWS. Perkinsus tiui;u\uli sp. nov. (incertae .sedis), a pathogenic protozoan parasite of Japanese scal- lops. Patinopecten yessoensis. cultured in British Columbia. Canada. Can. J. Zool. 76:942-953. Bower, S. M., J. Blackbourn. G, R. Meyer & D. W. Welch. 1999. Effect of Perkinsus qui;\\mli on various species and strains of scallops. Dis. Aqual. Org. 36:143-151. Burreson. E. M.. R. S. Alvarez, V. V. Martinez & L. A. Macedo. 1994. Perkinsus marinus (Apicomplexa) as a potential source of oyster Crcis- soslrea virginica mortality in coastal lagoons of Tabasco. Mexico. Dis. Aqual. Org. 20:77-82. Chagot. D.. M. Comps. V. Boulo. F. Ruano & H. Gnzel. 1987. Histological study of a cellular reaction in Ruditapes decussunn, infected by a pro- tozoan. Aquaeulture 67:260-261. Cheng. T. C. 1988. Strategies employed by parasites of marine bivalves to effect successful establishment in hosts. .4/;;. Fisli. Soc. Spec. Puhl. 18:112-129. Choi. K. S. & K. I. Park. 1997. Report on the occurrence of Perkinsus spp. in the Manila clams, Ruditapes plulippinarum. in Korea. J. Aquaeulture 10:227-237. Cigam'a. J.. C. Rodriguez & J. M. Fernandez. 1997. Impact of Perkinsus sp. on Manila clam Ruiiilapes plulippinarum beds. Dis. .Aquut. Org. 29:117-120. Feng. S. Y. 1988. Cellular defense mechanisms of oysters and mussels. Am. Fisli. Soc. Spec. Puhl. 18:153-168. Figueras, A.. J. A. F. Robledo & B. Novoa. 1992. Occurrence of haplospo- ridian and Perkinsus-Wke infections in carpet-shell clams. Ruditapes decussatus (Linnaeus. 1758). of the Ri'a de Vigo (Galicia, NW Spain). J. Shellfish Res. 1 1 :377-382. Figueras. A.. J. A. F. Robledo & B. Novoa. 1996. Brown ring disease and parasites in clams {Ruditapes decussatus and R. pliilippinarum) from Spain and Portugal. / Shellfish Res. 15:363-368. Figueras. A.. G. Lorenzo. M. C. Ordas. M. Gouy & B. Novoa. 2000. Sequence of the small subunit ribosomal RNA gene of Perkinsus at- lanticus-like isolated from carpet shell clam in Galicia. Spain. Mar. Biotechnol. 2:419-428. Fisher. W. S. 1986. Structure and functions of oyster hemocytes. hi: M. Brehelin. editor. "Immunity in Invertebrates." Berlin. Heidelberg: Springer-Verlag. pp. 25-35. Garrets. K. A.. J. F. La Peyre & M. Faisal. 1996. The effects of Perkinsus marinus extracellular products and puritled proteases on oyster defence parameters in vitro. Fish Shellfish Inunumil. 6:581-597. Goggin. C. L. & R. J. G. Lester. 1987. Occurrence of Perkinsus species (Protozoa. Apicomplexa) in bivalves from the Great Barrier Reef Dis. .Aqual. Org. 3:11.3-117. Gonzalez Herrero. V.. J. Pefias Rodriguez & A. Diaz Martinez. 1987. Dermocystidosis en almeja fina portuguesa de importacion. Cuad. Marisq. Publ. Tec. 12:695-700. Haniaguchi, M.. N. Suzuki. H. Usuki & H. Ishioka. 1998. Perkinsus pro- tozoan infection in short-necked clam Tapes ( = Ruditapes) philippi- narum in Japan. Fish. Pathol. 33:473—180. La Peyre. J. F.. D. Y. Schaffiauser, E. H. Rizkalla & M. Faisal. 1995. Production of serine proteases by the oyster pathogen Perkinsus mari- nus (Apicomplexa) in vitro. J. Eukaiyol. Microhirjl. 42:544-551. Lauckner. G. 1983. Diseases of Mollusca: Bivalvia. In: O. Kinne. editor. "Diseases of marine animafs", vol. 2. Hainhurg: Biologische Anstah Helgoland, pp. 532-550. Lester, R. J. G. & G. H. G. Davis. 1981. A new Perkinsus species (Api- complexa. Perkinsea) from the ahalone Haliolis ruber. J. Im-crtebr. Pathol. 37:181-187. Mackin, J. G., H. M. Owen & A. Collier. 1950. Preliminary note on the occurrence of a new protistan parasite. Dermocyslidium marinuin n. sp. in Crassostrea virginica (Gmelin). Science 111:328-329. Maeno, Y., T. Yoshinaga & K. Nakajima. 1999. Occurrence of Perkinsus species (Protozoa. .Apicomplexa) from Manila clam Tapes plulippi- narum in Japan. Fish. Pathol. 34:127-131. McLaughlin, S. M. & M. Faisal. 1998a. Histopathological alterations as- sociated with Perkinsus spp. infection in the softshell clam Mya arenaria. Parasite 5:263—271. McLaughlin. S. M. & M. Faisal. 1998b. In vitro propagation of two Per- kinsus species from the softshell clam Myu arenaria. Parasite 5:341- 348. Montes. J. F.. M. Durfort & J. Garcia Valero. 1995a. Cellular defense mechanism of the clam Tapes semidecussatus against infection by the protozoan Perkinsus sp. Cell Tissue Res. 279:529-538. Montes. J. F., M. Durfort & J. Garci'a Valero. 1995b. Characterization and localization of an Mr 225 kDa polypeptide specifically involved in the defense mechanisms of the clam Tapes semidecussatus. Cell Tissue Res. 280:27-37. Montes, J. F., M. Durfort & J. Garcia Valero. 1996. When the venerid clam Tapes decussatus is parasitized by the protozoan Perkinsus sp. it syn- thesizes a defensive polypeptide that is closely related to p 225. Dis. Aqual. Org. 26:149-157. Montes. J. F., J. A. Del Rio, M. Durfort & J. Garcia Valero. 1997. The protozoan parasite Perkinsus atlanticus elicits a unique defensive re- sponse in the clam Tapes semidecussatus. Parasitology 114:339-350. Navas. J. I.. M. C. Castillo, P. Vera & M. Ruiz-Rico. 1992. Principal parasites observed in clams. Ruditapes decussatus (L.). Ruditapes plul- ippinarum (Adams & Reeve). Venerupis pullastra (Montagu) and Venerupis aureus (Gmelin). from the Hueha coast (SW of Spain). Aquaeulture 107:193-199. Norton. J. H.. M. A. Shepherd. F. P. Perkins & H. C. Prior. 1993. Per- 1024 Ordas et al. kinsi(S-]ike infection in farmed golden-lipped pearl oyster Pinctada imixinw from the Torres Strait. Australia. J. Imertehr. Pathol 62:105- 106. Novoa. B.. M. C. Ordas & A. Figiieras. 2001. Two different protozoan parasites, Perkinsiis atlainicus and a Perkinsus-\\ke organism, coexist in hypnospores detected in clam (Rudiiapes decussatus) tissues incu- bated in RFTM. Aquacultiire. In press. Oliver. L. M.. W. S. Fisher, S. E. Ford, L. M. Ragone Calvo, E. M. Burreson, E. B. Sutton & J. Gandy. 1998. Perkinsiis mariims tissue distribution and seasonal variation in oysters Crassoxtrea virginica from Florida, Virginia and New York. Dis. Aquat. Org. 34:51-61. Oliver, J. L., T. D. Lewis, M. Faisal & S. L. Kaattari. 1999. Analysis of the effects of Perkinsiis nuiriniis proteases on pla,snia proteins of the east- ern oyster (Crassostrea virginica) and the Pacific oyster iCrussoslreti gigas). J. Invertehr Pathol. 74:173-183. Ordas. M. C. & A. Figueras. 1998. In vitro culture of Perkin.xus atlunticiis. a parasite of the carpet shell clam Riulitapes decussatus. Dis. Aquat. Org. 33:129-136. Perkins. F. O. 1988. Structure of proti-stan parasites found in bivalve mol- luscs. Am. Fish. Soc. Spec. Piibl. 18:9.3-111. Perkins, F. O. 1996. The structure of Perkinsus inariniis (Mackin. Owen and Collier, 1950) Levine, 1978 with comments on taxonomy and phylogeny of PerkiiiMis spp. J. Shellfish Re.s. 15:67-87. Ray, S. M. 1954. Biological studies of Dermncystidiiiui niariniini. Rue Inst. Paniph. 41(SI):1-1 14. Ray. S. M. 1966. A review of the culture method for detectmg Dermocys- tidiuin niariiuim. with suggested modifications and precautions (1963 Proceedings). Proc. Natl. Shellfish. Assoc. 54:55-69. Sagrista, E.. M. Durfort & C. Azevedo. 1995. Perkinsiis sp. (Phylum Api- coniplexa) in Mediterranean clam Riidilapes seinidecussatus: ultra- structural observations of the cellular response of the host, Aquacultiire 132:153-160. Shaw, B. L. & H. I. Battle. 1957. The gross microscopic anatomy of the digestive tract of Crassostrea virginica (Gmelin). Can. J. Zool. 35: 325-346. ,/,-/,/-;/■(/ -./ Slu-llli.sh R.'sruirh. Vol, 20. No. ?. |{12?-I().W. JOOl, EFFECTS OF PERKINSUS MARINVS ON REPRODUCTION AND CONDITION OF THE EASTERN OYSTER, CRASSOSTREA VIRGIN ICA, DEPEND ON TIMING* D. E. DITTMAN,' - S. E. FORD." AND D. K. PADILLA" ^Department of Zoology. University of Wisconsin-Madison. Madison. Wisconsin 53706: -Jiaskin Slu'Ufisli Research Laboratory. Rutgers University. Port Norris. New Jersey 08349: "'Deparinicnt of Ecology and Evolution, SUNY at Stony Brook. Stony Brook. New York 1 1794-5245 .ABSTRACT The protistan parasite Perl^insiis muriiuis has been respon.sible for high mortality of eastern oysters. Cra.ssostren vir>;inicii. along the Atlantic and Gulf Coast of the United States. It also causes sublethal effects, although an impact on reproduction has heretofore been difficult to demonstrate. We examined the effect of P. mariims infection on growth, reproduction (as measured by relative gonadal size, and the proportional representation of gametogenic tissue and gametes in the gonad), and condition index of eastern oysters. Oysters of known age and genetic history were exposed to P. marimis infection and sampled at 2-4 week intervals over an 11-month period during the 1990-1991 epizootic in Delaware Bay. During this time, 50 to lOO'/r of the oysters had detectable infections and nearly 55% died. Shell growth rates were inhibited as infections intensified during initial exposure in autumn 1990. but resumed in the spring and remained high, even after infection became heavy. The effect on reproduction varied with the stage of the gametogenic cycle. The relative gonadal size and the proportion of gametogenic tissue in the gonad, as well as the condition index, were most strongly depressed during the spring gametogenic period; however, there was no significant effect of parasite burden when the oysters were reproductively mature and ready to spawn. Data suggested that infected oysters, recovered, produced gametes, and spawned in early summer. Other studies have found inconsistent evidence of sublethal deleterious effects of P. maiiiuis on reproduc- tion. By sampling throughout the year and employing oysters of known genetic background, of the same age. reared in a common environment, and for which the timing of initial infection was known, we were able to show that P. muriinis does have significant negative effects on oyster reproduction, measurable primarily during gametogenesis. rather than at spawning. KEY WORDS: Pcikinsi{s nnirinus. oysters, reproductive stage, reproduction, parasite effects, condition index INTRODUCTION In Delaware Bay, two major protozoan parasites, Perkinsiis marinus {Mackin, Owen, Collier 1950, Levine 1978) and Hap- losporidiiiiii nclsdiii (Haskin, Slauber, Mackin 1966) infect oys- ters. Crasso.slreci virginica (Gmelin 1791) (Ford & Haskin 1982, Ford 1996). These parasites cause severe physiological disruption in the host oysters, often leading to death (Ray et al. 1953, Ford & Figueras 1988, Choi et al. 1989, Ford & Tripp 1996, Paynter 1996). They have been recognized as a cause of oyster mortality along the east coast of the United States for at least 40 years (Ford & Tripp 1996). While H. nelsoni has been prevalent in Delaware Bay since the late i950s. P. marinus did not become epizootic in until the late summer of 1990 (Ford 1996). Before that, only spo- radic and very low levels of this parasite were found in the Bay. The sudden upsurge oi P. marinus infections is probably due to a history of introductions of infected oysters into Delaware Bay and an environmental warming trend that allowed the parasite to be- come epizootic (Ford 1996, Cook et al. 1998). Decreased reproduction is one common effect of parasites on their hosts (Thompson 1983. Barber et al. 1988. Forbes 1993. Choi et al. 1994. Ford & Tripp 1996). While infection by H. nelsoni has been shown to have a significant sublethal effect on reproduction (Barber et al. 1988). similar evidence for P. minimis has not been definitive (Choi et al. 1994. Kennedy et al. 1995. Paynter 1996). Kennedy et al. ( 1995) were able to detect some deleterious effects of P. marinus infection on gametogenic characteristics, but not in a consistently predictable pattern. Barber (1996) found that C. virginica exposed to infections by both H. nelsoni and P. marinus in Virginia showed little gametogenesis at all; however, the rela- tionship between level of parasitism and gonadal stage in indi- vidual oysters was not reported. To explain variable effects of infection, there has been speculation that oysters can divert energy from growth to reproduction when infected with P. marinus (Choi et al. 1994. Kennedy et al. 1995. and Hofmann et al. 1995). Both reproduction and energy storage in bivalves have distinct seasonal cycles. Perkinsus imninus infection and subsequent disease devel- opment also have seasonal patterns intluenced by environmental temperature and salinity (Andrews 1988. Ford & Tripp 1996, Kennedy 1996). To properly assess the effects of parasites on reproduction and condition of their hosts under these conditions, it is necessary to sample at all stages of both the infection and re- productive cycles. In this study we measured the effects of P. marinus on the eastern oyster by examining individuals of known genetic background and age. reared in a common environment. We followed the effects of P. mariiuis tidiii the first signs of disease in early August of 1990 until the majority of the oysters had heavy infections at the end of June of 1991. Our goal was to quantify the effect of parasitism on the oysters at different stages of the repro- ductive cycle. Oysters are a commercially important species and a significant component of the estuarine ecosystem. Therefore, a direct quantification of the impact of the parasite on host repro- duction, and the timing of this impact, will provide useful infor- mation for management decisions (Powell et al, 1994. Kennedy 1996. Krantz & Jordan 1996. Powell et al. 1947). MATERI.XL .AND METHODS Corresponding author: D. E. Dittman. Tunison Laboratory of Aquatic Sci- ence. -W75 Gracie Rd.. Cortland. NY 13045-9.^57. E-mail; Dawn_ Dittman@usgs.gov "This paper is dedicated to F. James Rohlf on the occasion of his (i5th hnlhdav. Test Oysters Strains of C. virginica have been maintained in a common environment (lower Delaware Bay) for multiple generations since the 1960s, as part of a selective breeding project for resistance to 1025 1026 DlTTMAN ET AL. MSX disease, caused by H. nehoni (Haskin & Ford 1970. Ford & Haskin 1987). In 1988. oysters from Long Island Sound origin (Rutgers LA strain) and Delaware Bay origin oysters (Rutgers XF strain) were spawned to produce pure lines and reciprocal crosses. These 1988 year-class animals were seventh generation represen- tatives of the inbred MSX disease-resistant strains produced in the Haskin Shellfish Research Laboratory breeding program. All were produced and reared under identical conditions (see Dittnian et al. 1998). Sampling We determined reproductive state, condition index, and para- site infection intensities by sampling forty oysters at 2-week (spring, summer, and fall) to 4-\veek (winter) intervals from Au- gust 1990 to July 1991. Equal numbers (N = 10) of offspring from each pure line and their reciprocal crosses were sampled. Similar size distributions (as measured by maximum shell dimension) were sampled on each collection date to minimize the effect of absolute size on the estimates of reproductive measurements and condition index (Galtsoff 1964, Rainer & Mann 1992). Shell sizes of 400 haphazardly chosen individuals were measured every three weeks to monitor shell growth. Dead oysters were counted and removed at regular intervals and cumulative mortality was com- puted over the study period. Sample Processing The shell height (longest dimension from umbo to growing edge) of each oyster was measured with a calipers to the nearest 1 mm. Whole li\'e weight, (wet shell and soft body mass) was mea- sured to the nearest 0.1 g. The shells were opened and the body was removed. Soft body mass was measured to the nearest 0.1 g. Air-dried valves were weighed to the nearest 0.1 g. The rectum was removed and cultured for P. niariiiKs in Ray's fluid thiogly- collate medium (RFTM. Ray 1966). Oysters were individually tagged and fixed in Davidson's solution for 24 hours, and then placed in 70'7f ethanol for storage until histological analysis was conducted. To eliminate possible confounding effects of MSX disease on the analysis, all oysters were examined for H. iiel.st-ni infections by standard tissue section histology. The oysters were processed using the procedure of Barber et al. (1988) for combined condition index measurement and histology. In this procedure, the whole fixed tissue was weighed and a standard 5-mm-transverse section through the gills and visceral mass between the palps and the adductor muscle was removed. The section and the remaining tissues were then weighed separately. The remaining tissue was dried for the calculation of individual condition index. The section was dehydrated through an alcohol series, cleared in xylene, em- bedded in paraffin, and sectioned at 5 p,m. Tissue sections were stained with iron hematoxylin, acid fuchsin. and aniline blue. The slides were used for the measurement of reproductive condition, as well as to determine the presence and intensity of H. iichoni. Reproductive Measures The Gonadal Index (GI). which is the proportion of the cross- sectional visceral mass area occupied by the gonad, is a useful estimate of fecundity in oysters provided individuals with similar environmental histories are compared (Barber et al. 1988, Hef- fernan & Walker 1989. Morales-Alamo & Mann I9S9). In our study, the GI was estimated by image analysis. In addition, the proportion of the gonad occupied by all gametogenic tissue (gam- ete producing follicle tissue and gametes) and by gametes alone was estimated using point-count stereology on two randomly cho- .sen (x400) microscope fields of each oyster (Elias et al. 1971). Condition Index The Condition Index (CI) is a standard measure used in shell- fish growth and reproductive-cycle analyses (Widdows 1985). It is an indication of long-term changes in indi\idual nutrient status and is a useful measure of the metabolic state of bi\alves. including that produced by stress (Rainer & Mann 1992). It is often used to demonstrate seasonal physiological changes in natural populations (Lawrence & Scott 1982). We determined CI as the proportion of the internal shell volume occupied by dried soft tissue: CI = (dry body mass (g) / internal shell \olume (mm')) x 100. Internal shell volume was estimated as the difference between whole live weight and valve weight (Lawrence & Scott 1982). Total dry body mass was adjusted to account for the section re- moved for histology by assuming that the wet mass of the section had the same proportion of water as the rest of the tissue. Infection by Parasites Rectal tissues incubated in RFTM were stained with Lugol's Iodine, spread on a slide, and the level of P. mcuinns infection was rated on the Ray/Mackin scale of 0-5 using a compound micro- scope (Ray 1954, Mackin 1962). A 0 rating indicated that no were parasites detected, while a rating of 5 indicated that the oyster tissues were almost completely replaced by parasites. There is an exponential relationship between the Ray/Mackin scale and para- site abundance in the tissue (Choi et al. 1989), so when single or few parasites were found in the tissue samples, we gave a rating of 0.5 rather than I to allow us to distinguish, for the purpose of testing for parasite effects, between a very low level of detected infection and the relatively higher rating of 1 . For the sake of simplicity, we will term oysters in category 0 "uninfected", al- though we recognize that they have a high probability of being parasitized (Bushek et al. 1994). Sample infection levels are re- ported as pre\alence and the weighted prevalence (WP), which is the mean of all infection intensities including those rated 0. Tissue slides of all oysters were examined for the presence and intensity of H. nelsoni (Ford & Haskin 1982). Oysters with ad- vanced or systemic H. nelsoni infections were eliminated from analyses so as not to confound measurement of the effect of P. imiriinis (Ford & Figueras 1988, Ford et al. 1990). Statistical Analysis To facilitate analysis, we grouped the oyster samples into six periods according to published descriptions of the seasonal repro- ductive cycle (Loosanoff 1942, Ford & Figueras 1988, Thompson et al. 1996). The reproductive patterns of the oysters in this study fit the published descriptions. The first period was August and September of 1 990, the end of spawning for that summer, when the animals were reabsorbing gametes. The second period was Octo- ber to mid-December, when the oysters were reproduclively inac- tive. The third period encompassed early gametogenesis, from January to early April 1991. The fourth period was from mid-April through early May. when the oysters were undergoing rapid ga- metogenesis. The fifth period was mid-May to early June, when the oysters were fully ripe and beginning to spawn. The last period P/ :A'A/,V.S( S MARINUS EFFECTS ON O'iSTERS 1027 was from the middle to the end of June 1441. when the oysters were ni late spawning. When the samples uithin int'eetion elasses were <3. thev were pooled for the analysis. In addition to the oysters eliminated from analysis because of H. nelsoni infections, two animals with trema- tode. Buecephalus sp.. infections were excluded, because this para- site often castrates its hosts (Ford & Tripp 1996). A total of 679 oysters were analyzed. Two-way ANOVAs were performed to test the effects of sam- pling date and parasite burdens on reproductive measures and con- dition index (Sokal & Rohlf 1981). Preliminary tests indicated no significant effect of strain so that all 4 groups were pooled for further analysis. A separate ANOVA was performed for each of the six reproductive periods. Multiple comparisons among infec- tion-category means during each reproductive period were made using the Tukey-Kramer method (Sokal & Rohlf 1981. Gagnon et al. 1990). Correlation analysis was performed to determine the relationship between shell size and P. marinus intensity during infection peak in the autumn of 1990 and the early spring of 1991. Differences were considered significant at p = 0.05. RESULTS Perkiusiis inariiius infections were first detected in the experi- mental oysters in mid August of 1990, at which time prevalence was 18%. All but one infection were category 0.5; one oyster had a level 1 infection. Prevalence and intensity increased through fall and early winter, then declined gradually over the winter and spring, reaching a low in early May 1991 (Fig. I, Table 1). Both measures then increased dramatically from mid May to early June 1991. Mortality rose to 30% between the beginning of August and early November 1990, ceased over the winter and early spring, and resumed in mid May 1991 (Fig. 2). Between mid May and the end of June, when the linal sample was collected. MY'c of the remain- ing oysters died, raising the total cumulative mortality to about 55%. The high death rate continued after the study ended so that by September 1991. an estimated 97% of the oysters had died. Despite the high disease pressure, the mean size of oysters increased over the study period (Fig. 3). A lag in the rate of tH 4- P 3 5- J m S25- w*^ I ■ 0) 2- \ A 5 1 5- 1 Y'^ - 1- ^\ 1 ' 0 5- -»- Weighted Prevalence 1 Jf Prevalence 0- I '■ -; 1 , , ^ Figure 1. Seasonal pattern ot Perkins us marinus prevalence (percent) and weighted prevalence (mean ± se) during the study period. Each data point represents 411 oysters. increase occuned from mid October to mid November 1990. co- incident with the initial development of P. muriiuis infections, and a decrease in size was measured in late winter that may have been a combination of no growth and chipping of the shell edge. Rapid shell growth resumed in early April and continued through the early summer of 1991; even after infection intensities had become heavy (Fig. 1, Fig. 3). There was a marginally nonsignificant as- sociation between shell size and WP during the autumn. 1990. infection peak (p = 0.077. r" = 0.011); the correlation for the June 1991 peak was significant, although weak (p = 0.015. r" = 0.054). At both times, however, the conelation coefficient was positive, indicating that the larger oysters tended to be the more heavily infected. The GI decreased from a peak of 0.55 to about 0.15 as spawn- ing progressed during the summer of 1990. before the P. niarimis epizootic began. It began to rise again in mid April 1991 and peaked in mid to late May at 0.40. then declined through June. when our sampling ended (Fig. 4). Overall, we found a strong seasonal variation in the effects of P. marinus infection on the reproduction and condition of oysters (Table 2). During the gamete resorption phase of late summer 1990. most oysters were not de- TABLE 1. Prevalence and intensity ot Perkinsus marinus Infection in oysters sampled from lower Delaware Ba> from .\u(!ust 1990 to late June 1991. Prevalence is the percentage of oysters Infected in a given period. Intensity is the percentage and number of oysters in a period that had a given intensity of infection (Ray/Mackin ratings). Light Moderate Heavv Verv Heavv No Infection Infectiim Infection Infection Stage of the Reproductive Total % Infection (0.5 & 1) l2 & 3) 14) (5) Cycle (Sample Dates) N Infected (0) '7<- (N) 9c (N) 9c i.N) % (N) 9c (N) Gamete Reabsorplion (8/13 to 9/22/90) 62 50 50(31) 40(25) 8(5) 0 2(1) No Reproduction (10/6 to 12/10/90) 241 84 16 (,39) 13(32) 28 (67) 15(36) 28(67) Early Gametogenesis (1/1 7 to 4/2/91) 157 87 13(21) 45(71) 17(27) 11 (18) 13(20) Late Gametogenesis (4/17 to 5/1/91) SO 61 39(31) 40(32) 10(8) 4(3) 8(6) Early Spawning (5/16 to 6/4/91) 79 84 16(13) 28(22) 16(13) 8(6) 32(25) Late Spawning (6/16 to 6/30/91) 60 100 0 3(2) 23(14) 18(11) 53 (33) 1028 DiTTMAN ET AL. 3/1/90 6/1/90 9/1/90 12/1/90 3/1/91 6/1/91 9/1/91 12/1/91 Figure 2. Cumulative percent mortality of oysters before, during, and after the study period. The sampling period is indicated by the double ended arrow. tectably infected or had only light infections (Table 1 ). Ne\'erthe- less. there was a significant oserall effect of infection on the GI. caused by the few oysters with moderate to heavy infections. The average GI of these individuals was half that of the uninfected oysters (Fig. 5). We found no significant effects on the other reproductive measures or on the condition index (Table 2. Fig. 6. Fig. 7. Fig. 8). During the fall-early-winter period there was no obvious re- productive activity as judged by the absence of gametes and the very low quantity of gametogenic tissue. Nevertheless, we mea- sured a significant effect of parasite burden on both GI and CI (Table 2). Again, it was the worst infected oysters that were af- fected. The average GI of heavily infected oysters (Ray/Mackin rating = 5) was two-thirds that of uninfected oysters and the CI was three-quarters (Fig. 6). There was no significant effect on the proportion of gametogenic tissue (Table 2. Fig. 7). In both the gamete-resorption and the reproductive-inactivity periods, 2-way ANOVAs showed a significant effect of sample date, indicating that the measures of reproduction and condition changed over time within these periods. There were no significant interactions of parasite burdens and sample dates in any of the analyses, however, indicating that the relative pattern of the effects of these two factors did not change. During the winter period of early gametogenesis there was a significant effect of parasite burden on all measures of reproduc- tion and the CI (Table 2. Fig. 5. Fig. 6. Fig. 8). The average GI in heavily infected oysters was half that of uninfected ovsters. but there was a trend toward smaller indices in oysters beginning with category .^-4 infections (Fig. 5). Similarly, the average proportion of gametogenic tissue and gametes was also about half that of uninfected oysters (Fig. 7. Fig. 8). with a suggestion that gamete production was impaired in oysters with relatively light infections (Fig. 7). Even though average parasite burdens declined, as is typical during the spring (Fig. 1 ), reproductive activity during the period of late gametogenesis was severely impaired by P. mariinis. All measures of reproduction, as well as the CI. were affected (Table 2). The clearest impact was on the GI. which was signifi- cantly depressed at all infection levels (Fig. 5). The gonads of oysters with category 2-5 infections occupied, on average, only one-third the proportional area as those of uninfected or lightly infected individuals. The average proportion of gametogenic tissue and gametes in oysters with categories 4 and 5 infections, was only a quarter of that in uninfected oysters (Figs. 7 and 8). The CI of the most heavily infected oysters was two-thirds that of uninfected oysters (Fig. 6). Finally, during both the early spawning period of late May to early June and the later spawning period of mid-June to late June there were no detectable effects of parasite burdens on any of the reproductive indices or on CI (Table 2. Fig. .5, Fig. 6. Fig. 7. Fig. 8). There were marginally nonsignificant trends toward less ga- metogenic tissue (Table 2. p = 0.0.58) and fewer gametes (Table 2, p = 0.081) during the later spawning period, but statistical power was lost due to small sample sizes in the uninfected and lightly infected categories (Fig. 7, Fig. 8). DISCUSSION One of the most common sublethal effects of parasites on their host is a reduction in growth and reproduction (Thompson 1983. Price et al. 1986. Ford & Figueras 1988. Forbes 1993. Ford & Tripp 1996). Although an inhibitory effect of P. manuiis on oyster growth has been reported (Menzel & Hopkins 1955. Paynter & Burreson 1991 ). the consistent quantification of an impact on re- production has been elusive to date (Choi et al. 1994. Kennedy et al. 1995, Paynter 1996). We found clear evidence that P. inaiimis infection had a significantly negative effect on relative gonadal size, and on the proportions of gametogenic tissue and gametes in the gonad. Detection of an effect depended on when during the disease and reproductive cycles the sample was taken, the type of 6/1/90 9/1/90 12/1/90 3/1/91 6/1/91 Figure 3. Shell growth of oysters during study period. Each point represents the mean (± se) shell size of 4(l(t individuals. Figure 4. Seasonal changes in the gonadal index before and during the study period. Kach point represents the mean (± se) of 40 individuals. Pekkinsus marinus Effects on Oysters 1029 TABFE 2. Results of 2-v\a> ANOVA. Sample date was a fixed factor and intensity of /'. murimis on the Raj/.Matkin scale was a random factor. There were no significant interactions of parasite burden and sample date for any of the analyses. Parasite effect P values significant at <0.05 level are indicated in bold. Stage of the Reproductive Cycle ANOVA Model Factors (N) P \alues from 2-Hav ANOVAs % Cionad Area Condition Index C'f Cametogenic Tissue % Gametes Gamete Reahsorption Parasites (4) 0.018 (8/13 10 4/22/90) Dates (2) Total N (62) 0.026 No Reproductne Parasites (6) 0.006 Activity (10/(1 lo 12/10/yO) Dates (4l Total N(241) <0.(.)(11 Early Gametogenesis Parasites (7) 0.004 (1/17 lo 4/2/41) Dates (4) Total N(157) 0.124 Late Gametogenesis Parasites (5) . Paynter and Burreson (1991) reported that even very light infections in a population of oysters were sufficient to markedly slow the average growth rate. We, too, recorded a decrease in shell growth rate coincident with the on.set and development of infections in the autumn. When grow th resumed in the spring, infection intensities were apparently too low to affect shell deposition, but rapid growth continued even after parasite loads became high. This unexpected result did not occur because heavily infected oysters died, as they tended to be larger, not smaller, than lightly infected animals. Rather, it may be due to the abundant food supply at the Cape Shore test site, which has always been an excellent location for oyster growth (Dittman et al. 1998). High phytoplankton levels may have provided sufficient nutrients to fuel gamete production and somatic growth, as well as parasite growth. Whether "recoN- ery" would have occurred under less favorable conditions than we hypothesize is unknown. An interesting question is why has it been so hard to consistently quantify the sublethal effects of P. marinus infections on reproduction. Some studies suffered from low sample EARL\ GAMETOGENESIS 1991 EARL^ SPAWMNG 1991 ^ 0.25- ab "T b 11 23 iTiiTTi r-Ti. 1.0- 0 «- T 0.6- l).4- 11.2- ci n 11 10 ' ^ 12 , 0 1+2 3+4 LATE GAMETOGENESIS 1991 1.0 0.8- U.6- 0.4 0.2 0.0 I i 1.0- 0.8- LATE SPAWNING 1991 -r 0.6- 6 T 0,4- 9 T "T 0.2- n n 8 17 PAR.\.SITE LEVEL P.\RASITE LEVEL Figure 8. Mean (± se) percent gametes for oysters, plotted according to I'erkiiisus marinus infection category, during all 6 reproductive cycle periods, W hen parasite levels were pooled due to small sample sizes it is indicated. Letters above the columns indicate statistically significant differences. Pekkinsus makinl's Effects on Oysters 1033 sizes for the amount of reproductive and parasite level variation observed and there were scant data on the history of the oyster populations and their previous exposure to disease. Also, it is clear that for the most part. P iniiiiiuis intensities do not impair nutrient storage or reproduction until they become heavy (Ray/Mackin stages 4-5). Unless samples contain oysters with advanced infec- tions, an effect ai P. niariniis on reproduction will be difficult to detect. Our study has shown the need to consider the seasonal cycles of both protagonists in attempting to assess the metabolic effects of an endoparasite on its invertebrate host. Samples taken at a single season may fail to show an effect that is clear at another, nor will they provide the history of the populations needed to interpret any given sample result. For instance, the very drastic negative effect on reproduction predicted during the gametogenic phase did not materialize at the time of spaw ning because a period of diminish- ing parasite burdens coincided with gamete maturation. Our in- vestigation was also helped because all oysters were of known age and genetic history, were reared in a common environment, and were exposed to the disease agent at the same time. All of these factors and an experiment that involved sampling throughout the host's reproductive cycle and the parasite's infection cycle allowed us to detect and interpret the effects of infection intensity on oyster reproduction more easily than would otherwise have been possible. ACKNOWLEDGMENTS This research was supported by a New Jersey Graduate Excel- lence Fellowship to D. Dittman, USDA grant #90-38500-521 1 to S. E. Ford and by NSF awards BSR900907() and IBN93 17293 to D. K. Padilla. We thank T. J. Allen, A. O'Connell, M. Schenk, D. Byers, D. Bushek, R. Barber, K. Chalermwat, E. Vinje, everyone at the Haskin Shellfish Research Laboratory and the Director and staff of Friday Harbor Laboratories, University of Washington. This is Contribution No. 2001-25 from the Institute of Marine and Coastal Sciences at Rutgers University and NJ Agricultural Ex- periment Station Publication No. 0-32405-7-01. LITERATURE CITED Andrews, J. D. 1988. Epi/ootiology of the disease caused by the oyster pathogen Perkinsus nuiriiuis and its effect on the oyster industry. Am. Fish. Soc. Spec. Piihl. 18:47-63. Barber. B. J. 1996. Gametogenesis of eastern oysters. Crassosriva vir- ginica (Gmelin, 1791) and Pacific oysters, Cru.s.so.'ilreu giga.s (Thun- berg. 1793) in disease endemic lower Chesapeake Bay. / Shellfisli Res. 15:285-290 Barber, B. J., S. E. Ford & H. H. Haskin. 1988. Effects of the parasite MSX iHiipliispiiridiiim netsoni) on oyster (Crai'ioi/rea virginica) energy me- tabolism. I. Condition index and relative fecundity. J. Shellfish Res. 7:25-32. Bayne, B. L. & R. C. Newell. 1983. Physiological energetics of marine molluscs. In: K. M. Wilber & A. S. M. Saleuddin. editors. The Biology of Molluscs: Physiology. New York: Academic Press, pp. 407—148. Bushek, D., S. E. Ford & S. K. Allen Jr. 1994. Evaluation of methods using Ray's fluid thioglycollate media for diagnosis of Perkinsus mariiius infection in the eastern oyster. Cnissostreii virginica. Ann. Rev. Fish. Dis. 4:201-217. Choi, K. S., E. N. Powell, D. H. Lewis & S. M. Ray. 1994. Instantaneous reproductive effort in female American Oysters, Crassostrea virginica. measured by a new immunoprecipitation assay. Bio. Bull. 186:41-61. Choi, K. S.. E. A. Wilson, E. H. Lewis, E. N. Powell & S. M. Ray. 1989. The energetic cost of Perkinsus marinus parasitism in oysters: quanti- fication of the thioglycollate method. / Shellfish Res. 8:125-131. Cook, T., M. Folli. J. Klinck, S. Ford & J. Miller 1998. The relationship between increasing sea surface temperature and the northward spread of Perkinsus marinus (Dermo) disease epizootics in oysters. Estuarine Coastal Shelf Sci. 40:587-597. Dame, R. F. 1996. Ecology of Marine Bivalves: An Ecosystem Approach. CRC Press Inc. Dittman. D. E.. S. E. Ford & H. H. Haskin. 1998. Growth Patterns m Oysters from Different Estuaries. Mar. Biol. 132:461—469. Elias, H, A. Henning & D. E. Schwartz. 1971. Stereology: Applications to Biomedical Research. Physiol. Rev. 51:158-200. Forbes. M. L. 1993. Parasitism and host reproductive effort. Oikos 67:444- 450. Ford. S. E. 1986. Effect of repeated heniolymph sampling on growth, mortality, hemolymph protein, and parasitism of oysters, Crassostrea virginica. Comp. Biochem. Physiol. 85A:465— 470. Ford. S. E. 1996. Range extension by the oyster parasite Perkinsus marinus into the northeastern United States: response to climate change. ./. Shellfish Res. 15:45-56. Ford, S. E. & A. J. Fiaueras. 1988. Effects of sublethal infection hv the parasite Haplosporuhwn nelsoui iMS.X) on gametogenesis. spawning, and sex ratios of the oysters of Delaware Bay USA. Diseases of Aquatic Organisms. 41:121-133. Ford. S. E.. A. J. Figueras & H. H. Haskin. 1990. IntJuence of selective breeding, geographic origin, and disease on gametogenesis and sex ratios of oysters. Crassostrea virginica. exposed to the parasite Hap- losporidinm nelsoni (MSX). Aipiaculture 87:285-301. Ford, S. E. & H. H. Haskin. 1982. History and epizoology of Haplospo- ridium nelsoni (MSX). an oyster pathogen m Delaware Bay. / of Invert. Pathol. 40:118-141. Ford, S. E. & H. H. Haskin. 1987. Infection and mortality patterns in strains of oysters Crassostrea virginica selected for resistance to the parasite Haplosporidium nelsoni (MSX). J. Parasit. 73:368-376. Ford. S.. E. Powell, J. Klinck & E. Hofmann. 1999. Modeling the MSX parasite in eastern oyster {Crassostrea virginica) populations. I. Model development, implementation, and verification. J. Shellfish Res. 18: 475-500. Ford. S. E. & M. R. Tripp. 1996. Diseases and Defense Mechanisms. In; V. S. Kennedy, R. I. E. Newell & A. E. Eble, editors. The Eastern Oyster Crassostrea virginica. Maryland: Maryland Sea Grant Publications, pp. 581-660. Gabbot. P. A. 1983. Developmental and seasonal metabolic activities in marine mollusks. In: P. W. Hochachka. editors. The Mollusca. 2. New York: Academic Press, pp. 165-217. Gagnon. J.. J. M. Roth. M. Carroll. R. Hofmann, K. A. Haycock. J. Pla- mondon, D. S. Feldman & J. Simpson. 1990. SuperANOVA, Acces- sible general linear modeling. Berkeley, CA: Abacus Concepts Inc. Galtsoff. P. S. 1964. The American Oyster. Crassostrea virginica Gmelin. 64th ed. U.S. Fish Wildlife Ser. Bull Haskin. H. H. & S. E. Ford. 1979. Development of resistance to Miuchinia lu-lsimi (MSX) mortality in laboratory reared and native stocks in Dela- ware Bay. Mar. Fish. Rev. 41:54—63. Hedgecock. D. 1994. Does variance in reproductive success limit effective population sizes of marine organisms? In: Beaumont A. R.. editor. Genetics and Evolution of Aquatic Organisms. London: Chapman & Hall. pp. 122-134. Heffernan. P. B. & R. L. Walker. 1989. Quantitative image analysis meth- ods for use in histological studies of bivalve reproduction. J. moll. Stud. 55:135-137. Hofmann, E. E., E. N. Powell. J. M. Klinck & E. A. Wilson. 1992. Mod- eling oyster populations III. critical feeding periods, growth and repro- duction. J. Shellfish Res. 1 1:399-+ 16. Hofmann. E. E., E. N. Powell. J. M. Klinck & G. Saunders. 1995. Mod- 1034 DiTTMAN ET AL. eling diseased oyster populations I. modeling Perkinsiis maniuis infec- tions in oysters. J. Shellfish Res. 14:121-151. Kennedy. V. S.. R. I. E. Newell, G. E. Krantz & S. Otto. 1995. Reproduc- tive capacity of the eastern oyster Crassoslreu virgiiiica. infected with the parasite Peikiiisus marimis. Dis. Aqiiat. Org. 23:135-144. Kennedy. V. S. 1996. The ecological role of the eastern oyster. Crussosireu virgiiiica. with remarks on disease. J. Slu'lljisli Res. 15:177-183. Krantz, G. E. & S. J. Jordan. 1996. Management alternatives for protecting Crassosrrea virgiiiica fisheries in Perkiiisiis mariiiiis enzootic and epi- zootic areas. / Shellfish Res. 15:167-176 Levine, N. D. 1978. Perkinsiis gen. n. and other new taxa in the protozoan phylum Apicomplexa. J. Parusitol. 64:649. Lawrence. D. R. & G. I. Scott. 1982. The determination and use of con- dition inde.x of oysters. Estuaries 5:23-27. Loosanoff. V. L. 1942. Seasonal gonadal changes in the adult oysters. Oslrea virginica. of Long Island Sound. Bin. Bull. 82:195-206. Loosanoff. V. L. 1965. Gonad development and discharge of spavsn in oysters of Long Island Sound. Biol. Bull. 129:546-561 Mackin. J. G. 1962. Oyster disease caused by Dermocyslidiuiii iiniriniim and other microorganisms in Louisiana. Pitbl. Inst. Mar. Sci. Univ. Te.\. 7:132-229. Mackin, J. G.. H. M. Owen & A. Collier. 1950. Preliminary note on the occurrence of a new protistan parasite, Dennocystidium mariimin n. sp. in Crassostrea virginica (Gmelin). Science 111:328-329. Menzel. R.W. & S.H. Hopkins. 1955. Growth of oysters parasitized by the fungus Deiiwcysticliuin inariniim and by the trematode Bucephalus en- cuius. J. Parasitology 41:333-342. Morales-Alamo, R. & R. Mann. 1989. Anatomical features in histological sections of Crassostrea virginica ( Gmelin. 1 79 1 ) as an aid in measure- ments of gonad area for reproductive assessment. J. Shellfish Res. 8:71-82. Newell, R. L E.. K. T. Paynter & E. M. Burreson. 1994. Physiological effects of protozoan parasitism on the eastern oyster Crassostrea vir- ginica: feeding and metabolism. J. Shellfish. Res. 13:294. Paynter. K. T. & E. M. Burreson. 1991. Effects of Perkiiisus marinus infection in the eastern oyster. Crussosireu virginica: IL Disease and impact on growth rate at different salinities. / Shellfish Re<:. 10:425- 431. Paynter, K. T. 1996. The effects of Perkinsus marinus infection on physi- ological processes in the eastern oyster, Crassostrea virginica. J. Shell- fish Res. 15:119-115. Powell. E. N., J. M. Klinck. E. E. Hofmann & S. M. Ray. 1994. Modeling oyster populations: IV. rates of mortality, population crashes, and man- agement. f/.s/7. Bull. 92:347-373. Powell, E. N., E. E. Hofmann & J. M. Klinck. 1996. Modeling diseased oyster populations II. Triggering mechanisms for Perkinsus marinus epizootics. / Shellfish Res. 15:141-165. Powell. E. N., E. E. Hofmann & S. E. Ford. 1997. Varying the timing of oyster transplant: implications for management from simulation stud- ies. Fisheries Oceanography 6:213-237 Price, P. W., M. Westoby. B. Rice, P. R. Atsatt, R. S. Fritz, J. N. Thompson & K. Mobley. 1986. Parasite mediation in ecological interactions. Ann. Rev. Ecol. Syst. 17:487-505. Rainer, J. S. & R. Mann. 1992. A comparison of methods for calculating condition index in eastern oysters, Crassostrea virginica I Gmelin, 1791). J. Shellfish ««. 11:55-58. Ray, S. M. 1954. Biological studies of Dennocystidium inariniuu. a fungus parasite of oysters. Houston: Rice Institute. 1 14 pp. Ray, S, M„ J. G. Mackin & J. L. Boswell. 1953. Quantitative measurement of the effect on oysters of disease caused by Dennocystidium marimmi. Bull. Mar Sci. Gulf Carih. 3:6-33. Sokal. R. R. & F. J. Rohlf. 1981. Biometry. 2nd ed. New York: W, H, Freeman and Company. 859 pp. Thompson, S. N. 1983. Biochemical and physiological effects of metazoan endoparasites on their host species. Coiiip. Biocliem. Physiol. 74B(2): 1 83-2 1 1 . Thompson. R. J.. R. I. E. Newell, V. S. Kennedy & R. Mann. 1996. Reproductive Processes and Early Development. In: V.S. Kennedy, R. I. E. Newell & A. E. Eble. editors. The Eastern Oyster Crassostrea virginica. Maryland: Maryland Sea Grant Publications, pp. 335-370 Widdows, J. 1985. Physiological Measurements. In: B. L. Bayne, editor. The Effects of Stress and Pollution on Marine Animals. New York: Praeger Scientiric. pp. 3-45 Juiinuil oj Shellfish Kcscarch. Vol. 20, No. 3. 10.^5-1041. 2001. THE EFFECTS OF A REGULATORY GEAR RESTRICTION ON THE RECRUITING YEAR CLASS IN THE SEA SCALLOP, PIACOPECTEN MAGELLANICVS (GMELIN, 1791), FISHERY JEFFREY C. BRUST, WILLIAM D. DUPAUL, AND JAMES E. KIRKLEY Virt;i)ii(i Institute of Marine Science, School of Marine Science. College of William and Mary, P.O. Box 1346. Gloucester Point, Virginia 23062 .ABSTRACT In 1994. .Amendnu-m 4 lo Ihe slm scallop iPUicupccrcii iuagellcmkii.\) fishery managenient plan was adopted, which restricted fishing effort by controlling \essel days at sea. crew size, and gear size. Dredge ring size was increased from 76.2 mm (3.0") to 82.6 mm (.■!.2.'>") in March 1994. and again to ,SS.9 mm (3.5") in January 1996 to increase the age of entry of scallops into the fishery. Between Jime 1994 and April 1995. four trips were taken on commercial scallop vessels in the western mid-Atlantic to determine harvest efficiency of 88.9-mm dredge rings relative to 82.6-nim dredge rings used in the fishery at the time. Our study focused on the abundant and nearly ubiquitous 3-year-old. 1990 year class. At the time, individuals in this year class were approaching the size (70 mm) of both full recruitment to the gear and recruitment into the fishery. Relative harvest efficiency of this year class ranged from 607o to 72% over the study period. The 88.9-mm rings were found to be 90% efficient when scallops had grown to a size of 100-105 mm. The 88.9-mm ring dredge would therefore delay full recruitment of the 1990 year class for almost 1 y relative to the 82.6-mm ring dredge. Analysis shows that this delay could increase yield-per-recruit by almost 10% and spawning stock biomass by 40% to 60%. Benefits of the gear modifications will only be fully realized when used in conjunction with other measures that reduce or stabilize fishing effort. A'£l' WORDS: sea scallop, relative harvest efficiency, ring size, PUicvpeclen magelhmicus INTRODUCTION The fishery for the Atlantic sea scallop Placopectcn magcUaiu- ciis (Gnielin 1791 ) began in the United States in the mid-Atlantic region in the 1920s and expanded in the 1930s with the discovery of large concentrations of scallops on Georges Bank (Bourne 1964. Serchuk et al. 1979). Commercial landings increased rapidly fol- lowing World War II. and reached a peak of 17,174 metric tons of meats valued at over $145 million in 1990 (NOAA 1991 ). Despite high levels of exploitation and fluctuating annual land- ings, the first fishery management plan (FMP) for the sea scallop was not in place until 1982 (NEFMC 1982). Objectives of the original sea .scallop FMP were to be achieved by controlling the age at entry of the scallops into the fishery. An average meat count restriction of not more than 30 meats per pound (MP?) was en- forced for vessels that landed shucked meats. For vessels landing shell stock, a minimum shell size was set at 88.9 mm (3.5 inches). Several amendments were made to the FMP to correct per- ceived inadequacies and enforcement difficulties; however, the FMP remained relatively unchanged until 1994. Many of these problems with the original sea scallop FMP are discussed by Naidu (1987), Shumway and Schick (1987), Smolowitz and Serchuk (1987, 1988), DuPaul et al. (1989), Kirkley and DuPaul (1989), and Schmitzer et al. (1991). One major problem of the 1982 FMP was the attempt to manage a fishery that extends over a wide geographic area based on a single parameter of the animal, the adductor muscle size. Size of the adductor muscle has been found to vary widely between resource areas, water depths, seasons, and within the reproductive cycle (Haynes 1966, Shumway & Schick 1987, DuPaul et al. 1989), and this confounded managenient ef- forts. In March 1994. Amendment 4 to the sea scallop FMP v\as implemented to replace existing MPP restrictions. Amendment 4 restricted fishing effort by limiting vessel days at sea, crew size, the size of the fishing gear, and entry of new ves.sels into the fishery. Age at entry controls were implicitly imposed by increas- ing the size of the scallop dredge rings and the mesh of the scallop trawl nets (NEFMC 1993 ). The primary gear restriction of Amend- ment 4 increased the minimum ring size of the dredge from 76.2 mm (3.0") to 82.6 mm (3.25") for 1994 and 1995, and to 88.9 mm (3.5") beginning January 1. 1996. Previous gear trials have shown that increasing dredge ring size or trawl mesh size allows an in- creased escapement of smaller-sized scallops (Medcof 1952, Bourne 1965, Smolowitz & Serchuk 1988, DuPaul et al. 1988, DuPaul etal. 1989). In this paper we present an assessment of the effects of an increase in dredge ring size on aspects of selectivity, harvesting efficiency, and delay of entry into the fishery. The study targeted the very large 1990 year class of scallops (NEFSC 1993). We found that an increase in scallop dredge ring size will increase escapement of small, pre-recruit scallops less than 70 mm and will delay entry of scallops greater than 70 mm into the fishery. This regulatory measure could lead to substantial increases in yield and spawning potential of specific year classes of scallops. MATERIALS AND METHODS Data Collection Data for this study were collected during four trips aboard commercial scallop vessels (June, August, and November 1994 and April 1995), each lasting 7 to 14 days. Sampling was con- ducted aboard the fishing vessel (F/V) Carolina Breeze and the FA' Stephanie B in the Delaware/Maryland/Virginia region of the western mid-Adantic (NAFO statistical area 6, Fig. I ). Both ves- sels are approximately 23.01 in (75.5 feet) in length. Catches were sampled from 209 of the 759 tows conducted with the experimen- tal gear. The fishing gear used during the.se experiments was the stan- dard offshore New Bedford style scallop dredge used by most vessels in the fishery. Posgay ( 1957) provides a general overview of the gear, and a more detailed description is given in Bourne ( 1964). For each trip, the control gear was constructed from 82.6- min (3.25") rings, the size used in the fishery at the time of the study. The experimental gear was constructed from 88.9-mm ( 3.5") 1035 1036 Brust et al. Figure 1. Location of sampling efforts in the Mid-atlantic region. rings. Both vessels used are capable of towing two dredges simul- taneously. Since studies have shown that there is no statistical difference in catch for dredges towed in pairs (Bourne 1965, Du- Paul et al. 1989), the control and experimental gears were towed simultaneously. The Carolina Breeze towed two 4.27-m-( 14-foot) wide dredges, and the Stephanie B towed two 3. 96-ni-(l 3-foot) wide dredges. Data collection methodology was similar to that of Bourne (1965) and DuPaul et al. (1989). Sampling procedures were de- signed to allow usual commercial operating procedures, except that catches from the two dredges were kept separate throughout the trip. For sampled tows, the crew was allowed to cull out the commercial-sized scallops. The remaining bycatch was then sorted to retrieve undersized or discarded scallops. Up to two baskets ( 1 basket is approximately 1.5 bushels) of commercial-sized scallops, and at least a portion of the discards were retained from each dredge for shell height frequency sampling. Shell height, the maxi- mum distance from the umbo to the \entral margin oi the shell, was measured to the nearest 5-mm interval using National Marine Fisheries Service scallop measuring boards. Six one-basket samples of scallop shell stock were obtained from commercial scallop vessels between February and May 1 994 to collect shell height and meat weight data. Each sample repre- sented a single trip by a single vessel. When these scallops were shucked, the upper (left) valve was measured to the nearest mil- limeter using a measuring board, and the respective meat was weighed to the nearest 0.1 g using an Ohaus CT 600 electronic scale. Data Analysis To compare catches from the different trips, the data were standardized to a common unit of effort equal to 50 h of tow time using a 3.96-m wide dredge using the following equations: SPH„ = N„/{J'/J ST SC, = ^SPH,/ ('''/,,) (1) (2) (3) where yV„ = the total number of scallops at shell height i in Km j: n,i = the number of scallops sampled at shell height / in tow /.• h^ = the number of baskets sampled in tow/,- B, = the total number of baskets in tow j: SPH,j = the number of scallops caught per hour at shell height / in tow j: 7, = tow time in minutes of tow J: SC, = standardized number of scallops caught at shell height / for the trip; and ST = the number of tows sampled during the trip. For data from the trip using the 4.27-m-\\ide dredges, »,, was multiplied by 13/14 to standardize to a 3.96-m-wide dredge before using the above equations. Harvest efficiency of scallops from the l'-)90 year class was determined for the 88.9-mm ring dredge relative to the 82.6-mm ring dredge for each 5-mm shell height interval within the year class and for the year class as a whole. This was done by dividing the number of scallops caught in the larger ring dredge by the number caught in the smaller ring dredge. Efficiency estimates for the individual shell height intervals were then smoothed using a moving geometric mean of three (Pope et al. 1975, Serchuk & Smolowitz 1980, DuPaul et al. 1989). The range of the shell height intervals used for the year class was found using the Petersen method (Jearld 1983). Year classes were distinguished by different peaks in the shell height frequency distributions for scallops from all trips, except the April 1995 trip. In April 1995, the 1991 year class distribution overiapped slightly with the 1990 year class distribution. The 1990 year class was delineated by finding the modal shell height of the year class and visually estimating the right tail of the distribution. It was then assumed that the number of size classes present on each side of the mode was the same. The conventional allometric or shell height:meat weight model was estimated using SAS software (version 6.09). A log-log trans- formation was necessary to allow estimation of the relation using ordinary least squares. A random effects model was run, with time of collection being the random variable, in order to combine the samples and obtain a common shell height:meat weight relation- ship. The model can be expressed as: In(MW) = ln(a) + b x \n{SH) + » (4) where ""In" is the natural logarithm, MW and SH are meat weight and shell height respectively, and the error term k is N(0, cr"). Scallop growth was estimated by applying an exponential growth model to catch data collected during different trips, includ- ing data from a trip taken in November 1993 (DuPaul & Kirkley 1994a, DuPaul & Kirkley 1994b). It was during the November 1993 trip that the 1990 year class first started to recruit to the 76.2-mni ring dredge used at the time (DuPaul & Kirkley 1994a). The growth model can be expressed as GtAR Restriction in the Sea Scallop Fishery 1037 SH. SH„ X expIC X (/ - I,,)] (5) where SH, is shell height at time ;. SH,, is shell heiyhl at time 0. and G is the exponential growth coefficient. The mean shell height of the specified age class from each of the trips was plotted against number of days relative to the first sampling trip. It was assumed that daily growth of .scallops during each trip was minimal, and that for each trip, all scallops were collected on one day. The first day of each sampling trip was arbitrarily designated the collection day. All increments in days were counted from the tlrst day of the first trip to the first day of each successive trip. Results of the efficiency, growth, and shell height:meal weight calculations were used to examine the effects increasing ring size to 8S.9 mm might have on yield-per-recruit (YPR) of the fishery. The YPR calculations assume only scallops 70 mm and larger are retained by the fishermen. The June 1994 trip was used as a ref- erence point, and the catch in the 82.6-iTim ring dredge during that trip was assumed to be a representative sample of the shell height frequency distribution of the year class. Catch in the 82.6-mm ring dredge in June 1994 was multiplied by 10 to give an arbitrary initial abundance of the year class (A',;). Values of fishing mortality (F) and natural mortality (M) were set at F = 1.5 (NEFMC 1 993 1 and M = 0. 1 (Dickie 1955. Merrill & Posgay 1964) for a full year (0.75 and 0.05. respectively, over 6 mo). The year was broken into two 6-mo periods, and the equation N, = N„ X e- (6) was applied to find population size at time ; = 6 mo. At the end of each 6-mo period, the remaining population (A',) was reduced by the natural mortality rate to find the A',, for the next period. The growth coefficient was applied to the scallops comprising the new Ng to obtain the new shell height frequency distribution. These calculations were repeated four times to simulate harvest over a 2-y period (0-6, 6-12, 12-18, and 18-24 mo). Catch for each period was estimated from C = N,, - N,. The shell height:meat weight model was applied to the harvest at each shell height for each repetition and was summed over all repetitions to finds land- ings for the entire period. To estimate landings for the S8.9-mm ring dredge, similar methods were used except that the catch was adjusted by the relative efficiency at each shell height. A', was found as N,, - C. and then reduced by the natural mortality rate to give A',, for the next period. Spawning potential (in terms of overall fecundity of each ani- mal) was estimated using residual reproductive value (RRV) (MacDonald 1984), RRV takes into consideration the probability of a scallop surviving between successive spawning events. The probability of survival between ages X and X -i- 1 is multiplied by the fecundity at age X -I- 1 . Overall fecundity was found as the sum of fecundity at age X plus the RRV of fecundity at age X -i- 1 . RESULTS Relative Efficiency Harvest by each of the ring sizes during each of the trips is shown for each shell height standardized to 50 h of tow time in Table 1 . Estimates of relative harvest efficiency by the 88.9-mm ring dredge ranged from 60% to 72% over the study period (Table 2). Examination of the efficiency estimates by shell height (Table 3) indicated that scallops from the 1990 year class would be cap- tured v\ith 90% to 100% relative efficiency when they reach 100 to 105 mm shell height. Harvest by the larger ring dredge exceeded harvest by the smaller ring dredge for several of the larger shell heights, but then decreased. With the exception of the November 1994 trip, relative efficiency increased for larger scallops. The unusually large decrease in efficiency in November 1994 is not readily explainable, but has been observed and documented during other gear trial studies (DuPaul et al. 1989). Shell Height: Meat Weight and Growth Six one-basket samples of scallop shell stock were obtained between February and May 1994 for use in shell height:meat weight analysis. Each sample represented a single trip by a single vessel. A random effects model was run, with time of collection being the random variable, in order to combine the samples and obtain a common shell height:meat weight relationship. The rela- tionship found is shown below. The values under the equation are the / values for the coefficients. \n{MW) = -9.7776 + 2.6996 x \n[SH) (-73.604) (98.136) (7) The estimates of the daily growth coefficient between succes- sive trips are shown in Table 4. The overall estimate for the daily TABLE I. Shell hel};hl frequtncv distributions for the 82.6- and 88.'J-nim riiijj dredges for each trip, standardized to 50 h of tow time. The values in bold denote the 199(1 year class. (Scallops smaller than 55 iiini and larger than 110 mm caught by the gear are not shown.) Shell Height June 1994 June 1994 August 1994 August 1994 No> 1994 No> 1994 April 1995 April 1995 ( mm ) (82.6 mm) (88.9 mm 1 (82.6 mml (88.9 mm) (82.6 mm) (88.9 mm) (82.6 mm) (88.9 mm) 55-60 22 8 1069 545 S7() 502 989 610 bO-65 258 130 456 254 587 365 1.876 1. 161 65-70 2,107 866 1..198 813 1.238 755 3,987 2.585 70-75 10.128 5.796 5.253 3,359 2.S14 1.699 7,023 4.969 75-80 21,.M16 1.1.254 9.941 7,256 4,048 2.497 7,890 5,632 80-85 17,644 11.755 11.294 8,123 4,284 2.541 7,611 4,388 85-90 5.153 3.581 5.548 3.828 3,684 1.951 6,483 4,159 90-95 757 561 1.376 942 2,502 1,377 5,671 4.447 95-100 162 199 293 246 1,145 895 3,791 3.573 100-105 162 171 160 161 419 404 2.489 2.291 105-110 149 163 201 175 169 167 1.241 1.061 1038 Brust et al. TABLE 2. Relative harvest efficiency estimates for the 1990 year class during each trip and for all trips combined. Catch in Catch in Relative Modal Shell Height 82.6-mni Ring 88.9-mm Ring Harvest Trip Date (mm) Dredge Dredge Efficiency June 1994 August 1994 November 1994 April 1995 Overall 77.5 82.5 82.5 92.5 57.537 .^5,719 20,721 42.653 47.596 .36.510 24,982 12.484 30,9 1 6 3 1 ,542 0.63 0.70 0.60 0.72 0.66 growth coetTicient (G dilations in this paper. 0,000736) was used for all growth cal YPR and Spawning Patential Applying the shell height:ineat weight relationship to the shell height frequency distribution of the 1990 year class observed in June 1994 results in approximately 490 kg of meats. Delaying harvest by I y (assuming knife-edge selectivity) would yield ap- proximately 844 kg of meats, assuming M = 0.1 and relative efficiency is 909K This is an increase in yield of more than 70%. The offshore scallop dredge, however, does not perform with knife-edge selectivity. When size-specific relative efficiency is considered over a 2-y period, yield from the larger ring dredge was approximately 8.3% more than the smaller ring dredge (Table 5). Based on analysis of cumulative fecundity of scallops ages 3 and 4 using residual reproductive values, increases in fecundity were estimated to range from 43% to 59% higher when the 88.9-mm ring dredge is used and more scallops are present to spawn at age 4 (Table 6). DISCUSSION When Amendment 4 became effective in March 1994. the sea scallop resource was considered to be overfished if spawning stock bioinass (SSB) was below 3% of that of an unfished population. The optimum level of fishing mortality relative to the S.SB thresh- old definition (f^„. ) was determined to be 0.97 (NEFMC 1993). A 1997 re\iew of the sea scallop resource found that fishing mortal- ity in the Mid-Atlantic had been above the overfishing definition since 1985 (NEFSC 1997). Higher than sustainable levels of fish- ing effoiT decreased the abundance of sea scallops, and subse- quently, the number of exploitable year classes available to the fishery. This created the situation where the fishery is highly de- pendent on the recruiting year class (Serchuk et al. 1979, Brown 1987, NEFSC 1993, NEFMC 1993). When there are limited ex- ploitable year classes, a fishery may be detrimentally affected by continued exploitation. Gear regulations contained in Amendment 4 were implemented primarily to lower fishing mortality on small scallops and delay age at entry into the fishery. Data from this study suggest that the 88.9-mni rings will achieve these goals and also improve yield per recruit in the fishery and increase SSB of the scallop slock. The 88,9-mm dredge rings were anticipated to be an important factor in reducing fishing mortality to below the overfishing definition and meeting the management objectives, when used in conjunction with the other management measures. Relative Seleelivity, Efficiency, and Age al Entry The year class targeted during this study was the largest year class on record to recruit to the Delaware/Maryland/Virginia re- gion (NEFSC 1993), As many as 25 baskets (35-40 bushels) of these scallops were caught per tow in a single dredge during a sampling trip in November 1993 (DuPaul & Kirkley 1994a, Du- TABLE 3. Smoothed efficiency values by shell height for each trip and for all trips combined. The values in bold print denote the shell height range of the 1990 year class during each trip. Shell Height Imml June 1994 August 1994 November 1994 50-55 0.49 0.49 0.63 55-60 0.49 0.53 0.61 60-65 0.42 0.55 0.60 65-70 0.49 0.59 0.61 70-75 0.5.3 0.65 0.61 75-80 0.62 0.69 0.60 80-85 0.66 0.71 0.58 85-90 0.70 0.70 0.56 90-95 0.86 0.73 0.61 95-100 0.99 0.83 0.72 100-105 1.12 0.90 0.91 105-110 1.19 0.96 1.02 110-115 1.31 0.90 0.95 115-120 1.43 0.90 1.01 120-125 1.21 O.SI 1.01 April 1995 Overall 0.65 0.64 0.63 0.66 0.69 0.66 0.64 0.66 0.78 0.88 0.91 0.88 0.89 1.08 1.27 0.57 0.59 0.57 0.59 0.61 0.65 0.66 0.68 0,75 0,85 0,92 0.95 0.97 1,06 1,05 Gear Restriction in the Sea Scallop Fishery 1039 TABLE 4. Estimates of the expoiitntial jjriiHtli parameter (G) for llie IWd >ear class for each trip and for all trips combined. Mean Shell Height lnter> a! Between Trips Cumulative Interval Grow th Coefficient Trip Date 1 mm 1 (da) SI (daysl (X lO--") November 1993 62.16 0 June 1994 78.87 217 217 10.97 August 1994 80.17 67 284 2.44 November 1994 82.03 75 359 3.06 April 1995 92.50 181 540 6.64 Overall 540 54(1 7.36 Paul & Kirkley 1994b). The majority of these scallops were 60-6.5 mm and averaged 100 MPP. Larger individuals From this year class were already being retained by the I'ishermen (DuPaul & Kirkley 1994a). Amendment 4 increased the ring si/e used in ihe scallop dredge initially from 76.2 to 82.6 mm. and subsequently to 88.9 mm. DuPaul and Kirkley ( 1994a, 1994bl have shown that the 82.6-mm ring dredge decreased efficiency of the scallop dredge (in terms of number of baskets caught) by 12% on soft bottom (sand and mud) to 457c on hard bottom of cobble, relative to the 76.2-mm ring dredge. During the present study, harvest efficiency of the 88.9- mm ring dredge relative to the 82.6-mm ring dredge (in terms of number of scallops caught) ranged from 60% to 72% (Table 2). For this study, full recruitment (retention) to the experimental gear was considered to be reached when 90% efficiency relative to the 82.6-mm ring dredge was attained. For example, 90% to lOOVr relative efficiency was reached when scallops were 100-105 mm shell height (Table 3). Using the June 1994 modal shell height of 77.5 mm as a reference point, and by applying the estimated growth equation determined in this study, it would take the 1990 year class 346 days, or nearly a year to reach the size of 100 mm and thus be fnlly retained by the 88.9-mm ring dredge. YPR Many studies have examined the effects of delaying harvest as a means to increase YPR in terms of meat weight in the scallop fishery (e.g.. Posgay 1958, Posgay 1962, Caddy 1972a, Caddy 1972b, Posgay 1979. Serchuk et al. 1979. Sinclair et al. 1985). Posgay ( 1979) and Serchuk et al. ( 1979) estimated that maximum YPR is attained by harvesting 8-year-old scallops. Serchuk et al. (1979). however, suggest that scallops be harvested when they reach age 6. since delaying harvest past this age results in only minor additional increases in YPR. Previous studies have shown that an increase in age at first capture from age 5 to age 6 will TABLE 5. Estimates of yield (kilograms) over a 2-y time period using the 82.6- and 88.9-mm ring dredges. Time Period 82.fi-mm Rings 88.9-mm Rings 0-6 mo 2.2.%.53 1,463.82 6-12 mo 1,524.88 1.546.34 12-18 mo 984.35 1.639.49 18-24 mo 634.99 1.178.42 Total 5.380.75 5.828.07 result in an increase in YPR of 10% to 20% (Posgay 1962, Serchuk et al. 1979; Table 7). Delaying harvest from age 4 to age 5 will increase YPR by as much as 50% (Sinclair et al. 1985). By the early 1 990s. age at first capture had decreased to between ages 3 and 4. It is. therefore, important to investigate the changes in yield that could be expected by delaying harvest to age 4 or older. Assuming knife-edge selectivity and no changes in commercial culling size and shucking capacity, delaying harvest from age 3 to age 4 would increase yield in the fishery by nearly 75%. These conclusions are consistent with those presented in other studies that examined YPR for similar sized scallops (Table 7). The off- shore scallop dredge, however, does not perform with knife-edge selectivity, and partial recruitment results in the capture and shuck- ing of small scallops. When partial recruitment is considered in the analysis, and fishing mortality of F = 1.5 and natural mortality of M = 0.1 are assumed. Ihe 88.9-mm ring dredge would increase yield no more than 10% (Table 5). Our results indicate that advances in yield made possible by gear changes have to be optimized with additional measures to reduce fishing mortality and effort. Posgay ( 1979) concluded that a cull size of age 5 (95 mm) was too young and the fishing mortality (F = 0.7) was too high to achieve maximum yield for the Georges Bank scallop resource during the 1960s. Currently, in the mid-Atlantic fishery, age at recruitment is younger and fishing mortality is higher than on Georges Bank in the 1960s. Increasing ring size to 88.9 mm can increase the yield in the fishery, but additional measures to decrease fishing mortality and effort must be implemented in order to attain the maximum YPR. Spawning Potential anil Spawning Stock Biinnass Data from this stud> show the larger ring dredge has the po- tential to improve scallop SSB. which has been depressed due to high exploitation rates, drastically reducing the number of age 4 and 5 scallops. Scallops are sexually mature by the end of their third year (NEFMC 1993). and fall spawning generally occurs between late August and December (Posgay & Norman 1958, MacDonald & Thompson 1986, DuPaul et al. 1989, Schmitzer 1990). It is during this period that Ihe faster growing 3-year-old scallops begin to recruit to the 76.2-mm gear. Most scallops spawn at this time, but the 3-year-old scallops do not contribute signifi- cantly to the overall fecundity of the resource (McGarvey et al. 1993). Even considering partial recruitment, the larger dredge rings will decrease harvest of age 3 scallops and allow more scal- lops to spawn at age 4. Because fecundity of sea scallops increases exponentially with size for several years after first reaching sexual maturity (MacDonald & Thompson 1985. Langton et al. 1987, Carnegie 1994), the delay in harvest should increase the overall fecundity of the resource (Table 6). In addition, McGarvey et al. 1040 Brust et al. TABLE 6. Estimates of residual reproductive value (RRV) assuming 10% natural mortality and 66% harvest efficiency between ages 3 and 4. Reference Study Area Age RRV Cumulative Fecundity (y) (millions of eggsl (milll ions of eggs) J 4.95 4.95 4 2.59 7.54 3 9.30 9.30 4 3.96 13.26 3 5.49 5.49 4 3.24 8.73 MacDonald 1984 Langton et al. 1987 MacDonald and Thompson 1986 Sunnyside. Newfoundland Boothbay. Maine New Jersey (1993) found a statistically significant relationship between the number of spawners and recruits for Georges Bank scallop stocks. Delaying harvest by using the larger rings could therefore lead to increases in overall slock abundance and concomitant increases in SSB. Conclusions The effects of increasing dredge ring size are strongly depen- dent on the amount of fishing effort in the study area. A decrease in fishing effort would increase the benefits of the larger ring size, while increasing effort would reduce the positive effects of the larger rings. Analysis conducted for this study suggest that the 88.9-mm dredge rings should provide many benefits to the sea scallop fishery and resource, and combined with the other regula- tions defined in Amendment 4. should have helped achieve several of the objectives outlined in the FMP. The benefits of the larger ring size suggested by this study have not been realized, however. This is most likely due initially to management measures for other species, and subsequently, similar measures for scallops. In December 1994. the use of mobile fishing gear, including sea scallop dredges, was prohibited in certain areas of Georges Bank (Closed Area I and Closed Area II) and southern New En- gland (Nantucket Lightship Closed Area) in order to allow se- verely depleted groundfish stocks to rebuild (NEFSC 1997). Much of the fishing effort previously targeting scallops on Georges Bank was displaced to the southern New England and mid-Atlantic re- gions, which remained open. Fishing effort substantially increased in these regions (NEFSC 1997). In April 1998. at the request of the New England Fishery Management Council, the Secretary of Commerce closed two additional areas in the mid-Atlantic to scal- lop fishing to protect small scallops. Recently, the NEFMC ( 1997) determined that fishing mortality on scallops is still above the overfishing definition, and the closure of these 1 .900 square miles in the mid-Atlantic only concentrates fishing effort into the re- maining open areas. Modifications to scallop fishing gear designed to reduce the harvest of small scallops and advance the age of entry can be an important management tool, but are most effective when used in conjunction with measures to reduce or stabilize fishing effort and mortality. Data from this .study suggest that increasing scallop dredge ring size can decrease mortality of small scallops and post- pone recruitment of scallops to the commercial fishery. The delay in recruitment can lead to increases in fishery yield and spawning potential of the resource. The potential benefits of the recent gear modifications, however, have been diminished due to increased fishing effort in the area. In order to realize the full benefits of increasing scallop dredge ring size, the gear restrictions must be associated with measures to stabilize or decrease fishing mortality and fishing effort. ACKNOWLEDGMENTS This research was submitted to the Virginia Institute of Marine Science, College of William and Mary (Gloucester Point. VA) by J. B. in partial fulfillment of the Degree of Master of Science. The authors would like to thank Captains Andrew Benevidez {Steph- anie B) and Juan Araiza (Caiolina Breeze) and their crews for allowing us to work on their vessels, and the many captains and crews that provided shell stock for this research. J.B. would like to thank Colleen Brust. Geoff White, and Ryan Carnegie for their assistance in the laboratory and field, and for their continued sup- port throughout the research. We appreciate the constructive and insightful comments that were received from two anonymous re- viewers. This is contribution No. 2397 from the Virginia Institute of Marine Science. Posgay 1962 Serchuk et al. 1979 Serchuk et al. 1979 Sinclair et af 1985 Serchuk et al. 1979 Serchuk et al. 1979 Caddy 1972a Present study TABLE 7. Results of previous sea scallop yield-per-recruit (YPR) studies. Harvest Percent .Age (size) at Delayed Increase in Reference Study Area First Capture Until YPR Georges Bank Georges Bank Mid-Atlantic Georges Bank Georges Bank Mid-Atlantic Georges Bank Mid-Atlantic 5 5 5 4 (76 mm) (77 mm) (73 mm) 3-I- (72.5 mm) 6 6 6 5 (98 mm) (97 mm) (92 mm) 4+(100 mm) 18 11 15 55 37 39 65 74 Gear Restriction in the Sea Scallop Fishery 1041 Bourne. N. 1964. Si.allop,s and the oltshDre fishery of the Mantimes. Bull. Fish. Res. Bel. Canada No. 145. Bourne. N. 1965. A comparison of catches by .V and 4 inch rings on the offshore scallop drags. J. Fish. Res. Bd. Canada 22(2):.^Li-3.\^. Brown. B. E. 1987. The fisheries resource. In: R. H. Backus & D. W. Bourne, editors. Georges Bank. Cambridge: MIT Press, pp. 480-493. Caddy, J. F. 1972a. Size selectivity of the Georges Bank offshore dredge and mortality estimate for scallops from the northern edge of Georges in the period June 1970 to 1971. ICNAF Redljuok. Part 111:79-8.5. Caddy. J. F. 1972b. Some recommendations for conservation of Georges bank scallop stocks. ICNAF Res. Doc. 72/6. Carnegie. R. B. 1994. Size specific fecundity of the sea scallop, flu copecten magellanicus, during one spawning period in the mid-Atlantic resource area. Master's thesis. The College of William and Mary. Wil- liamsburg, VA. Dickie. L. M. 1955. Fluctuations in abundance of the giant scallop. Pla- cnpecten magellanicus (Gmelin). in the Digby area of the Bay of Fundy. J. Fish. Res. Bd. Canada. 12(6):797-857. DuPaul. W. D. & J. E. Kirkley. 1994a. A report to the sea scallop plan development team: preliminary assessment of the ,3.25" ring dredge. Virginia Marine Resource Report No. 94- 1 . DuPaul. W. D. & J. E. Kirkley. 1994b. Harvest efficiency and size selec- tivity of 3.00 and 3.25-inch sea scallop dredge rings. Virginia Marine Resource Report No. 94-5. DuPaul. W. D.. E.J. Heist, J. E. Kirkley. & S. Testeverde. 1988. A com- parative analysis of the effects on technical efficiency and harvest of sea scallops {Placopecten magellanicus) by otter trawls of various mesh sizes. Virginia Marine Resource Report No. 88-10. DuPaul W. D.. J. E. Kirkley. & A. C. Schmitzer. 1989. Evidence for a semiannual reproductive cycle for the sea scallop. Placopecten magel- lanicus (Gmelin. 1791). m the mid- Atlantic region. J. Shellfish Res. 8(1): 173-1 78. DuPaul. W. D.. E. J. Heist. & J. E. Kirkley. 1989. Comparative analysis of sea scallop escapement/retention and resulting economic impacts. Con- tract report. S-K No. NA 88EA-H-000I 1. Haynes. E. B. 1966. Length-weight relation of the sea scallop. Placopecten magellanicus (Gmelin). ICNAF Res. Bull. 3:32-48. Jearld. A. Jr. 1983. Age determination. In: L. A. Nielsen & D. L. Johnson, editors. Fisheries Techniques. Bethesda. MD: American Fisheries So- ciety, pp. 301-324. Kirkley J. E. & W. D. DuPaul. 1989. Commercial practices and tishery regulations: the United States northwest Atlantic sea scallop. Pla- copecten magellanicus (Gmelin. 1791). fishery. J. Shellfish Res. 8(1): 139-149. Langton. R. W.. W. E. Robinson & D. Schick. 1987. Fecundity and repro- ductive effort of sea scallops. Placopecten magellanicus. from the Gull of Maine. Mar. Ecol. Prog. Ser. 37:19-25. MacDonald. B. A. 1984. The partitioning of energy between growth and reproduction in the giant sea scallop. Placopecten magellanicus (Gme- lin). Doctoral dissertation. Memorial University of Newfoundland. St. John's. Newfoundland, Canada. MacDonald. B. A. & R. J. Thomp.son. 1985. Influence of temperature and food availability on the ecological energetics of the giant scallop Pla- copecten magellanicus. II. Reproductive output and total production. Mar. Ecol. Prog. Ser. 25:295-303. MacDonald. B. A. & R. J. Thompson. 1986. InlJuence of temperature and food availability on the ecological energetics of the giant scallop Pla- copecten magellanicus. III. Physiological ecology, the gametogenic cycle and scope for growth. Mar. Biol. 93:37^8. McGarvey, R., F. M. Serchuk & I. A. McLaren. 1993. Spatial and parent age analysis of stock-recruitment in the Georges Bank (Placopecten magellanicus) population. Can. J. Fish. Aquat. Sei. 50:564—574. Medcof. J. C. 1952. Modification of drags to protect small scallops. F/.s7i. Res. Bd Can. Atl. Prog. Rep. 52:9-14. LITERATURE CITED Merrill, A. S. & J. A. Posgay. 1964. Estimating the natural mortality rate of the sea scallop {Placopecten magellinticus). ICN.AF Res. Bull. 1:88- 106. Naidu. K. S. 1987. Efficiency of meat recovery from Iceland scallops lChlani\s islandica) and sea scallops (Placopecten magellanicus) in the Canadian offshore fishery. J. Northwest Ad. Fish. Sci. 7:131-136. New England Fishery Management Council. 1982. Fishery nianagenient plan, final environmental impact statement and regulatory impact re- view for Atlantic sea scallops [Placopecten niagelhuiicus). Saugus. MA. New England Fishery Management Council. 1993. Amendnienl 4 and supplemental environmental impact statement to the scallop fishery management plan. Saugus. MA. New England Fishery Management Council. 1997. Amendment 7 to the Atlantic Sea Scallop Fishery Management Plan. Saugus. MA. Northeast Fishery Science Center. 1993. Sea scallop survey report. Woods Hole. MA: NOAA/National Marine Fisheries Service. Northeast Fishery Science Center. 1997. Report of the 23rd Northeast Regional Stock Assessment Workshop (23rd SAW): Advisory Report on Stock Status. Woods Hole. MA: NOAA/National Marine Fisheries Service. NOAA. 1991. Fisheries in the United States. 1990. Woods Hole. MA: NOAA/National Marine Fisheries Service. Pope. J. A.. A. R. Margetts. J. M. Hamley & E. F. Akyiiz. 1975. Manual of methods for fish stock assessment. Part III. Selectivity of fishing gear. FAO Fisheries Technical Paper #41. Posgay. J. A. 1957. Sea scallop boats and gear. United States Department of the Interior Fish and Wildlife Service Fishery Leaflet 442. Posgay. J. A. 1958. Ma.ximum yield in the sea scallop fishery. ICNAF Document No. 2S. Posgay, J. A. 1962. Maximum yield per recruit of sea scallops. ICNAF Documetit No. 73. Posgay, J. A. 1979. Population assessment of the Georges Bank sea scallop stocks. Rapp. P-\: Reun. Cons. Int. E.xp. Met: 175:109-113. Posgay. J. ,A. & K. D. Norman. 1958. An observation on the spawning of the sea scallop. Placopecten magellanicus (Gmelin), on Georges Bank. Limnol Oceanog 3:478. Schmitzer. A. C. 1990. The gametogenic cycle of Placopecten magellani- cus (Gmelin) in the mid-Atlantic bight. Master's thesis. The College of William and Mary. Williamsburg, VA. Schmitzer. A. C. W. D. DuPaul & J. E. Kirkley. 1991. Gametogenic cycle of sea scallops [Placopecten magellanicus [Gmelin. 1791 1) in the mid- Atlantic bight. J. Shellfi.sh Res. I0( l):22l-228. Serchuk. F. M. & R. J. Smolowitz. 1980. Size selectivity of sea scallops by an offshore scallop survey dredge. ICES CM. 1980/K:24. Serchuk. F. M.. P. W. Wood, J. A. Posgay & B. E. Brown. 1979. Assess- ment and status of sea scallop (Placopecten magellanicus) populations off the northeast coast of the United States. Proc. Natl. Shell A.'^.wc. 69:161-191. Shumway. S. E. & D. F. Schick. 1987. Variability of growth, meLit count and reproductive capacity in Placopecten magellanicus: are current management policies sufficiently flexible'? ICES CM. 1987/K:2. Sinclair, M.. R. K. Mohn. G. Robert & D. L. Roddick. 1985. Consider- ations for the effective management of Atlantic scallops. Can. Tech. Rep. Fish. Aqitat. Sci. No. 1382. Smolowitz. R. J. & F. M. Serchuk. 1987. Current technical concerns with sea scallop management. In: Proceedings: The Oceans, an International Workplace. Halifa.\. Canada. William McNab and Sons. pp. 639-644. Smolowitz, R. J. & F. M. Serchuk. 1988. Developments in sea scallop gear design. In: Proceedings: World Symposium on Fishing Gear and Fish- ing Vessel Design. St. John's. Newfoundland. Canada: Marine Insti- tute, pp. 531-540. Journal ol SiH'lljhl, Research. Vol. 20. No. .^, :()43-l(b(l. 20(11. EFFECTS OF DEPLOYMENT TIME AND ACCLIMATION ON SURVIVAL AND GROWTH OF HATCHERY-REARED SCALLOP {PECTEN MAXIMUS) SPAT TRANSFERRED TO THE SEA GVDA CHRISTOPHKRSEN AND THOROLF MAGNESEN Centre for Studies of Envirouinciit and Resources. University of Beri^en. Norway ABSTR.ACT Halchery-proUuced great scallop iPecicn inaxiinuM spat ol I to 5 mm shell height «ere traiisterred to a sea-based nursery from March to August 1995. Because the growth season in Norwegian waters is limited by low sea temperature in the spring (5°C-10°C). acclimation to a colder temperature ( 10°C) than the 15°C in the hatchery was considered in order to enhance survival after deployment. Survival and growth of spat deployed directly to the sea were compared with spat acclimated for I and 3 wk. Results indicated that specific growth rates were 1% to i'y'c per day. Mean survival of acclimated groups was O'^f to 99c for small spat (0.7-2.3 mm) and 259c to 36% for bigger spat (4 mm). Acclimation improved mean survival by up to 79c of the small spat and 1 ]9c of the bigger spat. .\n extension from I to 3 wk acclimation did not further improve survival or growth. Duration of exposure to low temperatures (<10°C) and temperature at deployment time were main factors affecting survival and growth considering all spat groups, while si/e at deployment mainly determined the success for acclimated scallops. Less than 59c of the sinallest spat survived when directly transfeiTed to sea temperatures of 5 C to 7°C. Larger spat obtained 259c mean survival, which is comparable w ith 249c to 609c survival of spat transferred to sea temperatures alO^C. Shell growth rate likewise was elevated when the temperature exceeded IOX\ Acclimation to a lower temperature gave only limited increase in survival, and it was determined that small spat (<2.6 mm) should not be transferred to temperatures <7°C. By deploying bigger spat (>4 mm) to low sea temperatures, high survival can be expected, and thereby the production period is extended. KEY ]yORI)S: acclimation, nursery culture. Peclen maxiiinis. scallop spat, survival, growth INTRODUCTION Stable spat supply to growers is the tnain cotistraiiit in the development of a scallop cultivation industry. Due to overfishing of scallop populations around the world, and to the fact that scal- lops are high-value products, the establishment of hatcheries for ensuring a reliable production of spat has become itnportanl to the industry. Intensive larval growth in a hatchery is usually followed by growth to commercial-sized spat in semi-intensive growth sys- tems termed "'nurseries." A nursery can be sea- or land-based, depending on natural seawater supply or cultured algae as a food source. In cold water areas like the North Atlantic, the growth season in the sea is limited by environmental factors such as the low temperature in the spring. The longer the period of the year in which scallops can take advantage of the natural food production, the more beneficial it is to producers. Making spat a\ ailablc earlier in the year and providing growers with larger spat at the optimum time in the growth season would ease the constraint. Scallops are shown to be susceptible to small changes in tem- perature (Dickie & Medcof 1963. Strand et al. 1993. Chauvaud et al. 1998). but their tolerance litnits varies with species and natural habitat (Brand 1991 ). Scallops are more sensitive to environmental changes compared with intertidal bivalves like mussels, oysters, and clams because they are not able to completely close their valves when encountering unfavorable conditions. Changes in tetnperature may have a direct or indirect impact on survival. Temperature drops of 4°C to 7°C have been shown to cause im- mobility of sea scallops, which in their natural environment could increase vulnerability to predators and thereby increase mortality (Dickie 1958). Abrupt and considerable changes in environmental conditions are likely to occur in an aquaculture situation during handling and transfer operations. The transfer from hatchery con- ditions to colder temperature conditions in the sea for nursery growth may therefore be hazardous. In culture, both the thermal and the nutritional conditions v\ill affect scallop growth performance (Ventilla 19S2. Wallace &. Reinsties 1985. Andersen & Naas 1993. Couturier et al. 1995. Pilditch & Grant 1999. Grecian et al. 2000). Exposure to cold water temperatures of <10°C have been shown critical for the great scallop. Pectoi inaximiis, at different stages. High niortality of veliger larvae occurs at 7°C to 8°C. and at 5°C. total mortality may be expected (Davenport et al. 1975. Beaumont & Budd 1982). For P. iiKixiiniiiii juveniles of 20 to 30 mm shell height (SH). a reduc- tion in filtration rate of 507f to SOVr was recorded at 5°C compared with 9''C (Strand et al. 1993). and Laing (2000) found no growth or measurable uptake of food at 4.7°C by 5 to 14 mm spat. A 1 00% mortality at <4°C is reported from growth trials in suspended intermediate culture in Norway (Brynjelsen & Strand 1996). while a temperature of 9°C is suggested as a "biological zero" for shell growth in Irish waters (Wilson 1987). Acclimation may be a successful way to adjust an animal to different conditions. A three-phase process consisting of an im- tnediate response, a stabilization of this response, and a new steady state can explain the adaptation to new surroundings (Kinne 1963). Studies of the sea scallop have indicated that a rise or fall in acclimation temperature of about 5°C may result in a coirespond- ing change in lethal temperature of l°C (Dickie 1958). The rate of acclimation tends to follow the rate of metabolism, resulting in longer time needed for acclimation to a decrease in temperature than an increase. The animals' environmental history, genetic background, physiological condition, metabolism, age, and size are all factors affecting the capacity, rate, and effect of thermal acclitiiation (Kinne 1963. Schmidt-Nielsen 1990). One method for producing P. maxiimis spat in Norway is in- tensive hatchery rearing to approximately 2 mm SH at a tempera- ture of 15°C. before further growth to 15 nun in a sea-based nursery. Transfer to the sea is restricted to the period from June to August for the spat to reach a cotntiiercial size within one season. A prolonging of the production period would be possible if a method was found to successfully transfer the spat to the sea earlier in the spring when the temperature is lov\\ The prevailing method of transferring spat from the hatchery is by directly de- ploying the spat to the sea. Acclimation of spat to an environment of IO°C with subsequent transfer to tetiiperatures between 5^C and l(J43 1044 Christophersen and Magnesen 1()°C was considered a possible solution to ensure that scallop spat would tolerate transfer to the sea during the spring. The aim of this study was to determine whether exposing the spat to a temperature between the hatchery and the sea tempera- ture, before transfer to the sea in the spring, could enhance sur- vival. Spat groups that had been subjected to the prevailing pro- duction method of directly deployment to the sea were monitored throughout the season, and survival and growth rates were com- pared with spat groups acclimated for I and ."^ wk. MATERIALS AND METHODS The study was carried out during spring and summer 1995 in 0ygarden, Hordaland County, western Norway, with spat originat- ing from broodstock collected from the wild. During the period from October 1994 to June 1995, groups of 40 to 60 scallops were conditioned, induced to spawn, and cross-fertili/.ed. Spat of 1 to 5 mm SH were obtained by standard production methods at 0ygar- den Scallop Hatchery. Survival and growth of spat from 10 dif- ferent spawning groups (1-10) were determined after deployment to the sea throughout the production season. Spat from spawning groups 3 through 6 were subjected to acclimation treatments for 1 and 3 wk before transfer to the sea (Table 1 ). For some of the spawning groups, more spat groups of different age and size (A, B, and C) were deployed (Table 1 ). Prior to deployment, the spat were removed from 1 4()-|j.m mesh screens, which served as bottom of cylindrical growth containers or sieves f 1,225 cm"), by gentle brushing in a flow of water, and they were relayed onto plastic trays (60 x 60 x 8 cm) covered with 500 |j.m mesh. The trays were kept overnight in running seawater. allowing resettlement of the spat. Stacks of four trays, consisting of three experimental trays plus one as lid, were suspended from long lines in the sea at a depth of 8 to 10 m. Grow-out in the sea was within the period from March 1 to September 29 (Table 1 ). The growth time varied according to deployment date, as the spat grew to a size of 5 to 10 mm. big enough for grading and restock- ing. Acclimation took place under conditions of lower temperature and food concentration than were used prior to the treatment. The spat were kept at 10°C and were fed 10 cells per |xl of a mix of the monocultured algae Pavlova lutheri. Isochiysis galhaiia. and Skel- etoncina costatum in a 1:1:2 ratio, compared with I5°C and 15 to 20 cells per \x.\. respectively. The spat groups (3, 4A, 4B. 5. and 6A) subjected to acclimation were divided into three subgroups, each consisting of three sieves. One subgroup was deployed di- rectly to the sea without acclimation. The second subgroup was deployed after I wk of exposure to acclimation conditions, and the third was transferred after 3 wk of acclimation. Because of limi- tations in the hatchery, the spat were transferred directly from hatchery conditions to acclimation conditions without a gradual habituation to the lower temperature. Survival was estimated as the difference in numbers of animals at deployment time and the numbers retained on a 3-mm screen after 52 to 1 34 days grow -out in the sea. The initial number in each sieve was estimated by counting the spat of a 50 x 1 cm^ (4*^) area of the mesh bottom and multiplying to the total area. The final numbers were estimated by wet weight measurements based on counting subsamples of 100 spat from the trays. TABLE 1. Age, size, density, temperature, and grow-out time in the sea of cultivated scallop spat deployed directly to the sea from the hatchery (0), or transferred after 1 or 3 wk acclimation to a temperature of IOC. Trays Age at Size at Density at Sea Temperature Grow-Out Spat Acclimation Deployed Deployment Deployment Deployment Deployment at Deployment Time Group (»k) (no.) Date in 1995 (days) Inim) (no. cm -) ( C) (days) 1 0 3 Mar 01 131 2.6 2.8 5.9 133 2 0 3 Mar 01 106 2.6 3.1 5.9 134 3 0 3 Mar 01 86 2.3 3.0 5.9 133 1 3 Mar 06 91 2.2 2.8 5.7 128 3 3 Mar 2 1 106 2.0 3.6 5.1 113 4A 0 3 Mar 2 1 56 1.1 2.6 5.1 113 1 3 Mar 29 64 1.3 2.7 5.3 107 3 3 Apr 11 77 1.7 2.9 5.6 93 5 0 3 Apr 11 56 0.7 2.7 5.6 93 1 3 Apr 20 65 0.9 2.4 5.5 85 3 3 May 04 79 1.2 2.1 7.0 71 6A 0 3 Apr 11 56 0.8 2.8 5.6 92 1 3 Apr 20 65 1.3 2.9 5.5 85 3 3 May 04 79 1.6 2.9 7.0 71 4B 0 3 Apr 28 94 4.0 3.3 6,2 77 1 3 May 04 100 4.0 4.0 7.0 71 3 3 May 18 114 4.3 3.4 7.1 68 6B 0 2 May 31 106 3.6 2.5 10.0 55 4C 0 2 May 31 127 5.6 1.7 10.0 55 6C 0 2 May 31 106 5.0 1.7 10.0 55 7 0 3 Jiin 20 63 1.7 1.9 11.5 57 8 0 3 Jim 22 51 1.6 3.2 11.5 55 9 0 3 Aug 01 63 1.9 2.4 12.8 59 10 0 3 Aug 08 56 1.5 2.7 13.0 52 Recovery of Scallop Spat Transferred to the Sea 1045 Growth in the sea was based on mean SH nieasiirenients of 50 animals from each sieve at deployment time and Ifoni each tray at uptake date. Spat growth during the aeeliniation period was also measured. Final size of the bigger (4B. 4C. 6B. and 6Cl deployed spat was determined by using the weighted average of means from two size groups (3-10 mm and >10 mm). SH growth was calcu- lated as specific growth rate (% day"' ): G = (e" - 1 ) x 100, where the instantaneous daily growth coefficient g = (In SH,,,,,, - In SH„„„,,|) per days (Ricker 1979, Claus 1981). Sea temperature and salinity were measured weekly with a WTW Microprocessor Conductivity Meter (LF196) (Wissen- schaftlich-Technische Werkstiitten G.M.B.H. D-8120. Weilheim, Germany) connected with a WTW conductivity-measuring cell (TetraCon 96 A-4). Food concentration in the sea was measured occasionally using an electronic particle counter (Model ZM Coulter Counter. Coulter Electronics Limited, Luton, England). The statistical analyses were cairied out using STATISTICA * , version 5. and a significance level of 0.05 was applied to the tests. Analysis of variance ( ANOVA) was performed on the growth and survival data of the spat that were acclimated, and treatments showing significant differences were further characterized by the Tukey HSD test. The survival data (percentage) of the spat were arcsine transformed before statistical analysis to obtain variance homogeneity for all groups (Sokal & Rohlf 1995). Pearson prod- uct-moment correlation coefficients were calculated for shell pa- rameters and grow-out conditions in the sea. Multiple forward stepwise regression analysis was performed for all the spat groups transfen'ed to the sea and for acclimated groups separately. Sur- vival (percentage) and specific growth (percentage per day) were related to the variables; acclimation time. size, density, and sea temperature at deployment time, and the period of exposure to teinperatures <10'C during grow-out in the sea. RESULTS Seawaler Conditions Sea temperature at the grow-out depth of 8 m increased froin 5"C to 15°C during the experimental growth period (Fig. 1). The temperature reached 7°C in the beginning of May and exceeded 10°C in June. Salinity varied between 30 and 33 with occasional drops below 30%. A salinity of 27.6 was the lowest recorded during the period from June to October (Fig. 1 ). The number of 2- to 10-p.ni particles at 8 m was below 10 cells per p.1 in March and April. In the beginning of July, the natural food production had increased, and 50 cells per jjlI was measured. 16-] - 36 14- P o-o ■ 34 n 12 ■ • ••",• V ■32 7. 1')- » ° 1 s. H 4 - o oOo°°° V - 30 - 28 O li'nipcraturt' - 76 II - -•— salinity - tA Feb Mar Apr May Jun Jul Aug .Sep Oct Date Figure 1. Temperature and salinity at 8 m grow-out depth during the 1995 season, Ulvesundet, Oygarden, Norway. /u - 60 - 0 ■ XI - 50 - r 40 - > 1 1 ^ A + 2 X3 ♦ 4 • 5 06 I 30 -1 ^ 20- ♦ ■a 07 A8 ■ 9 A 10 10 - ♦ 0 — i- t rS- 1 1 1 1 1 1 Feb Mar Apr May Jun Jul Aug Sep Oct Deployment date Figure 2. Survival of scallop spat directly deployed to the sea during the 1W5 season. Different symbols represent the spat groups number I through 1(1, and each point represents the survival in one tray. Survival Stnvival ranging from 09r to 60% was observed for scallop spat transferred directly from the hatchery to the sea during the 1995 production season (Fig. 2). Dead scallops found upon retrieval passed through the 3-mni screen, indicating that no growth had taken place before mortality occuiTed. Spat in the size range 0.7 to 2.6 mm directly deployed to a sea temperature lielow 7°C showed inean survival of less than 5%. while the 4B group, which held a 4 mm SH at deployment, showed 25% survival (Fig. 2; Table 1). A large variation in survival between trays was shown for the spat groups deployed at a sea temperature of 10°C and above. The average survival was 38% (SD = 10.37) for these groups, ranging from 24% to 60% (Fig. 2). The spat groups (3, 4A. 4B. 5. and 6A) subjected to the accli- mation treatment were deployed from March I to May 18 when the sea temperature ranged from 5°C to 7'C (Table 1 ). Mean survival rates from 0% to 9% were obtained for the spat groups of initial average size 0.7 to 2.3 mm. For the 4-mm spat group (4B). mean survival rates of 25% to 36% were found (Fig. 3). Compared to 50 1 40 ^ 30 > u 3 !/3 20 10 Sno acclimation D acclimation I >\eek ^acclimation 3 weeks Jjii fTT 4A 5 6A Spat group 4B Figure 3. Mean survival of scallop spat subjected to acclimation treat- ments. Subgroups of spat were deployed directly to the sea (no accli- mation) and alter 1 and 3 »k of acclimation. Vertical bars show stan- dard deviation. 1046 Christophersen and Magnesen CO 3 •a a. ^ 2 o 0 i o o * ♦ o o « o XI + 2 X3 «4 • 5 Oh 07 AS ■ 9 A 10 Feb Mar Apr May Jun Jul Aug Sep Oct Deployment date Figure 4. Specific growth rate (percentage of SH per day) of scallop spat directly deployed to tlie sea during the 1995 season. Different symbols represent the spat groups nuniher 1 through III, and each point represents the mean growth in one tray. direct transfer from the hatcheiy to the sea. acchmation improved tlie mean survival by up to 7% for the smaller sized spat groups, while the bigger sized group gained a mean increase of 1 1% (Fig. 3 ). A one-way ANOVA was carried out for each spat group. The survival of the acclimated spat from the groups number 3 and 4A was not significantly different from the spat deployed directly to sea. For the other spat groups examined, there was a significant difference between the treatments. Survival of the spat transferred directly to the sea was significantly lower than the survival of acclimated spat. Extension of the acclimation time from 1 to 3 wk did not improve survival significantly for any of the spat groups. GruHili Specific growth rates of the spat deployed to the sea without foregoing acclimation were found to be in the order of 0.9% to 3.3% or 46 to 128 p-m increase in SH per day (Fig. 4). No sur- viving scallops were observed for spat group 5 (Fig. 2), which consisted of the smallest spat deployed during the experiment (Table 1 ). Hence, no growth could be calculated for this group. ? 4 c CJD □ no acL-limation D iicclimalion 1 week D jcclimation 3 weeks a d- ., J.. mi n 3 4A 5 6A 4B Spat group Figure 6. Mean SH at deployment lime of scallop spat subjected to acclimation treatments. Subgroups of spat were deployed directly to the sea (no acclimation I and after I and 3 wk of acclimation. Values indicated by different letters within each spat group are significantly different. Vertical bars show standard deviation. The mean final size of the spat groups surviving direct transfer to the sea was 9. 1 mm (SD = 1 .74). The SHs ranged from 3 to 2 1 mm. The mean specific growth rates obtained for the spat groups subjected to acclimation were between 1.0% and 2.5% per day during the grow-out period in the sea, as for the total experimental period, acclimation time included (Fig. 5). The statistical tests showed that acclimation of the spat group 6A significantly im- proved the growth rate during the total experimental period, and of the spat from group 3 acclimated for 3 wk, considering the grow- out phase in the sea. An extension of the acclimation time from 1 to 3 wk did not result in faster growth for any other spat groups. The SH at time of deployment to the sea differed between the subgroups, and except for group 3 and 4B, the spat groups sub- jected to acclimation treatment showed growth during the accli- mation period (Fig. 6). Upon retrieval, the differences in average SH were evened out. with the exception of the 6A subgroup not acclimated, which obtained a significantly smaller size compared with the acclimated ones (Fig. 7). Correlation and Regression Analyses Growth rate and survival are related to short grow-out time at higher temperatures (Table 2). Growth rate correlated with small spat, and survival correlated with lai'ge spat. Some of the shell and Dno uixlimation DdLilimalion I week Qacctimalion .1 weelvs 4A 6A 4B Spat group Figure 5. Mean specific growth rate (percentage of SH per day) of scallop spat during the total experimental period (acclimation lime included). The spat were deployed directly to the sea (no acclimation) and after I and 3 wk of acclimation. \ crtical bars show standard deviation. no ac'ctimaliun Dacclinialion 1 week Dacclimaluin .^«eeks 4A 6A 4B Spat group Figure 7. Mean shell height upon retrieval of scallop spat subjected to acclimation treatments. Subgroups of spat were deployed directly to the sea (no acclimation) and after I and 3 wk of acclimation. Values indicated by different letters within each spat group are significantly different. Vertical bars show standard deviation. Recovery of Scallop Spat Transferrld to the Sea 1047 TABLE 2. Pearson pioduct-niomenl lorrclation nuitrix for specitlc growth (% day"'), % survival (transforiiutl), acclimation time, age (days from spawning), size (shell height), densitv (no. cm"), and sea temperature al deplounent. average temperature during gro«-out. total grow-out time in the sea, and time exposed to sea temperatures 5°C would not be fatal. Our observations show that some P. iiHiximiix spat of <2.5 mm are able to survive transfer to 5°C to 7'C. while others suffer high mortality. The low lethal tolerance temperature for another cold water scallop species. Pci- riiiopecleii yessoensis. is also found to be about 5"C (Ventilla 1982). while the sea scallop. Placopecten iiuigi'lhiiiicKS. tolerate temperatures down to below ()"C (Couturier et al. 1995). The lower lethal temperature of scallop spat might not be an absolute tem- perature but rather a range. Each scallop species is distributed within a certain geographical and bathymetric range where the environmental conditions support survival (Brand 1991). Toler- ance limits, therefore, will vary between populations of the same species due to the exposure of different local temperature ranges and seasonal variations. Within a population, a high temperature shown to be lethal to the animal during winter inay be tolerated by the animal when exposed to summer conditions and vice versa (Schmidt-Nielsen 1990). A marked increase in shell growth rates was show n at tempera- tures above 10' C (Fig. 4; Table 1 ). Several studies have concluded that temperature rather than food availability is the most important environmental factor affecting scallop growth in the sea (Andersen & Naas 1993. Kleinman et al. 1996. Chauvaud et al. 1998. Laing 2000). Others (MacDonald & Thompson 198.5. Wallace & Reinsnes 1983. Thorarinsdottir 1994. Laing & Psiniopoulous 1998) stress the importance of food supply for scallop growth. Which of the two factors are most important could not be deter- mined in our study due to lack of regular food concentration mea- surements in the sea. The food rations in terms of particle counts were lower from March to May compared with later in the season, but the level was not believed to be growth limiting. Spring bloom events of algae probably ensured enough food since similar growth rates of surviving spat were found both early and late in the season. The low survival in spring found in our study may also be influenced by environmental factors other than temperature and food availability. More site-specific conditions like salinity, flow velocity, presence of predators, and fouling organisms are shown to affect growth and survival of scallop juveniles (Wilson 1987, Cahalan et al. 1989, Wildish & Saulnier 1992, Andersen & Naas 1993, Lodeiros & Himmelman 1996, Freites et al. 2000, Grecian et al. 2000). Likewise will the different environmental factors syn- ergistically affect the growth performance (Kirby-Smith & Barber 1974, Paul 1980, Strand et al. 1993, Grant 1996, O'Connor & Heasman 1998. Pilditch & Grant 1999). In suspended culture, the scallops are exposed to more fluctuations in food supply, salinity. and temperature compared with their natural habitat on the sea bottom (Tebble 1966). By keeping the animals higher up in the water column, as in our study, the temperature and food amount are elevated during spring and summer. The higher survival and growth observed for spat exposed to summer temperatures was probably a result of increased metabolism together with food avail- able in excess. Oxygen consumption and filtration rate of scallops are known to be temperature dependent (McLusky 1973. Bricelj & Shumway 1991). as are energy metabolism, feeding ability, and growth of bivalves in general (Walne 1972. Bayne et al. 1976. de Villiers et al. 1989. Rice & Pechenik 1992, Grant 1996). Relatively high growth rates were obtained for the spat de- ployed in the spring (Fig. 4). The growth was based on measure- ments of spat that had survived the total 7 to 19 wk grow-out in the sea. Seen together with the complementary low survival of the spat transferred early in the year, the good results are due to a few. but very vital, scallops. The growth rate calculations were based on the total grow-out period in the sea. A longer time span combined with a lower average temperature was expected to give a lower daily growth compared with animals allowed to grow at optimal time in the sea. This was confirmed by the significant negative correlation coefficients found between grow-out time and growth (Table 2). Acclimation from I3"C to 10' C reduced the negative impact of cold water temperature at deployment. According to the regression analysis, size was the most important factor explaining the varia- tion in survival and growth of acclimated P. iiia.xinuis spat (Table 3). The acclimated spat were transferred to the similar temperature range (5.1°C-7.1°C). which might account for the reduced effect of temperature at deployment. Larger spat (4 mm) showed better survival than smaller spat (<2.6 mm) when transfen-ed to low temperatures. Along with the size at deployment, a shorter time of exposure to temperatures below IC^C might have influenced the results, as this was also shown as an important factor explaining growth and survival variation. Grecian et al. (2000) likewise found initial size to have significant influence on growth and survival of P. iinigelldniciis spat transfeired to sea-based nursery, showing better results for spat of >3 mm SH compared with spat of 1.2 to 2 mm. For Piiiclachi margaritifera spat, on the other hand, it was found that transfer from the hatchery to the sea should occur as soon as possible after settlement in order to maximize growth (Pit & Southgate 2000). The scallops found dead upon retrieval in the present work were of a small size (<3 mm SH) compared with the majority of the surviving ones. Mortality is therefore likely to have occurred shortly after deployment time. According to the descriptions of the gill organogenesis by Beninger et al. (1994). the spat in our study were anatomically undeveloped. Beninger et al. (1994) found the gill function of P. maxinuis spat up to 4 mm to be different from the adult gill, and suggested that the development from one size stage to the next is critical in terms of survival. This may explain that the spat of 0.7 to 2.6 mm SH used in our study were incapable of rapid adjustment after transfer to the cold water environment. Poorer feeding ability of the spat deployed to the lowest tempera- tures in addition to the temperature stress may have caused the mass mortalities observed in spring. The acclimation to an intermediate temperature between the hatchery (15°C) and the sea (.'S°C-7°C) prior to deployment in spring enhanced the survival of scallop spat. The present study shows increased mean survival of up to 1% for small spat (0.7-2.3 Recover^- oe Scallop Spat Transferred to the Sea 1049 mm) and 1 1% for larger spat (4 mm). The overall survival rates of the larger spat were higher and comparable to those obtained for spat deployed later in the season (Figs. 2 and 3). Regarding growth rates of the larger spat, they were relatively low. and no improve- ment was gained from acclimation (Figs. 4 and 5). These results could be due to the fact that smaller individuals of scallops grow faster than larger ones (Bricelj & Shumway 1991. Parsons et al. 1993) (Table 2). or that growth in sea conditions, in spite of sub- optimal temperature, is more favorable than hatchery conditions. With respect to the SH of the spat subjected to acclimation, some growth compensation seems to have occurred in the sea since similar final sizes were achieved for the subgroups (Figs. 6 and 7). The spat of group 5 and 6A not acclimated and directly deployed to a temperature of 5.6''C appeared, on the other hand, not to reveal such growth compensation. One group suffered total mortality and the other showed a significantly lower final SH compared with the acclimated spat subgroups. These spat, which were 0.7 and 0.8 mm SH at deployment, belonged to the gill development stage 2 de- scribed by Beninger et al. ( 1994). The other spat in the study were stage 3 at transfer, further implying that the spat were transferred to the cold seawater conditions at a critical stage in life. Since no significant difference in survival and tinal SH was found between the groups acclimated for 1 and 3 wk, we believe that 1 wk is sufficient for small scallop spat to adapt to IO°C. Adaptation of mussels to a change in temperature is shown by an immediate response in oxygen consumption and filtration rate, followed by a 2-wk acclimation period for physiological compen- sation to initial level (Bayne et al. 1976). The observed growth during the acclimation period for most groups (Fig. 6) confirms that 1 to 4 mm P. iiuixiiinis spat can tolerate an abrupt change from 15°C to 10 C. Exposure to a decrease in temperature from 15°C to 5°C to 7 C, on the other hand, seemed to be too stressful for the small spat (0.7-2.6 mm). Compared with the survival (249^-60%) obtained for spat transferred to sea temperatures more equal to the hatchery temperature, the mean survival was low i<99c ). A gradual exposure to cold water before deployment early in the season might be a better solution for adjusting the temperature tolerance limit for scallop spat. Laing (2000) successfully acclimated spat of 5 to 14 mm from a rearing temperature of I7°C to 5°C by tem- perature reduction rates of no more than I °C per day. The constant and fluctuation temperature regime studied by Pilditch and Grant (1999) did not affect the shell growth rate of P. inagcllanicus differently, but limited ability to alter metabolic energy demands following temperature changes was shown. It is possible that P. imi.ximiis also has limited capability to regulate its metabolism to sudden changes of low temperatures. Thus, the success of spat production is highly dependent on deployment time, which was also shown for P. inagcllanicus (Grecian et al. 2(J00), and growth performance will depend on initial scallop vilality before abrupt transfer to colder water. Temperature at deployment and the period of exposure to low temperatures are foimd to be the main factors affecting survival and growth for small P. maximum spat ( 1-5 mm) transferred from hatchery to sea conditions. Improvement of tolerance to cold water environment is shown by acclimation to lower temperature, leav- ing size at transfer a critical factor. From an economical point of view, the result was of limited interest, while the maximum mean survival did not exceed 9'/f for the smallest acclimated spat. To ensure high survival rates, spat of <2.6 mm SH should not be transferred to sea temperatures below 7''C. Alternatively, enhanced survival of spat transferred to temperatures from 5°C to 7°C may be achieved by deploying bigger spat (>4 mm). If the spat could be kept in the hatchery to approximately 4 mm and acclimated to a colder temperature, we believe that an earlier start of the growth season in early spring is possible. ACKNOWLEDGMENTS We would like to express our appreciation to 0ygarden Scallop Hatchery for providing the spat and for placing their facilities at our disposal. Special thanks to the hatchery staff for technical assistance and to the Research Council of Norway who supported our work. LITERATURE CITED Andersen, S. & K. E. Naas. 1993. Shell growth and surMval of scallop (Pecreii iiki-\iiiii(\ L.l in a fertilized, shallow seawater pond. Aqiuiciil- hirc 1I():71-S6. Bayne. B. L.. R. J. Thompson & J. Widdows. 1976. Physiology I. In: B. L. Bayne, editor. Marine Mussels: Their Ecology and Physiology. Lon- don: Cambridge University Press, pp. 121-206. Beaumont, A. R. & M. D. Budd. 1982. Delayed growth of mussel {Mytiliis I'thilis) and scallop iPccren mctximiis) veligers at low temperature. Mar. Biol. 71:97-100. Beninger, P. G., S. A. P. Dwiono & M. Le Pennec. 1994. Early develop- ment of the gill and implications for feeding in Pecten ma.ximus (Bi- valvia: Pectinidae). Mar. Biol. 1 I9:40.';-H2. Brand, A. R. 1991. Scallop ecology: distribution and behaviour. In; S. E. Shumway, editor. Scallops: Biology, Ecology and Aquaculture. Am- sterdam: Elsevier, pp. 517-584. Bricelj. V. M. & S. Shumway. 1991. Energy acL|uisillon. In: S. E. Shum- way, editor. Scallops: Biology. Ecology and Ai.|iiacullure. Amsicidam: Elsevier, pp. 305-346. Brynjelsen, E. & 0. Strand. 1996. Provedyrking av storl kaniskiell i nicl- lomkultur - 1995-1996. Fisken og Havet No. 18. Institute of Marme Research, Norway. 34 pp. (In Norwegian with English sinnniaryi. Cahalan, J. A., S. E. Siddall & M. W. Luckenbach. 1989. EITecls of (low velocity, food concentration and particle flux on growth rates of juve- nile hay scallops Argopccwn irnulians. J. E.\p. Mar. Biol. Ecol. 129: 45-60. Chauvaud, L., G. Thouzeau & Y.-M. Paulet. 1998. Effects of ensironmen- tal factors on the daily growth rate of Pecten maxiinus juveniles in the Bay of Brest (France). J. E.\p. Mar. Biol. Ecol. 227:83-1 11. Claus, C. 1981. Trends in nursery rearing of bivalve molluscs. In: C. Claus. N. De Pauw & E. Jaspers, editors. Nursery Culturing of Bivalve Mol- luscs. Bredene, Belgium: European Maricullure Society Special Pub- lication No. 7. pp. 1-33. Couturier, C. P. Dabinett & M. Lanteigne. 1995. Scallop culture in At- lantic Canada. In; A. D. Boghen. editor. Cold-Water Aquaculture in Atlantic Canada, 2nd ed. Sackville. New Brunswick: The Canadian Institute for Research on Regional Development. Tribune Press, pp. 297-340. Davenport. J.. L. D. Gruffyd & A. R. Beaum. Mar. Biol. Ecol. 244:297-303. Grant. J. 1996. The relationship of bioenergetics and the environment to the field growth of cultured bivalves J. Exp. Mar. Biol. Ecol. 200:239-256. Grecian. L. A., G. J. Parsons, P. Dabinett & C. Couturier. 2000. Influence of season, initial size, depth, gear type and stocking den.sity on the growth rates and recovery of sea scallop, Placopeclen magcllanicus. on a farm-based nursery. Aqiiacult. Int. 8:183-206. Kinne, O. 1963. The effects of temperature and salinity on marine and brackish water animals. I. Temperature. Oceanogr. Mar. Biol. .Annii. Rev. 1:301-340. Kirby-Smith, W. W. & R. T. Barber. 1974. Suspension-feeding aquaculture systems: effects of phytoplankton concentration and temperature on growth of the bay scallop. Aquaculture 3:135-145. Kleinman. S.. B. G. Hatcher, R. E. Scheibling, L. H. Taylor & A. W. Hennigar. 1996. Shell and tissue growth of juvenile sea scallops {Pla- copeclen mugellanicus) in suspended and bottom culture in Lunenburg Bay. Nova Scotia. Aquaculture 142:75-97. Laing, I. 2000. Effect of temperature and ration on growth and condition of king scallop (Peclen ma.ximus) spat. Aquaculture 183:325-334. Laing, I. & A. Psimopoulous. 1998. Hatchery cultivation of king scallop (Pecten ma-ximus) spat with cultured and bloomed algal diets. Aqua- culture 169:55-68. Lodeiros, C. i. M. & J. H. Himmelman. 1996. Intluence of fouling on the growth and survival of the tropical scallop Euvola (Peclen) ziczac (L. 1758) in suspended culture. Aquacult. Res. 27:749-756. MacDonald, B. A. & R. J. Thompson. 1985. Influence of temperature and food availability on the ecological energetics of the giant scallop Pla- copeclen nuigellaniciis. 1. Growth rates of shell and somatic tissue. Mar. Ecol. Prog. Sen 25:279-294. McLusky. D. S. 1973. The effect of temperature on the oxygen consump- tion and filtration rate of Chlamys {.Aequipecten] opercularis (L.) (Bi- valvia). Ophelia 10:141-154. O'Connor, W. A. & M. P. Heasman. 1998. Ontogenetic changes in salinity and temperature tolerance in the doughboy scallop. Minuichlamys as- perrima. J. Shellfish Re.i. 17:89-95. Parsons, G. J., S. M. C. Robinson, RofL J. C. & M. J. Dadswell. 1993. Daily growth rates as indicated by valve ridges in po.stlarval giant scallop {Placopeclen magellanicus) (Bivalvia: Pectinidae). Can. J. Fish. .Aqual. Sci. 50:456^64, Paul. J. D. 1980. Salinity-temperature relationships m the queen scallop Chlamys opercularis. Mar. Biol. 56:295-300. Pilditch. C. A. & J. Grant, 1999. Effect of temperature tluctuations and food supply on the growth and metabolism of juvenile sea scallops {Placopeclen magellanicus). Mar. Biol. 134:235-248. Pit, J. H. & P. C. Southgate. 2000. When should pearl oyster, Pinchuia margarilifera (L.), spat be transferred from the hatchery to the ocean? Aquacult. Res. 31:773-778. Rice, M. A. & J. A. Pechenik. 1992. A review of the factors inlluencing the growth of the northern quahog. Merccnaria mcrccnarui (Linnaeus, 1758). / Shellfish Res. 1 1:279-287. Ricker, W. E. 1979. Growth rates and models. In: W. S. Hoar & D. J. Randall, editors. Fish Physiology: Bienergetics and Growth, Vol. 8. London: Academic Press, pp. 677-743. Schmidt-Nielsen, K. 1990. Animal Physiology: Adaptation and Environ- ment, 4th ed. London: Cambridge University Press. 602 pp. Sokal. R. R & F. J. Rohlf. 1995. Biometry: The Principles and Practice of Statistics in Biological Research. 3rd ed. New York: W.H. Freeman and Company, 887 pp. Strand, 0., P. Solberg, K. K. Andersen & T. Magnesen. 1993. Salinity tolerance of juvenile scallops {Pecten nui.xinuis L.) at low temperature. Aquaculture 115:169-179. Tebble, N. 1966. British Bivalve Seashells: A Handbook for Identification. London: Trustees of the British Museum (Natural History), 212 pp. Thorarinsdiittir, G. G. 1994. The Iceland scallop, ChUunys islandica (O. F. Miiller), in Breidaljordur, west Iceland. III. Growth in suspended cul- ture. Aquaculture 120:295-303. Ventilla, R. F. 1982. The scallop industry m Japan. .4(/r. Mar. Bml. 20: 309-382. Wallace, J. C. & T. G. Reinsnes. 19S5. The significance of various envi- ronmental parameters for growth of the Iceland scallop, Chlamys is- landica (Pectinidae), in hanging culture. Aquaculture 44:229-242. Walne, P. R. 1972. The influence of current .speed, body size and water temperature on the filtration rate of five species of bivalves. J. Mar. Biol. Assoc. U.K. 52:345-374. Wildish, D. J. & A. M. Saulnier. 1992. The effect of velocity and flow direction on the growth of ju\ enile and adult giant scallops. J. Exp. Mar. Biol. Ecol. 133:133-143. Wilson, J. H. 1987. Environmental parameters controlling growth of Os- trea edulis L. and Peclen maxiiiur'. L. in suspended culture. Aquacul- ture 64:119-131. JiHinuil of Shellfish Ki'scairh. Vol, 20. No. .^. lO.S 1-1().S7. ^(Kll. HETEROZYGOTE DEFICIENCIES AND GENOTYPE-DEPENDENT SPAWNING TIME IN MYTILUS EDULIS MIGUEL A. DEL RIO-PORTILLA* AND ANDY R. BEAUMONT School of Ocean Sciences, University of Wales. Bangor. Menai Bricli;e. Gwynedd. United Kingdom. LL59 5EY .ABSTRACT Deficiencies of heterozygoles agciinsl the expectations of tlie Hurdy-Weinherg model arc coninionly observed in wild populations of marine bivalves and genotype-dependent spawning time has been proposed as a possible cause of hetero/.ygote deficiency. Adult mussels, Mytihis ediilis L.. were collected every fortnight at two locations in the Menai Strait. North Wales during the spawning season in 1994 and induced to spawn artificially. Genetic analysis at ten allozyme loci revealed no significant differences in allele frequencies or heterozygosity between spawners and non-spawners during the season. It was concluded that the time that individual mussels spawned was not dependent on their genotype at these ten loci and could therefore not be an explanation for the significant heterozygote deficiencies observed at several loci in this .study. The presence of null alleles and/or selection against heterozygotes was suggested as the most likely causes of heterozygote deficiency. KEY WORDS: allozymes. heterozygote deficiency. Mxtilus ahilis. spawning time INTRODUCTION Deficiencies of heterozygotes against the expectations of the Hardy-Weiiiberg model are commonly reported in allozyme stud- ies of bivalves (Gartney-Kepkay et al. 19S(). Skihinski et al. 1983. Zouros & Foltz 1984. Gosling & Wilkins 1985. Volckaert &. Zouros 1989. Beaumont 1991. Gosling 1992) and several expla- nations have been offered. The extensive larval dispersal exhibited by bivalves would tend to rule out the Wahlund effect (the result of sampling differentiated sub-populations or mixtures of cryptic species, Raymond et al. 1997) or inbreeding and. although null alleles certainly occur in some populations of bivalves (Gaffney 1994. Hoare & Beaumont 1993). they need to be present at unex- pectedly high frequency to explain the observed levels of hetero- zygote deficiencies. There may be differential selection against heterozygotes during different life stages (Hu et al. 1993). Indeed. Beaumont ( 1991 ) has demonstrated the generation of heterozygote deficiencies at the post-larval stage in laboratory-reared mussels, presumably through selection against heterozygotes. but evidence for this is extremely difficult to detect in natural populations. One explanation proposed by Zouros and Foltz (1984) is that spawning within a population of bivalves may be genotype-dependent. If. for example, homozygotes for the common allele at a locus tend to spawn at a diffei-ent time, say a few days before or a few days after other genotypes in the population, then panmixia is disrupted and a deficiency of heterozygotes will occur at that locus among the offspring. Gosling and Wilkins (1985) have also proposed "that in species which have an extended spawning period, within a single locality, the biochemical genotype of individuals may be important in regulating spawning synchrony". Rios et al. ( 1996) record ap- parent differences in spawning time between individuals with dif- ferent heterozygosities in a Mediterranean population of Pecten jacobcieiis L. but the relationship between genotype and spawning time has yet to be tested in any other bivalve species. Temperature and food supply are considered to be the most important factors controlling gametogenesis and spaw ning in Mrli- *Present address: CICESE. Acuicultura. Km 107 Carr. Tijuana-Ensenada. Ensenada. Baja California. 22800 Mexico. Corresponding address: PO Box 4.^4844. San Diego. CA. 92143-4844. Fax: +52-61-75- 0534. E-mail: midelrio@cicese.mx. /((,v t'Jiilis. although there may be an interaction between these and other exogenous and endogenous factors (Seed & Suchanek 1992). Mussels may spawn following a variety of stimuli (Seed & Suchanek 1992). but only those mussels which have completed gametogenesis and are "■ripe" will spawn as a result of an artificial stimulus. The objective of the study presented here was to detect any possible allozyme genetic differences between those M. £'(//(//,v which would spawn and those which would not. throughout the spawning season, using artificial spawning induction as a ripeness indicator. MATERIALS AND METHODS Adults of Myriliis cdiilis were collected from a rocky shore at Gallows Point, near Beaumaris (53°15'N, 4°5'W). and by the is- land of Ynys Faelog. Menai Bridge (53°I4'N. 4°9'W). North Wales (referred to as 'Beaumaris' and "Menai Bridge', respec- tively). From January to June 1994. every fortnight during low tide. ca. 30 mussels were collected from these sites and brought into the School of Ocean Sciences (University of Wales. Bangor) laboratory and held at 6 ± T'C in tanks with a continuous flow of sea water until spawning induction (Table I). In mid-April 1994, samples were collected within three days of each other, but the samples from Menai Bridge were lost due to a technical failure in the holding tank. Usually spawning trials were attempted within three days of collection, but when trials were delayed for longer than three days, a tnixture of the micro-algae Pavlova hitheri (Droop) Green. Rhinomonas reticulata (Lucas) Novarino. Skele- toneina costatum (Greville) Cleve and Tetraselmis cliui (Kylin) Butcher was drip-fed into the tanks as a food supplement. Trials were always conducted within 15 days of collection. Spawning was induced by the injection of 2 ml of 0.5 M KCI into the mantle cavity of individual mussels that were then left out of water for one hour. Each mussel was then placed into a 250-nil glass jar and covered with 0.2 |j.m filtered and UV light treated seawater (FSW). and left at I6"C undisturbed for 24 hours. For each mussel in the trial, the maximum linear dimension of the shell was measured to the nearest 0.1 mm and the individual was categorised as a male, a female or a non-spawner. Samples of posterior adductor muscle and digestive gland were taken from each individual to be genetically analysed, placed to- gether in a microtube and frozen at -75°C. until electrophoresis. 1051 1052 Del Rio-PoRTiLLA and Beaumont TABLE 1. Mytiliis ediilis. Sampling date, sample size (n) and mean length (mm) of mussels collected in Beaumaris and Menai Bridge during the 1494 spanning season. Beaumaris Menai Bridge Date n mean se Date n mean se 2S-Jan nsp 23 63. 2& 1.74 28-Jan nsp 28 60.02 0.84 spa 7 57.71 2.88 spa 3 63.13 3.82 1 1 -Feb nsp 24 59.90 1.58 1 1 -Feb nsp 25 59.46 1.18 spa 6 60.32 2.85 spa 6 60.02 1.75 :ri-Feh nsp 24 66.16 1.44 25-Feb nsp 19 59.52 1.25 spa 6 59.07 2.60 spa 11 57.68 1.18 12-Mar nsp 19 64.56 1.38 15-Mar nsp 15 60.82 1.04 spa 11 61.93 2.17 spa 15 65.00 1,56 25-Mar nsp 14 63.03 1 .30 30-Mar nsp 23 58.89 1.06 spa 25 61.24 1.24 spa 18 56.61 1.52 lO-Apr nsp spa 19 17 60.11 57.32 1.42 1.05 13-Apr nsp spa 15 15 65.49 57.21 1.67 1.06 27-Apr nsp 12 61.42 2.34 27-Apr nsp 14 58.81 1.31 spa 19 64.04 1.58 spa 14 62.65 1.44 12-May nsp 10 64.83 1.64 12-May nsp 19 60.49 1.26 spa 20 64.40 1.68 spa 13 59.96 1.71 27-May nsp 22 62.53 1.28 27-May nsp 32 58.58 0.92 spa 19 61.21 1.66 spa 10 57.54 1 .06 1 2-Jun nsp 33 62.53 1.42 12-Jun nsp 33 59.29 1 .03 lisp = non- spawner; spa = = spawner; se = = standard error. All individuals which spawned (up to a maximum of 13) and about 10 non-spawners from each trial were scored. Starch gel electro- phoresis was performed for ten enzymatic systems: NADH- diaphorase (DIA. 1.6.2.2). esterase-D (ESD 3.1.1.1 ). glucose phos- phate isonierase (GPI .5.3.1.9). glutathione reductase (GSR 1.6.4.2). inorganic pyrophosphatase (IPP 3.6.1. 1), leucine amino peptidase (LAP 3.4.11.-). mannose-phosphate isomerase (MPI 5.3.1.8), octopine dehydrogenase (ODH 1.5.1.11). 6-phospho- gluconic dehydrogena.se (PGD 1.1.1.44). phospho- glucomutase (PGM 2.7.5.1). Gels stained for ESD. GPI and LAP were run in TME buffer (0.1 M Tris. 0.1 M maleic acid. 0.01 EDTA and 0.01 MgCU pH 7.4). Gels stained for DIA. GSR and IPP were run using a 0.1 M Sodium Citrate pH 7.4 buffer gel and run in TME pH 7.4 electrode buffer. TME buffer at pH 6.0 was used for ODH and PGM gel and electrode buffers (Beaumont & Beveridge 1983. Gentili & Beaumont I9S8). MPI was run as described by Mc- Donald et al. (1991) together with PGD. LAP and ODH were stained as described by Beaumont et al. ( 1983). and Beaumont et al. (1980). respectively. Other enzymes were stained following Harris and Hopkinson"s (1976) procedures, e.xcept that meldola blue was used instead of PMS (Turner & Hopkinson 1979). Allele nomenclature designated the commonest allele as "100" and the other alleles were given numbers according to their electrophoretic mobility relative to the 100 allele. To test shell length against spawning time, two-factor ANOVA for mussel length was performed using the sampling date and whether or not the mussel spawned as the two factors analysed (Fry 1993). A few trials early and late in the spawning season produced fewer than four spawners and these trials were excluded from the analysis. Spawners and non-spawners were considered as separated sub- populations. The data set was analysed using BIOSYS-I (Swof- ford & Selander 1981 ) to calculate allele frequencies and unbiased heterozygosities (Nei 1978). Agreement with the Hardy-Weinberg (HW) model was tested using the exact test (Genepop: Raymond & Rousset 1995) and the direction of any deviation from the model was indicated by the sign of Wright's (1951) fixation index Fis. Mean unbiased individual heterozygosities were compared for each sample using the /-test (Sokal & Rohlf 1995). Chi-square contingency table analysis was used to determine whether there was any relationship between the frequency of a specific allele at a locus and spawning. This analysis considered Beaumaris or Menai Bridge as populations, and spawners and non-spawiiei's as subpopulatioiis within them. The sequential BonfeiToni technique (Hochberg 1988) with an a = 0.05 was used to establish whether a particular test was significant when considering multiple tests of the same hypothesis. RESULTS Between the 28th of January and the 12th of June 1994. a total of 658 mussels was subjected to spawning induction treatment and 235 spawned (Table I ). There were minor differences in the num- ber of spawners from the Beaumaris (B) and Menai Bridge (MB) populations during the season. By the end of January the mussels were becoming sufficiently ripe to spawn following artificial in- duction. The number of spawners increased as the season pro- gressed with a peak between mid March and the end of April. Spawning activity had ceased by the beginning of June. The mean shell lengths of mussels used in these trials were between 57.2 and 66.2 for Beaumaris and between 56.6 and 65.0 mm for Menai Bridge (Table I ). Two-way ANOVAs for both sites showed no significant difference in mussel shell length between spawners and non-spawners for Menai Bridge (F, ^44 = 0.96. ns). Heterozygote Deficiencies and Spawning in M. epulis 1053 X td -J < o — C r- c vr O I /; — 'y; O c = I = ^1 = q 1' I I c p d r*-. c p c: 3 C C ri a- O^ — ^' 2£ ^. — ^ O ^ ^' — cO nO cO cO*-! o o o o c I I I r^ r- Z: I ri O r I — o o c f^ "^ s ~~ £ p ^ p ^ p ^ ^ £ _ c — . O (^ r- 3 c -O C '^' o o 5 = S I 5 I ■£ p — O ri O -t — o ^ I = p dp dp dp d ^ c Q ^ Ci rr-. n ir-. f^i C: do 3 d — r- d d d O' 1 n =0 :/: C — r', rj "-' — - I = S I elk. I k. c. ci ^ -^ ■- -a 1054 Del Rio-Portilla and Beaumont TABLE 3. Mytilus ediilis. Sample size (nl, agreement willi the Hardy-Weinbcrg expectancies (HW), fixation index Fis and unbiased heterozygosity (H) from mussels collected during the iyy4 spawning season in Menai Bridge, North Wales. Jan i28 Feb 11 Feb 25 Mar 12 Mar 25 Apr 27 May 12 May 27 Locus spa nsp spa nsp spa nsp spa nsp spa nsp spa nsp spa nsp spa nsp n 3 10 6 10 11 9 10 10 12 y 14 10 13 7 10 10 EsD HW Fis — ns -o.osq — — — — — 0.000 0.656 — ns -0.020 ns 0.000 ns 0.000 — — — Gpi HW ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns Fis 1 uno 0.118 0.318 -0.008 0.085 -0.116 0.204 0.273 0.106 -0.080 -0. 1 1 1 0.153 -0.168 -0.220 -0.426 Lap HW ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns Fi.s -11.50(1 -0.273 -0.2S2 -0.256 -0.176 0.510 -0.426 -0.016 0.009 -0.091 0.025 0.455 0.034 -0,075 -0 116 0.023 Dia HW — ns ns ns ns ns ns ns ns ns ns ns ns — ns ns Fis -0.125 -0. 1 1 1 -0.029 -0.081 -0.067 0.640 0.654 -0.158 -0.059 -0.213 -0.200 0.617 0.769* 0.737 Gsr HW — ns — ns — ns ns ns ns ns ns ns ns — — ns Fis -0.029 -0.059 0.000 -0.059 0.000 0.000 0.000 0.000 -0.029 -0.043 0.080 Ipp HW ns ns — ns ns — ns ns ns ns — ns ns — — ns fis n.ooo 0.000 0.000 0.000 0.000 0.000 0.304 0.000 0.000 0.000 0.000 Odh HW ns ns ns ns ns ns ns — ns ns ns ns ns — ns ns Fi.s 1.000 -0.059 0.000 -0.059 -0.07 1 0.158 -0.029 0.200 -0.059 0 172 1.000 O.OOO -0.108 0.000 Pgm HW ns ns — ns ns ns ns ns ns ns ns ns ns ns ns ns Fis 0.500 0.000 0.280 -0.176 0.294 0.446 0.100 -0.213 -0.001 -0.1)44 0.424 0.298 0.357 0.400 0.191 Mpi HW — ns — ns — — ns ns — ns ns ns ns — ns — Fis -0.080 0.000 -0.029 0.000 0.000 0.436 -0.077 -0.043 -0.059 Pg,l HW — ns ns ns — — ns ns — — ns — ns — ns ns Fis 1 .000 -0.250 0.000 0.000 0.000 0.000 0.000 -0.029 0.000 Overall Fis 0.444 O.UIO -0.037 0.012 -0.059 0.1 S7 0.124 0.145 0.067 -0.048 0.019 0.253 0.117 0.059 0.098 0.053 - ve 1 6 3 5 4 2 4 1 t 5 5 3 3 1 5 3 + ve 3 1 1 1 1 3 3 3 5 0 3 4 3 1 ■) ■) H 0,167 0.320 0.233 0.280 0.245 0.2 1 1 0.260 0.240 0.245 0.222 0.310 0.215 0.263 0.171 0.290 0.240 s.e. 0 lo: 0.07(1 0.094 0,076 0.082 0.073 0.076 0.083 0.072 0.064 0.078 0.05 1 0.083 0.095 0.097 0.082 ; -0 101 -0 1). i8 0.034 0.018 01) 25 n 093 0.099 0039 spa = spawner. nsp = non-spawner. ns = not significant. * = significant loilowing Bonferroni correction. — test not carried out due to presence of a tixed allele, s.e. standard error, r value of the difference between heterozygosities of spawners and non-spawners. but a significant tjifference for Beaumaris (F, ,117 = 10.091, /; = 0.001), with spawners being smaller. Overall (considering both sites), the mean shell length of the spawners was larger than that of the non-spawners in 6 instances out of 18. This ratio was not significantly different from a 1:1 ratio (/; > 0.1 ) using the sign test (Sokal & Rohlf 1995). No significant difference in mussel length between sampling dates was found in Beaumaris (Fy,,,^ = 1.88, ns), whilst there was a difference in Menai Bridge (F, -,4^, = 3.39, /; = 0.002): however, this difference was probably a sampling effect because mussels were only roughly size-selected during col- lection. There was no tendency for mussels of a particular size lo spawn for each sampling date and at each site, because there was no significant interaction in any of the two-factor ANOVAs (B: F9,.ici7 = 1.71, ns; MB: F7 24y = 1.76, ns). Agreement with HW expectations (by exact test ) was found at most loci in most samples, except in one out of 152 cases for Beaumaris (Dia. March 12 spawners). This value represents only 0.65% of the total number of tests of the Hardy-Weinberg. Also whenever Fis was significantly different from zero at a particular locus it was always positive (5 cases in Beaumaris and 1 in Menai Bridge, Table 2 and Table 3) indicating too few heterozygotes. In one case (March 12 non-spawners. Table 2) there was a significant positive value of Fix across all loci indicating a cumulative het- erozygote deficit in this group of 10 mussels. The mean unbiased heterozygosity (H) across loci ranged from 0.167 to 0.360 in Beaumaris and from 0.167 to 0.32 in Menai Bridge. There were no significant differences between the H of spawners and non-spawners in any trial for either population ( 10 tests for Beaumaris and 8 for Menai Bridge. Table 2 and Table 3). The sign test was used to determine whether there was a dif- ference between the number of negative and positive values of the Fis. Firstly, +Fis and -Fis values for spawners and non-spawners were .separately added for all loci at each sampling date (column wise - given at foot of Table 2 and Table 3). It was not possible to say that there were significant differences in any case. Secondly, the Fis values were added horizontally for each locus between spawners and non-spawners in both sites (Del Rio-Portilla 1996). Not a single significant difference was found after Bonferroni adjustment. Finally, total Fis values were assessed, and the nega- tive percentages were 44.1% for spawners and 51.6% for non- spawners in Beaumaris, whilst in Menai Bridge, the Fis + ve values were 43.8% and 37.2% for spawners and non-spawners, respectively. None of these was significantly different from 50%'. Furthermore, Wilcoxon tests showed that no median of the Fis values was different from a value of zero (B: spawners C (Wil- coxon statistic) = 639, ns: non-spawners C = 705, ns: MB: spawners C = 509, ns: non-spawners C = 452, ns) Chi-square contingency table analyses were carried otit on al- lele frequency data at each locus for each trial date and for both sites. In no case was there a significant difference in allele fre- Heterozygote Deficiencies and Spawning in M. edulis 1055 TABLE 4. Myliliis vtliili\. Alleli' rrc(|ut'iKy. tlxation inflt\ Fis and agreement HJth the Hardj Weinberg (H\\ I model from pooled data of mussel samples taken from Beaumaris. TABLE 5. Myliliis edulis. Allele Irequenew llNation index Fis and agreement with the Hardy-VVeinherg (HWi model from pooled data of mussel samples taken from Menai Bridge. Locus spa nsp Locus spa nsp Locus spa nsp Locus spa nsp EsD n 106 94 Ipp n 100 85 EsD n 79 73 Ipp n SO 75 74 0.000 0.000 85 0.(120 0.033 74 0.006 0.007 85 0.038 0.033 83 0.033 0.016 100 0.980 0.953 S3 0.019 0.020 100 0.936 0.960 100 0.953 0.963 119 0.000 0.012 100 0.962 0.973 119 0.006 0.007 117 0.014 0.021 117 0.006 0.000 133 0.000 0.000 133 0.006 0.000 HW ns ns HW ns ns HW ns ns HW ns ns Fis 0.174 -0.024 Fis -0.016 -0.034 Fis 0.322 -0.013 Fis 0.115 -0.029 Gpi n 106 94 Lap n 106 94 Gpi n 80 73 Lap n 79 75 77 0.033 0.016 82 0.019 0.032 77 0.006 0.027 82 0.006 0.033 83 0.038 0.043 90 0. 1 27 0.096 83 0.050 0.027 90 0.095 0.147 90 0.222 0.239 100 0.627 0.644 90 0.244 0.232 100 0.614 0.587 93 0.047 0.090 108 0.208 0.2(J7 93 0.069 0.044 108 0.234 0.207 100 0.604 0.548 117 (1.019 0.021 100 0.575 0.647 117 0.051 0.027 104 0.057 0.064 104 0.044 0.033 112 0.000 0.000 112 0.013 0.000 HW ns ns HW ns ns HW ns ns HW ns ns Fis 0.026 0.023 Fis -0.014 0,026 Fis 0.070 0.003 Fis -0.108 0.049 Odh n 106 94 Pkiii n 103 94 Odii n 80 75 Pgm n 80 75 58 0.005 0.000 66 0.005 0.000 58 0.006 0.000 66 0.006 0.000 73 0.009 0.000 78 0.(114 0.059 73 0.038 0.020 78 0.019 0.013 82 0.071 0.074 86 0.032 (.1.096 82 0.063 0.040 86 0.063 0.073 91 0.009 0.011 89 0.067 0.032 91 0.013 0.000 89 0.112 0.067 100 0.877 0.894 100 0.676 0.622 100 0.850 0.887 100 0.587 0.333 114 0.024 0.021 110 0.067 0.085 114 0.019 0.047 110 0.050 0. 1 1 3 130 0.005 0.000 115 O.I 00 0.085 130 0.006 0.000 115 0.I3I 0.173 144 0.000 0.0(10 121 130 0.010 0.010 0.021 0.000 144 0.006 0.007 121 130 0.031 0.000 0.007 0.000 HW ns ns HW HW ns ns HW ns Fis 0.290* 0.134 Fis 0.461* 0.210* Fis 0.132 0.180 Fis 0.196 0.229* Dia n 106 94 Gsr n 106 94 Dia n 78 75 Gsr n 80 73 92 0.137 0.128 57 0.028 0.016 92 0.160 0.093 57 0.006 0.007 100 0.816 0.840 75 0.033 0.053 100 0.782 0.860 75 0.025 0.047 113 0.047 0.032 100 121 0.920 0.019 0.915 0.016 115 0.038 0.047 100 121 0.962 0.006 0.920 0.027 HW * * HW ns ns HW ns ns HW ns ns Fis 0.309* 0.465* Fis 0.010 0.006 Fis 0.186 0.206 Fis -0.023 -0.054 Pgd n 100 85 Mpi n 100 85 Pgd n 80 74 Mpi n 78 73 64 0.020 0.059 70 0.025 0.065 64 0.019 0.007 70 0.045 0.034 100 0.950 0.924 80 0.000 0.006 100 0.944 0.966 80 0.019 0.007 131 0.005 0.006 100 0.960 (1.412 131 0.006 0.014 100 0.936 0.945 162 0.025 0.006 116 0.010 0.018 162 0.023 0.007 116 0.000 0.014 213 0.000 0.006 150 0.005 0.000 213 0.006 0.007 150 0.000 0.000 HW ns ns HW ns ns HW ns ns HW ns ns Fis -0.032 -0.060 Fis -0.026 -0.002 Fis -0.033 0.393* Fis 0.164 -0.035 Overall Fis 0.168* 0.104* Overall Fis 0.088* 0.104* spa = spawner. nsp = non-spawner. n = number of mussels scored for that particular locus, adjustment to HW model: ns = non significant, * significant after Bonferroni correction. quencies between samples of spawners and non-spawners suggest- ing no relationship between the possession of a particular allele and likelihood of spawning following artificial stimulus (Del Rt'o- Portilla 1996). Data from all spawners were pooled, as were data from non- spawners. for each sample site and genetic analyses are given in Table 4 and Table 5. These pooled data showed agreements with the Hardy-Weinberg model at most of the loci, except in Dia. and spa = spawner. nsp = non-spawner, n = number of mussels scored for that particular locus, adjustment to HW model, ns = non significant, * significant after Bonferroni correction. Pgiii in both spawners and non-spawners from Beaumaris, and Pgm of spawners in Menai Bridge. As with the unpooled data f w was positive in all these cases of significant deviation from the model. Overall, the number of loci with positive values of Fis was 24 out of 40 tests which is not significantly different from a 1:1 ratio. Heterozygosity was 0.23 (s.e. 0.061) and 0.26 (s.e. 0.063) for spawners and non-spawners in Beaumaris, and 0.26 (s.e. 0.070) 1056 Del Rio-Portilla and Beaumont and 0.24 (s.e. 0.065) for spawners and non-spawners in Menai Bridge respectively. Neither of these differences between spawn- ers and non-spawners was significant (/ = -0.353, ns and / = 0.038. ns for Beaumaris and Menai Bridge respectively). DISCUSSION Artificial induction of spawning in the laboratory is currently the best test we can make of the ripeness of mussels in the wild. It could be argued that if one mussel (A) spawns in the laboratory following an artificial stimulus while another (B) does not. this does not prove that mussel A would spawn, and mussel B would not spawn, following a natural stimulus in the wild, however, there are no published reports of attempts to detect spawning mussels in situ and the practical problems associated with detecting, and col- lecting spawning individuals of bivalves in the wild for a study such as this are probably insurmountable. It is worth noting that in our trials the results of artificial mussel spawning induction during the spring agreed with pre\'ious observations of spawning in the wild in the Menai Strait (Bayne 1963: Seed 1976). with mussels having both an artificially induced and a natural spawning peak between April and May. The data overall show a trend for mussels which spawned to be smaller than those which did not spawn and this effect is signifi- cant in the Beaumaris population. Possibly smaller mussels are more sensitive to the spawning stimulus but. more likely, this is a gender effect. Almost invariably more males than females spawn following the stimulus and this is particularly noticeable early in the season (Del Rio-Portilla 1996). Because of the reduced unit energetic cost of male vs. female gametes, males may spawn at a smaller size than females and this could account for the difference in size between spawners and non-spawners if we assume a 1 : 1 se.x ratio in the sampled population. Our results provide no evidence to support the hypothesis pro- posed by Zouros and Foltz (1984) that heterozygote deficiencies in natural populations of bivalves might be caused by genotype- dependent spawning. Allele frequencies at the scored loci were not different between spawning and non-spawning groups. Further- more, average heterozygosities across all loci did not differ be- tween groups and neither did an individual mussel's heterozygos- ity (the number of the scored loci at which it was heterozygous) relate to whether or not it spawned (Del Rio-Portilla 1996). This is of interest because, according to Rodhouse et al. ( 1986), reproductive output (the number of gametes released) is related to individual heterozygosity in adult mussels larger than 49 mm and highly heterozygous mussels should therefore produce more off- spring. The overall samples from the Menai Bridge and Beaumaris populations do exhibit a similar pattern of heterozygote deficien- cies to previous studies with other populations of mussels and with other species (Gosling & Wilkins 1985, Gaffney 1990). Although there were not significantly more instances of negative than posi- tive values of Fis. and three of the four mean values of Fis were not significatively different from zero (Beaumaris: spawners mean 0. 1 12 s.e. 0.036, p < 0.01 : nonspawners mean 0.067 s.e. 0.028 n.s.: Menai Bridge: spawners mean 0.076 s.e. 0.040 n.s.: nonspawners mean 0.069 s.e. 0.034 n.s.), whenever significant deviations from the Hardy-Weinberg model did occur, they were always the result of too few heterozygotes. The loci showing heterozygote deficien- cies in this study (based on exact test of H-W or significance of Fi.s. Tables 4 and 5 : Dia. Pgd. Odii and Pgm) have also shown heterozygote deficiencies in previous studies of mussels and other bivalves (Volckaert & Zouros 1989: Fairbrother & Beaumont 1993: Gaffney 1994: Hoare & Beaumont 1995: Ri'os et al. 1996). Although our results have demonstrated that the heterozygote deficiencies in these mussel populations have not arisen as a result of the timing of spawning being genotype-dependent, we have not proven that they are caused by any other factor. The presence of a differentiated sub-population structure (Wahlund effect), or con- sistent mating between close relatives (inbreeding) within our sampled populations seem very remote possibilities. More likely is the miscoring of null-allele heterozygotes as homozygotes. There is evidence that juvenile mussels which are heterozygous, or even homozygous for a null allele at the Oilli locus are able to survive and are apparently at no disadvantage compared with other geno- types possessing active Odh alleles (Hoare & Beaumont 1995) and null alleles have been found at other loci in laboratory crosses of mussels and other bivalves (Foltz 1986a & Foltz 1986b: Gaffney 1994: Del Rio & Beaumont, 2000). However, the Od!i locus is probably a special case because this enzyme operates in a bio- chemical pathway which may not be essential to the organism (Hoare & Beaumont 1995). Although null alleles may have caused the observed significant heterozygote deficiencies at the Odh locus in this study, null alleles are thought less likely to be present at sufficiently high frequency to account for the significant devia- tions from the Hardy-Weinberg model at other loci such as Pgm and Pgd which operate within critical biochemical pathways (Zouros & Foltz 1984). All four sub-samples of mussels (spawners and non-spawners at 2 sites. Tables 4 and 5) showed high positive values of Fis for the Odh locus (mean Fis = 0.184) which would be expected if null alleles were present in the populations. This was also true for the Pgm and Dia loci (mean Fis = 0.274 and 0.292 respectively), but not for the Pgd locus which had negative values of Fis in three sub-samples. Finally, selection against heterozygotes should also be consid- ered as a potential cause of the heterozygote deficiencies. If se- lection is operating at a locus we would expect it to have a similar force and direction in our four sub-samples of mussels and this is true for Odh. Pgm and Dia loci which all have large positi\'e Fis values, but not for the Pgd locus. We conclude that the heterozygote deficiencies against the Hardy-Weinberg model detected in this study are probably the result of the presence of null alleles and/or selection against het- erozygotes, but that they are not caused by genotype-dependent spawning time. ACKNOWLEDGMENTS One of the authors (MADRP) received grants for postgraduate studies from the Consejo Nacional de Ciencia y Tecnologi'a, Mexico and the Overseas Research Students Awards Scheme, United Kinadom. LITERATURE CITED Bayne, B. L. I'Jft.V The hehaviour and ecology of the young stages of Ins edulis: heterozygote deficiencies, heterozygosity and growth. Biol. Mytilus ediilis L (with some notes on the breeding of the same species). J. Linn. Sac. 44:273-28.*i. PhD Thesis. Bangor, UK: University College of North Wales, 131 pp. Beaumont, A. R. & C. M. Beveridge. I9S3. Resolution of phosphogluco- Beaamont, A. R. 1491. Genetic studies of laboratory reared mussels, Myli- mulase isozymes in Myiiliis edulis L. Mur. Biol. ten. 4:97-103. Heterozygotr Deficiencies and Spawning in M. edvlis 1057 Beaumont. A. R.. C. M. Beveridge & M. D. Budd. 1983. Selection and helero/ygo.sity within .single families of the mus.sel Mylihis edulis (L.). Mm: Biol. Leit. 4:1.^1-161. Beaumont. A. R.. T. R. Day & G. Gade. 19X0. Genetic \aiialion al the octopine dehydrogena.se locus in the adductor muscle of Cciasiodi-nihi cduk' (L.) and si.\ other bivalve species. Mar. Biol Letts. 1:1.37-14X. Del Rio-Portilla. M. A. 1996. Genetic aspects of Mytilus edulis in relation to spawning and hatchery culture. PhD Thesis. Bangor. UK: University of Wales. 213 pp. Del Rio-Portilla. M. A. & A. R. Beaumont. 2000. Larval growth, juvenile size and heterozygosity in laboratory reared mussels. Mytilus edulis. ./. Exp. Mar. Biol. Ecol. 2.'i4:l-l7. Fairbrother J. E.. & A. R. Beaumont. 1993. Hetero/ygote deficiencies in a cohort of newly settled Myiilii.\ edulis spat, ./, Mar. Biol. Ass. UK. 73:647-6.';3. Foltz. D. W. 1986a. Segregation and linkage studies of allozyme loci in pair crosses of the oyster Crassosirea i7ri;/«/( a. Biochem. Genet. 24: 941-956. Foltz. D. W. 1986b. Null alleles as possible cau.ses of lieterozygote defi- ciencies in the oyster Crassostrea virginica and other bivalves. Evolu- tion. 40:869-870. Fry, J. C. 1993. One-way analysis of variance. In: J. C. Fry. editor. Bio- logical data analysis: A practical approach. Oxford: IRL Press, pp. 3-39. Gaffney. P. M. 199t). Enzyme heterozygosity, growth rate and variability in Mytilus edulis: Another look. Evolution 44:204-210. Gaffney. P. M. 1994. Heterosis and heterozygote deficiencies in marine bivalves: more light? In: A. R. Beaumont, editor. Genetics and evolu- tion of aquatic organisms. London: Chapman and Hall. pp. 146-1.53. Gartney-Kepkay. K. E.. L. M. Dickie. K. R. Freeman & E. Zouros. 1980. Genetic differences and environments of mussel populations in the Maritime Provinces. Can. J. Fish. Aquat. Sei. 'il:llS-l%2. Gentili. M. R. & A. R. Beaumont. 1988. Environmental .stress, heterozy- gosity and growth rate in Mvlilus edulis L. J. Exp. Mar. Biol. Ecol. 120:145-1.53. Gosling, E. 1992. Genetics of Mytilus. In: E. Gosling, editor. The mussel Mytilus: Ecology. Physiology Genetics and Culture. Amsterdam: Elsevier, pp. 309-382. Gosling. E. M. & N. P. Wilkins. 1985. Genetics of settling cohorts of Mvlilus edulis L.: preliminary observations. Aquaculture 44:1 L5-I23. Harris. H. & D. A. Hopkinson. 1976. Hand book of enzyme electrophoresis in human genetics. Amsterdam: North Holland pub. Co. 259 pp. Hoare, K. & A. R. Beaumont. 1995. Effects of an Odh null allele and GPI low activity allozyme on shell length in laboratory-reared Mytilus edu- lis L Mar Biol. 123:775-780. Hochberg. Y. 1988. A shaiper Bonferroni procedure for multiple tests of significance. Biometrika 75:800-802. Hu. Y. P.. R. A. Lutz & R. C. Vrijenhoek. 1993. Overdominance in early life stages of an American oyster strain. J. Hered. 84:254-258. McDonald. J. H.. R. Seed & R. K. Koehn. 1991. Allozymes and morpho- metric characters of three species of Mytilus in the Northern and South- ern Hemispheres. Mnr, Biol. 111:323-333. Nei. M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics. 89:583-590. Raymond. M. & F. Rousset. 1995. Genepop (version 1.2) population ge- netics software for exact tests and ecumenicism. J. Hered. 86:248-249. Raymond. M.. R. L. Vaanto. F. Thomas. F. Rousset. T. de Meeils & F. Renaud. 1997. Heterozygote deficiency in the mussel Mytilus edulis species complex revisited. Mar. Ecol. Prog. Ser 156:225-237. Rios. C. J. Canales & J. B. Peiia. 1996. Genotype-dependent spawning: evidence from a wild population of Pecten jacobaeus L. ( Bivalvia: Pectinidae). ./. Shellfish Res. 15:645-651. Rodhouse. P. G.. J, H. McDonald. R. I. E. Newell & R. K. Koehn. 1986. Gamete production, somatic growth and multiple-locus enzyme hel- . erozygosity in Mytilus edulis. Mar. Biol. 90:209-214. Seed. R. 1976. Ecology. In: B. L. Bayne. editor. Marine mussels: their ecology and physiology. Cambridge: Cambridge University Press, pp. I-V65. Seed. R. & T. H. Suchanek. 1992. Population Community Ecology of Mvtilus. In: E. Go.sling. editor. The mussel Mytilus: Ecology. Physiol- ogy Genetics and Culture. Amsterdam: Elsevier, pp. 87-169. Skibinski. D. O. F.. J. A. Beardmore & T. F. Cross. 1983. Aspects of the population genetics of Mytilus (Mytilidae; Mollusca) in the British Isles. Biol. J. Linn. Soc. 19:137-183. Sokal. R. R. & F. J. Rohlf. 1995. Biometry. Third ed. New York: Freeman. 887 pp. Swoford D. L. & R. B. Selander. 1981. Biosys-I: a FORTRAN program for the comprehensive analysis for electrophoretic data in population genetics and systematics. J. Hered. 72:281-283. Turner. V.S. & D. A. Hopkinson. 1979. The use of meldola blue in isozyme stains after electrophoresis. F.E.B.S. Letters 105:376-378. Volckaert. F. & E. Zouros, 1989. Allozyme and physiological variation in the scallop Placopeclen magellanicus and a general model for the effects of heterozygosity on fitness in marine molluscs. Mar. Biol. 103:51-61. Wright. S. 1951. The genetical structure of populations. Ann. Eugen. 15: 323-354 Zouros. E. & D. W. Foltz. 1984. Possible explanations of heterozygote deficiency in Bixalve Molluscs. Malacologia 25:583-591. .loiinial ol Shellfish Rcseaich. Veil. :(). No. ,\ 1054-1(163. 2(1(11. ARTIFICIAL ENVIRONMENTAL CONDITIONS CAN AFFECT ALLOZYME GENETIC STRUCTURE OF THE MARINE GASTROPOD PATELLA CAERULEA ANNAMARIA MAURO, NICOLO" FARRINELLO, AND MARCO ACRULEO* DipitriimeiUo di Biologia Auiinalc via Airliirafi 18, 90123 Palermo. Italy ABSJKACT Siimples ot (he Modilenancan linipel Pahila ccicnilai collected troni lU .siles were examined lor genetic population structure. Six of the 20 identified enzymatic loci were polymorphic. The AAT* locus was polymorphic only in two samples froin an artificial environment (TI2 and PE|. The proportion of polymorphic loci ranged from 0.20 to 0.30. and the observed and expected mean heterozygosity varied between 0.098 and 0.076 and between 0.109 and 0.086, respectively. Mean Fis values were significantly positive in A47"*. ESTD*. PEPC-2*. and PEPD". showing heterozygosity deficiency. In all, mean Fst value of 0.007 indicated high genetic homogeneity between the samples analyzed, whereas single-locus Fsr analysis showed an interesting case of significant microscale genetic heterogeneity in AAT* {Fsl = 0.158, P < 0.(J0l 1. The specimens collected from the two artificial environments (T12 and PEl were responsible lor AAT- heterogeneity. These results, as suggested by other authors, could be related to the hydrodynamic conditions of artificial en\ironmcnts. KEY WORDS: Patella eaeriilea. senetic \ariatioii. allo/vnies. artificial environment. Mediterranean Sea INTRODUCTION Palclhi ccwriilea L. is a very coiiunoii gilslrdpiid along shel- tered Mediteiranean rocky shores, where it coexists with the con- generic species P. nislica L. and P. iilyssiponensis Gmelin. The Mediterranean sedentary P. cuemlca presents a planl 0.05) and was comparable to those obtained by Badino et al. ( 1986). who found F^, values lower than 0.01 in Thyrrenian and Adriatic P. caenilea populations (Fig. 1 ). This suggests that there is a high level of gene flow among different geographic areas and that P. cacndea popu- lation size is very large. Despite the sedentary adult stage supposed to favor genetic differentiation, the larval stage increases gene flow and could be responsible for the observed large-scale genetic ho- mogeneity. In contrast, single locus f „ analysis showed significant heterogeneity in SOD-l* (f,, = 0.021. P < 0.05) and highly sig- nificant heterogeneity in AAT* (f„ = 0.158. P < 0.001 ). Whereas SOD-l* heterogeneity was caused by two rare alleles (*90 and *I15) and could be related to geographical distance, the highly significant F,, value found in AAT* revealed an interesting case of microscale genetic heterogeneity. The AAT* locus occuiTed as a single allele (*100) in nalmal environments, but showed two alleles (*I00 and *115) in the samples taken from the two artificial basins (PE and TI2: Table 2). The *W0 allele was the same found by Sella et al. (1993), whereas Badino et al. ( 1986) did not analyze the AAT* locus. It is relevant that AAT* allelic frequencies were very similar between TI2 and PE samples (sites about 150 km apart), whereas genetic heteroge- neity was maintained even if the distance between samples did not exceed I km (Til -natural vs. TI2-artificial. Table 4). Hence, de- spite the large-scale genetic homogeneity observed for the major- ity of polymorphic loci analyzed, the AAT* locus showed a very highly localized heterogeneity. Significant heterogeneity among populations on a local scale, but little additional large-scale geo- graphical variation, have been observed in several marine gastro- pods molluscs (Johnson & Black 1984. Johnson et al. 1993. Hurst & Skibinski 1995). and different mechanisms such as mixing of larvae or post-settlement selection have been invoked. As we found a high level of homogeneity at the nuijority of polymorphic 1062 Mauro et al. TABLE 4. Pairwise Fst in .-lA 7" locus. TR AU SC CG Til CO UI GA TU T12 0.187** PE 0.111** 0.207** 0.128** 0.138** 0.066* 0.147** 0.070* 0.180** 0.120** 0.194** 0.117** 0.199** 0.121** 0.194* 0.117** 0.023 **P< 0.001: *P <0.05. loci, the reproductive isolation between coast and artificial envi- ronments as an explanation for the spatial differentiation in the AAT* locus can be rejected. The inconsistency among the loci analyzed indicates that the cause for local heterogeneity is not dependent on breeding systems, but is locus specific. As several authors suggest, allozymes do not always behave as neutral markers (Hilbish et al. 1982. Altukov 1990. Karl & Avise 1992. Riddoch 199.^. Johannesson et al. 1995, Mitton 1997) and even though we were not able to demonstrate the functional rela- tionship between allozyme variations and environmental param- eters of artificial basins, we think that the strong genetic hetero- geneity found in the AAV-'- locus may be related to the peculiar environmental conditions. The first artificial sample site was the channel collecting the discharge water used to cool down the Ter- mini Imerese thermoelectric plant (TI2). Salinity of the discharged water equals that of sea water {3il9(). whereas temperatures are slightly higher (l°C to 2°C) than tho,se measured in surrounding coastal waters. Hydrodynamic patterns are remarkably different as the site is sheltered from wave motion, though it is subjected to a continuous flow of discharge water. The second artificial site, located in an aquaculture plant in Petrosino (PE), is a tank col- lecting sea water to be directed to the breeding tanks. As result of the continuous exchange with the seawater, temperature and sa- linity do not differ much when compared to the natural sites, whereas the hydrodynamic patterns do, owing to the continuous fiow of incoming water. Even if several authors invoked tempera- ture (Powers et al. 1991, Sokolova & Portner 2001) and salinity (Koehn et al. 1980) as responsible for the particular genetic adap- tation found in marine organisms, we cannot demonstrate this tendency because in our case study, temperature and salinity of artificial sample sites are the saine of those of the sea. Microscale allozyme variation in the AAT* locus was also found in the marine snail Linoriiia sii.xatilis (Johannesson et al. 1995), and the authors concluded that a strong natural selection was acting. Therefore, it is probable that the AAT* locus was the specific target of complex environmental-genetic interactions. The //5* allele was probably maintained at low frequencies (below the limit of detection in a sample of 50-100 individuals) in natural environments by a mu- tation/selection balance, and it increases in frequency as a result of selection in artificial environments. Unfortunately, we do not know what kind of physiological consequences a locus could have. The AAT* enzyme is involved in the degradation process of nitro- gen-containing compounds and it is known that marine molluscs of the intertidal zone have developed different mechanisms of excre- tion of these compounds in relation to vertical zoning and to wave action exposure (Cognetti & Sara 1981). The microscale variation in the AAT* locus found in P. caendea and in L. sa.xalilis could be related to some of these adaptive processes. ACKNOWLEDGMENTS The authors wish to express their thanks to Dr. Marco Costan- tini (Riserva marina di Miraniare, Trieste) for the collection of Trieste samples. Research was supported by MURST (407^ and 60%). Altukhov, Y. 1990. Population Genetics: Diversity and Stability. .Amster- dam: Harwood Academic Publishers. Badino, G.. G. Sella & G. Celebrano. 19S6. Sistemi gene-en/inKi iiello studio delta differenziazione genetica di popolazioni naturali di mol- luschi a diversa vagilita. LavDri SIM. Atti Congr Palermo. 13-16 Sep- tember 1984. 22:241-258. CogneUi, G. & M. Sara. 1981. Biologia marina. Eil CiiUlcrini. Davis, B. J. 1964. Methods and applications to human serum protein. Ann. Acad. Sci. 121:404-428. Dodd, J. 1957. Artificial fertilization, larval development and metamor- phosis in PareUa vulgatci and PatcUa caernlai. Puhbl. Suiz. ZooL Mi/ot// 21:172-186. Goudet, J. 1995. FST AT. version 1.2. a computer program to calculate F statistics. J. Hered. 86:485^86. Guo, S. W. & E. A. Thompson. 1992. Pertbrming the exact test for Hardy- Weinberg proportion for multlipe alleles. Biometrics 48:361-372. Hilbish. T. J., L. E. Deaton cS: R. K. Koehn. 1982. Effect of an allozyme polymorphisn on regulation of cell volume. Nuuire 298:688-689. Hurst, C. D. & D. O. F. Skibinski. 1995. Comparison of allozyme and mitochondrial DNA spatial differentiation in the limpet PaieUa vid- auki. Mar. Biol. 122:257-263. Johannesson. K.. B. Johannesson & U. Lundgren. 1995. Strong natural LITERATURE CITED selection causes microscale allozyme variation in ;i marine snail. Proc. Nad. Acad Sci. USA 92:2602-2606. Johnson, M. S. & R. Black. 1984. The Wahlund effect and the geographical scale of variation in the intertidal lipel Siplionuiia sp. Mar. Biol. 79: 295-302. Johnson, M. S.. K. Holborn & R. Black. 1993. Fine-scale patchiness and genetic heterogeneity of recruits of the coralivorous gastropod Drupella cornus. Mar. Biol. 117:91-96. Karl, S. A. & J. C. Avise. 1992. Balancing selecUon at allozyme loci in oysters: implications from nuclear RFLPs. Science 256:100-102. Koehn. R. K.. R. I. E. Newell & F. Immemiann. 1980. Maintenance of an aniinopeptidase allele frequency by natural selection. Proc. Natl. Acad. Sci. USA 77:5385-5389. Manuins, Z., A. P. Apostolidis & C. Triantaphyllidis. 1998. Genetic pro- tein variation in red mullet {Miilliis barbatiis) and striped red mullet (M. siirmuletus) populations from the Mediterranean Sea. Mar. Biol. 130: 353-360. Mitton. J. B. 1997. Selection in Natural Populations. Oxford: Oxford Uni- versity Press. Pogson. G. H. & E. Zouros. 1994. Allozyme and RFLP heterozygosities as correlates of growth rate in the scallop Placupectcu maticllanicus-. a test of the associative overdominance hypothesis. Genetics 137:221-231. Powers, D. A., T. Lauerman, D. Crawford & L. Di Michele. 1991. Genetic Artificial Environmental Conditions Afkbct P. caervlea 1063 mechanisms for adapting Ui a changing ens ironment. Annu. Rev. Gcnel. 25:629-659. Raymond. M. & F. Rous.sel. 1995. GENEPOP (version 1.2): population genetic software lor exact te.sts and ecumenism. J, Hered. 86:248-249. Richardson. B. J.. P. R. Baverstock & M. Adams. 1986. Allo/yme elec- trophoresis. San Diego: Academic Press. Riddoch. B. J. 199.^. The adaptative significance of electrophoretic mobil- ity in phosphoglucose i.somerase (PGl). Biol. J. Linn. Sue. 50:1-17. Sella. G.. C. A. Robotti & V. Biglione. 1993. Genetic divergence among three sympatric species of MediteiTanean Patella ( Archaeogastropoda). Mar. Biol. 115:401^05. Shaklee. J. B., F. Allendorf. D. C. Morizot & G. S. Whitt. 1990. Gene nomenclature for protein-coding loci in fish. Trans. Ant. Fish. .Soc. 119:2-15. SokoloNa. I. M. & H. O, Former, in press. Temperature effects on key metabolic enzymes in Litlorina sa.xulilis and /,. ohnisata from different latitudes and shore levels. Mar. Biol. Volckaert, F. & E. Zouros. 1989. Allo/yme and physiological variation in the scallop Placopeelen maKcllanicns and a general model for the effects of heterozygosity on fitness in marine molluscs. Mar. Biol. 1(B:51-6I. Weir. B. S. & C. C, Cockerham. 1984. Estmiating F statistic for the analysis of population structure. Evolution 38:1 358- L'<70. Wright. S. 1978. Evolution and the Genetic of Populations. Vol 4: Vari- ability within and among Natural Populations. Chicago: University of Chicago Press. Zouros. E. & D. W. Foltz. 1984. Possible explanations of heterozygote deficiency in bivalve molluscs. Malaeolof^ia 25:583-591. ./(Hinuil cij Slh'llfi.Ui licscuirli. Vol. 20, No. 3. I()65-1()7(). 2001. EVALUATION OF MICROSATELLITE PRIMER CONSERVATION IN ABALONE BRAD EVANS.' - '* N. CONOD.' AND N. G. ELLIOTT' 'CSIRO Marine Research. GPO Box I5JIS Hohart. Tasmania 7001. Aiisiralia: \ScIuhiI of Zoology. University of Tasmania. GPO Box 252-05. Hohart. Tasmania 7001. Australia: "'Aqiuuultiire CRC Ltd. PO Box 123. Broadway. NSW 2007. Australia ABSTRACT This siudy investigated the interspecille aniphtication of 22 microsalelllte loci developed for HaliotL\ riihra across 12 other species within the Haliotidae. We reveal, through the discussion of three specific cases, a need for more thorough assays of microsatellite locus conservation, to ensure the utility of existing markers in foreign species. Optimization and analysis of PCR products revealed that of 12 loci examined in the Australian H. laevigata, only five were able to be reliably scored, while 6 of 10 for the South African H. midae were scoreable and none of the three for the North American species were useable. The assay examines five species from Australia, three from New Zealand, two from South Africa and two from North .America. Amplification success varied from 68% for H. conicopom, a possible sub-species, to 14% for the distantly separated and related species from North .America. The importance of cross-species amplification is discussed, but it is apparent that the proportion of markers that may be useful in other species of the same genera will vary greatly with the taxa being investigated. A'£'>' WORDS: Hulimis. abalone. microsatellites. aquaculliire. primer conservation INTRODUCTION Members of the genus Haliotis. commonly called abalone. are distributed in coastal waters of all continents. Many of the 37 recognized species (Geiger 2000) are harvested commercially or recreationally. and they are a highly valuable marine resource. Abalone populations, like those of most highly prized marine re- sources, have come under increased legal and illegal harvesting pressures in recent years as demand for the product continues to rise, and methods for their capture and distribution are refined. Along with the expanding effort within abalone fisheries world- wide, and in some cases the decline of those fisheries, there has been extensive development in the culture of many abalone species (Oakes & Ponte 1996. McBride 1998. Cook 1998). An important technological advancement that will benefit both the culture and wild harvest industries is the development and application of mo- lecular genetic markers. Molecular genetic markers are widely used in many seafood industries for both wild and aquaculture needs; for example, in Salnionids (Reilly et al. 1999) and oysters (McGoldrick & Hedge- cock 1996). They can be used for applications as diverse as track- ing the biological history of populations (Chambers & Mac- Avoy 2000). or as specific as detennining the parentage of individu- als in culture (O'Reilly et al. 1998). One such marker that has been applied in other genera (Dallimer 1999. Wuetal. 1999. Nesjeetal. 2()()()) but has only recently become popular in abalone research is microsatellite DNA. This marker consists of a nucleotide sequence of between two and six base pairs repeated in series at a set point on a chromosome (locus). The number of times that sequence is repeated at a single locus varies within (heterozygous individuals) and between individuals (intra-specific variation), and where that same locus is conserved across species, the number of repeats may vary widely (inter-specific variation) (Wright & Bentzen 1994). Microsatellite markers have been developed from partial geno- mic libraries of four Huliinis species: H. asinina (Selvamani et al. 2000), H. kamtshakana (Whithler. pers. comm.). H. riitira (Huang & Hanna 1998, Evans et al. 2000) and H. nifescens (Kirby et al. 1998). The development of microsatellite DNA markers, as de- *Corresponding author. E-mail: Brad. Evans(s'marine. csiro.au scribed in each of these papers, is a time consuming and expensive process (Wright & Bentzen 1994). For these reasons, the efficacy of markers between species within the same genus or family have been examined in both plant and animal groups with ambiguous results (Dowling et al. 1996. Huang & Hanna 1998). White and Powell (1997) tested 1 1 microsatellite markers de- veloped for the hardwood. Swietenia humilus for conservation within 1 1 members of the Meliaceae family, representing seven genera. They detailed four species specific, one genus specific and three family wide markers. This trend of good marker conservation within plant families is supported by other studies such as that by Thomas and Scott (1993) who found that primer sequence conser- vation existed among grapevine species, and more recently by Rosetto et al. (2000) who showed similar sequence conservation among members of the Myrtaceae family. Wirth et al. (1999) examined conservation of 1 1 microsatellite loci developed for the Walleye. Stizostedian vitrettm in four spe- cies representing two genera of the Percidae family. Three of the markers were conserved in all species tested, two were found to be specific to Slizostedion genus, four produced amplification from both genera, but not all species within them, and the remaining two markers amplified only in S. canadese. Primer sequences have also shown some conservation across 10 species of four genera of Lemur, endemic to Madagascar (Jekieiek & Strobek 1999). Huang and Hanna (1998) considered the cross-species ampli- fication of their three H. rubra microsatellite loci in species from USA (2 species). South Africa (2 species). South Korea (6 species) and Australia (5 species). Of the ten species tested from outside of Au.stralia. only two of the South Korean species produced any amplification product. Within Australian species the markers were more conserved, with at least two of the three loci producing an amplification product in all Australian species tested, except for Haliotis lacrif;ata. the greenlip abalone. which failed to amplify a product at any of the three loci. As H. riilva and H. laevigata are known to produce hybrids in the wild (Brown 1995) this latter result is unexpected, and required further examination as the hy- brid is being developed as an aquaculture product for which mo- lecular markers are required. In this article we describe the cross-species amplification of 21 microsatellite loci (22 primer pairs) developed for use in the Aus- 1065 1066 Evans et al. tialian Blacklip abalone. Haliolis rubra (Leach. 1814). Twelve species from Australia [5 species). New Zealand (3 species). South .Africa (2 species) and North America (2 species) were tested. These 12 species come from two discrete phylogenetic clusters within the genus based on sperm lysin DNA sequences (Lee & Vacquier 1995). We then expand upon this work by describing the optimization of some of these loci for genetic variation research in w ild and cultured South African Haliolis inidac. pedigree analysis and genetic variation research with the Australian H. laevigata. and genetic variation studies in the North American abalone. H. fulgens. MATERIALS AND METHODS Micrusalellile Amplifualion The twelve abalone species tested in this study were chosen to provide both close and distant evolutionary relationships to Hali- olis riilira. Four species (//. conicopora Peron. 1816; H. laeriiiata Donovan. 1808; H. roei Gray. 1826; H. scalaris Leach. 1814) share a temperate Australian habitat with H. rubra, while H. as- iiiiiia (Linnaeus. 1738) is a tropical species form Australian waters. Haliotis aiistralis (Gmelin, 1791), H. iris (Gmelin. 1791) and H. virgiiiea (Gmelin. 1791) are temperate species from neighboring New Zealand. The two most prevalent South African species H. inidae (linnaeus. 1758) and H. spadicea (Donovan. 1808) and two species of commercial importance in North America H. corrugata (Wood. 1828) and H. fulgens (Philipi. 1845) are examples of dis- tant temperate species. Twenty-two microsatellite primer pairs were developed for Haliotis rubra (Table 1) using methods described in Evans et al. (2000). Their potential for cross-species amplification was tested under standard PCR conditions with DNA extracted from gill or muscle tissue from at least two individuals of the 12 test species using a modified CTAB protocol (Grewe et al. 1993). Such low sample sizes are a problem for determination of diversity indices or population structure, but are sufficient for the detection of the locus in another species. TABLE L Characterization of 22 microsatellite primer pairs tested for cross-species amplification in this study. Previously published primers appear last, and a citation is given instead of full sequence. Locus Repeat Sequence Primer Sequence (5'-3') (F-Forward. R-Reverse) .\ccession Number -Approx. Size in H. rubra (bp) cmrHr].5 iCAGA), cmrHrl.6 (CA)4 (CA), cmrHr\.2?, (AC),, Ln,rHr23 (GTi.jTTtTG), cmrHr2.5 (GT),, *anrHr2.\5 (CA),, cmrHr2M (GT),s anrHr2.\V, (GAGTl, cmrHr2.2{) (AC),,(GCAC),s cinrHr2.22 (CA)„ aiirHr2.2y (AC),^ anrHr2.21 ( GT ) , ,( GCGT )„( GT), anrHr2.29 (CA),^ o»/7//i-li (AC) ,5 cmrHi\\-X (GT),3TT{GT),GA(GT), cmrHr\.2A (AT), cmrHr\.25 (CA),,( AT)6TT( AT),(TG), *cmrHr2.^) (GT),, cmrHr2A4 (GAGT)s (GAGT), anrHr2.2(y^ (ATTO.T^CtATTC), cmrHr2.30 (GT)s.,(GT),,(TG),,(AG), (TG), (TG),„ cmrHr236 (AC),, F-GGAAGAGGTATCGTAAACTG R-AGTCTCCCTGGTAAAACG F-GTTGTAAATGATGCCCTC R-CGTCTTTTTATTCAACGCC F-GCTGGGAA.\TCAATCTTC R-CCTCACTTTCAACACTCAC F-CCAGGCCCTATTCTTTCACA R-CGTCGCACTAAACACTGCAT F-GCGCAGACATTCATCGGATA R-GTCCATCGTCGACAGGTTTT F-TTTACATCGCATCGGCATTA R-TACTTAACGTTGCCCTGCCT F-AGGACTTGCCCAACCTTTTT R-TTACAGAACAAACACAAGTATTGAA F-GCTCCAGAATTCAAGGGTTG R-GCTGCTTAACCTCAGGATGC F-TTTTGAATGATTGTATTTCGTTTT R-TACCTTGCATCGTAATAAACAGACAC F-GGGTCGTCAGGTAGGTAGCA R-CCATAATCAGAGGGGAAGCA F-TGGAAGCTTTTCAAACATTGG R-TACA.AiTGGGGATTAAGAAGC F-GTCCAGGTCCACAGCTCATT R-GGAATTGAAGACCCTCCTCC F-TGATTGGTGTGTGAGGTGAAA R-CCGATGCCCTTATCATCACT Evans et al. 2000 Evans et al.. 2000 Evans et al.. 2(»0 Evans et al.. 2000 Evans et al.. 2000 Evans et al.. 2000 Evans et af. 2000 Evans et al.. 2000 Evans et af. 2000 AF .^02X24 126 AF 302X2-S 89 AF 302826 122 AF ,302827 100 AF I949.'i,'5 28.V299 AF l9-'i9.'i6 288 AF 302828 226 AF 302829 \}4 AF 302830 186 AF-W2831 117-193 AF .^02832 258-266 AF .102833 347 AF 3028.U 321 AF 194951 172-176 AF 195952 252-262 AF 195953 216-236 AF 1959.54 291-309 AF 195956 156-202 AF 195957 209-235 AF 195958 190-212 AF 195959 284-328 AF 195960 8.V121 The two primer pairs preceeded by the '"*" both amplify the same locus, with cmrHr2.9 being internal to ciiirHr2. MlCROSATELLlTE CONSERVATION IN ABALONE 1067 PCR reactions were performed in a volume of 25 (xl consisting of 67 niM TrisHCl, pH 8.8: 16.6 mM (NHjI^SOj: 0.45';i- Triton X-100: 0.2 mg/niL gelatin; 2.5 niM MgCK; 10 pmoles of each primer: 200 jiM dNTPs: 0.5 U Taq Fl polymerase (Fisher Bio- tech); and - 20 ng genomic DNA template. Amplification was in a Perkin Elmer 9600 thermocycler with one cycle of 94°C for 1 min. 50'C for 15 s and 72"C for 1 min. followed by 30 cycles of 94°C for 15 s. 50°C for 15 s and 72°C for 1 min. These cycles were followed by a final extension step of 72'"C for 10 min. Amplifi- cation products were visualized on 2% TBE agarose gels, and markers were scored as present when a single band of between 75 bp and 450 bp was detected (Table 2). Any amplification products above this size range, although possibly containing the microsat- ellite repeat unit, can not be scored on the AB1377 using the size standards commonly available. It is possible to score larger alleles with different systems, but such large products increase the pos- sibility of mutation in the regions flanking the microsalellite rather than within the repeat unit. In such cases it would be best to design new primers closer to the repeat unit. Amplification products less than 75 base pairs in size are difficult to reliably score due to their proximity to such PCR artifacts as primer-dimer. and the presence of unincorporated dyes. H. niidat Oplimizalion The loci that produced an amplification product for H. miclcie under standard conditions were further examined to optimize PCR parameters. Samples were initially subjected to PCR amplification at temperatures of 50°C and 55"C and with DNA template con- centrations of 10ng/|xl and 2ng/(a.l. Amplification products for all ten markers were not improved by increasing the annealing tem- perature from 50°C to 55°C. but were improved by increasing the DNA template concentration to 10ng/[jLl in all samples. The mark- ers were amplified in eight individuals of H. luidiie from Cape Hangklip on the southwest coast of South Africa. Products were diluted relative to amplification strength and mixed with formam- ide, loading dye and Genescan Tamra500 size standard (PE- Applied Biosystems). denatured at 95°C for 2 min and \.l\i\ loaded onto a Wc denaturing polyacrylamide gel. Samples were run on an ABI377 DNA autosequencer and genotypes determined using Genpotyper* software. Allele variation was scored between the eight individuals. H. laevigata and H. fulgens Optimization All loci that produced an amplification product in one of these two species under standard amplification conditions were sub- jected to further testing for optimization. This included a range of annealing temperatures from 48°C to 58°C. "Touchdown-PCR" to iinprove primer specificity, where the annealing temperature at the beginning of the cycling prograin was high and was lowered by either 0.5 or 1.0°C each cycle until the lowest selected annealing temperature was reached. In addition. DNA template concentra- tions tested ranged from Ing/|JL1 to 30ng/p.l. and MgCU concen- trations tested ranged from ImM to 5mM. All loci were tested on at least 30 H. iaeyigata or 8 H. fulgens individuals. RESULTS Micrusalcllitc Amplifuation Nineteen of the twenty-two primer pairs tested successfully amplified a product in at least one species other than H. nihia (Table 2). Not surprisingly, the species that appears to have re- tained the ino.st loci, at 15. is H. conicopora. a species that has been touted as perhaps a sub-species of H. nihra (Geiger. 20001. The three other temperate .Australian species (W. laevigata. H. .ualaris. H. roei) produced an amplification product from twelve of the 22 primer pairs. Halioiis asinina. the only tropical species included in the study showed sequence conservation in only five ot the markers tested. The three species from New Zealand showed conservation of 5, 5. and 9 markers for H. iris. H. aiistralis and H. virgiiiea respectively. As expected there was little cross-species amplification seen in the North American species H. ntfescetis and H. fulgens (three markers each), which were shown to be in a distant clade to H. rubra by Lee and Vacquier (1995). It should be noted that the three loci producing an amplification product in the two North American species were the only loci to amplify a prod- uct in all 12 species tested. Interestingly though, the South African species showed more sequence conservation than the species from New Zealand, with ten markers being conserved in H. midae and nine in H. spadiceae. None of the primer pairs was shown to be specific to all Australian species or to particular climatic regimes such as temperate and tropical species. Some primer pairs pro- duced an amplification product that was dramatically different in size to that expected. Where that product was greater than 450 or less than 75 base pairs the marker was denoted by an "a" for altered product size. Although these altered products may contain the same microsatellite as other amplification products, they can be of no use if they can't be scored reliably. Previously, researchers have simply reported agarose gel de- tection as the retention of a locus in a related species. In this article, we attempt to bridge the gap between identifying the pres- ence of a marker in a related species, and the use of that marker for further research. We present here, three case studies, in which we have taken the markers identified from the preliminary screening tests and attempted to optimize them for routine research in Hali- otis midae. H. laevigata, and H. fulgens. Haliotis midae — the South African Ahalime The 10 loci shown to produce an amplification product from H. midae in the preliminary screening were: cmrHrl.li. cmrHr\. 2-i. cmrHr2.9. cmrHrl.XS. emrHrl.lO. cmrHr2.2?>. cmrHr2.21. an- rHr2.29, cmrHr2.?0. cmrHr2.?>b (Table 2). At the initial PCR amplification conditions, both cmrHr2.21 and cmrHr2.30 pro- duced non-specific products when examined using the more sen- sitive automated detection techniques. These loci were re- amplified at annealing temperatures of between 50C and 58C and with [MgCl,] of between 2.0 and 3.0 mM. The resultant amplifi- cation products however were also non-specific and these markers were not examined further. Two of the markers were monomorphic in the eight individuals examined and were not tested further in this study. These markers were cinrHrl. 24 and cnuHr2.27. It is unlikely that these markers are actually the same locus as that amplified in H. rubra, as they were somewhat smaller than the product produced in that species, and did not show characteristic microsatellite amplification pro- files. The variation at each of the remaining 6 loci ranged from the minimum of 2 alleles at cmrHr2.30 (226-242 bp). 3 alleles at cmrHr2.2?< (244-252 bp). 4 alleles at emrHr2.3b (101-1 19 bp). 5 alleles at (:v)!;-Wr2.15 (2.'i()-280 bp). 6 alleles at cmrHr2.9 (173-205 bp) and a mavuiium of S alleles at idintitled as conserved in H. laeviijaUi, and reasons lor their exclusion Irom further research. Locus Accession numbers Repeal Sequence Reasons for exclusion cmrHi'\.b cmrHil.\4 cmrHil. 24 cmrHil.i cmrHr2.\4 cmiHrlM cmrHrl.lO cmrHrl.12 cmrHr2.23> anrHr2.21 cmrH 1-2. 29 cmrHr2.iOd AF 302828 AF 194952 AF 194953 AF 302827 AF 195957 AF 302828 AF 302830 AF 302831 AF 302832 AF 302833 AF 302834 AF 195959 (CAlj (CAI, (GT),,TT(GT),GA(GT), (AT)^ (GT)|4TT(TG), (GAGT)s.,(GAGT), {GT)_„ (AC),3{GCAC),« (CA)„ (AC),^ (GT),7(GCGT),,(GT), (CA)5s (GT)„ (GTl,,(TGl,,(AG) (Tl Monomorhic in H. laevigata Non-specific Monomorhic in H. laevigata Non-specific Suitable for research Non-specific Unreliable amplification Unscoreable-very messy peaks Suitable for research Unreliable aniplificaiion Unreliable amplification Suitable for research Sample size for testing was n = 12. ciurHr23Q. Two of these three markers have been used to deter- mine broodstock contributors in controlled spawning of H. laevi- gatalH. yiihni hybrids at a commercial culture facility (unpub- lished data). Locus awHrl.li was not useful in the hybrid study however, as although it was easy to score in H. laevigata, a third allele of equal intensity to the first two was detected in some H. rubra samples, and in many of the hybrid progeny. It should also be noted that the optimal conditions for PCR amplification would vary dramatically with different thermal cy- clers. This was exemplified in the transfer of 6 H. rubra micro- satellite loci that produced clean PCR products in H. midae in our study, but required extensive re-optimization when used in a ge- netic variation study in Cape Town, South Africa (E\ans et al. in prep. I. The most obvious reason for this discrepancy was the large variation in ramp times between the respective PCR machines. For this reason any attempt to transfer molecular marker technology between laboratories, and particularly between species will require additional PCR optimization at the new site. The substitution of specified reagents with those that are cheaper or more readily available may also affect amplification (Evans et al. in prep.). One thing that is often overlooked when testing microsatellite primers is the design of the primer sites. The failure of a particular locus to amplify in another species may not mean that the micro- satellite repeat is not present in that species, but simply that one or both of the primer sites have not been conserved. As the majority of microsatellite primers are published as part of a larger sequence on the Genbank (NCBI) database, the option of primer re-design is available. In this study, we have designed two pairs of primers for the cn}rHr2.9 clone, with the cmrHrl.9 primers being internal to those of ciurHr2.\5. What we have seen in this case is that the external primers {cnirHr2.\5) were conserved in the South African species, H. spadiceae. while the internal primers {cmrHx2.9) were not (Table 2). It could be argued therefore that the examination of published sequences, and if necessary the re-design of primer sites, could be a more affordable solution to marker development than the creation of a new microsatellite library. This study shows that microsatellite loci isolated from Austra- lian blacklip abalone, H. rubra, can be amplified in some related Haliotis species, but that the likelihood of marker conservation is reduced with increasing phylogenetic distance. Rosetto et al. (1999) suggest that the selection of a single species from a large genera for microsatellite locus development will result in a suite of markers for most laxa in that genus. They detail only minimal PCR optimization tor the transfer of markers between species of the Melaleuca genus. Scribner et al. (1996) provide examples of high levels of marker conservation in species ranging from whales to rodents to suppoil their results in salmon and trout from Alaska, North America and the United Kingdom. What we have seen here, and in the previous study of abalone niicrosatellites (Huang & Hanna 1998) however, is a much lower rate of marker conserva- tion between Halious species. This finding together with the very high levels of polymorphism encountered in most abalone species (H. midae, Evans et al. in prep. H. asinina. Selvamani et al. 2000; H. rubra. Evans et al. 2000), may point towards a more rapid mutation rate of microsatellite repeats in abalone than that seen in other organisms. Research on the cross-amplification of microsatellite loci within taxa should ensure that the markers are useful within that species, and do not simply produce an amplification product. Our results have clearly shown that the presence of a similar sized PCR product on agarose gels is not sufficient to report locus conserva- tion in another species. Future studies should therefore endeavor to test those products further by radioactive or fluorescent labeling methods similar to those that would be used in larger studies. This would ensure that only those markers that are likely to provide reliable genetic information would be considered by those com- mencing projects on these species. ACKNOWLEDGMENTS We thank Boze Hancock of Fisheries Western Australia for samples of H. roei and H. Conicopora. Liz O" Brien of the Uni- versity of Queensland for samples of H. asinina. Dr. Neville Sweijd of the University of Cape Town, South Africa for samples of H. midae and H. spadieea. Phil Critchlow for samples of H. scalaris. Dr. Miguel Angel Del Rio of CICESE in Ensenada. Mexico for samples of H. fulgens and H. corrugata. and Dr. Rod- ney Roberts of the Cawthron Institute. New Zealand for samples of H iris. H. rirginea and H. auslralis. We would also like to thank Dr. Bob Ward and Dr. John Benzie for helpful comments on the manuscript. 1070 Evans et al. LITERATURE CITED Brown, L. D. 1995. Genetic evidence for hybridisation between Hiiliohs nibni and H. laevigala. Marine Biology. 123:89-93. Chambers. G. K. & E. S. MacAvoy. 2000. Microsatellites: concensus and controversy. Cominirative Biocliemistiy and Physiology Pari B. 126: 455-^76. Cook, P. 1998. The current status of abalone famiuig in .South Africa. J. Shelljish Res. 17:601-602. Dallimer. M. 1999. Cross-species amplitlcation success of Avian micro- satellites in the red-billed quelea Qiielea quelea. Molecular Ecology. 8:695-697. Evans. B.. R. W. G. White & N. G. Elliott. 2000. Characten/ation of microsatellite loci in the Australian Blacklip abalone iHalimis rubra. Leach). Molecular Ecology. 9:1 1-83. Evans, B.. N. A. Sweijd, R. W. G. White. P. Cook & N. G. Elliott. (In prep.) Microsatellite DNA examination of stock structure in the South African abalone, Haliotis miclae. Ezenwa. V. O., J. M. Peters. Y. Zhu. E. Arevalo, M. D. Hastings. P. Seppa, J. S. Pederson. F. Zacchi, D. C. Queller & J. E. Strassmann. 199S. Ancient conservation of trinucleotide inicrosatellite loci in Polistine wasps. Molecular Phylogenetics and Evolution. 10:168-177. Geiger, D. L. 2000. Distribution and biogeography of the Recent Hali- otidae (Gastropoda: Vetigastropoda) worldwide. Bolletrino Malaco- logico. 35:57-120. Grewe P. M., C. C. Krueger. C. F. Aquadro. E. Bermingham. H. L. Kincaid & B. May. 1993. Mitochondrial DNA variation among lake trout ISalvelinus naniaycusli) strains stocked into Lake Ontario. Canailiiui Journal of Fisheries and Aquatic Sciences. 50:2397-2403 Huang, B. & P. J. Hanna. 1998. Identification of three polymorphic mic- rosatellite loci in blacklip abalone. Haliotis rubra (Leach), and detec- tion in other abalone species. J. Shellfish Res. 17:795-799. Isagi. Y., T. Kanazashi. W. Suzuki. H. Tanaka & T. Abe. 1999. Polymor- phic microsatellite DNA markers for Magnolia obovata Thunb. and their utility in related species. Molecular Ecology. 8:698-700. Jekielek, J. & C. Strobeck. 1999. Characterization of polymorphic brown lemur (Eulenmr fidvus) microsatellite loci and their amplification in the family Lemuridae. Molecular Ecology. 8:901-903. Kirbv. V. L.. R. Villa & D. A. Powers. 1998. Identification of microsat- ellites in the Californian red abalone. Haliotis rufescens. J. Shellfish Res. 17:801-804. Lee, Y. H. & V. D. Vacquier. 1995. Evolution and systematics in Hali- otidae (Mollusca:Gastropoda): Inferences from DNA sequences of sperm lysin. Marine Biology. 124:267-278. McBride. S. 1998. Current status of abalone aquaculture in the Californias. 7, Shell fi'^h Res. 17:593-600. McGoldrick, D. J. & D. Hedgecock. 1996. Microsatellite marker develop- ment in the pacific oyster (Crassostrea gigas): Variability transmis- sion, linkage and QTL mapping. ./. Shellfish Res. 15:512. Nesje. M., K. H. Roed, J. T. Lifjeld. Lindberg. P. & O. F. Steen. 2000. Genetic relationships in the peregrine falcon {Faico peregrinus) analy- sed by microsatellite DNA markers. Molecular Ecology 9: 53-60. Oakes. F. R. & R. D. Ponte. 1996. The abalone market: Opportunities for cultured abalone. .Acjuaculture 140:187-195. O'Reilly. P. T.. C. Herbinger & J. M. Wright. 1998. Analysis of parentage determination in Atlantic salmon iSalwo satar] using microsatellites. Aninud Genetics. 29:363-370. Picos-Garcia. C. F. L. Garcia-Carreno & E. Serviere-Zaragoza. 2000. Di- gestive Proteases in juvenile Mexican green abalone. Haliotis fiilgens. .\quaculture 181:157-170. Reilly A., N. G. Elliott, P. M. Grewe. C. Clabby. R. Powell & R. D. Ward. 1999. Genetic differentiation between Tasmanian cultured Atlantic salmon (Salino salar L.) and their ancestral Canadian population: com- parison of microsatellite DNA and allozyme and mitochondrial DNA variation. Aquaculture 173:457-467. Rossetto, M., F. C. L. Harriss, A. McLauchlan. R. J. Henry. P. R. Baver- stock & L. S. Lee. 2000. Interspecific amplification of tea tree {Mela- leuca alternifidia-MyniKcae] microsatellite loci-potential implications for conservation studies. Australian Journal of Botany. 48:367-373. Selvamani. M. J. P.. S. M. Degnan. D. Paetkau & B. M. Degnan. 2000. Highly polymorphic microsatellite loci in the Heron Reef population of the tropical abalone Haliotis asinina. Molecular Ecology. 9:1184- 1185. Thomas. M. R. & N. S. Scott. 1993. Microsatellite repeats in grapevine reveal DNA polymorphisms when analysed as sequence tagged sites (STSs). Theoretical and Applied Genetics. 86:985-990. White. G. & W. Powell. 1997. Cross-species amplification of SSR loci in the Meliaceae family. Molecular Ec<}logy. 6:1 195-1 197. Wirlh. T.. B. Saint-Laurent & L. Bernatchez. 1999. Isolation and charac- terization of microsatellite loci in the walleye iStizosiedion vitreiim). and cross-species amplification within the family Percidae. Molecular Ecology. 8:1960-1962 Wright. J. M. & P. Bentzen. 1994. Microsatellites: genetic markers for the future. Reviews in Fish Biology and Fisheries. 4:384-388. Wu. L.. L. Kaufman iS: P. A. Fuerst. 1999. Isolation of microsatellite markers in .Asatoreochromis alluaudi and their cross-species amplifi- cations in other African cichlids. Molecular Ecology. 8:895-897. Jininial of Shellfish Risvanh. Vol. 20. No. 3. 1071-1075. 2001. INDUCTION OF TRIPLOIDY IN PACIFIC RED ABALONE (HALIOTIS RUFESCENS) ROSALIO MALDONADO.' ANA M. IBARRA,' * JOSE L. RAMIRP:Z.' SUSANA AVILA,' J. ENRIQUE VAZQUEZ,- AND LETICIA M. BADILLO" ^ Aqiiaciiltiiral Genetics Labonitory. Ceiitm dc Investigaciones Biologicas del Nuroeste, S.C. A. P. 128. La Paz B.C.S. Mexico 23000: -Abulanes Cuhivados. S. de R. L dc C.V. Calle Doce # 23S Fondepoit El Sauzal Ensenada B.C. Mexico 22760. ABSTRACT Induction oftriploidy was evaluated in red abalone. Halintis nifescens. using cylochalasin-B (CB). Three experiments v\ere done at dilferenl times of the year and with different CB concentrations, but triploids were produced only in May and Noveinber. The largest percentages of triploids were obtained in the November induction (85%, 100%, 100%, for 0.5, 0.75, 1.0 ing/L CB, respectively), and poorer triploidy success was obtained in the May induction (3%, 1 1%, 26%, for 0.3. 0.5, 0.7 mg/L CB). When the induction was done in November CB concentration did not play a significant role in success of triploidy, although it did play a role on relative survival to trochophore, with the two highest concentrations resulting in lower survivals. For the May and September inductions, treatment with CB had a significant effect on survival, but this was independetit of CB concentration. The results on triploidy success at each time of induction paralleled those for larval viability at each of those times of the year when evaluated in untreated control larvae, with the largest survival to trochophore (55%) and the lowest coefficient of variation {CV = 12%) for survival of untreated larvae produced in November. Untreated larvae produced in May had an intermediate survival (37%) and CV (57%). whereas that produced in September had the lowest survival to trochophore (1%) and the largest CV (143%). KEY WORDS: red-abalone, HaHotis mfl'scens. triploid. egg quality, larval \ lability INTRODUCTION Pacific red abalone {HaHotis nifescens) is fotiiid from southern Oregon-northern California in the USA to the north of the Baja California Peninsula in Me.\ico (Cox 1962, Lindberg 1992). Aqua- culture production of this species began during the last decade in Mexico, where the whole life cycle is managed under controlled condition.s. The growth rate in this abalone is slow, reaching 1 20 mm to 160 mm in three to four years (Tegner & Butler 198.5), and ways to improve production need to be defined. Triploidy, among other methods, has been proposed to improve production of aba- lone (Fujino 1992, Elliott 2000). Triploidy has been induced in a wide variety of marine mol- lusks (Beaumont & Fairbrother 1991), with the pioneer work hav- ing been on oysters (Stanley et al. 1981 ). For most species, induc- tion of triploidy in mollusks is done by inhibiting the extrusion of the second polar body during the conclusion of meiosis in fertil- ized eggs. Several induction methods have been tested; chemical treatment with cytochalasin-B (CB) or 6-dimethylaminopurine (6- DMAP). and physical treatment with hydrostatic pressure or ther- mal shock (cold or heat). In abalone species, Arai et al. ( 1986) and Kudo et al. ( 1991 ) were the first to report successful production of triploid abalone (HaHotis discus liannai and HaHotis diversicolor diversicolor) using thermal shock. Curatolo and Wilkins (1995) produced triploids of HaHotis discus liannai using hydrostatic pressure. Recently. Zhang et al. ( 1998) produced triploids of that same species using 6-DMAP. Yang et al. (1998a). Yang et al. (1998b) produced triploids of HaHotis diversicolor using thermal shock and cytochalasin-B. and Stepto and Cook (1998) produced triploids of HaHotis midac using cytochalasin-B. Powers et al. ( 1996) mentioned their success in producing and growing triploids of Pacific red abalone. HaHotis rnfescens. although no details are given in the abstract. In this study we report on the effects of induction to triploidy using different CB concentrations in east Pacific red abalone. HaH- otis nifescens. during three times of the year. *CoiTesponding author. Fax: +52-612-125-3625; E-mail; aibaiTaCscibnor.mx M.ATERIAUS AND METHODS Three experimental inductions to triploidy were done at differ- ent times during the year 2000 in a private hatchery; late May. early September, and late November. Spatininf; Spaw ning and gamete collection procedures were the same for all experimental inductions. Mature spawners (10-15 cm shell length) of red abalone. HaHotis rufescens. were selected from the broodstock kept in the commercial fartii 'Abulones Cultivados" at Erendira. Baja California. Mexico, and used to obtain eggs for the treatments. Spawning was induced following methods developed by Morse et al. (1977), by adding hydrogen peroxide dissolved (final concentration O.OlVc) in seawater for a period of approxi- mately two hours, or until spawning began. Gamete Collection Four sets or blocks were used for each of the three experimental induction times. Because the number of eggs shed by red abalone is not large enough to be able to divide the eggs from one female into different CB treatment concentrations, each set was composed of a mixture of eggs from .3-5 females. The sets were made se- quentially, that is. eggs of the first three to tlve females that spawned were mixed to make the first set. and the other sets were made sequentially, each of them with 3-5 spawners. Eggs were fertilized by adding abiiut 30 sperm per egg. Triploid Induction For each of the three experimental induction times, each set was divided into the same four treatment groups (one untreated control, and three CB treatment concentrations). For the May in- duction, the CB concentrations tested were 0.3, 0.5. and 0.7 mg L '; for the September induction 0.6. 0.8. and 1.0 mg of CB L ' were tested; and for the November inductions the CB concentra- tions tested were 0.5. 0.75. and 1 mg L"'. During the May induc- tion. cvlochalasin-B was added at the time at which the eiiizs in the 1071 1072 Maldonado et al. control group showed the beginning of the extrusion of the second polar body ( PB2j and ended when 50% of the control group eggs showed the second polar body. On average, the treatment lasted 20 minutes. Water temperature during the induction was 15°C. Egg density for each treatment was 2 x KF L"'. For the inductions in September and November. CB was added when 50% of the eggs in control group showed the first polar body (PHI), and ended when 50%' had the second polar body. For the September induc- tion, treatment lasted from 19 to 55 minutes, depending on when 50% of the eggs in each set showed extrusion of PB2. Water temperature during treatments was 20°C. and egg density for all treatments was 1.2 x 10^ L"'. During the November induction, treatment lasted between 32 to 48 minutes, depending on the set. Water temperature during the induction was I4.5°C. and egg den- sity for the treatments was 2.3 x 10'' L*'. For all three experimen- tal induction times, after treatment with the different concentra- tions of CB. eggs were rinsed in DMSO ( I niL L"' ) for 1 5 minutes to remove any remaining CB (Allen et al. 1984). Absolute and relative survival from egg to trochophore larvae (about 20 hours post fertilization) and absolute and relative sur- vival from egg to postlarvae (about 120 hours post fertilization) were estimated for each treatment group and control for all ex- perimental inductions. Relative survivals were estimated within each set and induction time as the ratio between the survival of the treated group to the survival in its control group, and expressed in percentages. Ploidy success was estimated by flow cytometry (Ploidy Ana- lyzer II. PARTEC Germany) of postlarval stages (approximately 200 / treatment / block), with the exception of the experimental induction in September, for which trochophore larvae (for those sets with surviving larvae) were analyzed. The flow cytometry methodology was that originally developed by Allen ( 1983). with modifications for larvae in Allen & Bushek ( 1992). In short, larvae were concentrated in a 1-ml suspension, centrifuged, and the pellet was stained with 0.5 ml of a DAPI/detergent/DMSO solution (Allen & Bushek 1992). re-suspending by vortex. Trochophore and postlarvae were disaggregated as in Allen and Bushek ( 1992); trochophore larvae by repeated aspiration with a 1 ml syringe, and postlarvae were disaggregated by means of crushing the meat/shell with a glass rod. In both cases, the obtained cell suspensions were passed through a 3()-|j.m screen, adding 1.5 ml of the DAPI/ detergent/DMSO solution before flow cytometry assay. luirval Ciillure After treatment, the embryos were kept in 100-L tanks until hatching. No feeding, aereation. or water flow was pro\'ided during this period (20 h). After hatching, swimming trochophore larvae was transferred to 1 50-L tanks at a stocking density of 4 larvae / ml. providing continuous flow. Trochophore larvae were induced to settle after 6 days following Morse et al. (1979) methodology, modified by Searcy-Bernal and Anguiano-Beltran (1998) by add- ing GABA (gamma-aminobutyric acid) at a final concentration of 1.6 |j,M. After settling, po.stlarvae was grown in 200-L circular tanks (1 12 cm diameter, 28 cm height) by placing 30,000 postlar- vae / tank, feeding by inoculating each tank with 3000 cells/cm" of the benthonic microalgae Navkiiki inccrlu. Statistical Analyses For each experimental induction, the effect of CB concentration on mean larval survival, and mean triploid success (n = 4), was analyzed using a randomized block design analysis of variance (Neter et al. 1985). Set (block) was considered a random effect. Percentage data were transformed to arcsine for analysis (Zar 1999). Significance was set as at P < 0.05 for all analyses. Tukey pairwise comparisons were used to test for differences between means (Neter et al. 1985). To establish whether there were differences in larval viability between dates of spawning regardless of CB treatments, the effect of experimental induction time on absolute sm^vival of the control, untreated groups, was analyzed with the same model above. RESULTS The results on sur\ i\al to trochophore and to postlarvae. and on success in triploid production for each experimental induction time are in Table 1. Triploidy Triploid abalone were successfully produced only during the May and November inductions. During the September induction, no triploids were detected in any of the tested concentrations. For the May induction, the highest concentration (0.7 mg CB L"') gave the best success in triploid production (26%;), whereas the lowest concentration (0.3 mg CB L~') resulted in the lowest suc- cess in triploids pioduced (3%) which was not significantly dif- ferent (P > 0.05) from the control (0% triploids). The intermediate concentration (0.5 mg CB L"') resulted in 11%- triploids, which was not significantly different from the results using the lowest or the highest concentrations (Table I A). There were no triploids detected in the control groups of any of the sets. For the November induction, all tested concentrations resulted in high success in triploid induction (Table IC), all being significantly different from the control group, but not from each other. Only one of the four sets in the 0.5 mg L*' treatment had less (40%) than 100% trip- loids. Success in triploidy ranged from 85% for the lowest con- centration (0.5 mg CB L"' ) to 100% for the intermediate (0.75 mg CB L"') and highest concentration ( 1 mg CB L"') tested. Suniral Treated and Control Groups — Within Experimental Induction Times For the May induction, there were no significant differences in absolute survival to trochophore or postlarvae between the treated groups and the control (Table lA). However, there were signifi- cant differences in relative survival to trochophore (P < 0.05). with the control group having a survival larger than both the 0.3 and 0.5 mg CB L"' treated groups, but not than the 0.7 mg L"' treated group. There were no significant differences between treated and control group for relative survival to postlarvae. For the September induction, absolute survival to trochophore larvae was not different between concentrations or the control, but relative survival was significantly larger (P < 0.05) for the control group than for any of the treated groups, which did not differ significantly from each other. There were no surviving postlarvae in any group (Table IB). For the November induction, there were no differences in ab- solute survival to trochophore larvae between control and CB treated groups, but there were differences (P < 0.05) in absolute survival to postlarvae, with the control group showing the largest survival, and the CB treated groups not being different from each Triploha' in Pacific Riio Abalone 1073 TABLE 1. Absolute and rtlativt mean survival I" tnicliophore larvae and postlarvae treated «ith different concentrations of CB, and percent of triploidy success in Pacific red abalone. Haliotis rufesccns. induced lo Iriploidv in (A) May, (B) September, and (C) November, n = 4 for all induction times. Absolute Relative CB .Survival to Sui ivival to Survival to Survival to Percent Treatment Trocbophorc (%) Posllarvae (%) Trocbophorc ( '7r I Posllarvae (%) Triploidy (A) Control 37" 10-' 100" 100" 0" 0.3 mg L-' 20' 5-" 56" 54" 3" 0.5 1112 L-' 2I-' 5" 58" 50" 1,.,,, 0.7 mg L^' 23-' 2" 63'" 24" 26" (B) Ciuitrol 0.%' 0 100° 0 0" 0.6 111" L-' 0.44-' 0 20" 0 0" 0.8 mg L-' 0.46' 0 34" 0 0" 1.0 mg L-' 0.30' 0 10" 0 0" (C) Control 55" 46" 100" 100° 0" 0.5 mg L-' 48^ 2,b 84"" 47" 85" 0.75 mg L"' 31'' 12" 56" 24'' 100" 1.0 mgL-' 32'' 13" 6(1" 27" 100" Within columns and expcnnicntal induction time. dilTerent letters arc used when means are significantly dilferent (ANOVA. P < 0.05). Other. Relative survival to trochophore larvae and to posllarvae was also different from the control group. The lowest relative survival to trochophore was seen for the two highest tested CB concentrations (0.75 and I mg L"'). whereas the lowest CB con- centration did not show significant differences with the control or the other tested CB concentrations. All treated groups had signifi- cantly lower relative survival to postlarvae than the control group (Table IC). Control Groups — Behteen Experiiiiental Induction Times Survival of trochophore and postlarvae in the control groups was .significantly different (P < 0.05) between experimental induc- tion times (Table 2). The highest survival was seen for larvae produced in November (559^ to trochophore larvae, 46% to post- larvae), followed by larvae produced in May (37% to trochophore larvae, 10% to postlarvae). For the September spawn, there was only 1% survival to trochophore larvae, and no surviving postlar- vae. The coefficients of variation for mean survival was high for the September induction, followed by May. and with the lowest CV the survivals in November. TABLE 2. .Absolute means survival Cr ) and coefficient of variation (CVl from ffifi lo Irochopbore and from e^f! to posllarvae of Ibe control untreated groups, for each of the experimental inductions. Induction Survival Mean % (CV) time Egg to Trochophore Survival Mean % (CV) Egg to Postlarvae May 36.7" (57) September 1.0" (143 November 55.3^12) y.6'(62) 46.0" (28) Different letters within columns indicate significant differences between induction times. DISCUSSION Our results show that triploidy can be induced in red abalone, Haliotis rufescens, using cytochalasin-B (CB) to inhibit release of the second polar body, and that factors other than CB concentra- tion can have an effect on success of triploid induction as evi- denced by the different results obtained between May and Novem- ber in spite of having used two equal CB concentrations. The cause for the variable success rate in triploids produced between the three experimental induction times is not known, but three possible causes might have been influencing the results, either indepen- dently or combined: different teinperatures during the spawning and inductions, differences in the CB concentrations and duration of the inductions, and differences in egg quality of abalones used in each of those months. During the three induction times the temperature was different only for the September induction, when a unusually high tempera- ture of 20°C was recorded and a low survival to trochophore with no triploids produced, and no surviving postlarvae were observed. The effect of increased water temperature on larval viability during the spawning for this species is not known, and no inferences can be drawn with regard to this factor alone. Furthermore, in May and November the results on triploidy success were dilTerent between both inductions, but water temperature was about the same for both inductions. I4.5°C and 15°C. Whereas it is known that at least in oysters meiotic rates are increased as temperature increases (Downing & Allen 1987. Eudeline et al.. 2()0()b). in our study meiotic rates were actually slower during the September induction (40 min for 50% PBI ) than for the November induction (27 min for 50% PBI) in spite of the higher water temperature occurring during the September induction. This points toward another factor rather than temperature on itself being the cause of no success in triploid production during September. The second possible cause. CB concentration and duration of induction, can not explain com- pletclv the difference between induction times in September and 1074 Maldonado et al. Ndvemher becuuse similar concentrations were tested, and in both cases the biological criteria defined by Allen and Bushek (U)y2) were used for the inductions; starting when 50% of the eggs showed the first polar body (PB). and ending when 50% showed the second PB. When the May and November induction results are compared, they indicated that CB concentration was important for the percentage of triploids produced only during the May induction but not during the November induction. However, this might be explained by differences in the duration of the induction between May and November, because induction in May was not started until the appearance in the first eggs of the extrusion of the second polar body, whereas in November it was started before, when 50% of the eggs had the first polar body. Given the shorter treatment duration in May resulting from this (20 min) when compared with the other induction times (>30 min), increasing CB concentrations might have been important in inducing a larger percentage of triploids during the short induction time in May. Supan et al. (2000) found that whereas increasing CB concentration was im- portant for improved success in triploidy induction for the Ameri- can oyster. Cnissostrea viri^inica. a longer treatment time might have resulted in greater triploidy success. They used a fixed treat- ment time (10 min). which as in our case for the May induction, did not follow completely the biological criteria recommended by Allen and Bushek (1992) for triploid induction. The third possible cause for the differences in triploidy success between induction times could have been differences in reproduc- tive condition of the abalones between May. September, and No- vember. There are few studies on the reproductive cycle of red abalone native to the study area. From the two found it is known that the reproductive cycle of red abalone in northern Baja Cali- fornia is active from May to December, reaching maturity around September (Ortiz-Quintanilla et al. 1990). with the spawning peak occurring between October and December (Molina-Martinez 1983). If during the year of the experiment the reproductive cycle followed that described, mature individuals should have been available at all three induction times, and in fact, they were. How- ever, an observation during the induction times was that, even if available, mature spawners were more difficult to find at the farm during the May induction, and still more difficult to find in Sep- tember than in November, when the largest success was seen in triploid production. Then, whereas the variable abundance of ma- ture spawners appears to indicate that the egg quality was variable between those induction times, biochemical determination of egg quality was not done. However, alternative or indirect measure- ments of egg quality that do not involve biochemical determina- tions of reserves have been previously used in association with evaluations of triploid success. For example. Aldridge et al. ( 1990) used survival in the control groups as a rneasurement of egg qual- ity, finding an as,sociation between survival in the controls and success in triploid induction in bighead carp. They proposed that egg quality, measured as larval viability in the control groups, could provide a good indicator of success in triploid production, hi the present study survival in the control groups indicated also an association with triploidy success, with the highest survivals to trochophore larvae having seen during November, when the largest success to triploidy was obtained, and the lowest triploidy success was observed when the lowest survival occurred in the control groups. Furthermore, the coefficients of variation (CV) for mean survival to trochophore larvae in the control groups for each in- duction time, with the lowest CV seen for the November data, and the highest CV for the September data, are also indicative of large variability in survival being associated to low success in triploid production and low variability in survival with high success. The importance of evaluating survival in the control groups rather than in the treated groups as a potential indicator of success in triploidy is demonstrated by the fact that Allen and Bushek ( 1992) did not find a correlation between sur\ ival of treated batches or families of oyster eggs and success in triploidy. nor did Ruiz-Verdugo et al. (2001 ) when working with scallop eggs. However, the last authors found that the correlation between number of eggs and number of D-larvae in untreated groups was highly significant, whereas in the CB treated groups that correlation was lost, indicating that sur\ i\ al in treated groups is affected by factors other than egg quality, as for example, toxicity of the CB treatment itself. The importance of egg quality in triploid induction of mollusks have been stressed by Utting and Doyou ( 1992), who demonstrated that variability in egg quality among the treated eggs of the Manila clam, Riiclitapt's pliUipp'u\aniin. can affect success during triploid induction. This is presumably caused by a lack of synchrony in meiotic rates, which are known to be highly important in triploid and tetraploid induction (Allen & Bushek 1992, Eudeline et al. 2()()0a). In the present study, there was evidence of lower synchro- nization in meiotic rates between the sets used during September than those used during November. That evidence comes from the coefficient of variation (CV) of meiotic rate between sets, mea- sured as time to reach 50% of PBl. which was twice during the September induction (CV = 33%) than in November (CV = 16%). In conclusion, success of triploidy in Pacific red abalone can be obtained when using CB concentrations from 0.5 to 1 mg L~'. and when the induction is started when 50% of the eggs show extrusion of the first polar body. Larger concentrations of CB will result in lower survival of treated eggs, possibly as an effect of toxicity of the chemical. Variation among induced batches at different times of the year can be expected, and this might be a consequence of differences in the egg quality of the spawners at each time, al- though this has to be further evaluated using biochemical data. ACKNOWLEDGMENT,S We thank the board of trustees and specially Mr. Benito Alta- mira-Rodriguez from the commercial farm Abulones Cultivados for their support in doing this research. We also thank Dr. S. K. Allen and Aimee Howe from ABC at the Virginia Institute of Marine Sciences, for the May induction analyses by flow cytom- etry. This research was supported by CONACYT grant 28256-B to A.M. Ibarra. The senior author is a CONACYT and SEP (DE- CYTM) PhD fellow, and the results presented here are part of his thesis. Dr. Ellis Glazier edited the Enclish lansuace text. LIRERATURE CITED AkIriJuc. F, J.. R. Q. MarMon & J. V. Shirenian. 1990. Induced triploids and tetriiploids in bighead carp. Hypophthalmichthy.s nohilis. verified hy multi-embryo cytonuorometric analysis. Aqmiciilnire 87:121-131. Allen Jr. S, K. 148.^. Flow cytometry: assaying experimental polyploid fish and shellfish. Aqmiculture 33:317-328. Allen Jr. S. K. & D. Bushek. 1992. Large-scale production of triploid oysters. Triploid'i' in Pacific Rei5 Abalone 1075 Cnis.sii>;!ri'ii viri^iiiicii (Gmelin), using "stripped" gametes. Aquaciilture 103:241-251. Allen Jr. S. K.. S. L. Downing & K. K. Chew. 1989. Hatchery manual for producing triploid oysters. Seattle: University of Washington Press. Arai, K., F. Naito & K. Fujino. 19X6. Triploidi/ation of the Pacific abalone with temperature and pressure treatments. Nippun Suixiin Gcikkiii'.lii 52:417-422. Beaumont. A. R. & J. E. Fairhrothcr. 1991. Ploidy nianipuialion ui mollusk shellfish: a review. J. Shellfish Res. 10:1-18. Cox. K. W. 1962. California abalones. family Haliotidae. Calilornia Fish and Game. Fishery Bulletin No. 1 18. [i? pp. Curatolo. A. & N. P. Wilkins. 1995. Indu/ione della iriploidia perme//o della presione idrostatica nell'abalone giapponese {HalitHis discus han- nai). Binlogia Marina MediWinmean 2:3 15-.'' 1 6. Downing. S. L. & S. K. Allen Jr. 1987. Induced triploidy in the Pacific oyster Crassostrea gigas optimal treatments w ith cylochalasin B depend on temperature. Aqiiaculnire fil:l-16. Eudeline. B. & S. K. Allen Jr, & X. Quo. 2000a. Optimization of tetraploid induction in Pacific oysters, Ciassnstrea gigas. using first polar body as a natural indicator. Aquaciiliiire 187:73-84. Fudeline, B., S. K. Allen Jr. & X. Quo. 2000b. Delayed meiosis and polar body release in eggs of triploid Pacific oysters. Crassostrea gigas. in relation to tetraploid production. ./. Exp. Mar. Biol. Ecol. 248:151-161. f.lliott, N. G. 200(1. Genetic improvement programs in abalone: what is the future. Ai/iiaculliire Res. 31:51-59. Fujino, K. 1992. Review of genetics and slock management of the Pacific abalone. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo. editors. .Abalone of the world. Biology, fisheries and culture. O.xford: Fishing News Books, pp. 491-503. Kudo. M.. K. Arai & K. Fujino. 1991. Triploidization of Halintis diversi- color diversicolor by cold shock. Nippon Suisan Gakkaishi 57:1263- 1267. Lindberg. D. R. 1992. Exolution. distribution and systematics of Hali- otidae. In: S. A. Shepherd. M. J. Tegner & S. A. Gu/man del Proo. editors. Abalone of the world. Biology, fisheries and culture. Oxford: Fishing News Books, pp. 3-18. Molina-Marti'ne/. J. 1983. Estimaciones de fecundidad en Huliuiis nife- scens de El Bajo, Bahi'a Rosario, Baja California. Institute Nacional de la Pe.sca. Secretarfa de Pesca. Mexico. Ciencia Pesqiieru 4:77-95. Morse, D. E.. H. Duncan. N. Hooker & A. Morse. 1977. Hydrogen per- oxide induces spawning in molluscs, with activation of prostaglandin endoperoxide synthetase. Science 196:298-300. Morse. D. E.. N. Hooker. L. Jensen & H. Duncan. 1979. Induction of larval abalone settling and metamorphosis by -aminobutyric acid and its con- geners from crustose red algae: 2. Applications to cultivation, seed- production and bioassays; principal causes of mortality and interfer- ence. In: J.W., Avault Jr.. editor. Proceedings of the tenth annual meet- ing. World Maricult. Soc. Baton Rouge: World Mariculture Soc. pp. 81-91. Neter. J.. W. Wasserman & M. H. Kutner. 1985. .Applied linear statistical models. Illinois USA. Richard D. Irwin. Inc. Orliz-Quintanilla. M.. G. Leiin-Carballo. M. A. Reinecke-Reyes. A. Lelevier-GrijaUa. J. R. Turrubiates-Morales & J. G. Gon/alez-Aviles. 199(1. Proceso reproductivo de las especies de abulon (Huliotis spp.) y su relacion con los niecanismos regulatorio-pesqueros en la costa oc- cidental e islas de la Peninsula de Baja California. Mexico. In: M. Dalley & H. Berisch editors. Proceedings of the VIII Simposium in Marine Biology, June 4-8 1990, Baja California Mexico: Universidad Autonoma de Baja California, pp. 156-163. Powers, D. A.. V. Kirby & M. Gomez-Ghiarri. 1996. Genetic engineering abalone: gene transfer and ploidy manipulation. ./. Shellfish Res. 15: 477. Ruiz-Verdugo. C. A., S. K. Allen Jr. & A. M. Ibarra. 2001. Family dif- ferences in success of triploid induction and effects of triploidy on fecundity of catarina scallop iArgnpecten \-eniriciisiis). At/iiaciiltiire 201:19-33. Searcy-Bernal. R. & C. Anguiano-Beltran. 1998. Optimizing the concen- tration of ganima-aminobutyric acid (GABA) for inducing larval meta- morphosis in the red abalone Haliaris rufescens (Mollusca: Gaster- opoda). J. World Aqiiaculmre Soc. 29:463—170. Stanley. J. G., S. K. Allen Jr. & H. Hidu. 1981. Polyploidy induced in the American oyster Crassostrea virginiea with cylochalasin B. Aqiuuul- tiire 12:1-10. Stepto. N. K. & P. A. Cook. 1998. Induction of triploidy in the South African abalone using cylochalasin B. Aqiiaculnire Int. 6:161-169. Supan, J. E.. C. E. Wilson & S. K. Allen Jr. 2000. The effect of cylocha- lasin B dosage on the survival and ploidy of Crassostrea virginiea (Gmelin) larvae. J. Shellfish Res. 19:125-128. Tegner. M. J. & R. A. Butler. 1985. The survival and mortality of .seeded and native red abalones. Haliotis riifiscens. on the Palos Verdes Pen- insula. California Fish Game 71:150-163. Utting. S. D. & J. Doyou. 1992. The increased utilization of egg lipid reserves following induction of triploidy in the Manila clam (Tapes philippinarum). Aquacullure 103:17-28. Yang. H. S.. H. C. Chen & Y. Y. Ting. 1998a. Induction of polyploidy and embryonic development of the abalone. Haliotis diversicolor. with temperature treatment. A/h. Malaculogicul Bull. 14:139-147. Yang. H. S.. Y. Y. Ting & H. C. Chen. 1998b. Blocking polar body with cylochalasin B in the fertilized eggs of the small abalone, Haliotis diversicolor supertexta (Lischke). and the development and ploidy of the resuUant embryos. Aquacullure Res. 29:775-783. Zar. J. H. 1999. Biostatistical Analysis. Englewood Cliffs. New Jersey: Prentice-Hall international. Zhang, G.. Z. Wang, Y.Chang. J. Song. J. Ding, Y. Wang & R. Wang. 1998. Triploid induction in Pacific abalone, Haliotis discus hannui Ino. by 6-dimethylaminopurine and the performance of triploid juveniles. / Shellfish Res. 17:783-788. Joiirmil ofSlh-lljUh Rvscanh. Vol. :(), No. 3, I077-IO.S7. 20(11. THE EFFECTS OF ENVIRONMENTAL FACTORS ON THE BIOCHEMICAL COMPOSITION OF THE BIVALVE TAGELUS DOMBEIl (LAMARCK, 1818) (TELLINACEA: SOLECURTIDAE) FROM THE INTERTIDAL FLAT OF COIHUIN, PUERTO MONTT, CHILE G. X. URRUTIA. J. M. NAVARRO, E. CLASING, AND R. A. STEAD lustituto dc Biohtiiia Marina "Dr. .liirgcn Winter". Universidad Austral dc Chile. Valilivia. Chile Casilla 567. ABSTRACT Samples of Tageliis domlH-ii (eight adults of each sex ranging between 5.5 cm and 7.0 cm) were collected monthly from the tidal flat of Coihufn. Puerto Montt. Chile. hetv\een June 1995 and July 1996. The soft tissue was divided into gonad, digestive gland and remaining somatic tissue, which were dried before determination of dry weight and biochemical composition (protein, lipid, carbohydrate and ash). Several environmental \ariables were measured at the same time (temperature, salinity, seston. clorophyll a. phytoplankton). Data were standardized to one individual of 6.0 cm shell length and presented as monthly means ± s.d. for each sex. There was no seasonal pattern in the dry meat weight of T. domheii. The spring diatom bloom of 1995 was not reflected in an increase in dry body weight. The increases in the weights of the somatic fraction and the digestive gland during the winter of 1996. together with increases in all the biochemical components, were attributable to an unusual diatom bloom and a reduced metabolic expenditure associated with lower temperatures. The cycle in the biochemical components is reflected in the energy content, which follows the trend in the weight of the corresponding tissue. There is an increase in the gonad weight and in the gonad index in T. domlieii in spring and a further increase in summer. Gametogenesis occurs initially at the expense of reserves accumulated earlier in the .soma (principally carbohydrate and protein) or in the digestive gland (mostly lipid), resulting in decreases in the storage tissues as the reserves are transferred to the gonad. Subsequently, energy for gametogenesis is obtained directly from ingested food as this source becomes adequate. Nevertheless, whether gametoge- nesis is associated with high body condition or with an abundant food source, there is no decrease in the weight of the animal. The annual variation in the biochemical components of the tissues is closely coupled to the availability of food, to temperature cycles that regulate metabolism and to the processes of gamete synthesis and release. Temporal tluctuations in environmental factors, especially temperature and food supply, drive the cycles of storage and utilization of metabolic energy reserves, which in turn govern growth and gametogenesis in this species. KEY WORDS: bivalves. Tagelu.s domlieii. environmental \ariables. biochemical composition INTRODLCTION The aiituial and interaiiiuial variation in the a\ailahilit_\ of nu- trients, in both qualitative and quantitative tertiis. seems to deter- mine intraspecific variation in growth, reproduction and energy storage in bivalves (Newell et al. 1982, MacDonald & Thompson 1985a, MacDonald & Thompson 1985b, MacDonald & Thompson 1986. Emmett et al. 1987). In general, the synthesis and storage of protein, carbohydrate and lipid in tnolkiscs occurs during those months in which food availability is gieatest (Pieters et al. 1980. Ruiz et al. 1992. Hawkins et al. 1985. Emmett et al. 1987), these reserves being utilized when food becomes scarcer (Ansell 1974) or when tnetabolic demands such as those of gametogenesis must be met (Gabbott & Bayne 1973. Bayne 1976. Soniat et al. 1984. Soniat & Ray 1985. Emmett et al. 1987). Depending on the availability of food and the size of the indi- vidual, energy may be allocated \o the synthesis of body tissue, i.e.. growth, or to the production. Ansell and Trevullion (1967) found that in Tellina tenuis dry body weight can increase by 44*^ when food is abundant and the gonad is developing, and decreases sharply during spawning. When the decrease in the weight of one or more tissues coincides with gonad development, one mav con- clude that the reproductive process is at least partly dependent on accumulated energy reserves. The weights of the digestive gland and adductor muscle decrease as gonad weight increases in Pati- nopeeten yessoensis (Fuji & Hashizume 1974) and Placopeeten maficUanieiis (Robinson et al. 1981 ). Thus reproduction is a pro- cess which requii-es a considerable expenditure of energy, obtained *E-mail: gurrutiiCsmercurio.uach.cl frotn i-ecently ingested food or from the calabolism of reserves in the body tissues (Gabbott 1976. Gabbott 1983). In some species, such as Pecten maximits (Comely 1974) and Chlamys opereularis (Taylor & Venn 1979). energy reserves are laid down in summer for utilization during gametogenesis in the autumn, whereas in others, e.g.. Chlamys septemradiata (Ansell 1974) -dnd Plaeopeeten mageUaniciis (Thompson 1977), the maximum food supply coin- cides with gonad growth and development. There are also species which can adopt either strategy, depending on the population (Thompson 1984. Gabbott 1983). or which utilize both strategies, depending on the tinie of year. Consequently, the seasonal varia- tion in the biochemical composition of bivalve tissues is the result of complex interactions between environmental factors and meta- bolic processes (Gabbott 1975. Gabbott 1976. Gabbott 1983. Be- ninger & Lucas 1984, Thompson & MacDonald 1990), which are reflected in fluctuations in the dry weight of the different body tissues (Ansell 1972, Barber & Blake 1981. Shafee 1981). In cold temperate environments the food supplv to bivalves and physical factors exhibit considerable spatial and temporal \ ariation (Vahl 1980. Dellarosa et al. 1993, Navarro et al. 1993, Stead et al. 1997, Toro et al. 1999). Navarro and Jaramillo (1994) concluded that the principal contribution of phytoplankton to the nutrition of suspension-feeders in the south of Chile, occurs during spring and summer, values being tninimal during the rest of the year. Conse- quently. Venus antiqua exhibits a markedly .seasonal growth pat- tern with high rates of growth in spring and sumtiier (Clasing et al. 1994). Since the patterns of growth and reproduction in bivalves reflect environmental conditions (Trevallion 1971. Incze et al. 1980. Chaparro & Winter 1983). it is likely that temporal changes in temperature and food supply are important in the regulation of 1077 1078 Urrutia et al. biochemical cycles and reproductive processes. According to Slna- fee (1981) and Thompson and MacDonald (1990). these environ- mental variables can affect the biochemical components of the gonads of males and females in different ways, owing to the dif- ferences in the biochemical composition of the gametes. The objective of the present study was to characterize and quantify the temporal variation in the biochemical composition (lipid, carbohydrate, protein and inorganic material) and weights of various body tissues in the bivalve Tai>f/iis dumbeii (Lamarck 1818), and to establish their relationships with several environ- mental variables (temperature, salinity, seston. pigments, phy- toplankton) over the intertidal flat of Coihum. Puerto Montt. Chile. MATERIALS AND METHODS Samples of T. dombeii were collected monthly between June 1995 and July 1996 from the intertidal flat of Coihuin (4r29'S; 72°54'W), located approxitnately 8 km SE from the town of Puerto Montt (Fig. 1). At each sampling time, water temperature and salinity were recorded and water samples (3 replicates of 2 liters volume) collected for determination of chlorophyll a, phaeopig- ments, particulate organic matter and particulate inorganic matter (Strickland & Parsons 1972). Additional water samples were fi.xed in Lugol's iodine for quantification of the principal phytoplankton taxa (diatoms and microtlagellates). For determination of tissue dry weight and subsequent bio- chemical analyses. 8 individual clams of each sex (determined by microscopic inspection of gonad smears), ranging between 5.5 cm and 7.0 cm valve length, were opened and the soft tissues sepa- rated into digestive gland, gonad and the rest of the tissue (soma). Separated tissues were dried at 60°C for 48 hours, weighed, ground in a ball mill (Retschmiihle type MM) and stored at -20°C for biochemical analysis. Gonad and digestive gland indices were calculated as the dry weight of each tissue as the percentage of the total dry weight of soft tissues. Carbohydrate, lipid, protein and ash were determined on sub- i ■„ i' \ CHIL ( ( / vA - samples of dried, homogenized tissue. Carbohydrate was estimated b\ the phenol-sulphuric acid procedure (Dubois el al. 1956) and lipid gravimetrically according to Bligh and Dyer (1959). Total nitrogen was determined with a CHN analyser (Perkin Elmer 2400) standardized with acetanilide. and protein calculated as 5.8 X N (Gnaiger & Bitterlich 1984). Ash-free dry weight was deter- mined by combustion of dry samples for 12h at 500'C. Biochemi- cal data were expressed as absolute values per standardized indi- vidual clam. To express the biochemical data in terms of energy, values for the different tissues were multiplied by the appropriate conversion factors (protein: 24 KJ-g"'. lipid: .^9.5 KJg''. carbohydrate: 17.5 KJg"'. Gnaiger 1983). Data for the dry weight, biochemical composition and energy content of the tissues were presented for males and females sepa- rately as monthly means with standard deviations. Values were corrected for a "standard" clam of shell length 6 cm. as described by Bayne et al. (1987): W, ( Ls/Lp)" X We Where: W^ is the dry weight of the standard clam, Lg the standard shell length (6 cm), Lp and W^ the recorded shell length and dry weight respectively of the experimental clam, and b the weight exponent of the relationship between length and weight. Student's /-test was used to test for differences between mean values for males and females. The relationships between dry weight, biochemical composition, food availability and physical factors were examined by conelation analysis and principal com- ponents analysis (Statistica 4.2 for Windows). RESULTS EinintiiimnUil Factors Figure L Map of the tidal Hal at Cdiluiin, Puerto MonK, Chile, show- ing tlie location of the sampling area. A seasonal cycle in water temperature was observed, with mini- mum values in winter (9°C in July 1995, 10°C in July 1996). increasing gradually until summer (17°C in January 1996). Salin- ity remained stable, varying between 24.99(< in December 1995 and 32.4':;;( in May 1996 (Fig. 2a). Total particulate matter (seston) was highest in autumn- winter, reaching 10 mgl"' in July 1995 and 12 mgl"' in July 1996. During spring and summer values remained around 5 mg 1"'. The organic fraction of the seston exhibited a similar cycle (Fig. 2b), with minimum values (0.63 mg-P' ) in September 1995 and maxi- mum values (5.57 mg4"') in June of the same year. The concentration of chlorophyll a was lowest in summer (0.36 |j.g r') and highest in April and May of 1996 (7.57 |j.gr' and 8.61 |xg-r' respectively). There were two minor peaks in July and September 1995 (Fig. 2c). Phaeopigment concentrations were lower than 2 (J.g 1"' throughout 1995. In April 1996 a value of 7.8 (xg-r' was recorded, probably as result of high zooplankton ac- tivity. Diatoms showed two main peaks (Fig. 2d), the first in November 1995 (5129 cellsinr') and the second in April and May 1996 (6954 cellsmP' and 5626 cellsmP' respectively), whereas high densities of microtlagellates were observed only in the sum- mer 1995-1996 (1833 cellsmP'). Dry Tissue Weight Somatic Tissue No large variations in the dry weight of somatic tissue of T. doinheii were observed during the study, although values were Effects of Environmental Factors on Tagelus domhhii 1079 M 20 oa lU ^^— ^''''^^^•"*"^ • - O - Salinily 20 jj IS ;; I" I JJASONDJFMAMJJ 5^ c — O — Phacopigmenl T A I ^Cn/^ x:w 6 5 JJASONDJFMAMJJ ♦ Diatoms - O — Microflagellales A— a^ JJASONDJFMAMJJ 1995 1996 Figure 2. Temporal variation of the temperature and salinity (a) total and organie seslon: Ih) chlorophyll a and phaeopigments: (cl diatoms and niicronagellates; (d) present in the water column. Values are. means ±s.d. slightly higher during uutiiniii and winter, especially during 1996 (Fig. 3a). With a tew e.xceptions. males and females did not differ in somatic dry weight iP s 0.05). This observation was confirmed by the lack of a significant difference between sexes when the overall annual mean was compared. In some clams it was not possible to determine the sex owing to the scarcity of gonad tissue. These were classified as "indeter- minate", and constituted the entire sample in June 1995 and May 1996. When indeterminate individuals were present as well as males and females, their somatic weight was similar to that of the females but significantly different {P £ 0.05) from that of the males (Fig. 3a). Gonad JJASONDJFMAMJ.l 600 GONAD -r b 500 * 1 400 300 U Jrk ion 100 r 0.01) in overall inean total dry weight be- tween males, females and indeterminates. 1080 Urrutia et al. (>onad and Digestive Gland Indices Minimum values were observed for the gonad index at the beginning of the study (June 1995. Fig. 4a). i.e., all individuals were indeterminate. During the spring, gonad index increased to peak values of 22.3% in inales (November) and 18.3% in females (October) before decreasing in December to 9.8% (males) and 12.8% (females). A second peak was observed in February of 18.9% for males and 26.5% in females, to decrease to minimum values of 2% in autumn (April). Only in November 1995 there was a significant difference (P < 0.05) in gonad index between sexes. In those months in which clams of indeterminate sex were ob- served (Fig. 4a). such individuals exhibited the lowest gonad in- dices, suggesting that at these times the gonad was regressing and/or that gametogenesis was about to begin. The digestive gland index remained relatively constant throughout the year (Fig. 4b). Values varied between 7.5% and 10.6% in males and 7.5% and 10.8% in females. Comparisons of monthly means showed significant differences (P s 0.05) between males and females in 6 months (Fig. 4b). but there was no signifi- cant difference overall (P > 0.05) between males and females. Proximate Biochemical Composition Composition of Somatic Tissue The protein content varied from 55 1 mg to 865 mg in males, from 521 mg to 871 mg in females, and from 623 to 868 mg in indeterminate individuals, with slightly higher values in winter of 1995 and autumn of 1996 (Fig. 5a). Significant differences were observed (P £ 0.05) between males and females in September. November and December 1995 and in April 1996. and between lASO^DJFMAMJJ ;oo (1 T 150 -A- ■I s 100 H •"^t^*"^ .i.^^ -^ 50 0 * * 1995 1996 JJASONDEFMAMJJ 1995 1996 Figure 4. T. domheii. Temporal variation of the gonad index and the digestive gland index to one Individual of fi.O-cm shell length. Data presented as. monthly means ± s.d. of males (•). females (O) and in- determinates (A). *: Significant differences between males and females at the P < 0.(15 level &: Significant differences between females and indeterminates at the P < 0.05 level. Figure 5. T. domheii. Temporal variation in the biochemical compo- nents of the somatic tissue, (a) Protein; (b) Lipid: (c) Carbohydrates; and (dl ash; to one Inditidual ofA.O-cm shell length. Data presented as, monthly means ± s.d. of males (•). females I : ) and indeterminates (A). ®; Significant differences between males and females at the P < 0.05 le\el #: Significant differences between males and indeterininales at the P < 0.05 level. males and intermediates in Ju!\ 1995 and April 1996. There were no significant differences (P > 0.05) in annual means between males, females and indeterminate clams. Lipid content varied be- tv\een 44 mg and 63 mg in males. 38 mg and 74 mg in females, and 39 and 71 mg in indeterminate clams (Fig. 5b). No significant differences (P > 0.05) were observed between males and females in the annual means, although values were significantly different (P s 0.05) in March and April 1996. Carbohydrate content was lowest in spring-summer and highest in autumn- winter (Fig. 5c). reaching a maximum in July 1996. Intermediate values were ob- served in indeterminate individuals. Significant differences were only observed (P s 0.01 ) between annual means for indeterminate clams and those for males or females. Ash content varied between 101 mg and 154 mg in males and 84 mg and 151 mg in females, with means of 1 15 ± 28 mg, n = 83 and 1 14 ± 23 mg, n = 81. respectively (NSD. P > 0.05). Significant differences were ob- served (P s 0.05) between males and females in September 1995 and between males and females and males and indeterminate clams in April 1996 (Fig. 5d). The annual mean ash weight in Effects of Environmental Factors on Tagelus dumbeii 1081 indeterminate clams was significantly different from that fcir males {P < 0.01 ) and females (P s; 0.05). Composition of the Gonad In sonic miintlis (June and .luK l'-)'-)3. .'\pril and May 1996) insufficient tissue was available to undertake the biochemical analy.ses (indicated by "T syinbol. Fig. 6). Protein content showed two rises (Fig. 6a); the first in October 1995 with the highest value for males ( 186 mg) and the second in February 1996 with the highest value for the females (221 mg). The lowest protein content was observed in August 1 995 for both se.xes (48 mg in males and 51 mg in females). Values for both se,\es from September to November 1995 and from January and March 1996 were significantly different (P < 0.05). The average annual protein level in males was significantly different (P s 0.01 ) from that obtained for females. Lipids were notoriously higher in October and in February (Fig. 6b). at the time when gonad weight showed the highest values (Fig. 3b). There were no sig- nificant differences (P < 0.05) between the monthly averages nor in annual averages obtained for this component. The carbohy- z u i- o T 300 a 1 1S« 200 >-v 1 \\ ISO 100 50 ? ? 5 PA * * * * ^ s ? ? \ JJASONDJFMAMJJ J J A S O N D J F M A M .1 .1 30 C * T 1 20 Tj / T , T 6 10 7 7 ^ /'^i\^ ^- 0.05). Ash content (Fig. 6d) increased significantly in No- \ ember 1995 and in February 1996. Significant difference between males and females was only observed in July 1996. No significant differences were observed when compared annual means (P > 0.05). Composition of the Digestive Gland Protein was highest in Juh 1996 with a minimum \alue found in July 1995 in males and in December in the females; in indeter- minate individuals protein fluctuated between 30 mg in July 1995 and 57 mg in .^pril 1996 (Fig. 7a). There were significant differ- ences (P < 0.05) between males and females in August 1995 and o at m a I 6 I w 5 x/ts5^ ^^-N^ ^->- ^■i 40 l 20 # # s * * * b at' F i» & * * I M Q 2tt T ■i 0. I 5. T L^i" 10 I ■■■f-^ ^i-^-^j^^ .^j^'^-i 40 d T 30 if \ z < 20 1U I - |/^^xt^ 1 ^ \^ 4i I JJASONDJ FMAM.I J 1995 1996 Figure 7. T. domheii. Temporal variation in the biochemical compo- nents of the digesti\e gland, (a) Protein; (b) Lipid: (c) Carbohydrates; and (d) ash; to one indi\idual of 6.()-cm shell length. Data presented as monthi) means ± s.d. of males (•!. females ( : I and indeterminates (A). *: Significant differences between males and females at the P < O.OS le>el #; Significant differences between males and indeterminates at the P < i)M5 le\el &: Significant differences between females and indeterminates at the P < 0.05 level. 1082 Urrutia et al. 3 I I .= II O © O O c >r, r- w* c ^ U-, — r- n — rj O r-l ri o o o o o o o o i r I r^ ?, -t £ 1 1 T 1 r^ -r ^1 ^. ~ ~ T T "■ T T 5 T ITi -* * s ^' r 1 3. -! =: o ^ c o j: crt &S i'h- H U H O CL c/5 Q c U 2 X 2 ri 2 ~- 0^ o U 1 ^a. Vt «-'. 3 i/ ^ i E T L. c II a 3 t/2 m ic <"l Wj "/"i '*! O oc o rl -t o sc r^* i/"i n 1^ — r) (-1 r^i O — =' =' o c o' o o o o o' o «C n >c t^ *^< ^ O (^1 ri r^, n r*-i r^^ — X r*-, r«-, « w- ^ sC ^ >c r^ -i- 0.05) were observed in annual averages of this component between males and females, although there were significant differences between the former and the indeterminate clams. Lipids increased in winter, being very high in July 1996. Minimum values occurred in spring-summer. Indeterminate clams showed values between 6 mg and 15 nig. Significant differences (P s 0.05) between males and females were observed in September 1995 and April 1996. between males and indeterminate clams in April 1996, and be- tween females and indeterminates in July 1995. No significant differences {P > 0.05) were observed when comparing annual averages between males and females, although difference was ob- served (P s 0.01 ) in comparing these with the average lipid in the indeterminate individuals. Carbohydrates (Fig. 7c) showed a maxi- mum in females in April 1996. a value which was significantly different from those obtained for males and indeterminate clams. Both sexes showed a minimum value in January (ca. 7 mg). with indeterminate clams showing a lower concentration in winter 1995 and a maxiinum in autumn 1996. Significant differences iP < 0.05) existed between sexes in October 1995 and April and July 1996. between males and indeterminate individuals in July 1995. and between females and indeterminate individuals in July 1995 and April 1996. No significant differences were observed (P > 0.05) when comparing annual means of males and females. The -10 OS 0 0 0.5 25 SOMA a :o J. . ^— Q— «-^«--HJUilKT(.JN N SOMAW£IGln 9 GONAD WElGin n) 010 G wrrGKi II TOTAL WEinm i; (iONADlC INDFX II DIO GL INDEX 14 SOMA PHtTTEIN n SOMA LIPID 16 SOMA CARfl 17 SOMAEM-JtGY 18 SOMA ASH ly GONAD PROTCIN 20 GONAD LIPID :i GONADCARB ;2 GONAD ENERGY 13 GONAD ASH :j DIGG PROTEIN ;i DIG r, LIPU> ;e. DIG GL CARB :7 DtO GL ENFRGY ;»DIO Gl ASH ai=2i.6% 011=7.90-/. - 0 ^5 0 0 OS 1 D -10 -0 5 0,0 O.S 1.0 g 1^3 a t=t s 1^ a «^iH 1=1 ^^■■H D 1=^ ■■■^k C2 1 1 ■^^^ > 1^^^^ q I^^^B Q ^^^^m li 1 0 g a 1 i 4 ii a « Figure 9. T. dombeii. Principal Components analysis indicating the charge of each variable and the variance in each component. The black bars present charge >(l.7. indicating significant association among variables. only significant difference (P s 0.05) was observed when males and females were compared with indeterminate individuals. The lowest ash content occurred in December and the maximum in January. Indeterminate individuals showed values ranging from 7 to 1.'! mg. There were no significant differences (P a 0.05) be- tween males and females, with significant difference (P s 0.05) only between males and indeterminate individuals in April 1996 (Fig. 7d) . There were no significant differences (P > 0.05) when comparing the annual averages for males and females, but there was a difference observed iP < 0.05) between the average of both males and females with that of the indeterminate clams. z u H Z o p Id o W z ? ? D JJASONDJKMAMJJ 25 DIGESTIVE GLA.ND c 20 15 T.-p'h-^i^''^"-^ -^ ^ ^•-a ^ 10 5 n F^nergelic Content No variation in the energetic content of the somatic tissue was observed (Fig. 8a). with no significant differences between males and females. When annual averages were compared, no significant differences {P > 0.05) were obtained between males, females and indeterminate clams. The energetic content of the gonad tissue increased in males, beginning in August 1995 (Fig. 8b). Significant differences (P < 0.05) were observed between males and females in October. November and December 1995 and February 1996. The annual average for males was significantly different (P < 0.05) than the average obtained for females. The energetic content of the digestive gland increased notably in July 1996 (Fig. 8c). with the minimum value for both sexes in January of this year. No significant differences (P > 0.05) were observed between males, females and indeterminate clams. Annual averages for energy con- tent in the digestive gland did not show significant differences (P > 0.05) between males, females and indeterminate individuals. 1995 1996 Figure 8. T. tlnniheii. Temporal fluctuation In the energetic content of the dilTerenl tissue. Data presented a.s, monthly means ± s.d. of males (•). females ( ) and indeterminates (A). *: Signiricant differences between males and females at the P < 0.05 level ?: Without information by tissue scarcity. Principal Components Analysis The principal components analysis included 28 variables (Fig. 9). with 140 values estimated for 70 males and 70 females. Sex was not identified as an inlluencing factor and thus results from males and females were combined in the analysis. The variables with a charge sO.7 are associated with the variable component. 1084 Urrutia et al. and as well, associated among themselves {Fig. 9). The three com- ponents explain 67.3% of the total variance. The first component associate positively chlorophyll a. total seston. organic seston, body weight, digestive gland weight, total weight, protein, carbo- hydrates and body energy content and protein, lipids, and energetic content of the digestive gland. In the second component are posi- tively associated gonad weight, gonad index, protein, lipids, car- bohydrates, energetic content and gonad ash. In the third compo- nent are positively associated diatoms and phytoplankton. DISCUSSION Eiiyiniiinuiiltil Variables The annual cycles of temperature and salinity measured were similar to those previously described for the south of Chile (Na- varro et al. 1993. NavaiTO & Jaramillo 1994, Stead et al. 1997, Toro et al. 1999). The concentrations of total and organic seston were lower than those reported by Navarro et al. (1993) for the tidal flat at Yaldad, (Chiloe). The total seston reached 12 mg P' and the organic seston 5.6 mgl"'. while at Yaldad, total seston reached 48 nigP' and organic seston 18 mg P'. These differences were due to variations in primary productivity and teirigenous imports, as well as characteristics of sediments which facilitate resuspension. The highest concentrations of total and organic seston measured at Coihuin in winter 199.^ were related to resus- pension phenomena, in contrast with autumn-winter 1996 where they were related to the high primary production. Chlorophyll a rises in spring, as described for southern Chile (Navarro et al. 1993. Navarro & Jaramillo 1994. Stead et al. 1997. Toro et al. 1999) and for similar latitudes in the northern hemisphere (Mac- Donald & Thompson. 1985a; Navano & Thompson Jr.. 1995). A major rise was observed from April to July 1996. related with a diatom bloom. Phaeopigments shov\ed a similar pattern to chloro- phyll (/ but with lower values. In February and April 1996 high concentrations of phaeopigments were observed, probably attrib- utable to excretion products due to an increase in grazing activities of zooplanklon or. as a result of senescence of the phytoplankton (Hallegraeff 1981, Welschnieyer & Lorenzen 1985). The phytoplankton consisted primarily of diatoms and mi- croflagellates. The diatoms, primarily represented by Skclcloiiciini costaliiiii. showed major abundance in spring and fall, which was reflected by the increase in chlorophyll a as well as for total and organic seston. The greatest abundance of microflagellates oc- curred in December 1995 however, this was not reflected by chlo- rophyll CI nor by the seston values, given the low biomass of this group. Fliicliialion in Weiiilil and lineigy Reserves Numerous studies have denioiistralcd thai the variation in weight in bivalves is related to reproductive cycles and environ- mental conditions (Thompson 1977, Sundet & Valh 1981, Jarainiilo & Navarro 1995). Temperature is commonly mentioned as a factor which influences growth and accumulation of metabolic reserves in bivalves (Bayne & Worral 1 980. Vahl 1980. Mac- Donald & Thompson 1985a). However, Shafee (1981) concluded that the increase in weight of tissues was directly linked to the abundance of food available, Sastry 1966 and Sastry and Blake (1971), showed that temperature affected the initiation of game- togenesis in Aeqidpecten irradians. regulating the transfer of nu- tritive reserves from other tissues into the gonad. In Myiilus edulis (Thompson & Bayne 1974"). Doihix iniiniiliis (Ansell et al. 1980) and Modiolus modiolus {NavaiTO 1990), the body weight demon- strated marked .seasonal cycles with minimum values in winter and maximum in spring-summer in periods of high food availability. Conversely, Taf>ehis doinlieii did not show a clear seasonal pattern, only presenting a rise in tissue weight in winter 1996. which mav have been related to the unusually high concentration of diatoms occurring since autumn of that year. The situation was reflected by increase in weight of all body tissues, rise in carbohydrates and protein in somatic tissues, and in all biochemical components of the digestive gland. Interannual differences in tissue weights and cycles of metabolic reserves have been described for other species (Newell et al. 1982, Navarro 1990), and may be the result of yeariy changes in environmental conditions. The decrease in gonad weight and all their biochemical com- ponents from October to December, indicates the occurrence of a massive spawning during this period, as it was reported previously for this population by Clasing et al. { 1998). The lack of increased body weight during these months, in spite of the abundance of phytoplankton. may be interpreted as a reduction in filtering ac- tivity during spawning periods as an adaptive mechanism for avoidance of ingestion of their own gametes (Newell & Thompson 1984, Thompson 1984). Also, during this period the natural rise in temperature may have resulted in higher metabolic demands. Based on these results, the reproductive cycle of T. dombeii did not have a major influence on the body weight of this species. However, the initiation of gonadal development in winter 1995 ciiincided with a decline in the weight of the somatic tissue, sug- gesting that the initiation of gametogenesis occuned at the expense of reserves in this tissue. This is supported by a significant nega- tive correlation between gonad weight and somatic carbohydrate (r = -0.59). Even when they were not significant at P > 0.05, (Table 1 ) a negative correlation was also observed between the gonad weight and the somatic protein (r = -0.26) and the lipid content of the digestive gland (r = -0.20). The digestive gland of T. donilwii did not show a decline during this season, although its lipid content decreased. Both conditions have been described for Chkimys opercidahs (Taylor & Venn 1979) and Placopecten nuLxiniKs (Faveris 1987) where the gonadal development in winter caused a decline in carbohydrates and protein in the adductor muscle, and a decline in lipids in the digestive gland. Under con- ditions of food scarcity, it has been suggested that glycogen acts as the primary energy reservoir for the formation of gametes in bi- valves. Also a reduction of this reserve in storage organs is com- monly correlated with an increase in gonadal lipids (Barber & Blake 1981. Benninger & Lucas 1984). The considerable inciease for gonad weight, gonadal index, and all biochemical components of the gonad of T. doniljeii in February 1996 was related to the absorption of nutrients from the food, with significant positive correlations with chlorophyll <(. organic seston and phytoplankton (Table I ). This suggests that this clam is able to take advantage of particulate food matter resuspended by tidal movement, waves, and winds. Minor water movement is sufficient to resuspend food particles at the bottO)ii/water interface where this species remains buried with its siphons open at the sedi)iient sur- face. Clasing et al. (1998) suggest there may be high concentra- tions of food material for this species at the sediment surface, gi\cn that the anioiMil of chlorophyll ci on the surface of the Coihuin tidal flat is somewhat greater than that encountered in the water column. The majority of studies on the biochemical composition of bivalves describe temporal variations in components of different Effects of Environmental Factors on Tagelus noMisEii 1085 tissues, comparing different populations, stages of maturation, stages in the life cycle, size classes, etc. Some studies have com- pared biochemical composition between males and females (Sha- fee 1981. Davis & Wilson 19S.\ Ruiz et al. 1992. Thompson & MacDonald 199(1). Shafee 1981 suggested thai during gonad de- velopment of Cltlaiiiys viiria. protein was more abundant in males than in females, while these contained more lipids given the dif- ferent composition of the gametes. Davis and Wilson ( 1983) de- scribed different levels in lipids between males and females of Nuciilii tiii'i^iclti only during the spawning .season, vsilh females having higher levels than males. Thompson and MacDonald ( 1990) found that lipids of Placopt'ctcn luagcUaiiicits varied from 3% to 19*7^ in females in the spring, with no variation in males. Conversely. T. dinnheii showed no significant differences in aver- age monthly values for lipids between gonads of males and fe- males. Content of protein in the male gonad was significantly higher than in the females. Carbohydrates in male and female gonads were significantly different only during the second period of gonad development which may be related to differential trans- formation of carbohydrates to lipids, as has been reported for other species (Gabbott 1975. Zandee et al. 1980). Results obtained in April 1996 on comparative tissue weights between males and females suggested that the females were able to take better advantage of food availability than males, as during a large diatom bloom the females preferentially increased in weight of the fraction somatic and the digestive gland (thus total body weight). Also in this month, protein, lipid, and ash increased in female body tissues, as well as in organic componeiUs of the digestive gland. Thus, differences observed between sexes may occur during particular situations, rather than as part of annual cycles of storage and use of tissue reserves. During May. our sampling included only indeterminate individuals that had spawned were in regression, or which were initiating gametogen- esis. A parallel histological study on T. domheii (Clasing et al. 1998) concluded that during this month 579f of males and 80% of females were found in early active of gametogenesis. while 43'/r of the remaining males and 20% of the females were in the recovery stage. Also. 10% of the population was in an "indeterminate" stage, post regression and prior to gametogenesis. where sex de- termination could not be made. In May. indeterminate individuals underwent increases in somatic tissue weight and lesser increases in gonad weight in the presence of high diatom concentrations. The weight of the digestive gland of indeterminate individuals in- creased compared with males from the previous month. At that time, an increase in all biochemical components was observed in the somatic tissue, except for the lipids (gonad could not be sampled due to scarcity of this tissue). Also, a decrease was ob- served in proteins and carbohydrates of the digestive gland, with an increase in ash content. These observations suggested that nu- trients coming from the phytoplankton were rapidly transferred from the digestive gland to other tissues as maintenance energy, for growth and to sustain incipient gametogenesis. This agrees with Gabbott ( 1976). who suggested that when sufficient food was available during gametogenesis. there was rapid transfer of assimi- lated food from the digestive gland to the other tissues. According to Thompson (1972). this transfer may occur within about seven days. The energy available for the production of gametes may be affected by metabolic stress, such as produced by parasitism (Sanders 1966, Sanders & Lester 1981). Periodical sampling of the T. cldiiihc'ii population at Coihuin revealed a high incidence of parasitism by a digenetic trematode (Familia Philophthahnidae). infestation could reach at times 100% of the adult population, but it rarely caused castration of individuals (<2.5%) (Yafiez 1998). In ihe principal component analysis, the first component asso- ciated positively the chlorophyll a and seston with the weight of somatic tissue and that of digestive gland, as well as with energetic content of these tissues. In the second component, the variables related to the gonad. This situation was expected, considering that the somatic tissue and digestive gland are organs of storage, con- stituting the main body of the organism, and that the gonad only develops in certain seasons of the year and depends as much on body reserves as on food availability. The cycle of biochemical components in body tissues and di- gestive gland of 7". domheii at Coihuin is related to the production and liberation of gametes. Initiation of gametogenesis in winter takes place al the expense of stored tissue reserves, until food availability at the beginning of spring is able to continue to support this process. This situation has been described for Chlaiiiys oper- cuhiris (Taylor & Venn 1979). Plucopecten magcllanicus (Comely 1974. Faveris 1987) and Mylihis edidis (Zandee et al. 1980). Given that the increase in weight of the gonad coincides with a high condition of the organism, or elevated of food availability, game- togenesis does not provoke a decrease in body weight of individu- als. Environmental variables such as temperature, which regulates metabolism, and food availability which provides necessary en- ergy, play fundamental roles during the process of storage, distri- bution and utilization of energy reserves of T. domheii. Temporal and spatial differences in these variables are reflected in the growth and gametogenesis of this species. ACKNOWLEDGMENTS We thank Blanca Vargas. Marco Lardies and Yennie Yaiiez for collaboration in fieldwork and processing of laboratory samples. We also thank the Ocean Sciences Center. Memorial University of Newfoundland, Canada, for help with nitrogen analyses, and especially aid from Dr. Ray Thompson. This work was supported by the Chilean National Fund for Science and Technology (FONDECYT # 1951202) and the Programa Acuiciiltura y Biotecnologia Marina. I (97). Fondap. Chile (Subprograma Invertebrados). LITERATURE CITED Ansell. A. D. 1972. Distribution, growth and seasonal changes in biochemical composition for the bivalve Donax vitlatiis (dii Costal tiom Karnes Bay. Millport. J. Exp. Mar. Biol. Ecol. 10:1.^7- 150. Ansell. A. D. 1974. Seasonal chanjies in biochemical composition tor the bivalve CliUimys septciiiradiahi from llif Clyde Sea Area. Mar. Biol. 25:85-99. Ansell. A. D. & A. Trevallion. 1967. Studies on Tclliiia tenuis (da Cosla)— Seasonal growth and hiocheniical cycle. / Exp. Mar. Biol. Ecol. 1: 220-235. Ansell, A. D.. L. Frenkiel & M. Mouc/a. 19S0. Seasonal changes m the tissue weight and biochemical composition for the bivalve Donax Irnn- ciiln.s on Ihe Algerian coast. J. Exp. Mar. Biol. Ecol. 45:105-1 16. Barber. B. J. & N. J. Blake. 1981. Energy storage and utilization in relation to gametogenesis in Argopecien irrailians concenlricus (Say). J. Exp. Mar. Biol. Ecol. 52:121-1.^4. 1086 Urrltia et al. Bayne. B. L. 1976. Aspects of reproduction in bivalves molluscs. In; M.Wiley, editor. "Estuarine Processes". New York: Academic Press, pp. 432-14S. Bayne, B. L. & C. M. Worral. 1980. Growth and production of mussels Myriliis edulis from two populations. Mar. Ecol. Prog. Ser. 3:317-328 Bayne. B. L.. A. J. S. Hawkins & E. Navarro. 1987. Feeding and digestion by the mussel Mylilus edulis L. (Bivalvia: Mollusca) in mixtures of silt and algal cells at low concentrations. J. Exp. Mar. Biol. Ecol. 111:1 -22. Beninger. P. G. & A. Lucas. 1984. Seasonal variations in condition. le- productive activity and gross biochemical composition of two species of adult clam reared in a common habitat: Tapes decussatus L. (Jef- freys) and Tapes philiphinaniiii (Adams & Reeve). / E.\p. Mar. Biol. Ecol. 79:19-37. Bligh, E. G. & W. J. Dyer. 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 32:91 1-916. Chaparro. O. & J. Winter. 1983. Effects of poor existing nourishment on grouth and reproduction of Mviiliis chilensis and existing stock in the muUipopulation of Yaldad (Chiloe. Sur de Chile). Menis. .Asoc. Lati- noani. .Aciiicidt. 5:203-214. Clasing. E., T. Brey, R. A. Stead. J. M. Navarro & G. Asencio. 1994. Population dynamics of Venus antiqua (Bivalvia:Veneracea) in the Bahfa de Yaldad, Isla de Chiloe. Southern Chile. J. Exp. Mar. Biol. Ecol. 177:171-186. Clasing. E.. J. M. Navarro. M. A. Lardies & R. A. Stead. 1998. Alimen- tacion suspensi'vora-detritivora en bivalvos de la infauna: una estrategia dual para optimizar el aprovechamiento de la oferta alimentaria. In- forme Final Proyecto Fondecyt N° 1951202. Comely, C. A. 1974. Seasonal variation in the flesh weights and biochemi- cal content of the scallop Pecten maximus (L.) in the Clyde Sea area. J. Cons., Cons. Int. Explor. Mer. 35:281-295. Davis. J. P. & J. G. Wilson. 1983. Seasonal changes in tissue weight and biochemical composition of the bivalve Nucula turgida in Dublin Bay with reference to gametogenesis. Neth. J. Sea Res. 17:84-95. Dellarosa. V.. R. Quifiones, O. Ulloa & T. Antezana. 1993. Principal annual productivity of some locations of Archipielago de Chiloe. Re- sumenes Xlll Jornadas de Ciencias del Mar. p. 62. Dubois, W.. K. A. Gilles. J. M. Hamilton. P. A. Rebers & F. Smith. 1956. Colorimetric method for the determination of sugars and related sub- stances. Anal. Chem. 28:350-356. Emmett, B.. K. Thompson & J. D. Popham. 1987. The reproductive and energy storage cycles of two populations of Myrilus edulis (Linne) from British Columbia. / Shellfish Res. 6:29-36. Faveris. R. 1987. Studies on the evolution of glycogen content of somatic and germinal tissues during the annual reproductive cycle in Pecien maximus. In: Sixth International Pectinid Workshop. Menai Bridge. Wales. 10 pp. Fuji. A. & M. Hashizume. 1974. Energy budget for a Japanese common scallop. Pannopeclen yessoensis (Jay), in Mutsu Bay. Bull. Fac. Fish. 25:7-19. Gabbott. P. A. & B. L. Bayne. 1973. Biochemical effects of temperature and nutritive stress on Myrilus edulis (L.). J. Mar. Biol. As.soc. UK. 53:269-286. Gabbott. P. A. 1975. Storage cycles in marine bivahes molluscs: a hy- pothesis concerning the relationship between glycogen metabolism and gametogenesis. In: H. Barnes, editor. Proceedings of the 9th European Marine Biology Symposium. .Aberdeen: Aberdeen University Press, pp. 191-211. Gabbott. P. A. 1 976. Energy metabolism. Marine Mussel: Their Physiology and Ecology. In: B. L. Bayne. editor. London: Cambridge University Press, pp. 293-355. Gabbott. P. A. 1983. Developmental and seasonal metabolic activities in marine molluscs. In: P. W. Hochachka. editor. The Mollusca: Environ- mental biochemestry and Physiology. New York: Academic Press, pp. 165-217. Gnaiger. E. 1983. Calculation of energetic and biochemical equivalents of respyratory oxygen consumption. In: E. Gnaiger & H. Forstner. editors. Polarographic oxygen sensors Appendix C. Berlin: Springer- Verlag. pp. 337-345. Gnaiger, E. & G. Bitterlich. 1984. Proximate composition and caloric content calculated from elemental CHN analysis: a stoichiometric con- cept. Oecologia 62:289-298. Hallegraeff, G. M. 1981. Seasonal study of phytoplankton pigments and species at a coastal station off Sydney: Importance of diatoms and the nanoplankton. Mar. Biol. 61:107-118. Hawkins. A. J.. P. N. Salkeld. B. L Bayne. E. Gnaiger & D. M. Lowe. 1985. Feeding and resource allocation in the mussel Mylilus chilensis: evidence for time-average optimization. Mar. Ecol. Prog. Ser. 20:273- 287. Incze. L. S.. R. A. Lutz & L. Watling. 1980. Relationship between of en\ ironmental temperature and seston on growth and mortality of Myri- lus edulis in a temperate northern estuary. Mar. Biol. 57:147-1156. Jaraniillo, J. R. & J. M. Na\arro. 1995. Reproductive cycle of the Chilean ribbed mussel .Aulcu-oniya aier (Molma. 1782). ,/. Shellfish Res. 14: 165-171. Macdonald. B. A. & J. R. Thompson. 1985a. Influence of temperature and food a\ ailability on the ecological energetics of the giant scallop Plu- copeden nuigellanicus. I. -Growth rates of shell and somatic tissue. Mar. Ecol. Prog. Series 25:279-294. Macdonald. B. A. & J. R. Thomp.son. 1985b. Influence of temperature and food availability on the ecological energetics of the giant scallop Pla- copecten nuigellanicus. II. -Reproductive output and total production. Mar Ecology Prog. Series 25:295-303. Macdonald, B. A. & J. R. Thompson. 1986. Influence of temperature and food availability on the ecological energetics of the giant scallop Pla- copeclen nuigellanicus. Ill.-Physiological ecology, the gametogenic cycle and the scope for growth. Mar. Biol. 93:37—18. Navarro. J. M. 1990. Influence of the natural food supply on the physi- ological energetics and biochemical storage cycles of the horse mussel Modiolus modiolus (Linneaus). Ph.D. Thesis Memorial University of Newfoundland. Canada. 131 pp. Navarro. J. M., E. Clasing. G. X. Urrutia, G. Asencio. R. A. Stead & C. Herrera. 1993. Biochemical composition and nutritive value of sus- pended particulate matter over a tidal flat of southern Chile. Estuarine, Coastal and Shelf Sci. 37:59-73. Navarro. J. M. & J. R. Jaramillo. 1994. Evaluation of the available nour- ishment filtered to the organisms at the bay of 'V'aldad (43 "08'73' -W'), sur de Chile. Re^: Biol. Mar 29:57-75. Navarro. J. M. & J. R. Thompson. 1995. Seasonal fluctuations in the size spectra, biochemical composition and nutritive value of the seston available to a susoension-feeding bivalve in a subarctic environment. Mar. Ecol. Prog. Ser 125:95-106. Newell, R. I., T. Hilbish, R. Koehn & C. Newell. 1982. Temporal variation in the reproductive cycle of Mylilus edulis L. (Bivalvia:Mytilidae) from localities on the east coast of the LInited States. Biol. Bull. 162:299- 310. Newell, R, I. & J. R. Thompson. 1984. Reduced clearance rates associated with spawning in the mussel. Mylilus edulis (L.) (Bi\al\ia:Mytilidae). Mar. Biol. Ull. 5:21-33. Pieters. H., J. H. Kluytmans & D. 1. Zandee. 1980. Tissue composition and reproduction of Mylilus edulis in relation to food availability. Neili. J. Sea Res. 14:349-361. Robinson. W. E.. W. E. Wehling. M. P. Morse & G. C. Mcleod. 1981. Seasonal changes in soft-body components indices and energy stores in the Atlantic deep-sea scallops. Placopeclen nuigellanicus. Fish. Bull. 79:449-458. Ruiz. C. D. Martinez. G. Mosquera. M. Abad & J. L. Sanchez. 1992. Seasonal variations in condition, reproductive activity and biochemical composition of the flat oyster. Osirea edulis. from San Cibran (Galicia, Spain). Mar. Biol. 112:67-74. Sanders. M. J. 1966. Parasitic castration of the scallops Pecten alha (Tate) by a Bucephalid trematode. Nature 212:307-308. Sanders, M. J. & R. J. G. Lester. 1981. Further observations on a Bucepha- Effects of Environmental Factors on Tagelus domheii 10X7 lid irematode infection in scallops (Pecren alba) in Port Phillip Bay. Victoiia. Aiist. J. Mar. Freslmaier Res. 32:475-478. Sastry. A. N. 1966. Temperature effects in reproduction of the bay scallop. Acquipecten irradians (L.I. (Woods Hole). Bid. Bull. 130:1 lS-134. Sastry. A. N. & N. J. Blake. 1971. Regulation of gonad development in the bay ^caWoxi. Aeqiiipecteii inailiaits (L.). (Woods Hole). Biol. Bull. 140: 274-282. Shafee. M. S. 1981. Seasonal changes in the biochemical composition and caloric content of the black scallop Chlaims variu (L,l from Lanveoc. Bay uf Brest Oceanol. acta. 4:331-341. Soniat. T. M.. S. M. Ray & L. M. Jeffrey. 1984. Components of the seston and possible available food for oysters in Galveston Bay. Te.xas. Con- trib. Mar. Sci. 27:127-142. Soniat. T. M. & S. M. Ray. I9S5. Relationships between possible available food and the composition condition and reproductive state of oysters from Galveston Bay. Texas. Contrih. Mar. Sci.. Univ. Te.x. 28:109-121. Stead, R. A., E. Clasing. J. M. Navarro & G. Asencio. 1997. Reproductive cycle and cohort formation of Venus antiqua (Bivalvia: Veneridae) in the intertidal zone of southern Chile. Rev. Chilena de Historia Natural. 70:181-190. Strickland, J. D. H. & T. R. Parsons. 1972. A practical handbook of seawater analysis. Bull. Fish. Res. Bel. Can. 167:1-310. Sundent, J. H. & O. Vahl. 1981. Seasonal changes in dry weight and biochemical composition of the tissues of sexually mature and imma- ture Iceland scallops, Chlamys isUuulica. J. Mar. Biol. .\ss. UK. 61: 1001-1010. Taylor. A. C. & T. J. Venn. 1979. Sea.sonal variation in weight and bio- chemical composition of the tissues of the queen scallops. Chlamys opercularis. from the Clyde Sea area. J. Mar. Biol, .\ssoc. UK. 59: 605-621. Thompson. R. J. 1972. Feeding and metabolism in the mussel Myiilus edulis. L. PhD. Thesis, University of Leicester. Thompson. R. J. 1977. Blood chemistry, biochemical composition, and the annual reproductive cycle in the giant scallops Placopecten magellaiii- cus. from Southeast Newfoundland. J. Fish. Res. Board. Can. 34:2 10-t- 2116. Thompson. R. J. 1984. The reproductive cycle and physiological ecology of the mussel Mvtilus edulis in a subartic. non-estuarine environmenl, (Berlin). Mar. Biol. 79:277-288. Thompson. R. J. & B. L. Bayne. 1974. Some relationships between growth, metabolism and food in the mussel Myiilus edulis. Mar. Biol. 27:317- 326. Thompson, R. J. & B. A. Macdonald. 1990. The role of environmental conditions in the seasonal synthesis and utilization of biochemical en- ergy reserves in the gigant scallop. Placopecten ntai;cllanicus. Can. J. Zool. 68:750-756. Trevallion, A. 1971. Studies on Tellina tenuis (da Costa). IIL- Aspects of general biology and energy flow. J. Exp. Mar. Biol. Ecol. 7:95-122. Toro, J. E.. P. I. Paredes. D. J. Villagra & C. M. Senn. 1999. Seasonal variation in the phytoplanklonic community, seston and environmental variables during a 2 years period and oyster growth at two maricultuie sites, southern Chile. Mar. Ecol. 20:63-89. Vahl. O. 1980. Seasonal variations in seston and the growth rale of the Iceland scallop, Chlamys isUmdica (0. F. Miiller) from Balstjord. 70°N. J. Exp. Mar. Biol. Ecol. 48:195-204. Welschmeyer, N. A. & C. J. Lorenzen, 1985. Role of herbivory in con- trolling phytoplankton abundance: annual pigment budget for a tem- perate marme fjord. Mar. Biol. 90:75-85. Yafiez. Y. 1998. Descripcion y efecto de la prevalencia de un digeneo sobre tejido somaticoy gonadal de T. (/wii/ie;/ (Lamarck. 1818). Thesis. School of Biologia Marina. Facultad de Ciencias. Universidad Austral de Chile. Valdivia. Chile. Zandee, D. I.. J. H. Kluymans. W. Zurhurg & H. Pieters. 1980. Seasonal variations in biochemical composition of Mytilus edulis with reference to energy metabolism and gametogenesis. Neth. J. Sea Res. 14:1-29. Joiifihil nl Shellfish Research. VciK 20. No. .^. 1()S4-I(W4. :(){11. ENERGETICS VARIATION OF THE STRIPED CLAM HURHOMALEA EXALBIDA (CHEMNITZ, 1795) IN USHDAIA BAY, BEAGLE CHANNEL (54 50'S) BETINA J. LOMOVASKY,' * ELBA MORRICONL' AND JORGE CALVO' " ^Centra Austral cle Invesligadones Cientificas (CADIC-CONICET).C.C.92 - Ushiuiia (VQ4IOBFD).Tieira del Fiicf-o. Arficntiiui: 'UNPSJB. Sale Usiniuui ABSTRACT The processes of energy Iranster in Ihe clam Eurhamatea exalbida Chemnitz. 17').^ can be analyzed by measuring the seasonal variations in energy content with respect to the reproductive characteristics of the population. The energy content of three organ groups of £. exalbida (gonad, digestive gland and gastrointestinal tract-foot (FV); adductor muscles (M); and gills-mantle- siphons (CMS), was measured using calorimetry in monthly samples of the Ushuaia Bay population for a year. The relative condition index was also measured monthly for the three organ groups. No relationship was found between the energy content per gram (kJ/g ash free dry weight) of the organ groups FV. M and GM,S and the si/.e of the adult organisms. The mean values for Ihe twelve month experimental period for FV were very similar for males and females with 20.55 ± 1 . 1 0 kJ/g for males and 20.26 + 0.9 1 kJ/g for females: for M they were 20.6S ± 0.65 kJ/g for males and 20.76 ± 0.76 kJ/g for females; and for CMS they were less than for the other two organ groups, with values of 20.02 ± 1.09 kJ/g for males and 20.02 ± 0.84 kJ/g for females. An analysis of the energy content data and the relative condition index (RCI) measured at monthly intervals throughout the year revealed significantly lower values for FV in November ip < 0.05), which is believed to be linked to the greater gametic emission in this month. In contrast, no significant differences over time were found in the energy content or relative condition index of either M or GMS. The relative stability of the energy content and RCI of M and GMS over time suggests that there were no changes in the proportion of their biochemical components and no changes in the mass of the organs, hence Eurhomalea exalbida does not have an energy reservoir and the energy used in reproduction may be extracted directly from the food consumed. KEY WORDS: energetics, clams, calorimetry, Eurhoiiialea e\alhida. bivalve. Beagle Channel INTRODUCTION An understanding of the energy balance iif an organism is based on knowledge of the temporal variation in energy acquisi- tion and its use in maintenance, growth and reproduction. Physi- ological processes in temperate and cold water bivalves are also strongly influenced by seasonal variations in temperature and the availability of food (Mann 1979, Newell & Branch 1980. Smaal et al. 1997), thus patterns of growth and reproduction vary seasonally for each species. Variation in these activities has been shown to be correlated with changes in the energy content of different organs (Jobling 1994. Lucas 1996), thereby reflecting the spatial and tem- poral energy distribution within the organism. This distribution closely reflects the seasonality of the cost of reproduction and the capacity to accumulate reserves (Zandee et al. 19iS(), Robinson et al. 1981, Navarro & Torrijos 1995. Martinez & Mettifogo 1998). Many investigations have examined variations in the energy content of different bivalve organs in relation to reproductive cycles, growth and basal metabolism. In such studies, the bio- chemical analysis of tissues (Beukema & De Bruin 1977, Beukema & De Bruin 1 979, Zandee et al. 1 980, Robinson et al. 1 98 1 , Sundet & Vahl 1981, Herald & Deslous-Paoli 1983, Martinez & Metti- fogo 1998), methods involving measures of the metabolic rates of live organisms (Newell & Branch 1980, MacDonald & Thompson 1986, MacDonald & Bourne 1987, Widdows & Johnson 1988, Sukhotin 1992, Smaal et al. 1997) and dii-ect measures of caloric content by complete combustion (Beukema & De Bruin 1979, Griffiths & King 1979, Herald & Deslous-Paoli 198,^) have been used. Bomb calorimetry is an accurate method for measuring the energy content in aquatic organisms (Beukema & De Bruin 1979, Hartman & Brandt 1995, Lucas 1996), Eurhomalea exalbida (Chemnitz 1795) is a species with a wide geographic distribution, from the Beagle Channel (54-50'S) to the *Corresponding author. E-mail: betinaKS'hotmail.com island of Chiloe (42 Si in Ihe Pacific littoral zone (Soot-Ryen 1959. Dell 1964, Osorioetal. 1979) and to the province of Buenos Aires (36° S) in the Atlantic littoral zone (Carcelles 1944, Car- celles 1950). E. exalbida is commercially exploited in various regions, although management practices are not generally used as very little is known of the clam's biology and the environmental factors that influence its populations. The Beagle Channel popu- lation represents the extreme south of the species' distribution, withstanding large temperature variations (4'C-11"C), with inarked seasonal variation in the biomass of phytoplankton (Her- nando, pars. comm.). These conditions may cause temporal fluc- tuations in the energy content of this population. The purpose of this study was to analyze the temporal variation in the energy content of different organs in E. exalbida in relation to its rep]oducti\e cycle. METHODS Samples were collected in Ushuaia Bay (54°50'S, Beagle Channel), in a subtidal flat with a depth between two and four meters at low tide. Monthly sampling was conducted by SCUBA diving from October 1998 to September 1999. Monthly mean sur- face temperature was recorded. Clams with shell lengths greater than 38 mm (n = 246) were caught and kept in aquaria for 24 hours. Shell length (anterior-posterior axis, SL), measured with electronic calipers to the nearest 0.01 mm, and total weight (TW, ±0.1gl were recorded for each individual. Calorimetric satnples were identified and stored at -20" C until piocessing. After removing the valves, the sex of each individual was determined using gonad smears and the soft parts were sepa- rated into 3 groups: foot-visceral mass (gonad, digestive gland and gastrointestinal tract) (FV), adductor muscles (M), and gills- mantle-siphons (GMS). These di\'isions were made based on the presence or absence of gonads in the tissues and on their potential capacity to fulfill a storage function: FV includes all the tissues where gonads are present: M plays an important role as a reserve 1089 1090 LOMOVASKY ET AL. organ in some bivalve species: and CMS is made up of all the other tissues where gonads are not present. Although it was not possible to separate the gonad tissue from the visceral mass, the calorimetric analysis of the soft parts separated as described above was considered to give more information about the energy balance of the species than an analysis of the organism as a whole. The soft parts were dried at 80 C to constant weight. The relationships between shell length and FV dry mass. M dry mass and GMS free dry mass are represented by 1 ) log (Mass) = a + b X (log (SL)). Monthly condition values were analyzed using 2) condition index (CI) = Mass/Shell length '\ where b is the slope in (1). and were normalized using 3) (RCI) = (CI - mean CI)/ (S.D. of CI) to give a I'clative condition index. The caloric contents of approximately 20 individuals per month were obtained by burning pellets (25-200 mg) in a Parr model 1425 micro-bomb calorimeter, as described by Beukema and De Bruin (1979) and Lucas (1996). The values obtained were cor- rected for ash and acid contents and were expressed as kJ/g ash free dry weight (AFDW). Stalistical A nalysis A regression analysis was peiformed with the energy content of the different organ groups as the dependent variable and the size of the organisms (SL) as the independent variable. The monthly differences in relative condition index and energy content in each of the groups of organs (FV. M and GMS) were analyzed using an analysis of variance (ANOVA). The assump- tions of normality and homogeneity of variances were tested and the appropriate transformations were applied when necessary. When this was not possible, a non-paratnetric test was used (Kruskal-Wallis). LInplanned comparisons were made (STATISTICA softwaie) when significant differences were found. After verifying Tukey's assumption of non-additivity (BIOM-STAT software), an analysis of variance (ANOVA) for randomized block design (RBD) was used to determine if there were diffei-ences in energy content (kJ/g AFDW) among FV. M and GMS for each month and both sexes. When significant dif- ferences were found among organ groups, unplanned comparisons were made (Zar 1984. Sokal & Rohlf 1995). RESULTS An analysis of the relative condition index for FV o\er time showed the highest values in January and February in both sexes, whereas the lowest values occurred in November and April for fetnales and November and August for males. These results were highly significant (one-way ANOVA; F^, ,,,,„ n ,,,1, = 2.84/7 < 0.002 females and F^,^ II „5 ,, n, = 3.96 /;< 0.001 males: modi- fied Tukey test, p < 0.05). Theses results coincide with the game- togenic cycle observed for this species (Morriconi et al. unpubl.). No significant changes in the lelative condition index over the experimental period were found for either M or GMS in either sex (one-way ANOVA. p > 0.05, Fig. I ). The minimum mean seawater temperature (4.5°C) was recorded in August and the maximum (8.7°C) was recoi'ded in January (Fig. 2). Linear regressions were used to determine the possible relation between the energy content (kJ/g AFDW) of the three organ groups (FV. M and GMS) and the size of the clams (SL). The results were not significant (/> > 0.05) for FV (R- = 0.001 ) and GMS (R- = 0.0002) and wei'e significant for M although only 6'/r of the total variation is explained by the regression (R" = 0.06). o 4 3 - 2 1 0 -1 -2 -3 -4 4 3 2 1 0 -1 -2 -3 -4 4 3 2 1 0 -1 ■2 -3 Foot-Visceral mass Adductor muscle — • — Males Females Gills-mantle-siphons Males Females Sep Oct Nov Dec Jan FebMar Apr May Jun Jul Aug Sep Oct 1998 1999 Figure I. Mean and standard error of the relati\e condition index for fool-\isceral mass, adductor muscles and );ills-niantle-siphons o\er the experimental period lor females and males of £. exalbida. As caloric content per gram was found not to be related to the size of the organisms analyzed in this study, an analysis of variance (ANOVA) was used to analyze the variations in energy content. Foot- Visceral Mass (gonad, digestive gland and gastrointestinal tract) The mean energy content of FV over the experimental period for females (n = 121) and males (n = 125) were very similar (Figs. 3a and 3b respectively), at 20.26 ± 0.91 kJ/g AFDW for females and 20.55 ±1.10 kJ/g AFDW for males. The differences among the months were highly significant (one-way ANOVA) in both males (F„ „,,„;; n ,,, = 4.32. /) < 0.001) and females (F^^,,,,, ,,, i.w = -'*-^9. p < 0.001). It was shown in the unplanned comparisons (modified Tukey test) that the energv content for females was significantly lower in Novem- ber and March (/) < 0.05). and in October and February (the months prior to the decrease) the maximum energy content was observed although statistical significance was not obtained. The t)nly significant variation in energy content among males occurred in November and the energy content was significantly less (/) < Energetics of Eurhomalea exalbida 1091 2 3 2 0) a E Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct 1998 1999 Figure 2. Monthly means of sea Hater temperature in I'shuaia Bay, Beagle Channel. 0.05) ih.iii 111 the rest ol the year. Another decrease oeetirred in March, and in the niniuhs before these declines the greatest \aiues were iihtained. Aitdiicliir Muscle The mean energy content of M over the experimental period (Fig. 4a. Fig. 4b) was 20.76 + 0.76 kJ/g AFDW (n = 121) for 25 1 24 23 22 21 20 19 18 17 16 -I < o < 15 25 24 23 22 21 20 19 18 17 16 15 Ml,ul|( I', till SONDJ FMAMJ JASO 1998 1999 Figure .'. Mean and standard error of the energy content for fool- \isceral mass (kJ/g .\FD\\ ) over the experimental period for (a) fe- males and (b) males of E. exalbida. Q U. < 3 Q < 25 24 23 22 21 20 19 18 17 j 16 15 24 23 22 21 20 19 18 17 16 15 I H •••••••• ^ t^jun^M** SONDJ FMAMJ JASO 1998 1999 Figure 4. Mean and standard error of the energy content for adductor muscles (kJ/g AFDW) over the experimental period for (al females and (b) males of E. exalbida. tcinales and 20.68 ± 0.65 kJ/g AFDW (n = 125) for males. No significant differences were found in the energy content of M attributable to either sex (F„„„,„ ,212 = 0.08, p > 0.05). time (F,< = iiii.s. 11.222 = 0.77. /)> 0.05). or their interaction (Fj,^ II 1,5 i, 22: = (1.47. /) > 0.05) (two-way ANOVA). Gills, Mantle, Siphons The mean energy content of GMS over the experimental period (Fig. 5) was 20.02 ± 0.84 kJ/g AFDW for females (n = 121 ) and 20,02 ± 1.04 kJ/g AFDW for males (n = 125). No significant difference was found in the GMS energy content for males over time (Kruskal-Wallis; H = 16.29, p = 0.1.^07). In contrast, for females there were significant differences in energy content (one- way ANOVA, F„_„,|3 I, |„y = 2.58. /; < 0,01 ), with the lowest values in December and April and the maximum value in June (modified Tukey test: p = 0.05). Variation AniDiif; ()ri;aii Groups The monthly measures of energy content in the different groups of organs in both sexes (Fig. 3, Fig. 4. Fig. 5) show a marked stability in energy content for M, while for FV and GMS the variation was greater. A one-way ANOVA for RBD revealed sig- nificant differences {p < 0.05) among the energy content of FV, M and GMS (Table 1 ). This result suggests a difference in the bio- chemical composition of the organ groups. In order to determine the organs responsible for these differ- 1092 LOMOVASKY ET AL. 5 Q < Q < D) 25 24 23 22 21 20 19 18 17 16 15 25 24 23 22 21 20 19 18 17 16 15 K. M i n I M M O N D 1998 J F M A M J J A S 1999 O Figure 5. Mean and standard error of the energy content lor gills- niantle-siphons (kj/g AFDW) over the experimental period for (a) females and (h) males of £. exalbida. ences. unplanned comparisons were made using the Tukey test. The energy content tor both sexes was lower tor CMS than for FV and M for most of the year, and in many instances these differ- ences achieved statistical significance (Table 2; p < 0.05). The greatest difference in energy content occun'ed in November and TABLE 1. Results of ANOVA (RBDl on the differences in energy content (kJ/g AFDW) among foot-visceral mass (FV), adductor muscle (M) and gills-niantle-siphons (CMS) of £. exalbida in Ushuaia Bay. Males Females Date October 1998 November 1998 December 1998 January 1999 February 1999 March 1999 April 1999 May 1999 June 1999 July 1999 August 1999 September 1999 9.14 0.0012* 2.62 0.1133 4.47 0.0248* 8.02 0.0027* 3.38 0.1382 5.09 0.0135* 3.54 0.0424* 5.81 0.0088* 2.16 0.1377 0.39 0.6817 0.41 0.6698 0.38 0.0059* 3.42 0.0553 4.56 0.0476^' 11.57 0.0003* 4.17 0.0307* 0.31 0.7382 7.46 0.0062* 3.13 ().()X(14 0.21 0.8 i:i) 0.13 0.S603 4.55 0.0235* 5.25 0.0136* 15.23 0.0000* March and is attributable to the large decrease in FV energy con- tent, with lower values than for M and GMS. DISCUSSION Calorimetric analysis of the energy content of different organs allows the proces.ses of energy transfer at the individual and popu- lation levels to be understood for a given species, yet such studies in clams are scarce. In E. exalhidu. as in clams in general, it was not possible to separate each organ, and the organ groupings used here may mask some differences among the individual organs. Notwithstanding, the results obtained have yielded valuable infor- mation about annual energy variations in specific locations w ithin the clam. In this study off. exalbida no significant regression was found between energy content per gram and the size of the individuals being studied for FV and GMS, and for M the regression was significant although only 6% of the variation in energy content was explained by size. As the amount of energy used for growth and reproduction is different for juveniles and adults (Griffiths & King 1979. Heral & Deslous-Paoli 1983. Smaal et al. 1997). only adult individuals larger than the size at first sexual maturity (SL = 40 mm) (Morriconi et al. unpubl.) were used. Thus, the energy avail- able for different metabolic processes per gram of tissue (AFDW) in sexually mature individuals in this study was independent of their sizes. The energy content measured (Figs. 3. 4 and 5) in this study is consistent with the range indicated in the literattire (Brey et al. 1988). In indi\ iduals of both sexes, the gonads contained mature ga- metes throughout the year, indicating partial spawning (Lo- movasky et al. 2000). Coinciding with a decrease in gametic abun- dance in November (Lomovasky et al. 2000). declines in the rela- tive condition index (Fig. 1) and energy content (Fig. .3) of FV were observed in both males and females. Although the decline in the energy content of FV cannot be unequivocally attributed to an increase in sexual activity, it should not be rejected as a likely cause as the proportion of lipid content in the gonads can be expected to decrease due to the emission of gametes. The increase in the relative condition index and the increase in the energy con- tent during the summer months (December. January. February) TABLE 2. Tukey Test for unplanned comparisons among the energy content (kJ/g .AFDW I of foot-visceral mass (FV). adductor muscle (Ml and gills-mantle-siphons (GMS) for males and females of E. exalbida in Ushuaia Bay. Only months with significant differences are shown (p<0.05). Males Females FV M FV M GMS M F = calculated F statistic; p ference with a = 0.05. probability level and (*) signillcarK dil- October October December December December January January January March March May May July September September August Septenibei March November March September Energetics oe Eurhomalea exalhida 1093 may be directly related to the increase in the number of mature yametes in these months (Morriconi et al. unpubi.). The higher surface temperatures in the sea and the greater avaihibihty of phy- toplanlif;a.\ estlmee par mesures microcalorimetrlques et par dosages biochimiques. Oeeanologica Acta. 6:193-199. Jayabal. R. & M. Kalyani. 1986. Biochemical Studies in the Hard Clam Meretrix merelrix L from Vellar Estuary, East Coast of India. Indian J. Mar Sei. 15:63-64. Jobling. M. 1994. Fish Bioenergetics. London: Chapman and Hall Press. 309 pp. Lomovasky. B. J.. E. Mon-iconi & J. Calvo. 2000. Variacion energetica relacionada con el ciclo reproduclivo de la alnieja rayada Eurhomalea exalbida de la Bahia Ushaaia. Canal Beagle. In proceeding of IV Jor- nadas. Nacionales de Ciencias del Mar: Puerto Madryn. Argentina. 82 PP Lucas. .\. 1996. Bioenergetics of aquatic anunals. London: Taylor and Francis Press. 169 pp. MacDonald. B. A. & R. J. Thompson. 1986. Influence of temperature and food availability on the ecological energetics of the giant scallop Pla- copecten magellanicus. III. Physiological ecology, the gametogenic cycle and scope for growth. Mar. Biol. 93:37—18. MacDonald. B. A. & N. F. Bourne. 1987. Growth, reproductive output, and energy partitioning in weathervane scallops. Paiinopecten caurinns, from British Columbia. Can. J. Fish. Aquat. Sei. 44:152-160. Mann. R. 1979. The effect of temperature on growth, physiology, and gamelogenesis in the manila clam Tapes philippinarum (.\dams & Reeve. 1850). J. E.vp. Mar Biol. Eeol. 38:121-133. Martinez. G. & L. Mettifogo. 1998. Mobilization of energy Irom adductor muscle for gametogenesis of the scallop, .Argopecten purpiiratus Lamarck. / Shellll.di Res. 17:1 13-1 16. Navarro, E.. J. I. P. Iglesias & A. Larranaga. 1989. Interannual variation in the reproductive cycle and biochemical composition of the cockle Cerastodenna edule from Miuidaca Estuary (Biscay. North Spain). Mar Biol. 101:503-511. Navarro. J. M. & R. A. Torrijos. 1995, Fisiologi'a energetica de Conchole- pas concholepas (Bruguiere, 1789) (Gastropoda: Muricidae) en la ba- hfa de Yaldad. sur de Chile. Rev. Chil. Hist. Nat. 68:61-77. Newell. R. C. & G. M. Branch. 1980. The influence of temperature on the maintenance of metabolic energy balance in marine in\ ertebrales. Adv. Mar Biol. 17:329-396. Osorio. C. J. A. Cifuentes & S. M. Fischer. 1979, Moluscos marinos de importancia economica en Chile. Biol. Pesq. Chile 1 1:.^— 17. Racotta. I. S.. J. L. Ramirez, S. Avila & A. M. Ibarra. 1998, Biochemical composition of gonad and muscle in the catarina scallop. Argopecten ventrieosus. after reproductive conditioning under twii feeding systems. Aquacnilure 163:111-122. Robinson. W. E.. W. E. Wehling. M. P. Morse & G. C. McLeod. 1981. Seasonal changes in soft-body component indices and energy reserves in the Atlantic deep-sea scallop. Plaeopecten Magellanicus. Fish. Bull. 79:449-+58. Smaal. A.C.. A. P. M. A. Vonck & M, Bakker, 1997. Sea.sonaI variation in physiological energetics of Mytilus ednlis and Cerastoderma edule of dilferent .size classes. ,/. Mar. Biol. Ass. UK. 77:S17-S3S. 1094 LOMOVASKY ET AL. Sokal. R. R. & F. J. Rohlf. \995. Biometrv: The principles and practice of statistics in biological research. New York: W. H. Freeman and Com- pany. 887 pp. Soot-Ryen, T. 1959. Pelecypoda. Acta Univ. Lmul. Avd.2. 55:1-86. Sukhotin. A. A. 1992, Respiration and energetics in mussels {Myliliis cdu- lis L.) cultured in the White Sea. Aqiiaciilliire 101:41-57. Sundet, J. H. & O. Vahl. 1981. Seasonal changes in dry weight and bio- chemical composition of the tissues of sexually mature and immature Iceland scallops. Chlamvs islandica. J. Mm: Biol. A,?.v. UK. 61:1001- 1010. Widdows, J. & D. Johnson. 1988. Physiological energetics of Mytilus ediilis. Scope for growth. Mar. Ecol. Prog. Ser. 46:113-121. Zandee, D. I., J. H. Kluytmans. W. Zurburg & H. Pieters. 1980. Seasonal variations in biochemical composition of Mytilus edulis with reference to energy metabolism and gametogenesis. Neth. J. Sea Res. 14:1-29. Zar. J. H. 1984. Biostatistical analysis. Englewood Cliffs: Prentice-Hall, Inc. 718 pp. Journal oj Slwlllisli Rcscairh. Vi.l. 31). No. ?. 1(W5-1()W. :(1()1. BROODSTOCK MAINTENANCE AND EARLY GONADAL MATURATION OF PHOLAS ORIENTALIS (BIVALVIA: PHOLADIDAE) E. T. MARASIGAN AND L. V. LAURETA Institute of Aquaculture. CoUej^e of Fisheries and Ocean Sciences. Universiiv of the i'hilippines in tlie Visayas. Miciiiuo. Ilo'do. Pliilippines 502.^ ABSTRACT The angelwing?.. Pluilus oncnhiiis were maintained and conditioned in indoor tanks/basins. They showed high survival in both static s)stem and flow-througli system proxided with a muddy sand substrate. Enhanced gonad development and earher maturation were observed for angelwings fed a mixture of Cluieloceros calchram and Terraselmis suecicci al a rate of about 294 milMon cells per broodslock per day. Angelwings fed a mi,xture of two algal species spawned one to two months earlier than those led one algal species. Results of the study demonstrated the hatchery potentials of angelwings. KEY WUKDS: I'lwlas (inviiuilis, angelwings. broodstock. gonadal maturation INTRODUCTION Several years ago, the high demand for angelwings. Phohis (Monotlnia) (inciualis locally ktiown as diwal in Patiay and Ne- grcs Islands in Central Philippines resulted in the overexploitation of the species. The once dense beds are now almost devoid of this resource. Certain actions or interventions must be undertaken so that these areas would again become an important sustainable source of income for coastal fishermen. The most active strategy in assisting improvement and reha- bilitation of coastal mollusk fisheries is through stocking of aili- ficially produced seedlings. This has been demonstrated in many eastern Asian maritime countries. In Japan, the production of HaH- Otis spp.. Pciphia spp., and Pecteii sp. have increased tremendously (Honma 1980, Inoue 1984, Oshima 1984). In Taiwan, increased commercial productions were attained for shellfishes like Meretiix lussoriii. Gdinpliiiia venehfonnes. Haliotis diversicolor var. supei- texta. Soleteiliiia diphos and tsnadani granosa (Chen & Yang 1979. Chen 1984. Ting 1984). The Philippines lag behind in techniques of seed production and restocking of coastal areas to increase and sustain shellfish production compared to its neighboring countries. The main reason is the absence of molluscan hatchery, which is crucial and comple- mentary to a rehabilitation program. We view that even hatchery- related studies are limited and mainly focused on oysters, mussels and scallops. Research on broodstock maintenance, spawning and nursery techniques for endemic bivalve species such as angelwings has not been given priority. This study is the first effort to determine the hatchery potential of angelwings, P. ohenialis. The feasibility of maintaining spawn- ers of this species under laboi-atory conditions was evaluated. Fur- ther, we tested the influence of different types of algal species to bring about gonadal maturation and spawning in the laboratory. MATERIALS AND METHOD Spawiier Maiulcnance Adult angelwings (50-130 mm in shell length) were collected from the intertidal flats of Barotac Nuevo. lloilo in Central Phil- ippines (122"47'N and 10°55'E) that earlier was described by Laureta and Marasigan (2000). The specitnens were scrubbed free of fouling organisms and debris prior to stocking. Two holding techniques were evaluated for spawner mainte- nance in the laboratory: (a) static water system with aeration and without substrate sediments, and (b) flow-through water system with muddy sand substrate. All incoming water was filtered through sand. Three 40-1 capacity circular white basins were used in the static system and three 30-1 capacity fiberglass tanks were used in the flow-through system. Ten pieces of angelwings were placed in each container. The bivalves were laid horizontally in static systems and those in the tlow-through systems were buried in the muddy sand substrate. During the two months of holding, the angelwings were fed either Cluwtoceros calcitrans. Tetniscl- mis siiecica or their combination. The algae were provided to the angelwings in the morning and afternoon at the rate of about 100 million cells per adult bivalve per feeding. The angelwings were reared in ambient light and temperature (range: 24-27°C) condi- tions of the labot-atory with minimal disturbance. Survival of an- gelwings was evaluated after two months. Survival of the angel- wings between the two holding techniques was compared using analysis of variance (Zar 1984). Gonadal Development Experimental Animals and Set-up A total of 160 angelwings (>70 mm shell length) collected at five different dates in Barotac Nuevo were scrubbed free of fouling organisms and acclitnated to feed on test phytoplankton species, T. suecica or C. calcitrans. Ten specimens were dissected and sub- jected to histological preparations to determine angelwings' stages of maturity and gametogenic cycles before the onset of the feeding experiment. Permanent mounts of the gonads were prepared fol- lowing the inodified Bell and Lightner (1989) method. The stages of gonadal development were described following stages for other clams (Rossell 1979. Jones 1981. Hesselman et al. 1989. Shafee & Daoudi 1991, Panurovsky & Yakovlev 1992) and the earlier de- scription on reproductive stages of the angelwings by Laureta and Marasigan (2000). All ten specimens were found to be spent. Fifteen cylindrical concrete tanks (diameter = 0.5 m; height = 0.75 m) with standpipes at the center filled with 0.3 m compact muddy sand and supplied with seawater to a maximum depth ot 0. 1 7 m by means of a flow-through water system, were used in this study. Each tank was stocked with ten angelwings by burying them in the muddy sand substrate. Water physico-chemical parameters in the culture tanks were monitored regularly. Water temperature was measured using mercury thermometer and salinity was moni- tored using Atago refractometer. 1095 1096 Marasigan and Laureta Experimental Treatments and Protocol The stud) tesled the effects of different algal food on the go- nadal maturation of angelwings. The algae used were (a) C. cal- cilrans: (b) T. siiecica: and (c) combination ( 1 : 1 ) of the two algal species. Algae were given to the angelwings twice a day (morning and afternoon) at the rate of approximately 1.47 x 10" cells tank"' feeding"' or 2.94 x 10" cells tank"' day"'. Since 10 angelwings were .stocked in each tank, each angelwing was presumed to re- ceive 294 million cells day"' during the first month. The daily provision of algal food for each tank remained the same througli- out the experiment but algal food allotment tor each angelwing increased every month because of decreased number of angel- wings in each tank due to sacrifice sampling. Each of these feeding regimes was replicated five times and assigned at random to the 15 concrete tanks. Water flow was stopped for about two hours during feeding. Evaluation of (ionadal Development Gonad developments of angelwings fed algal diets were moni- tored every month for five months. One angelwing v\'as extracted and dissected from each tank representing five replicate samples for each type of algal food. The histological preparation and method of determining gametogenic cycles as described earlier were used. The gonadal stages were: early active, late active, ripe, partially spent and spent. The percentage distribution of gonadal stages in each treatment diet was calculated. The gonad index was estimated by assigning rank for each stage modified from Wilson ( 1987) and King et al. (1989): 3 for early active. 4 for late active. 5 for ripe. 3 for partially spent, and 1 for spent. The gonad index of angelwings in each treatment diet per monthly observation is the sum of products of percentage of each stage (PJ and its rank (R^,. giving value from 100 (all spent) to .500 (all ripe). Gl = P, X R^ RESULTS Spawner Maiiilenaiice High sur\i\al of angelwings was obtained for broodstocks maintained in both .static system with aeration (90 ± 0.0%) and the flow-through system with muddy sand substrate (93.3 ± 5.8%). There was no significant difference (p > 0.05) between the two systems. The spawners in both holding containers were agile and active as exhibited by constant protrusion of their siphons espe- cially during feeding time. Gonad Development The percentage distribution of gonadal stages and the gonad index of angelwings after feeding different algal species in the laboratory are shown in Table 1 and Figure 1 . The same table contains Laureta and Marasigan's (2000) percentage distribution of the combined gonadal stages of male and female angelwings collected at different sampling dates in their natural habitat for comparison purposes with the above. Gametogenesis started in angelwings indoor one month (Jan.) after the start of the experi- TABLE L Percentage distribution of gonadal stages off. orienlalis fed Chaetoeenis ealeilrans (Cc), Telraselmis sueeiea (Ts) and their combination (CT) in the laboratory, and samples from a natural population *(Barotac Nuevo, lloilo. Central Philippines). (lOnadal Stages Partially Date Site Diet N Early Active Late Active Ripe Spent Spent Dec. '94 1 iitloor Cc 5 0 0 0 0 100 Ts 5 0 0 0 0 100 CT 5 0 0 0 0 100 Barotiic Nuevo 15 20 47 27 6 0 Jan. 9.S indoor Cc 5 20 0 0 0 80 Ts 5 20 0 0 0 80 CT 5 20 0 0 0 80 B^iiotiiC Nuevo 14 29 50 21 0 0 Feb. '95 indoor Cc 5 60 0 0 0 40 Ts 5 60 0 0 20 20 CT 5 60 20 0 0 20 Barotac Nuevo 12 16 42 42 0 0 Mai. '9.^ indoor Cc 5 40 60 0 0 0 Ts 5 40 60 0 0 0 CT 5 0 40 60 0 0 Barotac Nuevo 9 0 44 56 0 0 Apr. '95 indoor Cc 5 0 80 20 0 0 Ts 5 0 60 40 0 0 CT ? 0 0 80 20 0 Barotac Nuevo 6 0 33 67 0 0 May -95 indoor Cc 5 0 20 60 20 0 Ts 5 0 0 80 20 0 CT 5 0 0 40 40 20 Barotac Nuevo 16 n 25 56 19 0 * Data extracted and recomputed from Laiirela and Marasigan 1 2000). Hatchery Potentials of Angelwings 1097 500 450 400 350 300 250 • 200 ■ 150 ■ 100 ■ 50 ■ 0 -Cc -Is -CT -BN D94 J95 A95 M95 F95 M95 MONTH Fiyurt' I. Cioiiad indtx of /'. oricnlalis led Chueloecnis calcilrans (Cc). Telraselmis suecica (Tsl and their combination (CT), and samples t'nini a natural popniation in Barotac Neiivo, Iloilo. Central Phil- liplnes. mem when 20'^r of the animals showed early active development. As the conditioning progressed from February through May. there was an increased percentage of mature angelwings that spawned later. Angelwings fed combination of T. siwcicn and C culcitriins showed faster gametogenesis. earlier maturation and spawning than angelwings fed one algal species. Sixty percent of the angel- wings fed the mixed algal diet had ripe gonads by March while majority of angelwings fed either C. calcitrans or T. suecica did not attain the same gonadal stage until May. Angelwings fed com- bined algal species began spawning in April and continued in May while angelwings fed one algal species did not spawn until May. The highest histological gonad index of 460 was attained in March b\ those fed the combined species and in May by those fed single algal species Angelwings in the natural habitat had ad\anccd gonadal stages (Table 1 ) and higher gonad index (Fig. 1 ) compared to angelwings in indoor tanks at the start of the e.xperiment. Although a similar proportion was ripe in the laboratory and the field in March, an- gelwings in indoor tanks spawned one month earlier (Apr.) than in natural habitat. Some of the angelwings fed combined algal species were already spent in May while in the natural population were onh partially spent. Water I'hysicii-clumicul Parameters Water temperature in the culture tanks ranged from 24-26°C and salinity ranged from 31 to 33 (%c) (Fig. 2). No marked fluc- tuation was observed in the xalues of the parameters obtained during the culture period. DISCUSSION In this study, we reared and carried out gonad,il niatuiation ot angelwings in the laboratorx. The similarity of sur\i\al rates be- tween angelwings reared in circular basins without mud substrate and angelwings in tanks with mud substrate that accommodate the burrowing behavior of the species could be attributed to the mini- mal disturbance in both holding systems. Both systems minimized stressful conditions to the angelwings resulting in high sur\i\al and relatively faster gonadal development. Optimum holding con- ditions such as filtered seawater, gentle aeration, reduced water turbulence and ample supply of algal food enabled the bivalves to be active during the daytime for feeding as evidenced by continu- (°C) 27 26 5- 26 255 25 24 5- 24 • 235 (%o) 40 • 30 20 10 D94 J95 F95 WI95 A95 M95 MONTH Figure 2. Variations in the temi)erature and sahnity readings water in the indoor tanks from December 1W4 to Ma> 1995. of the ous protrusion of their feeding siphons. In their natural habitat, the angelwings are subjected to en\ ironmental perturbations brought about by substrate instability caused by water turbulence, tidal fluctuations and fishing activities. In the field, we observed that angelw ings in the coasts of Barotac Nuevo moved up and extended their siphons out of the sediment only during calm periods to feed. If dislodged by fishing i.e., trawls and dredges or turbulence due to typhoons, the angelwings could no longer get back to their bur- rows/homes (Ablan 1938, Laureta & Marasigan 2000). Further, we also noted that they were unable to create another buiTOW by themselves once exposed, thereby subjecting them either to physi- cal stress such as water current and strong waves that usually damage and break their shells or predation. Our results showed that angelw ings could be conditioned to an earlier gametogenesis by feeding them high ration of single-cell algal diets. Earlier studies were successful in conditioning bivahes to sexual maturity through manipulation of temperature and in- creased food supply (Loosanoff & Davis 1963, Paon & Kenching- ton 1995). Apparently, the two types of algal food at a high ration fed twice a day per adult angelwing possibly satisfied the energy demanding process of gonadal maturation in shellfish (Sastry 1979). In an eariier study, Sastry (1968) reported that post- spawning adults of Argopecten irnulttius required an abundant food supply for initiation of gonadal growth and gametogenesis. The ration of 294 million cells/animal/day used in our study was higher than the highest feeding ration i.e., 67 million cells/animal/ day of Chaetoceros gracilis and c-iso used by Villalaz (1994) to condition scallop. Argopecten ventrici}siis. The amount of algae fed to the angelwings approximated the recommended ration by Utting and Spencer (1991). However, our ration was apparently lower than the feeding ration totaling 400 million cells/day of Thalassisiora pseudoiuina fed to every broodstock of Crassostreu gigas by Muranaka and Lannan (1984). The feeding mixed (1:1) algal species of C. calcitrans and T. suecica enhanced gonadal de\elopment and maturation of P. ori- eiitalis. and spawning which occurred one to two months eariier compared to individuals fed only one algal species diet. This may be attributed to the complementation of polyunsaturated fatty acids provided by two or more algal species as compared with a single algal species (Walne 1970. Helm 1977. Webb & Chu 1982. Gal- lardo et al. 1992). Robinson ( 1992) reported that mixing of differ- 1098 Marasigan and Laureta ent algal species of unicellular algae as feed for broodstock oysters provide greater \'ariation of fatty acid than from one species of algae. Thus, the reduction of environmental stresses and the provision of combination of two species of single cell algae to angelwings reared indoor influenced its faster gonadal development and matu- ration compared to angelwings found in the natural habitat. Al- though we did not include the interaction of temperature and food supply. (Hir results showed earlier maturation of laboratory reared angelwings despite the lower temperature compared to the tem- perature in the natural habitat (Laureta & Marasigan 2000). It appears that variation in water temperature ranging from 24 to 29°C was tolerable to the gonadal maturation of the angelwings due to high food concentration, which satisfied the angelwings requirements for development process. Earlier studies by Mac- Donald and Thompson ( 1985) also found that rapid tissue growth and shell increment of giant scallop Placopecten magellankus were correlated to food availability rather than warmer tempera- ture of the water column of Newfoundland and New Brunswick coasts. Our observations on the earlier or faster gonadal develop- ment of angelwings provided with high algal ration in the labora- tory were in agreement with the observations of MacDonald and Thompson (1985) that growth of the giant scallops largely de- pended on reduction of nutriti\'e stress and not necessarily influ- enced by warmer seawater temperature. Thus, it could be assumed that slower gonadal development and maturation of angelwings in the natural environment might be limited by the food supply. .ACKNOWLEDGMENTS This study was supported by the Fisheries Sector Program- Bureau of Agricultural Research of the Philippine Department of Agriculture. We would like to thank the staff of Institute of Aqua- culture. College of Fisheries and Ocean Sciences, University of the Philippines in the Visayas. for their moral support and technical assistance. We are grateful to Drs. Arnulfo Marasigan and Ar- mando Tamse for reviewing and improving the manuscript. LITERATURE CITED Ablan. G. L. 1938. The diwal fishery of Occidental Negros. Philipp. J. Sci. 66:.379-3S5. Bell, T. A. & D. V. Lighlner. 1989. A handbook of normal penaeid his- tology. World Ai/iiaciill. Soc. 114 p. Chen. H. C. & H. H. Yang. 1979. Artificial propagation of the ahalone Haliotis diversicolor var. siiperlexta. China Fish. Montlily 3l4;3-9. Chen, T. P. 1984. Mariculture in Taiwan with reference to sea ranching. In: I. C. Liao & R. Hirano. editors. Proceedings of the ROC-Japan Sym- posium on Mariculture. TML Conference Proceedings. 5 pp. Gallardo, W. G., M. T. R. de Castro. R. T. Buensuceso, C. C. Espegadera & C. C. Baylon. 1992. Gonad development of Placima placenta Lin- naeus fed Isochrysis aalhana Parke. Tetrasebnis teliahele (G.S.) West Butch or their combination. Aqiiacnlttire 102:357-361. Helm. M. M. 1977. Mixed algal feeding of Osrrea eihilis larvae with Isochrysis galhana and Tetrasebnis siiecica. J. Mar. Biol. Assoc. U.K. 57:1019-1029. Hesselman. D. M., B. J. Barber & N. J. Blake. 1989. The reproductive cycle of adult hard clam. Mercenariu spp. in the Indian River lagoon, Florida. J. Shellfish Res. 8:43-49. Honma. A. 1980. Aquaculture in Japan. Tokyo:Japan. FAO Association. SI pp. Inoue, M. 1984. On the present state of our knowledge of propagation and culture of shellfishes, especially of culture of Haliotis. In: 1. C. Liao & R. Hirano. editors. Proceedings of ROC-Japan Symposium on Mari- culture. TML Conference Proceedings. 6 pp. Jones, D. S. 1981. Reproductive cycles of the Atlantic surf clam, Spisiila .soliilissima. and the ocean quahog, Artica islandica off New Jersey. J. Shellfish Res. 1:23-32. King, P. A., D. McGiath & E. M. Gosling. 1989. Reproduction and settle- ment of Mytilus edulis on exposed rocky shore in Galvvay Bay. West Coast of Ireland. ,/. Mar. Biol. Assoc. UK 53:530-684. Laureta, L. V. & E. T. Marasigan. 2000. Habitat and reproductive biology of angelwings Pholas orientalis (Gmelin). J. Shellfish Res. 19:19-22. Loosanoff, V. L. & H. L. Davis. 1963. Rearing of hivabe mollusks. Adv. Mar. Biol. 1:1-136. MacDonald. B. A. & R. J. Thomson. 1985. Influence of temperature and food availability on the ecological energetics of giant scallop Pla- copecten magellaniciis. I. Growth rates of shell and somatic tissue. Mar. Ecol. Prog. Ser. 25:279-294. Muranaka, M. S. & J. E. Lannan. 1984. Broodstock managemeni of Cn/v- sostrea gigas: Environmental mtluences on broodstock conditioning, Aquacuhia-e 39:2 1 7-228. Oshima. Y. 1984. Status of fish farming and related technological devel- opment in the cultivation of aquatic resources in Japan. In: I. C. Liao & R. Hirano. editors. Proceedings of ROC-Japan Symposium on Mari- culture. TML Conference Proceedings. 1 1 pp. Panurovsky, S. K. & Y. M. Yakovlev. 1992. The reproductive biology of the Japanese littleneck. Tapes philippinarwn (Bivalvia:Veneridae). J. Shellfish Res. 11:265-277. Paon, L. A. & E. L. Kenchington. 1995. Changes in somatic and repro- ductive tissues during artificial conditioning of the sea scallop, Pla- copecten magellanicus (Gmelin. 1971). J. Shellfish Res. 14:53-58. Robinson. A. 1992. Dietary supplements for the reproductive conditioning of Crassostrea gigas Kumamoto (Thunbergs): II. Effects on glycogen, lipid and fatty acid content of broodstock oysters and eggs. J. Shellfiish Res. 11:443^47. Rosell. N. C. 1979. A study on the biology and ecology of Placuiia pla- centa Linne. Nat. Appl. Sci. Bull. 31:204-251. Sastry. A. N. 1968. Relationship among food, temperature and gonadal development of the hay scallop. Aqiiipecten irradians Lamarck. Phys- iol. Zool. 41:44-53. Sastry. A. N. 1979. Pelecypoda (excluding Ostreidae). In: A. C. Giese & J. S. Pearse. editors. Reproduction of marine invertebrates. Vol. 5. Mol- luscs: Petecvpods and lesser classes. New York: Plenirni Press, pp. 1 1 3-292 Shatee. M. S. & M. Daoudi. 1991. Gametogenesis and spawning in the carpet shell clam. Riiditapes decussates (L) (Mollusca:Bivalvia), from the Atlantic coast of Morocco. .Aqiuicult. and Fish. Manag. 22:203- 216. Ting. Y. Y. 1984. Shellfish culture in Taiwan. In: I. C. Liao & R. Hirano. editors. Proceedings of ROC-Japan Symposium of Mariculture. TML Conference Laboratory, pp. 129-142. Utting. S. D. & B. E. Spencer. 1991. The hatchery culture of bivalve Hatchery Potentials of Angelwings 1099 mollusk larvae and juveniles. LoweMot'f, UK: Lab. Leftl. 68. MAFF Fish. Res. Villalaz. J. G. 1994. Laboratory sludy of food concentration and tempera- ture effect on the reproductive cycle of Argapeaen ventricosiis. J. Shellfish Res. 13:513-519. Walne. P. R. 1970. Studies on the food vakie of nnieteen genera of algae to juvenile bivalves of the general Osuea. Crassostreii. Mercenaria and Mylihis. In: Ministry of Agriculture, Fisheries and Food: Fishery Investigation. Series II. Vol. 26. No. 5. London: Her Majesty's Sta- tionary Office. 61 pp. Webb, K. L. & F. L. E. Chu. 14X2 Phyloplankton as a food for bivalve larvae. In: G. D. Pruder. C. J. Langdon & D. E. Conklin, editors. Proceedings of the Second International Conference on Aquacullure Nutrition: Biochemical and Physiological Approaches to Shellfish Nu- trition. Baton Rouge: Louisiana State University Press, pp. 272-291. Wilson. J. H. 19X7. Spawning of Pecleii ina.ximiis (Pectinidael and the artificial collection of juveniles in two bays in the west of Ireland. Aqiiaculture 61:99-1 1 1. Zar. J. H. 1984. Biostatistica! analysis. New Jersey: Prentice Hall. 718 pp. Jiiiirihil of Slwllfish Research. Vol. 20. No. 3. 1 lOI-l lOS. 20(11. THE REPRODUCTIVE CYCLE OF THE NEW ZEALAND VENUS CLAM RUDITAPES LARGILLIERTI PAUL E. CRIBBKN.' ROBKRT G. CREESE,' AND SIMON H. HOOKER" School of Eiiriniiiiiu'iitiil and Marine Sciences and Lei^h Marine Laboratory. University of Auckland, Private Bui^ 92019. Auckland. New Zealand: ~National Institute of Water and Atmospheric Research Ltd {NIWA). PO Box 109695. Newmarket. Auckland. New Zealand ABSTRACT The reproducti\e cycle of the sublidal native New Zealand venerid clam Riulitapcs turgillieni was studied over a 16-nionth period in the Whangateau Harbour, northeastern New Zealand. Gametogenic development between males and females was synchronous. Riidinipi's Uirfiillierti was found to have an annual cycle, with gametogenesis beginning in autumn (March to May), and spawning over an extended period commencing in August and ending in early summer (December). Successful fertilization of gametes was verified by observations of settlement of postmetamorphosed spat onto spat catching bags suspended in the water column in the vicinity of the Whangateau Harbour at the end of the spawning season. The study population was dominated by large adults (45-55 mm shell length). Sex ratios between males and females were equal, and there was no evidence of hermaphroditism. The reproductive cycle described here is similar to other \enerid and mesodesmid clams occurring in northern New Zealand. KFA WORDS: clam, molliisca. Ruditapcs largillierii. reproduction, veneridae INTRODUCTION Although the export of cultured bivalves (made up almost ex- clusively of the Pacific oyster. Crassostrea gigas. and the green lip mussel. Periw canaliculus) is worth NZ$1 80m annually, the po- tential for culturing New Zealand clams for export remains rela- tively unexplored. In addition, there is known to be a significant amount of recreational harvesting, the extent of which is often underestimated (Kearney 1999). Venerid clams, in particular, are likely to play an important part in the growth of future shellfish industries because of the large and well established international markets for clam species from this family. At present, the only venus clam commercially harvested in New Zealand is the cockle. Austroveuus stutchhiuyi. for which a total of approximately lOOl were exported in 1999 harvesting season (NZ Seafood Industry Council 2000). It is also the most commonly harvested recreational species. However, another species of venerid clam. Ruditapcs largillierti. may be suitable for commercial exploitation (Maguire 1991 ) because of its similarity to the Manila clam. Tapes pliilip- piiwrum. the only clam for which ubiquitous global markets exist (Manzi & Castagna 1989). Ruditapcs largillierii was collected by indigenous (Maori) harvesters. Investigation of the seasonal reproductive cycle of any murine bivalve is essential in developing a management strategy for a fishery, whether it is commercial or recreational (Shaw 1965, Manzi et al, 1985), The success of any hatchery-based commercial venture is also ultimately dependent on knowledge of reproductive cycles and an ability to spawn broodstoek (Eversole 1989, Hooker & Creese 1995). By following the progress of gonad and gamete production, researchers can determine the timing and duration of spawning events (Ropes & Stickney 1965. Corni et al. 1985. Hooker & Creese 1995). A knowledge of spawning periods is also needed for assessing larval abundance (Ropes & Stickney 1965). predicting periods of annual recruitment (Kecket al. 1975. Man/i et al. 1985) and for interpreting growth rates and mortality data (Keck et al. 1975. Baron 1992). Ultimately, all these factors influence the dis- tribution and abundance of juvenile and adult populations (Shaw 1965. Hooker & Creese 1995). Histological techniques are the most commonly applied meth- ods for determining the reproductive cycle of clams. Although they can be expensive and time consuming, histological techniques are the only reliable way of documenting gametogenic processes (e.g.. Shaw 1962 and Shaw 1965. Ropes & Stickney 1965. Adachi 1979. Eversole et al. 1980. Mann 1982. Robinson & Breese 1982. Ropes et al. 1984, Manzi et al. 1985, Rosenblum & Nie.sen 1985. Knaub & Eversole 1988, Rowell et al. 1990, Ponurovsky & Yakovlev 1992. Hooker & Creese 1995). Histological sections can be used for both qualitative assessment and quantitative analysis of gamete development, typically using measures of oocyte size. Eversole (1989) contends that comparisons of gametogenic cycles among species would be easier if more quantitative measures of gonadal development were used (e.g.. Kennedy & Battle 1964. Heffernan ct Walker 1989). In the present study, both qualitatixe and quan- titative measures were used to assess gonad development ii\ R largillierti in a northeastern New Zealand harbor. The information presented in this paper provides the first de- scription of the reproductive cycle of/?, largillierti in New Zealand using both qualitative and quantitative methods, and explores whether or not the patterns observed are part of an annual cycle. MATERIALS AND METHODS The reproductive cycle of R. largilhcrii in the Whangateau Harbour (northeastern coast of New Zealand's North Island; Fig. I ) was documented using histological analysis of samples col- lected monthly from October 1996 to October 1997. An additional sample was collected in January 1998. Thirty adult clams were collected subtidally in 3-7 m of water each month using SCUBA. and transported to the Leigh Marine Laboratory, approximately 4 km from the Whangateau Harbour (Fig. I ). Clams were processed within two hours of arriving at the laboratory. The length (anterior-posterior axis) of each clam was measured to the nearest millimeter using vernier calipers. Following this the shell was pried open using a small knife and the visceral mass removed and fixed in Bouin's solution. After allowing five days for the sample to harden the visceral mass was extracted, and the siphons, mantle and gills removed. The remaining visceral mass, with associated gonad, gut and attached foot, was then placed in a histological cassette and preserved in 709( ethanol. Samples were dehydrated using a graded ethanol series, blocked in paraffin wax and sectioned al 7 |jLm. Three longitudinal 1 101 1102 Gribben et al. Figure I. Map showing the location of the VVhangateau Harbour, where clams were collected for reproducti\e analysis from October |y96 to Januar> 1998, and the location of spat catching bags placed at Motuora Island and Anchor Ba\, from August 1997 to January 1998. Scale bar = 3 km. sections (along the anterior-posterior axis) were taken from each sample. All sections were stained using Haematoxylin and coun- terstained with Eosin. The histologically prepared slides were then examined using a compound microscope at x40. xlOO and x400 magnification. Gonads from both male and female clams were placed into five qualitative categories adapted from Porter (1964). Keck et al. ( 1975) and Ropes (1968): early active, late active, ripe, partially spawned and spent (Table 1; Fig. 2A-J). The gonadal state of each clam was described as one of the five stages based on the most dominant stage present in 10 randomly selected follicles from each sample. The sex of each clam was determined from microscopic ex- amination of the histological slides. Clams were deemed sexually mature if any mature gametes were present. A Chi-squared good- ness of fit test (a = 0.05) was used to test the hypothesis that there was an equal representation of males and females in this popula- tion. Clams were al.so examined for any evidence of hermaphro- ditism as some species of venerid clams are known to ha\ e male and female gametes co-occurring within the same individual. Monthly mean oocyte diameters were determined using \'ideo image analysis (Mocha Image Analysis 1.2; Jandel Coip. 1994) in order to validate the gametogenic development of female clams. The diameters of all oocytes within three haphazardly selected follicles from each of the three slides were measured for 10 female clams sampled in each month. Only oocytes with \isible nuclei were measured. Interpretation of spawning events from histological analysis is based on the qualitative assessment of gonad staging. In order to \alidate spawning events deduced from the histological stagings, a spat catching experiment was established in Anchor Bay (174°50.5'E. 36°30'S) approximately 4 km from the Whangateau Harbour, and adjacent to Motuora Island (I74°48'E. 36°30'S). approximately 15 km from the Whangateau Harbour (Fig. 1 ). An- chor Bay was chosen as water exiting the Whangateau Harbour circulates through this area (Parr 1993). Motuora Island was cho- sen as it is known to be a sink for bivalve larvae (Morrison 1998). It was not possible to put spat catching bags in the Whangateau Harbour because of the strong tidal cunenis and the high recre- ational use of the harbor. Three 25 m drop lines were anchored to the seafloor in each region. Three 0.5 ni"^ replicate mesh bags were placed at intervals of 5 m. 10 m and 15 m above the seafloor on each drop line by SCUBA in August 1997. The meshbags were collected and re- placed in September 1997 and October 1997. The final set of bags was removed from the lines in January 1998 and not replaced. All bags were transported back to the laboratory and indi\idually .searched lor the presence of newly settled clams. RESULTS Reproductive Cycle Despite slight differences in gamete development between males and females, the reproductive development of both male and female clams followed similar patterns (Fig. 3a. b) which appeared to be related to changes in sea-surface temperatures (Fig. 3c). During 1997, gametogenic development began in March/April (early autumn): a period of decreasing water temperatures. Ga- metes continued to develop through August and September (win- ter) with ripe individuals first observed in the coldest winter months. Spawning began in August with a small number of both male (20%) and female (7%) gonads appearing partially spent. Spawning activity increased as sea-surface temperatures continued to rise. By October 1997. 70% of males and 82% of females appeared in a partially spent or spent condition. This pattern was also observed a year earlier in October 1996, although the spawn- ing event seems to be stronger with all clams in either a spent or partially spent condition (Fig. 3a. b). During the 1996/1997 spawn- ing period, all clams were completely spent by December 1996. Although clams were not sampled in December 1997. this pattern appears to be repeated during 1997/1998 with all clams having spent gonads by January 1998. Oocyte Development The mean monthly diameter of oocytes ranged from 21.3 jjliii (February 1997) to 31.9 (Am (March 1997) (Fig. 4). The peak value in March 1997 was due to the small number of residual oocytes that hadn't yet been resorbed. From October 1996 to January 1998 mean monthly diameters alternated regularly between high and low values. Although there does not appear to be a strong rela- Reproductivr Cycle of Ruditapes largillierti 1103 TABLE 1. Criteria used to stafje histologically prepared slides. Adapted from Porter (l%4). Keck et al. <|y74) and Ropes (1968). Stage Males Females Early acli\e Follicle walls thick, lined with a thick layer ol spermatiigonia occupying up to a third of follicle area. Spermatocytes and spermatids develop in the middle of the follicle. Sperinato/oa. with tails pointing into the lumen, otien occupy the center of the follicle. Gonad volume is small. Late active Follicle walls are thin. Spermatogonia restricted to lining the follicle walls. Follicle dominated by dense areas of spermatids and spermatocytes. Spermatozoa more abundant than early actixe. Ripe Spcmiatogoma as for late active. Follicle dominated by very dense spermatozoa with tails pointing into the lumen. Spermatids and spermatozoa occupy less follicle volume. Gametes occupy nearly all the gonad volume. Partially spawned Spermatozoa still occur in the follicle but with large gaps. Center of the lumen often appears empty. Spermatogonia intrude further into the follicle although no-more abundant than in the previous stage. Spermatids and spermatocytes less dense but still fairly common. Spent Follicle walls are thickened. Few sperinatogonia still present, with few unspawned spermatids, spermatocytes and spermatozoa still remaining. Gametes occupy very little of the 'jonad volume. Follicle walls thick, often contracted. A lot of ovogenic activity with inany oogonia and primary oocytes attached to the follicle walls. Mature oocytes and ova present. Gonad volume is small. There is often a lot of connective tissue visible within the gonad. Follicle walls are not as thick. There are a lot more o\a and mature oocytes than in the early active stage, often attached to the follicle wall by a thin stalk. There are fewer primary oocytes. Mature oocytes are often rectangular or polygonal in shape. Follicle walls are thin and full of ova often lying free in the lumen. There is little ovogenic activity within the follicle except for a few mature oocytes. Ova are usually spherical in shape. Follicles are large with walls that are very thin and often ruptured. There are usually large spaces within the lumen. although free ova are still frequent within lumen of the follicle. There are still mature oocytes present. Spawned: The follicle is essentially empty with a few ova still free m the lumen. Follicle walls may be contracted and thickened with connective tissue. tionship between stage of reproductive developnient and mean monthly oocyte diameter, peak mean oocyte values occurred dur- ing June 1997, a month dominated by developing late active oo- cytes, and September 1997. which is dominated by ripe and late active oocytes. Analysis of the monthly frequency histograms for oocyte di- ameters indicates egg size ranged from 10 to 27 |a.ni and that a broad range of egg sizes was present in all months sampled (Fig. 5). Again, there does not appear to be a strong association between the relative freqtiencies of egg diameters in each mode and stage of reproductive development. Sex-Ralio Adults collected during monthly sampling ranged in size from 29-58 mm shell length, and all could be se.xed. The Chi-squared test performed on the monthly and total number of males and females sampled failed to detect any significant difference from a ratio of 1:1 (Table 2). However, the sex ratio of the 10 clams collected <35 mm (3 males and 7 females) was 1:2.33. Size al Sexual Maturity Length (longest anterior-posterior distance) was used to assess size-at-sexual maturity. All clams collected were mature. Males and females had alinost identical size ranges and mean shell lengths in all months sampled (Table 3). Males ranged in size from 29-57 mm and females ranged frotn 29-58 mm (Table 3). The mean monthly length of males ranged from 44.3 (±7.51) to 51.3 (±1.5) and from 44.0 (±9.4) to 5 1. 1 (±2.9) for females. Occurrence of .Vi-it/y Melamorpluised Juveniles Twelve R. Idrgillicrii spat were recovered from the series of meshbags retrieved from Anchor Ba\ in .lanuary 1998. These ranged in size from 0.5-1.5 mm in length. The presence of R. largillierti spat at this time correlates well with the spawning pe- riod evident from the histological stagings in the 4 months prior to January 1998. However, no post-metamorphic clams were found on any of the bags recovered from Motuora Island. DISCUSSION Temperature is often regarded as the tnain factor determining the general patterns of reproductive development and spawning. Indeed, many species experience latitudinal gradients in the timing of reproductive development and spawning events, and also in the number of spawning events associated with changes of tempera- ture (e.g.. Keck et al. 1975, Eversole 1989. Heffeman et al. 1989). As with other species of clams, the reproductive cycle of R. largillierti in the Whangateau Harbour appears correlated to changes in sea-surface temperature. Gamete development began in autumn (March-April) as seawater temperatures started falling, with ripe individuals first present when temperatures were at their lowest during July. Spawning commenced in August as seawater temperatures began rising, although the majority of spawning ap- peared to occur during the months of October and November. Paturusi (1994) found a similar pattern of reproductive develop- ment for R. largillierti in Georges Bay, Tasmania. The native geoduck, Panopca zelaiulica. collected from the Coromandel Pen- insula from June 1999 to March 2()0I also went through a similar reproductive cycle to that of R. largillierti. with spawning begin- ning in early spring and ending in early summer (Gribben unpub. data). Spawning of the pipi, Paphies australis (Hooker & Creese 1995), and the tuatua, P. subtriaiigiilata (Grant & Creese 1995), in the Whangateau Harbour also began in early spring, however, it contintied right through summer. Although the reproductive cycle was correlated to sea-surface temperatures, low temperatures during the winter, particularlv in 1 104 Gribben et al. B >>. "^^mm^ F j>" ^ WcjM^ 2^ P^^^«o^o;'^sn#' /A 4r.A D *,— V'-- ^'^m:i .^" «!--!95',&aH', E H ^i4-^«^^^i ? I s'^o_ f^* f-.> ^ %^ «4 0 c.*?^ ms^ Figure 2. Photomicrographs olreprochictivu staj-es; Males (A) earl> aili\e. (H) lale aeli\e. ((_ I ripe. ID) partiallj spawned. (E) spent: and females (F) earl) active, (G) lale active, (Hi ripe, (1) partially spawned, (Jl spent. Scale har on F = KM) pni, and applies to all micrographs. Rbpr()DL'cti\e C^cle of RUDir.\l>ES l\kg/luerti 1103 I ! Early Aclive F7s3 Ripe I ; 1 Lale Aclive ^^| Parlial ^|lawlli^l: Spent 60 40 20 0 60 40 - 20 0 Dec 1S96 n=124 60 - 40 - 20 - 0 May 1997 n=333 n^^^PI I ilX 60 40 20 0 SONDJFMAMJ JASONDJF (c) Monthly sea-surface temperatures. 21 „ 18 O °- 15 3 12 ~r ~r ^ >. ^ = ro 03 !i Q. n z! nj < 1 1 ' i 1 ^ i 1 1 c -3 X3 n D) F o F F 3 =3 < CI) (11 01 ^if Q. O > o o 01 W z o 1998 ■^ 0) QJ GJ •" X3 -Q JD 2 E E O OJ C) o 01 1996 Month 1998 Figure 4. Mean monthly oocyte diameters (±sel from histological analysis of female R. largillicrti (October iy96-Januarj 19981; n = I(t-I7 indi\iduals. ^ csj en 60 ^ 40^ 20 A n J Size class (um) Jan 1998 n=90 Figure 5. Frequency histograms for all oocyte diameters measured from nine haphazardly selected follicles for each of 10 females from each monthly sample collected for histological analysis (October 1996- January 19981. August, do not appear to restrict developinent, as early active, late active and ripe clams were evident as early as July, when water temperatures were at their lowest (13°C: Fig. 3c). Ruditupes largillierti occurs throughout New Zealand's waters with its dis- tribution extending to some sub-Antarctic islands (Powell 1979). The Whangateau Harbour is near the northern limit of its distri- bution. If R. Uirgillierti is a species more suited to cooler tempera- tures, then a period of inactivity following spawning, such as that which occurred from December to March after the 1996/1997 spawning period, may be a result of high summer water tempera- tures restricting ganietogenic development. Early male develop- ment, however, commenced in March, soon after the maximum average monthly water temperatures were recorded. In gonochoristic species, such as the majority of venerid clams, the synchronous development and spawning of gametes in local populations is essential for successful fertilization to occur (Ever- sole 1989). The fact that small numbers of post-metamorphic clams were found on spat catching bags in Anchor Bay. confirms that spawning and successful fertilization occurred prior to January 1998. Despite successful fertilization occurring, there were slight differences in the liming and duration of development, maturation and spawning between the two sexes. Female reproductive devel- opment began later than in males, possibly because the energetic !l06 Gribben et al. TABLE 2. Numbers of male and female R. largillierli, and range and mean (±s.d.l of lengths (mml for clams collected during monthly sampling in the Whangateau Harbour (n = 30). x" values test for sex ratios differing from 1:1 (XT = 3.84 at a = 0.05). Length Mean Month Males Females (mm) (±s.d.) P value Ocloher 1996 20 10 29-56 44.2 (±7.8) 0.(J6S November 1996 13 17 45-57 50.9 (±2.8) 0.465 December 1996 IS 12 43-54 ,50.4 (±2.9) 0.273 January 1997 13 17 45-53 .5(17 (±1.9) 0.465 February 1997 15 15 39-57 47.9 (±4.8) 1.000 March 1997 15 15 33-53 46.7 (±5.1) 1.0(10 April 1997 16 14 34-58 47.4 (±6.2) 0.715 May 1997 14 16 37-54 48.4 (±3.9) 0.715 June 1997 15 15 34-54 48.3 (±4.0) 1.000 July 1997 14 16 42-54 48.8 (±2.9) 0.715 August 1997 16 14 32-55 48.0 (±5.7) 0.715 September 1997 14 16 36-56 46.9 (±4.3) 0.715 October 1997 15 15 29-53 46.9 (±6.1) 1.000 January 1998 15 15 35-56 47.5 (±4.5) 1.000 Total 213 207 0.761 costs of reproduction lire often greater for fonales than males (Ropes et al. 1984). Therefore, following a major spawning e\enl. females may need to accumulate greater energy reserves relative to )nales in order to produce gametes once the costs of growth and ]espi)'ation have been met. However, female gametes develop mo)'e quickly through the early active to the late active stage, whereas males develop steadily with a long period dominated by the early active stage of the gametogenic cycle. A pattern of simi- lar overall development with subtle differences between males and fetnales is well documented for many species of clams (e.g.. Porter 1974. Manzi et al. 1985. Coi'ni et al. 1985. Heffeman et al. 1989. Hesselman et al. 1989. Rowell et al. 1990. Sbrenna & Catiipioni 1994). The appearance of a large nu)nber of spawned and partially TABLE 3. Range and mean length (±s.d.) (mm) for male and female R. largillicrti collected monthly (n = .MO from the Whangateau Harbour, from October 1996 to .January 1997. Male Female Males Mean Females Mean Range Range Length (±s.d.) Length (±s.d.) Month (mm) (mm) (mm) (mm) October 1996 29-54 32-56 44.3 (±7.51) 44.0 (±9.4) November 1996 45-54 46-57 50.7 (±2.7) 51.1 (±2.9) December 1996 46-53 43-54 50.2 (±2.7) 50.8 (±3.28) January 1997 48-53 45-52 51.3 (±1.5) 50.2 (±2.0) Februarv 1997 48-57 39-53 48.6 (±5.3) 47.1 (±4.4) March 1997 33-53 40-53 46.9 (±5.9) 46.6 (±4.4) April 1997 34-55 39-58 47.3 (±6.7) 47.6 (±5.7) May 1997 37-54 4.3-52 48.2 (±4.9) 48.5 (±3.0) June 1997 34-53 4.3-54 47.3 (±4.8) 49.3 (±2.9) July 1997 4,3-53 42-54 49 (±2.7) 48.6 (±3.2) Au2ust 1997 32-55 32-54 48.5 (±5.3) 47.5 (±6.3) September 1997 36-56 42-55 46.9 (±4.7) 46.9 (±3.9) October 1997 32-53 29-53 44.9 (±5.1) 45.5 (±4.6) January 1998 35-56 41 -.54 48. 3 (±5.4) 48.5 (±6.2) spawned clams in both October 1996 and 1997 suggests that re- production followed an annual cycle. Although samples were not collected during November and December 1997, by January 1998 all ckuiis had completely spawned. Thts was also observed i)i Januai-y 1997. The small number of unspawned late active and ripe individuals that occurred in October 1997 but not in October 1996 are most likely due to inter-annual vaiiability in reproductive development. Eversole ( 1989) suggests that cotnparisons of reproductive de- veloptnent amo)ig species would be easier if more quantitative )neasures of gonadal development were used. Both Kanti et al. ( 1993) and Villalejo-Fuerte et al. (1996) found a close relationship between oocyte diarneters and the gametogenic cycle in Spisuhi solidissima simihis in California, and the cockle, Laevicaidiiiiu elatmn. in Me.xico, respectively. Xie and Burnell ( 1994) also found that oocyte frequency histograms appeared to support qualitative data on reproductive cycles for both Tapes philippituiriiiii and T. dtriissatus on the south coast of Ireland. In general, studies that have employed quantitative analyses indicate that periods of matu- ration and spawning tend to coincide with maximum oocyte di- ameter values. However, the association between oocyte diameters and the remainder of the reproductive cycle (excluding spawning events) is unclear (e.g.. Eversole et al. 1980. Heffernan & Walker 1989, Hesselman et al. 1989). This also appears true for R. largilli- crti. Peak mean oocyte diameters are associated with months con- taining eggs in later developmental stages (e.g., October and No- vember 1996. and June and Septetnber 1997), and troughs with the pj'csence of higher numbers of primary oocytes (e.g.. January, May and August 1997). Although tnonthly mean oocyte diameters sup- port the findings of qualitative stagings, the patterns are not strong enough to be used on their own as a quantitative indicator of gonad state. Much of this can be attributed to the inherent variability in gonadal state both within and among individuals collected from the same sample, evidenced from the large size range of oocyte diatiieters that exists for each monthly sample (Fig. 5). Several authors have tried to resolve this problem by reducing the nutnber of stages used to assess gonadal state (e.g.. Keck et al. 1975, Eversole et al. 1980). With the exception of the hardshell clatn. Mcixciuiiia mene- naria. sex ratios in venerid clams rarely differ significantly fiom a ratio of 1:1 (Eversole 1989); M. mercenaiia is protandric with tnales dominating juvenile stages until sexual maturity is reached. However, the sex-ratios of adults are equal. Overall, this study showed equal numbers of adult male and female clams. However, the )estricted size range sampled (only clams over 29 rnm) ex- cludes investigation of sex-ratios of juveniles. The sex-ratio of 1 tiiale to 2.-3.3 fonale clams in the lower end of the sample (29-33 mm) indicates that the ratio may be quite different for smaller cknns. However, this is probably more a reflection of the small satnple size rather than an indicator of any real trend. The lack of small size classes also makes it difficult to determine the size at which R. Icirgillierti reaches sexual maturity. However, all clams between 29-35 mm in length were matute and contained gametes. Analysis of tag and recapture data and experimental growth rates (Gribben 1998) suggest that R. largillierti 29-35 mm in length are in their second year of growth. It is likely that sexual maturity is reached between 25-28 mm, as clams grown under experimental cultui'e conditions grew to a tnaxtnnun size of 24 mm in one year and contained no gatnetes (Gribben 1998). The cotiDiiercial culture of venerid clams is well developed overseas with several species (e.g., M. mercenaria and T. philip- piiuiniiii] relying on the production of hatchery reared spat (Manzi Reproductive Cycle of Rlditai'ls l\kgil.ufmti 107 & Castagna 1989). Kent et al. (1999) have successfully produced an ongrown hatchery-reared R. largilUerti spat in Australia. Given that R. laii^illicrti spat can be pniduced in a hatchery, and Gribhen (199SI has achieved good growth rates ongrowing juvenile R. !ar)iillieni intertidally and subtidally in mesh bags on racks, sug- gests that this species may be amenable to aquaculture. The results from our research suggest R. largillieni in the Whangateau Harbour are ripest from early spring to early summer. Collecting bloodstock during this time will maximize egg viability and fertilization success: two of the most important issues facing aquaculturists attempting to produce commercial iitiantities of spat. ACKNOWLEDGMENTS This study was supported by a Graduate Research in Indus- try Foundation scholarship, made possible with the help of Jim Dollimbre and Jon Nicholson. Many thanks to Beryl Davy for her invaluable instruction and guidance with the preparation of histological samples. Thanks also to the numerous people at the Leigh Marine Laboratory who assisted with collecting sam- ples from the Whangateau Harbour. This research formed part of a MSc thesis in Marine Science from the L'niversity of .Auck- land. LITERATURE CITED .^dachi. K. 1974. Seasonal changes of the protein level ui the adductor muscle of the clam. Tapes pbilipi'inarum (Adams & Ree\ei with ref- erence to (he reproductive seasons. Comp. Biochein. Physiol. 64:85-89. Baron. J. 1992. Reproductive cycles of the bivalve molluscs Auiciodea striata (Gmelln). Gafrarium liimittiim Roding and Anadara scaplia (L.) in New Caledonia. Aiist. J. Mar. Fresliwaler Res. 43:393—102. Corni. M. G., M. Fameti & E. Scarselli. 1985. Histomorphological aspects of the gonads of Chamelea galliiui (Linne) (Bivalvia; Veneridae) in autumn. J. Shellfish Res. 5:73-80. Eversole. A. G. 1989. Gametogenesis and spawning in North American clam populations: implications for culture. In: J. J. Manzi & M. Castagna. editors. Developments in .Aquaenlrure and Fisheries Science. Clam Marieulture in North .America. Vol. 19. Amsterdam: Elsevier, pp. 75-109. Eversole. A. G.. W. K. Michener &. P.J. Eldridge. 1980. Gametogenesis and spawning in North American clam populations: implications for culture. In: J. J. Manzi & M. Castagna. editors. Developments in .Aqua- culture and Fisheries Science, Clam Marieulture in North .America. Vol. 19. Amsterdam: Elsevier, pp. 111-125. Grant, C. M. & R. G. Creese. 1995. The reproductive cycle of the Tualua — Paphies suhtriangulata (Wood. 1828). in New Zealand. ./. Shellfish Res. 14:287-292. Gribhen. P. E. 1998. Demography and life history characteristics of the Venus clam Ruditapes largilUerti. Unpuhl. MSc. Thesis. New Zealand: University of .\uckland. 147 pp. Heffernan, P. B. & R. L. Walker. 1989. Quantitative image analysis meth- ods for use in histological studies of bivalve reproduction. ./. Moll. Stud. 55:135-137. Heffernan. R. L.. J. Walker & L. Carr. 1989. Gametogenic cycles of three bivalves in Warsaw Sound. Georgia: 1. Mercenaria mercenaria (Lin- naeus. 1758). J. Shellfi.^h Res. 8:51-60. Hesselman. D. M.. B.J. Barber & N.J. Blake. 1989. The reproductive cycle of adult hard clams. Merceiuiria spp. in the Indian River Lagoon, Florida. / Shellfish Res. 8:43-49. Hooker. S. H. & R. G. Creese. 1995. The reproductive biology of pipi. Paphies australis (Gmelin, 1790) (Bivalvia: Mesodesmatidae). I. Tem- poral patterns of the reproductive cycle. J. Shellfish Res. 14:7-15. Kami. A.. P. B. Heffernan & R. L. Walker. 1993. Gametogenic cycle of the southern surfclam. Spisula solidissiina siinilis (Say. 18221. from Si. Catherines Sound. Georgia. / Shellfish Re.s. 12:255-261. Kearney, M. 1999. Ecology and management of Austrovenus siutchhuryi in the Whangateau Harbour. Unpuhl. MSc. Thesis. New Zealand: Uni- versity of Auckland. 1 35 p. Keck. R. T.. D. Maurer & H. Lind. 1975. A comparative study of the hard clam gonad development cycle. Biol. Bull. 148:243-258. Kennedy, A. V. & H. I. Battle. 1964. Cyclic changes in the gonad of the American oyster. Crassostrea virginica, (Gmelin). Can. J. Zool. 42: 305-322. Kent, G. N., G. B. Maguire. I. Dulhie & R. Pugh. 1999. Spawning, settle- ment, and growth of the New Zealand venerid Ruditapes largilUerti (Philippi 1849) in culture. ,V. Z J. Mar. Freshwater Res. 33:55-62. Knauh. R. S. & A. G. Eversole. 1988. Reproduction of ditferent slocks of Mercenaria mercenaria. J. Shellfish Res. 7:371-376. Maguire. G. B. 1991. Small clams and cockles — are lhe> valuable for fisheries and aquaculture? Austasia Aquaculture 5:48-49, Mann, R. 1982. The seasonal cycle of gonadal development in .Arctica islandica from the Southern New England Shelf. Fish. Bull. 80:315- 326. Manzi. J.. M. Y. Bobo & V. G. Burrell Jr. 1985. Gametogenesis in a population of the hard clam. Mercenaria mercenaria (Linnaeus), in North Santee Bay, South Carolina. Veliger. 28:186-194. Manzi, J. & M. Castagna. 1989. Introduction. In: J.J. Manzi & M. Ca- stagna. editors. Developments in Aquaculture and Fisheries Science. Vol, 19. Amsterdam: Elsevier, pp. 1-22. Morrison, M. 1998. Population dynamics and exploitation of scallops. Pecten novaezelandiae. in the Hauraki Gulf Unpublished Pli D. thesis. New Zealand: University of Auckland. New Zealand Seafood Industry Council. 2000. New Zealand Seafood Ex- ports: Calender Year to December 1999. Report No. 7. 43 pp. Parr. S. W. 1993. Bivalve larvae of the Whangateau Harbour. Unpubl. MSc. thesis. New Zealand: University of Auckland. 150 pp. Paturusi, A. 1994. Effects of stocking density and initial size on the per- formance of the clam. Venerupis largilUerti. (Phillipi. 1849) in experi- mental cages in Georges Bay, Tasmania. Unpubl. MSc. thesis, Austra- lia: University of Tasmania. Ponurovsky. S. K. & Y. M. Yakovlev. 1992. The reproductive biology of Japanese Littleneck. Tapes Philippinarum, (Adams and Reeve. 1850) (Bivalvia: Veneridae). / Shellfish Res. 1 1:265-277, Porter. H. J. 1964. Seasonal gonadal changes of adult clams. .Mercenaria mercenaria IL.) in North Carolina. Proc. Natl. Shellfish .Ass. 55:35- 52. Porter. R. G. 1974, Reproductive cycle of the soft-shelled clam. .Mya arenaria. at Skagit Bay. Washington. Fish. Bull. 72:648-657. Powell, A. W. B. 1979. New Zealand mollusca. Collins Auckland. 500 pp. Robinson A. M. & W. P. Breese. 19S2. The spawning season of four spe- cies of clams in Oregon. ./. Shellfish Res. 2:55-57. Ropes. J, W. 1968. Reproductive cycle of the surf clam, Spisula solidis- sima. in offshore New Jersey. Biol. Bull. 1 35:349-365. Ropes. J. W. & A. P. Stickney. 1965. Reproductive cycle of Mya arenaria in New England. Biol. Bull. 128:315-327. Ropes. J. W., S. A. Murawski & F. M. Serchuk. 1984. Size, age, sexual maturity and se.x ratio in ocean quahogs, Arctica islandica Linne. off Long Island. New York. Fish. Bull. 82:253-267. Rosenblum, S. E. & T M. Niesen. 1985. The spawning cycle of the soft- shell clam. .Mya arenaria. in San Francisco Bay. Fish. Bull. 83:403- 412, Rowell, T. W.. DR. Chaisson & J. T McLanc. 1990. Size and age of sexual maturity and annual gametogenic cycle in the ocean quahog. .Arctica islandica (Linnaeus, 1767), from coastal waters in Nova Scotia. Canada. J. Shellfish Res. 9:195-203. Sbrenna. G. cS: D. Campioni. 1994. Gametogenic and spawning patterns of I10 bottoms. In Ensenada de La Paz (Fig. I), the highest densities of about 48 clams m"" have been observed in the upper portion of the mid- intenidal area between 0 and 3 ni deep (Garci'a-Domi'nguez et al. 1994). Distribution is determined by the type of bottom, with the clam living on sandy or rocky-sandy bottoms ( Garcia- Domi'nguez 1991 ). Catch takes place in shallow waters during low tides, when fishermen collect clams by hand or using some dredging instru- ments. In several locations along the southern Peninsula of Baja California, clam stocks have decreased and some have disappeared almost completely (Garcia- Domi'nguez 1991). There are two ap- parent reasons for this: they have been overexploited because of their high vulnerability and fisheries operate under an open access scheme, or because of bottom disturbances caused by human ac- tivities along the coasts. For both, the absence of knowledge to support appropriate management measures has been an additional problem. In this paper, the stock assessment of a venus clam C califoniicnsis in Ensenada de La Pa/ is developed as a contribution to the know ledge of the population dynamics. Some management aspects are discussed. MATERIAL AND METHODS Ensenada de La Paz has an area of 4.*^ kiir. with a communi- cation with Bahia de La Paz through a channel 1.5 km long and 4 km wide (Fig. 1 ). The stock we studied is located in a place know n as Las Palmitas (24°10'N and 110°24'W). Samples were taken from July 29. 1988 to September 20. 1989 on a transect perpen- dicular to the shoreline, passing across the high clam-density area. Venus clams were collected every 25 m along the transect, where 20 cm"" of sediment (a .square of I m per side and a depth of 20 cm) was taken and sieved through a I -mm mesh. The total length of the clams was measured with u precision of 0.01 mm within a range of 4 to 4.^.5 mm. A number of clams were also weighed to compute the length-weight relationship. The venus clam stock is exploited twice a year, within the period of time we studied. September to October 1988 and Feb- ruary to April 1989. Catch records were collected from Sepesca ( 1988. 1989) for these periods. Growth The ELEFAN I method (Pauly and David 1981. Pauly 1987) was used to estimate the parameters of the von Bertalanffy growth equation (VBGE). For this, length frequency distributions were grouped in 3-mm length-class sizes. Age Sinictiire Age structure was estimated as follows. First, we selected the Gaussian components in the length-frequency distributions, using the Bhattacharya (1967) method, to obtain mean length and vari- ance for some age groups. Then a function of how variance of the mean length-al-age group changes with relative age was fitted. The relative age from modal lengths was estimated according to the VBGE. With this function we were able to estimate mean and variance for all ages in the stock. Normal distributions representing age groups were converted to a standardized normal distribution N(0,ff') to obtain a length-age key in terms of probability, which 1109 1110 Wright-Lopez et al. MEXICO 110° 26' 110° 20' ; ^ 724° 27' PACIFIC OCEAN A 24° 23' 24" 16 LA PAZ BAY ENSENADAS r^ ^ LA PAZ \ ^^/^^ f^ '■L V,-**^"^ LA PAZ Figure 1. Location of Ensenada de La Paz in Baja California Sur, Mexico. was used to assign the probability for a given length to belong to a given age. Here, we named it the length-age key probability matri.x. PM. Then, the product of the transposed vector of a length frequency distribution and the PM resulted in an age-structured vector. Natural Mortality-at-Leni'th Natural mortality. M. was estimated through an iterative pro- cess based in the transition matrix following Arregui'n-Sanchez et al. (2000). This procedure computes M varying with length and time, and is described by the relationship N((. t 4- At) = SJG((. k) X Sik)\ X N(^. t) (1) where N represents the length sti'ucture \ectors. in this case ex- pressed as stock density, and k and / represent successive length classes. G is the growth matrix whose elements represent growth probabilities per length class according to the VBGE (following Shepherd 1987). S is the survival matrix, whose elements represent an average survival estimate of the k length class, and t represents time. Because the venus clam is a relatively short-lived species, age was measured in months; At represents time between samples. When At * 1 mo. time was measured as a proportion, taking the month as a unit. Elements in the survival matrix are expressed as S(k) = exp(-M)At when there is no fishing, and Sik) = exp(-(M -I- F)) At when exploitation is being done, with F being the instan- taneous rate of fishing mortality. Because exploitation occurred in two periods, September and October and February to March, Eq. 1 was iteratively solved for those months when there was no fishing. For this, the growth matrix was computed using the estimated parameters of the VBGE; N(k. t) and N((, t -I- At) are known, and M was varied. Fitting process used a least squares algorithm with the form MIN 2'N.,bs('- 1 + At) - N^JlA = At))- (2) Catchability-al-l.englh The average yield per fishermen in a regular fishing season is about 200 clams per hour, corresponding to an average density of 33.5 clams m'- in a searching area of 10 kni~. This information provides an index of fishing effort at time t, E,. Catchability was estimated following Arregui'n-Sanchez (1996) and Arreguin- Sanchez and Pitcher ( 1999). whose method is also based on Eq. 1. but using catch per unit effort length frequency data for the ex- ploited periods, but giving the M-at-length values estimated as above. In such case, survival is expressed as s(i) = exp - |M(t t) -I- q[k. t)f(t)|, with q and f the catchability coefficient and fishing effort, respectively. Then, iterations were done \arying catchabil- ity coefficients in elements of the survival matrix. The fitting process also used a least-square algorithm as in Eq. 2. M-at-length and catchability-at-length were easily transformed to age by computing relative ages through the VBGE. Slock Size and Fishing Mortality Stock size was estimated through the use of the Gulland ( 1465) and Murph\ ( 1965) methods as follows: where N„, = N„,_,xexp-(M,,-hF„,) (3a) (3b) which represents the stock survival process for two successive ages. E_, , represents the exploitation rate-at-age, with a the index for age. and E , , expressed by {1 -exp(-|M.,,-HF,,])} (4) IM,,-i-F,,) The ratio of catch at successive times is represented by C,-h \A+ I {exp[-(M,,-hF„,])} X E,.,,,,^! C "'^■"" E ' ' *-.i,i .1.1 Given M, and F,. a forward solution can be obtained for two situations. For E.,,, ^0. E, , I ,, , , = R,., X E,, X exp( - ( M..., + F,, (6a) and ForE„, = 0, E,,,,,,=R,_,.,_, x E,_,.,_, X exp|-([t-x]M,_,.,_,-hF„_,.,_J| (6b) Where (a - x) and (t - x) represent age and time between the exploited age at time t and the nonexploited age at time x. _ 4 ••s- •••• " 1 10 40 50 20 30 age (months) Figure 2. Standard deviation(s) of length (millimeter) per age group, as a function of age (months) for the venus clam C. californiensis from Ensenada de La Paz, Mexico. Stock Assessment eor Venus Clam II 1 1 Once estimates for E^ , are obtained, an estimate tor N,, , for the last age group can be obtained from Eq. 3a. and for ottier age groups from Eq. .3b. In this case, the fitting process was a least- squares algorithm based on catch with the tbrm MIN 2 (C„hJa.t + Al)-C,„(a.t + Atil (7) where C,,,,^ = observed catch and C^.,, = estimated catch, which was obtained from Eq. 3 (with estimated values for N., , and E., ,). Murphy (1965) suggested a forward solution to a catch equa- tion in a cohort analysis: for this, an estimate of recruits should be a\ailable. Here, recruits per month were estimated from survey and commercial catch as follows. During catch periods. Eq. 6a was applied using a seed value for F_,, as in the virtual population analysis (Gulland 1965. Gulland 1983. Pope 1972. Pope & Shep- herd 1982. Pope & Shepherd 1985). and using Eq. 7 as the fitting criterion. Results were compared with survey abundance data, and the average ratio was used to estimate recruitment over those months with no fishing. Once recruits were obtained, the stock assessment process, following Muiphy's algorithm, was devel- oped. RESULTS Growth GrovMh parameters for the von Bertalanffy growth equation were L-c = 48.2 mm. K = 0.735 year"', and to = -0.64 vears. Parameters for the length (millimeters) - weight (grams) relation- ship were a = 0.00016 and the slope b = 3.234. These estimates resulted in a value for asymptotic weight of W'.< = 46 g. Age Structure Because the venus clam is a relatively short-iived species, with a life span ^1\' around 4 y (Castro-Orti'z & Garci'a-Domi'nguez 1993). we decided lo express age in month tniits. Once Gaussian components were obtained from length-frequency data, a plot rep- resenting variance per age group as a function of age was ap- proached as a parabolic function (Fig. 2). which was used to es- timate variance-at-age and for the construction of the PM and the age structure, as described above. M-at-LeiigtIi The algorithm used to estimate M-at-length provided values for those ages and months where fishing was closed (Table I ). where stock declination is only from natural causes. In general. M-at-age declines with age. from an average of M^ , = 1.48 mo~' at ages between 1 and 6-7 mo. to M^, = 1-15 mo~' for clams around 20-24 mo old (Fig. 3a). A time variation was also observed that suggested a seasonal behavior (Fig. 3b) with higher values after mid-year. M^, , estimates for exploited periods were obtained in- terpolating values from tendency over time. Catchability-al- Length Catchabilily patterns with length ha\e a similar shape, but a different magnitude between the two catch periods (Fig. 4). with TABLE I. Natural niortality-at-age and tinit in \ears (indicated b) ciilumn heads as pnipiirtiiin of the vearl lor the venus ilam C. californieiisis from Ensenada de La Paz, Mexico. Ksllmallon folloHs .Vrreguin-Sanchez et al. (unpubl. data). Age (Months) (1.578 (1.682 0.726 0.786 0.841 (1.874 0.456 0.025 0.082 0.118 0.189 0.260 0.304 0.463 0.548 1 3.438 2.891 2,308 1,603 1,215 1.113 1,411 1,538 1,825 1 ,567 1 ,036 0.847 0,888 2, 1 .^9 3,223 2 2.928 2.435 1.983 1.447 1,157 1.086 1.312 1 .423 1 ,603 1 .363 0.911 0.766 0.815 1,927 2.783 3 2.678 2.213 1.821 1 .366 1.126 1.071 1.259 1 .362 1.491 1.260 0.848 0.724 0.111 1.819 2.565 4 2.,S2() 2.073 1,719 1,314 1,105 1 .060 1.225 1 .323 1,419 1.195 0.807 0.697 0,752 1,748 2.426 5 2.4(J7 1.975 1,646 1.276 1 .090 1,053 1,200 1.294 1 .368 1.148 0.778 0.677 0.734 1.697 2.326 6 2.322 1 .900 1,591 1,247 1.078 1,047 1,180 1.272 1.329 1.113 0.756 0.662 0.719 1 .657 2.251 7 2.2.'i.s 1,841 1 ,547 1.223 1.068 1 .042 1.164 1.254 1.297 1 .085 0.739 0.6.50 0.708 1.626 2.191 8 2.200 1 .793 1,511 1,204 1,060 1.038 1.151 1.239 1.272 1.062 0.724 0.640 0.699 1 ,600 2.142 9 2.1.'^4 1,753 1.481 1.188 1 .053 1 .035 1.140 1.226 1,250 1.042 0,712 0,632 0.69 1 1 .578 2.101 10 2.115 1.719 1,456 1.174 1.047 1.032 1.131 1.215 1.232 1 .026 0,702 0.625 0.684 1 .560 2.067 11 2.082 1.690 1.434 1.162 1.042 1.029 1.123 1.206 1.216 1.012 0.693 0.619 0.678 1.544 2.037 12 2.053 1 .665 1.415 1.152 1.038 1.027 1.116 1,198 1,203 0.999 0,685 0,613 (1673 1 .530 2.011 13 2.028 1,643 i,.w; 1,143 1 ,034 1 .025 1.109 1.191 1.191 0.989 0,679 0.608 0,669 1.518 1.988 14 2.005 1,624 1,384 1 . 1 35 1 .030 1,023 1.104 1.184 1 . 1 80 0.979 0.673 0,6(.)4 0,665 1,507 1.968 15 1.985 1 .607 1,371 1,127 1.027 1.022 1.099 1,179 1 . 1 70 0.971 0.667 0,600 0.66] 1,497 1 .950 16 1.968 l,.5yi 1,359 1.121 1,024 1 .020 1 ,094 1.173 1.162 0.963 0.662 0.597 0.658 1.488 1 .934 17 1.951 1 ,577 1 ,349 1,115 1.022 1.019 1 .090 1.169 1,154 0.956 0.658 0.594 0.655 1 .480 1 .920 18 1 .937 1 .565 1.339 1,110 1,019 1.018 1 .086 1 . 1 65 1.147 0.950 0.654 0.591 0.653 1 .473 1.907 19 1 .924 1.553 1,331 1.105 1,017 1.017 1 ,083 1,161 1.141 0.944 0,65 1 0.589 0.6.50 1.467 1 .895 20 1.912 1 .543 1.323 1.101 1.015 1,016 1,080 1,157 1 , 1 35 0.939 0.647 0,586 0.648 1,461 1,884 21 1.901 1.534 1.316 1.097 1.013 1.015 1,077 1 . 1 54 1,1.^0 0.935 0,644 0.584 0.646 1 .455 1,874 22 1.891 1 ,525 1 ,309 1 .093 1,012 1,014 1.075 1.151 1 , 1 25 0.930 0.642 0.582 0.644 1 ,450 1 ,865 23 1.882 1,517 1 .303 1.089 1,010 1,013 1.072 1,149 I.I2I 0.926 0,639 0,581 0.643 1.446 1.857 24 1.S74 1,510 1.298 1.086 1 .(J(J9 1.013 1 ,070 1.146 1.117 0.923 0.637 0.579 0,641 1.442 1 ,850 Average 2,184 1,781 1 ,500 1.195 1.055 1 .035 1.144 1.2.W 1.262 1.053 0,719 0.635 0.694 1 ,588 2,126 Standard deviation 0.381 0.336 0.247 0. 1 28 0.052 0.025 0.085 0.098 0.174 0.158 0.098 0.066 0.062 0. 1 73 0.336 CV 0.175 0,188 0,165 0,107 0,049 0,025 0,074 0.079 0. 1 38 0.1.50 0, 1 36 0.104 0,089 0.109 0, 1 58 1112 Wright-Lopez et al. 10 15 Length (mm 10 15 20 25 4.0 00 02 04 06 0£ time (fraction of the year starting in January)) Figure 3. Natural mortaliU as a function of lengtli (topi and time (bottom) for the \enus clam C. californiensis from Ensenada de I-a Paz, Mexico. Iiiizhei" \;lllle^ lor clams aged 7-8 mo. In general. catchabilit_\-at- length values were lower during September and October where relatively small values were found for clams between 4 and 12 mo. and ranging from q = 0.0003 to q = 0.00137. Higher calcliability 0 002 0 002 0 001 - 0.001 0.000 -^- — 0 5 10 15 20 25 0 5 10 15 20 25 Length (mm) Length (mm) Figure 4. Catchability-at-lengtii patterns estimated for the tv\o fishing periods for the \enus clam C. califi>niieii\is in Ensenada de Ea Paz, Mexico. Note that catchahilit> scale for all plots is equal «ith exception of February (top left) «here catchability is higher by one order of magnitude. values were found during the first fishing period of the year, par- ticularly in February where catchability varied between 0.008 to 0.036 for clams between 2 and 12 mo. Catchability decrease later during March and early April in one order of magnitude (Table 2). The main feature of calchability-at-length patterns is that catch- ability values for the venus clam are considerably higher for adults just before spawning (Garci'a-Dominguez & Levy-Perez 1995). TABLE 2. Catchability-at-length estimated for the venus clam C. californiensis from Ensenada de Ea Paz, Mexico. Estimation follows Arreguin-Sanchez (1996), Arreguin-Sanchez and Pitcher ( 1999). and Martinez-Aguilar et al. (1999). (Column heads represent dates as proportiim of the year.) Age (Months) 0.12 0.19 0.26 0.68 0.73 0.79 1 0.0 10809 0.000051 O.OOOOOI 0.0001 Ih 0.000275 0.000156 2 0.014626 0.000584 0.000041 0.000278 0.000592 0.000252 3 0.02 1 525 0.00 1038 0.000093 0.000388 0.000879 0.000331 4 0.028372 0,001399 O.0O0135 0.000459 0.001112 0.000393 5 0.0335 1 8 0.001664 0.000169 0.000499 0.(X)1275 0.000439 6 0.036246 0.001837 0.000197 0.0005 1 8 0.001362 0.000470 7 0.036406 0.001924 0.000220 0.000520 0.001368 0.0004S7 8 0.034173 0.001937 0.000239 0.000512 0.001294 0.000493 9 0.029886 0.001885 0.000255 0.000496 0.001 144 0.000488 10 0.023948 0.001780 0.000269 0.000476 0.000921 0.000476 11 0.016771 0.001632 0.000280 0.000454 0,000633 0.000457 12 0.OOS736 0.001450 0.000291 0.000431 0.000286 0.000432 13 0.000177 0.001242 0.000300 0.000408 O.OOOOOI 0.000404 14 0.000001 o.ooion 0.000308 0.000387 0.000001 0.000372 15 O.OOOOOI 0.000780 0.000316 0.000367 O.OOOOOI 0.000338 16 0.000001 0.000536 0.000323 0.000349 0.000001 0.000303 17 0.000001 0.000290 0.000330 0.000334 0.000001 0.000267 18 0.000001 0.000046 0.000337 0.00032 1 0.000001 0.000231 19 O.OOOOOI O.OOOOOI 0.O00343 0.000310 0.000001 0.000195 20 O.OOOOOI O.OOOOOI 0.000349 0.000301 O.OOOOOI 0.000159 21 O.OOOOOI O.OOOOOI (1.000355 0.000295 O.OOOOOI 0.000124 22 O.OOOOOI O.OOOOOI 0.000360 0.000290 0.000001 0.0()()()9() 23 O.OOOOOI O.OOOOOI 0.00036(1 0.000288 0.000(.)01 0.000058 24 O.OOOOOI 0.000001 0.(X)037I 0.000287 0.000001 0.000026 Stock Assessment for Venl's Clam 1113 80 CO 6.0 0,68 0,0 + 15 20 0 5 10 Length (mm) Figure 5. Fishing niortaiitv-at-leiigth patterns estimated for tlie vcnus clam C. califnrniensis in Knsenada de La Paz. Tliin lines correspond to the period of February to earlj April, while hold lines are for Septem- her and Octoher. Fishing Morlalily and Simk Size Fishing niort;ilit\-at-;ige. F,. shows a similar pattern as catch- ability, with higher values for clams age 6-8 months (Fig. 5), during September (F^ = 6.51 mo~') and October (F, = 5.04 mo"'; Table 2). Garcia-Dominguez and Levy-Perez (1995) re- ported a reproductive period from April to December with a peak from August to November. These high values of fishing mortality impact spawning stock during the peak of the reproductive season. Even though there are two spawning seasons, stock size esti- mation (Table 3) indicates that dynamics are strongly governed by one strong cohort that is recruited during September and October: remaining more lime ni the stock than the early Spring cohort. Stock size was over 60 million clams per year, whereas catch amounted to approximately 5 million clams. Average annual har- vest rate (HR = -) was 14%. but in September it reached a value of 0.8%. The average exploitation rate, -. varied from 0.45 to 0.47 *- p z between fishing periods, but with - = 0.72 in September. The F — / average proportion of the stock under exploitation. (-)( 1 - e ). was 0.43, but in September it reached 0.72, DISCISSION Most clam stocks are highly vulnerable to fishing and to habitat degradation. Many stocks are located in shallow waters, and spe- cies generally are associated with specific habitats. Exploitation of these stocks require special attention by managers, u.sers. and sci- entists if sustainable yields are expected. For growth estimates, length-based methods ha\ e been used for bivalves with success as shown by Vakily ( 1990) and Defeo et al. (1992). C. californiensis is a relatively short-li\ed species, with a lifespan of about 3 to 4 y. M, as estiinated here, indicates a decreasing pattern with age, as occurred with other clam stocks (McLachlan et al. 1996). For the clam, this behavior is expected because of density-dependent ef- fects such as space competition during settlement and food com- petition once individuals ha\e settled. M of adults decreased slowly because they have .some advantages over small individuals for space and feeding. McLachlan et al. (1996) analyzed 15 clam fisheries around the world and suggested that the unharvested frac- tion of exploited sandy-beach stocks may be directly related to the amount of fishing because of sediment disturbance and incidental damage during the collection. For the venus clam, higher values of fishing mortality were present during September, coinciding with high values of M. as expressed by the seasonal pattern. In this case, however, higher values for M also coincide with the peak of the spawning period. Catchability shows a large variation with size and time, repre- senting a large variation in vulnerability to fishing. Results suggest higher catchability values are present during the prereproductive and reproductive seasons. This is of importance because great vulnerability to fishing can negatively impact the reproductive success and, together with M, it could be a potential cause for a collapsing stock. Fishing mortality reflects catchability variations, with higher values for young adults than for older clams. McLachlan et al. (1996) suggests M be directly related to stock density. For the venus clam, this is not clear, but instead, M ap- pears to be related to spawning and the recruitment process. Stock size estimates and the dynamics described in pre\ious paragraphs suggest there is a dominant cohort during the year, and the population dynamics, fishing success, and yields will depend upon the size and survival of such cohort. The first aspect about stock size is that even though C. californiensis has a longevity of 3 to 4 y, dominant cohort practically disappears after I y. Survival after 1 y resulted in a poor number of individuals (less than \%). This condition of the stock suggests that current exploitation is TABLE 3. Stock-size estimation (millions) for C. culijiirnieiisis from Knsenada de La Paz, Mexico, during the first year of life of the cohorts (see text for explanation). .\ge (Months) 0,58 «.68 (1.73 0.79 0.84 0.87 0.96 0.02 0.08 0.12 0.19 0.26 o..^o 0.46 0.55 1 18.1.18 y.(.)43 0.1 SO 0. 1 83 28.666 0.041 0.015 0,018 0.037 0,031 0.335 21.942 0.138 0.083 0.558 2 23..s().S 8..S66 0.912 1.209 0.097 8,646 0.016 0,006 0,009 0.030 0.237 4.208 4.367 0.056 0.025 3 1.^.327 -^.037 1.742 2.3.S7 0.683 0.031 3..543 0,007 0.00.^ 0.03.^ 0.470 4.743 0.977 1 .903 0.014 4 y.'iiio 3.746 2.448 3.322 1.375 0.225 0.0 1 3 1 .485 0.004 0.044 0,773 5,6)8 1.192 0.443 0.684 5 4.489 2..S24 2.. ^08 3,174 1.978 0.462 0.044 0.O06 0.844 0.045 0.931 5.143 1 .487 0.5.54 0.167 6 1..S77 1.26.^ 1.-S26 2.163 1.919 0.675 0. 1 95 0.041 0.003 0.035 0.858 3.685 1.413 0.704 0.215 7 0.438 0.465 0.778 1.141 1.322 0.663 0.286 0.086 0.024 0.022 0.642 2.210 1 .042 0.674 0.279 8 0.108 0. 1 34 0.342 0.503 0.704 0.461 0.281 0.128 0.050 0.013 0.410 1.176 0.640 0.506 0.274 y 0.027 0.034 0.143 0.196 0.3 1 3 0.248 0.197 0.127 0.075 0.007 0.2.34 0.579 0347 0.314 0.207 10 0.007 0.009 0.062 0.072 0.123 0. 1 1 1 0.106 0.089 0.075 0.004 0.126 0.272 0.174 0.172 0.1. W 11 0.002 0.002 0.030 0.02y 0.046 0.044 0.047 0.048 0.053 0.002 0.065 0.124 0.083 0.086 0.072 i: 0,001 0,001 0.020 0.016 0.018 0.016 0.014 0.022 0,024 0.001 0.033 0.056 0.038 0.041 0.036 1114 Wright-Lopez et al. done on young and prereproductive clams and it is probably not the best choice to sustain the fishery. These facts also contribute to make the venus clam an extremely fragile stock in those cases when the fishery is developed under an open access scheme or a poorly controlled exploitation, as commonly occurs with many artisanal fisheries. These types of problems have been experienced by other clani species. Gallucci ( 1982) studied the fishery of Spisiila solidis.siiiui. where the rate of loss of biomass was greater than the stock growth rate. Sasaki (1993) evaluated the state of exploitation for the S. sachaUnensis fishery and showed that stock abundance varied strongly during the year, making this an unstable fishery. Defeo (1989) reported a depleted open-access fishery of Mesodesina mactroides in Uruguay that was temporally closed and reopened under a controlled management scheme after it recovered. Tarr ( 1994) also reported a declining fishery on the white sand mussel Donux serra in southern Africa, where even though several mea- sures to control open access were attempted since 1968. all li- censes were finally revoked. In Weave Bay, New Zealand, McLachlan et al. (1996) also reported a depleted stock for the toheroa clam Papliies ventricosa. which has not recovered despite efforts made over 20 y. Experience with several beach clams exploited in the intertidul areas shows they are highly vulnerable to harvest because of the low costs involved in their exploitation. Additionally, loss of habi- tat because of human activities increases mortality. All of these aspects must be considered by managers in order to sustain fish- eries. For C. califomiensis. we also showed that highest mortality (both fishing and natural) occurred during reproduction and spawning seasons. This is an important impact on the stock, par- ticularly under high fishing pressure in an open-access fishery. Previous experience with clams in Bahia de La Paz and adja- cent regions indicate that the lack of knowledge of population dynamics as described here was a large contributor to the decline of clam stocks. This report atteinpts to advise fishery managers about the relevant aspects of the population biology of the venus clam to maintain biomass and, if possible, current yields. Our advice is that this type of analysis must be made for the most recent years as a basis to simulate experiments resulting in man- agement scenarios. ACKNOWLEDGMENTS We would like to thank the National Polytechnic Institute for support given through CEGEPI— 980056 project, COFAA, EDI, and PIFI. H.W.-L. also thanks CONACyT. Thanks to Dr. Ellis Glazier for editing the Ensilish lanauaae text. LITERATURE CITED Arregui'n-Sanchez, P. 1996. Catchability, a key parameter for fish stock assessment. Rev. Fish Biol. Fish. 6:221-242. Arregui'n-Sanchez. F. & T. J. Pitcher. 1999. Catchability estimates account- ing for several .sources of variation: application to the red grouper fishery of the Campeche Bank. Mexico. Fish. Bull. ^liiy.lAb-l^l . Baqueiro, C. E. & H. Guajardo. 19X4. Analisis de la pesuqueri'a de almejas y caracoles en Baja Califrnia Sur. En: Memorias III Simposium sobre Biologi'a Marina de la Univ. Auton. B.C.S. La Paz B.C.S. 27-29 de Mayo 1981. Mexico, pp. 9-23. Baqueiro, C. E., J. A. Masso & H. Guajardo. 1982. Distribucitin y abuii- dancia de moluscos de importancia comercial en Baja California .Sur. Secretaria de Pesca. Serie de Divulgacion. pp. 1 1-32. Bhattacharya. C. G. 1967. A simple method of resolution of a dislrihution into Gaussian components. Biomelrics 23:1 15-135. Castro-Ortiz. J. L. & F. Garcia-Dominguez. 1993. Crecimiento de Chionc califomiensis (Broderip, 1835) (Veneridae) en la Ensenada de La Paz. Baja California Sur. Mexico. Im: Mar. CICIMAR. 8(2):45-52. Defeo, O. 1989. Development and management of artisanal fishery for yellow clam Mesodesina mactroides in Uruguay. Fishbyte 7:21-25. Defeo O., F. Arregui'n-Sanchez & J. A. Sanchez. 1992. Growth study of the yellow clam Mesodesina mactroides: a comparative analysis of three length-based methods. Scientia Marina 56:53-59. Gallucci, V. F. 1982. Strategies for managing the harvest of bivalve stocks. /;).■ J. C. Castilla, editor. Second workshop: biological basis for the use and management of renewable natural resources: Mar. Biol. Resinirces No. 2.T69-179. Garci'a-Domi'nguez. F. 1991. Dislribucion, abundancia, repioduccion y fauna asociada de la almeja rofiosa, Chione califomiensis. en la Ensenada de La Paz, B.C.S., Mexico. Tesis de Maestn'a en Cicncias CICIMAR, La Paz, B.C.S. Garci'a-Domi'nguez, F. & A. Levy-Perez. 1995. Cicio reproductor de ki almeja rofiosa Chione califomiensis Broderip, 1835 (Veneridae) en la Ensenada de La Pa/. Baja California Sur, Mexico. Rev. Inv. Cicni. Scr. Cienc. Mar.. UABCS, Mexico 5(2):37^2. Garcia-Domi'nguez, F., J. Arvizu-Marti'nez, S. Rodriguez-Astudillo & R. Rami'rez-Sevilla. 1994. Dislribucion v abundancia de Chione califin-ni- ensis (Pelecypoda: Veneridae) en La Ensenada de La Paz. Baja Cali- tornia Sur, Mexico. Rev. Biol. Trop. 42(3):533-538. Gulland, J. A. 1965. Estimation of mortality rates. Annex to Rep. Arctic Fish. Working group. ICES CM 1965. 3:9. Gulland, J. A. I9S3. Fish stock assessment, a manual of basic methods. London: FAO/John Wiley and Sons. Holgui'n, O. 1976. Catalogo de especies marinas de importancia comercial en Baja California Sur. SIC. Subs. Pesca. Mexico, Insl. Nal. De Pesca. Martinez-Aguilar, S., E. Morales-Bojorquez, F. Arregui'n-Sanchez & J. A. De Anda-Montafiez. 1999. Catchability: programa computarizado para estimar el coeficiente de capturabilidad en funcion de la longitud. Cen- tro Regional de Investigacion Pesquera de La Paz del INP, Centro Interdisciplinario de Ciencias Marinas del IPN, Centro de Investiga- ciones Biologicas del Noroeste S.C. La Pa/. Baja California Sur, Mexico. McLachlan, A.. J. E. Dugan. O. Defeo. A. D. Ansell. D. H. Hubbard, E. Jaramillo & P. Penchaszadeh. 1996. Beach clam fisheries. Oceanog- raphy .Mar. Biol. .Anna. Rev. 34:163-232. Murphy, G. I. 1965. A solution of the catch equation. ./. Fish. Rc\. Board Can. 22:191-202. Pauly, D. 1987. A review of the ELEFAN system for analysis of length- frequency data in fish and aquatic invertebrates. In: D. Pauly & G. R. Morgan, editors. Length based methods in fisheries research. ICLARM Conf Proc. 13. pp. 7-34. Pauly, D. & N. David. 1981. ELEFAN I: a BASIC program for the ob- jective extraction of growth parameters from length frequenc\ data. Meeresforsch. 28(4):2U5-2I 1. Pope. J. G. 1972. An investigation of accuracy of Virtual Population Analysis. /;(/. Comin. NW Allan. Fish. Res. Ball. 9:65-74. Pope. J. G. & J. G. Shepherd. 1982. A simple method for the consistent interpretation of catch-at-age data. Journal du-Conscil. Cons. Intern. Explor. Mer. 40:176-184. Pope. J. G. & J. G. Shepherd. 1985. A comparison of the performance of various methods for tuning VPA using effort data. / Cons. Int. E.xplor. Mer. 42:129-151. Prado-Ancona, J. D. 1998. CicIo reproductivo de tres poblaciones de Stock Assessment for Venus Clam 1115 Chinne caUfomiensis Broderip. 1835 (Bivalvia; Veneridae) en Baja California Sur. Mexico. Tesis de Maestria CICIMAR. La Paz. B.C.S. Sasaki, K. 1993. Ecology and stock property of the Sakhalin surf chnn. Japan-Fish. Rcsotir. Coiiseir. Assoc. 42. 82 p. Shepherd. J. G. 1987. A weakly parametric method for estimating growth parameters from length composition data. In: D. Pauly & G. Morgan. editors. Length-based methods in fisheries research, Manila Philippines ICLARM Conf Proc. 13. pp. 113-119. Sepesca. 1988. Anuario Estadistico de Pesca 198S. Secretaria de Pesca. Mexico. Sepesca. 1989. Anuario Estadistico de Pesca 1989. Secretan'a de Pesca. Mexico. Tarr. R. 1994. Aspects of the west coa.st Dona.x population and fishery. In: L. E. McGwynne. editor. White sand mussels: ecology, status and con- servation. Institute for Coastal Research, University of Port Elizabeth. Report 36:10-12. Vakily, J. M. 1990. Determination and comparison of growth in bivalves with emphasis on the tropics and Thailand. Doctoral dissertation. Christian-Albrechts-Universitat, Germany. .lournal o/ Slwlll'ish Research. Vol. 20, No. 3. I 117-1 125. 2(101. MODELING GROWTH OF THE NORTHERN QUAHOG, MERCENARIA MERCENARIA CRAIG L. APPLEYARD AND JOSEPH T. DEALTERIS Depiirtiiifiit of Fisheries. Aiiiiiuil and Veteriiuiiy Science. University of Rliodc Islcind. Kini^sioih Rhode Ishind 02881 ABSTRACT Growth ol the noilhcni qiuihog, Mcrceiiaiiu iiicni'iuiriu (Lmiie). was deterministii.all,v modeled over the life span of the bivalve as well as during its first growing season (nursery stage) in Rhode Island waters. Specifically, the von Bertalanffy growth equation was used to predict increases in shell length (niillimelers). weight (grams), and the relative growth rate (9r increase in weight or volume per day) at various instantaneous annual growth coefficients (K). The relative growth rate (RGR) was also determined by averaging over different time intervals including I. 4. 7. 14. and 28 days and annually. The age at which the maximum shell length and weight was reached varied with K. A higher K (0.30) resulted in more rapid initial growth and an earlier asymptote, while a lower K (0.20 and 0.10) resulted in slower initial growth and a later asymptote. RGR averaged over an annual time interval (annual RGR) decreased rapidly as the northern quahog aged, approaching 0.5% increase/day after age 2. The annual RGR's at different K values were similar, indicating that RGR was relatively insensitive to changes in K. During the first growing .season (210 days in the northeast), the increase in shell length predicted by the von Bertalanffy equation was linear with a slope determined by K. that is, a higher K resulted in a greater slope. RGR. however, varied greatly during the first growing season, decreasing from 1 Wc increase/day at 90 days after spawning to 2% increase/day at 210 days after spawning. The RGRs at different K values were also insensitive to changes in K. There were no detectable differences between RGR determinations at a A.' of 0.10. 0.20, and 0.30 with varying time averaging intervals {T)\ however, the value of RGR at a specific time varied with the time interval used to calculate RGR. The a and /) coefficients estimated for the weight4ength relationship from the adult and nursery stage northern quahogs differed from each other, and published measures from Narragansett Bay northern quahogs. This suggests that researchers should use data collected from northern quahogs in a size range similar to that being modeled when estimating biomass from length and abundance data. Predicted lengths and RGRs were compared to observed lengths and RGRs from a field experiment growing northern quahogs in an experi- mental-scale upweller (nursery stage). Our northern quahogs grew at a /f of 0.25 indicating favorable conditions for growth. Early in the experiment (between 70 and 100 days after spawning), the experimental RGR differed markedly from the predicted measure; however, after 100 days post spawning the experimental RGR followed the general trend of decreasing RGR over time, but was higher than predicted. KEY WORDS: northern quahog. Mcrcenaria mcrcfnaria. growth modelling INTRODUCTION The success of a shellfish aquaeulture operation depends on optimizing production, specifically, on maximizing growth and survival. Growth is defined as an increase in the size of an indi- vidual or the mean increase in the size of a population (Malouf & Bricelj 1989). Growth is usually expressed as a change in shell length, weight, or volume. The change in size (shell length, weight, or volume) per unit time is defined as the growth rate. The growth rate is usually expressed as an absolute growth rate, a relative growth rate, or an instantaneous (specific) growth rate (Ricker 1975). The absolute growth rate describes an increase in size (shell length, weight, or volume) over a specific time interval, usually a month or year. The absolute growth rate does not account for differences in the initial size of the animal. Two northern quahogs may have the same absolute growth rate, but different initial sizes; therefore, the smaller northern quahog is growing more relative to its initial size than the larger northern quahog. To incorporate the effect of initial size on growth, the relative growth rate (RGR) or specific growth rate (SGR) is commonly employed. The RGR measures the change in size of the animal relative to its initial si/e. The SGR estimates the growth rate (change in weight over time) and then uses a log or natural log transformation (Jobling 1994). A number of growth models are presented in the literature to describe the size (length or weight) of a northern quahog overtime. The von Bertalanffy model is the standard for marine species (King 199.5) and has been successfully applied to the northern quahog (Jones et al. 1989). Traditionally, northern quahog growth studies have focused on aging northern quahogs in the natural environtnent based on sclerochronoloizv or shell iirowth riniis (Kennish & Loveland 1980. Peterson et al. 1984. Jones et al. 1989. Rice et al. 1989. Arnold et al. 1991. Slattery et al. 1991). These studies have documented the effect of environmental conditions (Ansell 1968. Jones et al. 1989), habitat (Peterson et al. 1984, Slattery et al. 1991 ). and fishing pressure (Rice et al. 1989) on the growth history of the animal. In aquaeulture research, sclerochro- nological measurements are not required for documenting growth because the age is based on time from spawning or settlement. Growth of the noilhem quahog varies both temporally and spatially. Ansell (1968) documented growth of the northern qua- hog over its natural geographic range. Jones et al. ( 1 989) compared AnselPs data to their own and found that northern quahogs from Narragansett Bay grew exceptionally fast during the first 2 years of life. They also reported that growth varied widely throughout the bay. They found that the von Bertalanffy estimates of the maxi- mum shell length (L-^J varied from 67 to 100 mm, the growth coefficient (K) varied from 0. 1 6 to 0.30, and the time at zero length (;,,) varied from -0.05 to -0.81 years. Mercenaria mercenaria have been known to live as long as 40 years, although only 10% i)f the northern quahogs fi'om Narragansett Bay live longer than 30 years (Jones et al. 1989). Rice et al. (1989) also documented dif- ferences in growth throughout the bay. Northern quahogs sampled from Greenwich cove exhibited an Z.., of 87 mm and a A" of 0.09. while northern quahogs from the West Passage exhibited an L, of I 1 1 mm and a A' of 0.10. There are a variety of factors that could account for these differences including fishing pressure (density), environmental conditions (temperature, salinity, oxygen, and food availability), sediment type, food concentration and quality, nutri- ent loading, and current speed. The diversity of growth estimates used in shellfish aquaeulture .117 118 Appleyard and DeAlteris research is daunting. Malinowski (1988) described northern qua- hog growth as an increase in shell length (mm) over the number of days from spawning, an increase in microns per day. as well as a percent weekly increase in packed volume. Manzi et al. (1984) defined northern quahog growth in terms of increases in shell length and settled volume (biomass). Since the time intervals \ar- ied between measurements, both authors converted biomass in- creases into an equivalent monthly relati\e growth rate. Similarly, a number of studies have transformed biomass increases into daily, weekly, and monthly growth rates (Manzi et al. 1986, Malinowski & Siddall 1989. Hadley et al. 1999). Although the majority of researchers have utilized RGR to quantify northern quahog growth, Rheault and Rice (1996) successfully applied SGR to a shellfish experiment, characterizing oyster and scallop growth rates to varying degrees of food limitation. The variability present among the above experiments makes it extremely difficult to compare growth rates between studies, be- cause growth is being expressed during different time periods and over varying sizes. Furthermore, investigators have failed to pre- dict or model growth under optimal conditions. More importantly, without an estimate or expectation of growth, researchers lack a baseline for comparison. In other words, observed differences in growth due to experimental treatments could be confounded by differences in the predictable growth of the animal depending on the measure used. Thus, the purpose of this paper is to: 1. present a set of expectations for northern quahog growth over both a life time and the first season of growth based on measures of shell length, weight, and volume; 2. describe the sensitivity of RGR to changes in the growth coefficient (A') and the time averaging period over which RGR is measured; and 3. to finally compare the theoretical growth models to obser- vations from the field and froni a nursery upweller. MATERIALS AND METHODS The von Bertalanffy growth model was used to characterize growth of the northern qtiahog in the coastal waters of Rhode Island. The von Bertalanffy equation in teriDs of shell length is; L, = LA\ -t--^"-'"'] where L, is the shell length at age /, L^, is the maximum shell length attained by the animal, K is the instantaneous annual growth co- efficient. / is the age in years, and /„ is the theoretical age (years) at zero length (King 1995). Length in centimeters at time t (L,) was converted to weight in grams at time I (W,) by using; SGR ^MVV„„„,)-LMW„„„„,) T (100) w. at where a is a unit conversion factor and /) is the volumetric expan- sion factor. The growth rate was defined as the daily relative growth rate (RGR) and as the specific growth rate (SGR). RGR was calculated as; where LN is the natural log, IV^,,,,,, is the final weight (g), VV„„,._,, is the initial weight (g), and T (days) is the intervening time period. The SGR is also expressed as a % increase per day. Growth Over the Lifetime The increase in the length and weight of the northern quahog was characterized over the course of the bivalve's lifetime at in- stantaneous annual growth coefficients (K) of 0.10, 0.20, and 0.30. These growth coefficients were selected to encompass the range observed in Nairagansett Bay. The maximum age (lifespan) of the northern quahog was assumed to be 40 years (Jones et al. 1989) and r,, was assumed to be -t-0. 1 0 years or 36 days. A value of -i-0. 1 0 was chosen because the time from spawning to settlement can take anywhere from three to five weeks depending on water tempera- ture (Rice 1992). The L., for the von Bertalanffy and the a and h weight coefficients for the length-weight relationship were based on samples collected in Narragansett Bay and are summarized in Table 1 (Jones et al. 1989, Rice et al. 1989). The daily RGR and SGR averaged on an annual basis were calculated over the life span of the quahog. Growth During the First Growing Season To document growth of the northern quahog during the nursery stage (2 to 14 mm) shell length, weight, and RGR were modeled during the first growing season. In Rhode Island waters, the grow- ing season is approximately seven months (210 days) and lasts from mid April to mid November (Ansell 1968). The model was initiated at the time of spawning and f„ and L-, values of 36 days and 14.0 cm were used. The increase in shell length, weight. RGR. and SGR were modeled during the first growing season (36 to 210 days) at K values of 0. 10. 0.20. and 0.30. The RGR during the first growing season was further investigated by varying the growth interval (7^ between determinations. Specifically, growth intervals of 1. 4. 7. 14. and 28 days were used to calculate daily RGR. Application of the Growth Model to Field and Experimental Data The theoretical models for growth oxer the lifetime and during the nursery stage were applied to northern quahog growth data from Narragansett Bay and Point Judith Pond. Rhode Island. Weight-length data for northern quahogs in the size range of 60 to 140 mm were collected with a dredge in upper Nairagansett Bay and analyzed for the a and /) coefficients of the weight-length relationship. Data collected during the summer 1999 from an ex- perimental-scale upweller in Point Judith Pond were used to define growth of the northern quahog during the first growing season. TABLE \. Parameters used to model [jrowth of the northern quuho^ in Rhode Island waters. RGR: W,, W,„ vv,. (7- where W,,,,,,/ is the final weight (g) and W„„„„, is the initial weight (g) and r(days) is the intervening time period (Ricker 1973). The RGR is expressed as a % increase per day. The SGR was calcu- lated as; Coefficients t„ a b K Samples Collected 1 40 mm 0.10 years {36 days) 0.0004 2.80 0.1.0.2, and 0.,^ Growth Modeling of Mekcenakia mercenaria 1119 Specifically, ihe growth coefficient [K) was dctcirnincd hy non- linear regression methods (DeAlteris & Skiobe 2000); the lime at length zero (r,,) was determined from a linear regression of length versus time after spawning; and the length-weight parameters ia and /)) were calculated using linear regression of natural log (LN) transformed length and weight data. RESULTS Growth Over the Eifetimc The increase in shell length over the course of the northern quahog's 40-year life follows an asymptotic relationship (Fig. I A). The shell length increases in an apparent linear pattern during the first two years of the northern quahog's life and then increases at a decreasing rate until the asymptote or maximum shell length is attained. The rate that the northern quahog reaches its maximum shell length varies substantially with varying K. The northern qua- hog reaches a maximum length of 140 mm in 1 1 years with a K of 0.30. in 1 7 years with a A' of 0.20. and in 33 years with a K of 0.10. Unlike the growth function describing length over time, the weight (g) function resembles an apparent sigmoid-shaped relationship. The northern quahog's weight increases slowly during the first two years and reaches a maxmiuni of 647 grams in 15. 23. and 39 years with a A' of 0.30. 0.20, and 0.10. respectively (Fig. IB). As the growth coefficient increases, the time required for the northern quahog to reach its maximum shell length and weight decreases. RGR over an annual time interval decreases from 3.3% increase/ day during the first year to 0.5% increase/day during the juvenile years (age 1-2 years) and becomes negligible during the adult years (age 2-I-) (Fig. 2A). Although the same general trend was apparent for the SGR averaged over an annual time interval, the SGR during the juvenile years was about half the value of the RGR (Fig. 2B). Growth During the First Growhii^ Season The increase in shell length during the first seven-month (210 day) growing season was linear, as would be expected from the von Bertalanffy growth curve (Fig. 3A). The maximum shell length reached during the first growing season varied considerably with K. The northern quahogs reached a shell length of 27 mm with a A" of 0.30. a shell length of KS mm with a A' of 0.20, and a shell length of 9 mm with a A' of 0.10. The increase in weight during the first growing season was exponential with a consider- able lag during the first 40 days. Weights of 6.20, 2.20, and 0.30 20 25 Time (years) B 800 1 700 600 500 -3 -400 300 200 too 0 K=0.1 -K=0.2 -K=0.3 ' / / ^ . / '' ■ / '' / '' .' / '' - / »' • / ^ : / '' ' / ^ ' / ^ .' / ^ ) 5 10 15 20 25 30 Time (years) 35 40 45 Figure 1. Length (A) and weight (B) at time for A' values of 0.10, 0.20. and 0..10 omt the lifespan of Ihe northern quahog in Rhode Island waters. 1 1 20 Appleyard and DeAlteris 3.0 2.5 i' f 2 0 1.0 0.5 K=0.1 K=02 K=0.3 10 20 25 Tiine(yeais) 40 45 K=0.1 K=0.2 K.-0.3 20 25 Time (yeais) 30 Figure 2. Daily relative growth rate (RGR) t\) and specific growth rate (SGR) (B) averaged on an annual hasis at A' value.s of (1.1(1, 0.20, and 0.30. g were reached during the first growing season with K values of 1 1 0.3. 0.2. and 0.1. respectively (Fig. 3B). The differences in final shell length and weight observed at A' values of 0.10. 0.20. and 0.30 were not apparent when growth was conveiled to a RGR (Fig. 4A). Although there was a considerable change in the RGR during the first growing season, the northern quahogs displayed similar RGRs at varying K values. In particular, at 90 days after spawning, the northern quahogs grew at a rate of 1 \% increase/day. while at \5() days after spawning they grew at 3% increase/day. The RGR was high early in the northern quahogs first growing season, but decreased substantially and leveled off after 1 80 days post spawning. The SGR during the first growing season followed a similar pattern to the RGR (Fig. 4B). The RGR was also determined using a variety of growth inter- vals. There were no detectable differences between RGR determi- nations at Ks of 0.10, 0.20, and 0.30 with varying time intervals (Fig. 4A. Fig. 4B, Fig. 5A. Fig. 5B. Fig. 5C. and Fig. 5D); how- ever, the value of RGR at a specific time \'aried with the time interval used to calculate RGR (Fig. 6). At 90 days after spawning, the northern quahogs grew at 26% increase/day based on a 28-day growth interval, at 155^ increase/day based on a 14-day growth interval, at 13% increase/day based on a 7-day growth interval, at % increase/day based on a 4-day growth inter\al, and at I07f increase/day based on a 1-day growth interval. In other words, the larger the growth interval that RGR was averaged over, the larger the RGR. This relationship becomes less obvious when the north- ern quahogs reach 120 days after spawning and disappears after the first growing season (Fig. 6. Fig. 2A. and Fig. 2B). Application of the Growth Model to Field and Experimental Data For adult northern quahogs collected in upper Narragansett Bay. the linear regression of the LN of weight (g) versus the LN of length (cm) resulted in an R" of 0.94 (Fig. 7A). From this relationship, the a and b coefficients of the weight-length relation- ship were determined to be 0.00023 and 3.15. respectively. For the nursery stage, the northern quahogs were spawned in 1999 April 8. at Bluepoints Company. Inc.. West Sayville. New York. The linear regression of shell length versus the number of days after spawning was significant (f( 1.6) = 132.87. p < 0.05) with an R" of 0.96 (Fig. S). From this relationship. /,, was estimated as 63 days after spawning. The linear regression of the LN of weight (g) versus the LN of length (cm) resulted in an R" of 0.99 (Fig. 7B). The a and /' coefficients of the weight-length relation- ship were determined to be 0.00027 and 2.81, respectively. Growth Mooeling of Mercenaria mercenar/a 121 A 30 28 26 24 22 20 K.=0,3 K=0.2 •^ 18 ^^^^'-^''^ ^ t 16 14 ^^^^^^^ ^^^^ 12 10 8 6 4 2 0 60 ^ ^ ^:=^ K=0 1 go 100 120 140 160 180 200 220 Time (Dumber of days after spawning) B 7.00 1 6.50 K=0.3 6.00 J .-' 5.50 / 5 00 4.50 .'' 3 4.00 1, 3.50 1 3.00 / 2.50 y K=0,2 2.00 J ^_,^^ 1.50 ^^ ^^^^ -"""^ 1.00 0.50 000 _^^<::;Il----^^^^ K=0 1 1 1 60 80 100 120 140 160 180 Time (number of days after spawning) 200 220 Figure 3. Length (A) and weight (B) at time for K values of 0.10, 0.20, and 0.30 during the northern quahogs first growing season. A non-linear regression method estimated K to be 0.25 for the nursery stage growth experiment completed during the summer 1999 (Appleyard 2000) (Fig, 9). The predicted RGR was com- pared to the observed RGR based on a 4-day growth interval (Fig. 10). The experimental RGR varies considerably from the predicted RGR during the first 25 days of the experinienf. the northern quahogs grew slower than expected. At approximatel> 100 days after spawning, the experimental RGR intersects the predicted curve; thereafter, the northern quahogs grew faster than expected, but tended to follow the general trend of decreasing RGR o\er time. DISCUSSION Growth Over llif Lifelime The results of this growth modeling exercise indicate that the growth rate of the northern quahog over its life span changes considerably (Fig. I A and Fig. IB) as a function of A". In particu- lar, first year growth of the northern quahog is apparently linear for length and exponential for weight and both reach an asymptote as the animal ages. The age that growth reaches an asymptote de- pends on the particular growth coefficient (K) used. A higher K results in rapid growth and an earlier asymptote in the animal's maximum shell lensth and weiiiht. while a lower K results in slower growth and a later asymptote: therefore. A" is a useful in- dicator of en\ironniental suitability for growth. Ansel! (1968) re- ported the growth rates of adult northern quahogs throughout its geographical range based on annual growth rings. Within an ac- ceptable temperature range. Ansell concluded that growth was determined by a variety of factors. Although the author success- fully documented growth throughout the northern quahog" s geo- graphical range, he was unable to quantify why the northern qua- hogs were growing differently. A more useful measure would be the instantaneous annual growth coefficient (A'). K incorporates the gamut of environmental conditions into one measure. Researchers could use K estimates to determine the health of the animal in a particular area and aquaculturists could utilize K estimates to de- termine the most suitable area for culture. RGR averaged over an annual time interval decreases rapidly as the northern quahog ages, approaching 09c increase/day after age 2. RGR was relatively insensitive to changes in K. This has serious implications for experiments hoping to quantify differ- ences in growth between treatments. Manzi et al. (1984) investi- gated northern quahogs in an experimental-scale uptlow nursery system. The experimenters varied density to attain a range of chlo- rophyll-a effective flow rates and compared the growth, as ex- pressed as RGR. between treatments. Although the investigators found qualitative differences in growth between treatments they ii: Appleyard and DeAlteris 20 18 16 14 12 6 4 2 0 If 1^ K=0.1 K=0.2 K=0.3 60 80 100 120 140 160 Time (number of days after spawning) 200 220 - K=0.1 — K=0.2 - -K=0.3 80 120 140 160 180 Time (number of days after spawning) 200 220 Figure 4. The relative growlli rate (RGR) (A) and specific growth rate (SGR) (B) expressed as % increase/day at A' values of 0.10, 0.20. and 0.30 during the northern quahogs first growing season in Rhode Island waters (calculated over a l-daj time interval). did not find quantitative ditt'erences. Based on this modeling ex- ercise, the authors would have been hard pressed to elucidate differences in growth between treatments based on RGR. A better measure to quantify differences between treatments would have been a change in length or change in weight. SGR closely followed the patterns observed for RGR; however, because of the LN transformation. SGR during the juvenile years was about half the RGR. The LN transformation essentially com- presses the growth .scale and reduces the growth rate during the first two years. Comparisons between RGR and SGR are only appropriate for adult northern quahogs and should be avoided for juvenile animals. For example. Rheault and Rice ( 1996) measured SGR for oysters and scallops with an initial shell height of 45 and 4.^ mm, respectively. Their experiment monitored SGR during a six-week experiment. Since these bivalves are almost past the juvenile stage. SGR can be compared to RGR. To ensure an ac- curate comparison between this study and other similar experi- ments. SGR can be easily converted to RGR. Growth During the First Growing Season To predict growth during the nurserv stage, growth of the northern quahog was investigated during the tnst growing season (21(1 days in Rhode Island). The increase in shell length and weight was linear. The final shell length and weight reached during the first growing season varied considerably with A'; however, the RGR at different K values was almost identical. Again, the mea- sure of RGR was relatively insensitive to changes in A'. The RGR did vary during the first growing season. The RGR was extremely high early in the growing season (at 90 days after spawning the predicted RGR was 119i' increase/day I. but then decreased sub- stantially and leveled off (at 150 days after spawning the predicted RGR was less than 4% increase/day). The RGR was further investigated b> altering the time period between RGR determinations. There were no observable differ- ences between RGR at A's of 0. 10. 0.20. and 0.30 with varying tinie intervals of I. 4. 7. 14. and 28 days (Fig. 4. Fig. 5A, Fig. 5B. Fig. 5C. and Fig. 5D). In other words. RGR does not detect differences in the instantaneous growth rate (A'l. The value of RGR at a spe- cific time varied with the time interval used to calculate RGR. Specifically, at 90 days after spawning there was a considerable difference between the RGR calculated over a 28-day interval and the RGR calculated over a 4-day interval (Fig. 6). Investigators should keep in mind that when comparing the results of growth experiments with diffeient time intervals, observed differences in Growth Modeling of Mercenaria mercenaria 1123 A 20 . 18 i \ - -K=01 1 -K = 02^ - K=0.3 S - 16 \ \ n 14 \ li 12 10 8 > V S -9 6 4 S ^ "g S? ''v^ a 2 . 0 t- 60 80 100 120 140 160 180 200 220 Tune (Dim ber 0 fdays after spawning) 20 S '* 3 „ 16 a: a- ij g .s 60 80 100 120 140 160 180 200 220 Tune (number of days after spawning) c K = 0.1 K = 0.2 20 . 1 K. = 0.3 18 ^ \ s 16 1 \ <^ & 14 - 1 1 12 - \ 1 1 10 8 \ S S 6 ^^ S 5S ^-^^^ •33 — 4 - *>i.^„^ b: 2 -— '--^^ 0 1 6 0 80 iOO 120 140 160 180 200 220 Time (Dunber of days after spawn) D K = 0.1 K=0.2 20 1 V K. = 0.3 18 J I a 16 1 I a i> 14 ^1 12 \ ^i 10 \ g s 6 \. § ^ 'Vs^ 4 oi 2 ^ — -^ 60 80 100 120 140 160 180 200 220 Time (number of days after spawn) Figure 5. The relative growtli rate (RCRl expressed as % increase/day over time at \arvin}; A calculated over: 4-da> time interval (A): 7-dav time interval (Bl: 14-dav time interval (C): and. 28-da\-time inter\al (D). RGR may be due to ditferences in the time inter\ai. not due to actual differences in growth. Application of the Growth Model lo Experimciilal Ihila The a and /) coefficients estimated for the weight-length rela- tionship from the adults and nursery stage northern quahogs were different from each other and from other values published for Narragansett Bay northern quahogs. This suggests that researchers should always use data collected from locally available northern quahogs in a size range similar to that being modeled when esti- mating biomass from shell length and abundance data. The relative growth rate changes considerably during the first growing season. The grov\'th rate depends on the size of the animal as well as the time period between measurements. Growth studies on upwellers have not taken into account these differences. In addition, the variability between growth measurements has pro- hibited meaningful comparisons between research studies. Manzi et al. ( 1984) converted their biomass increase to a monthly percent increase because the tniie interval between volume determinations varied. The investigators failed to take into account changes in the growth rate at different time intervals. A study completed by Hadley et al. ( 1999) found a relationship between the daily growth rate (DOR) and flow ratio at varying northern quahog shell lengths. In particular, the authors found that the DGR increased as the size of the animal decreased. Based on the results of this 40 35 30 iS ?; 25 If %r le 15 a: 10 5 0 60 P-ldW — • — -P = 4d«vi ■P = 7inica (Gmelin 1791). and the bay scallop. Arsopecten irrailians (Laniarch 1819). / Sliellfisli Res. 15(2):271-283. Rice. M. A. 1992. The northern quahog: the biology of Mercenaria mer- cenaria. Rhode Island Sea Grant. University of Rhode Island. Nar- ragansett. 60 pp. Rice. M. A., C. Hickox & I. Zehra. 1989. Effects of intensive fishing effort on the population structure of quahogs. Mercenaria mercenaria (L.. 1758). in Narragansett Bay. / Shellfish Res. 8(2):345-354. Ricker. W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bulletin of the Fisheries Research Board of Canada. Bulletin 191. Ottawa, Canada. 382 pp. Slattery. J. P.. R. C. Vrijenhoek & R. A. Lutz. 1991. Heterozygosity, growth, and survival of the hard clam. Mercenaria mercenaria, in seasirass vs. sandflat habitats. Mia. Hinl. 1 1 1:335-342. Jounwl ofSlu-Ujhh Rc.seanh. Vol. 20, No, 3. I 127-1 1.^4, 2001, REPRODUCTIVE STRATEGIES IN TROPICAL BIVALVES (PTERIA COLYMBUS, PINCTADA IMBRICATA AND PINNA CARNEA): TEMPORAL COUPLING OF GONAD PRODUCTION AND SPAT ABUNDANCE RELATED TO ENVIRONMENTAL VARIABILITY H.-JORG URBAN Alfred Wegener Institute for Polar unci Marine Research. Section for Conjjuirative Ecosystem Research. Postfach 12 01 61. 27515 Bremerhaven. Germany ABSTRACT The effects of environmental variability (El Nino 1997/1998) on three Caribbean bivalves (Preriu colynilnis. Pimuuki imtviiutu and Piniui caiiwti) were studied based on time series (1994 to 199S) of temperature, salinity, particulate organic matter (POM), larval and spat abundance. Monthly condition and gonad production cycles over a 12 months period are used to investigate whether these processes are coupled with recruitment. The findings clearly show that El Nino 1997/1998 increased POM. decreased salinity and had little effect on temperature. In contrary to the Pacific, these changes in the study area were not caused directly, but by increased precipitation and no significant effects on larval or spat abundance were observed. The findings show similarities for the two closely related pearl oysters (Pteria colymlms and Piiuicuhi imhricum). such as correlated larval and spat abundance cycles, and a negative correlation between temperatures and spat (whereas temperature was positively correlated with Pimm cciriwa spat). However, due to fewer spat peaks and a lower mean spat abundance, the uncoupling of gonad production and spat abundance, as well as less continuous gametogenic activity, a different reproductive strategy for Pwria colymbiis is indicated. KEY WORDS: bnalves, Niiio. Caribbean pearl oysters, fan shells, reproductive strategy, gonad production, larval abundance, spat abundance. El INTRODUCTION Environmental variability can be seen in the context of latitu- dinal dependent environmental changes (i.e., seasonality) or inter- annual climate changes. (e,g„ the climate anomaly El Nino), These environmental changes affect biological processes stich as the re- production and recruitment of marine species. Regarding season- ality it is generally accepted that as an adaptation to decreasing seasonality of environmental factors from polar to tropical regions (Lalli & Pearsons 1993) marine invertebrates from high latitudes have shorter spawning and recruitment periods, while at low lati- tudes longer and more continuous periods dominate. In the liteia- ture we find little published results on the relation between energy allocated to reproduction and temporal larval or spat variability in tropical benthic invertebrates. This tnight be explained by the fact that the reproductive output of animals with planktonic eggs and continuous reproduction are difficult to assess, and this reproduc- tive strategy is likely to be very common among warm temperate and tropical benthic species (Crisp 1984). El Niiio might be regarded as an oceanographic phenomenon regionally restricted to the South American Pacific coast princi- pally off Peru and North Chile which on statistical average occurs every three to seven years. Here, under El Nino conditions, the cold and nutrient rich upwelling system changes towards warm and nutrient-poor conditions that can lead to catastrophic conse- quences for the marine ecosystem (Arntz & Fahrbach 1991 ), Apart from effects on the marine ecosystem, impacts of El Niiio on the terrestrial ecosystems of the South American continent, such as increasing rainfall in deserts and droughts in otherwise humid regions, can be observed. For example in the South American Caribbean, which is not affected directly by El Nino, different precipitation patterns lead to secondary (indirect) oceanographic changes, and thus effects upon the marine ecosystem can be as- sumed. Among the many reports of El Niiio impacts on marine ecosystems, the results of Urban (1994) and Urban and Tara/ona (1996) are most relevant to this study. They postulate that for several bivalve species, El Nino normally only affects the energy balance (such as metabolism, growth rate, gonad output), larval survival and reciuitmcnt. These two studies so far seem to be the only long-term approaches quantifying the effects of El Nino on marine benthic invertebrates of South America. Pearl oysters (family Pteriidae) and fan shells (family Pinnidae) are of commercial importance, as reflected by the quantity of publications devoted to aquaculture and spat settleinent (for pearl oysters see review in Urban 2000a), For fan shells (different Atriiia species and Pinna rugo.ui). see studies on ecology, fishery and aquaculture by Arizpe and Felix (1986). Cendejas et al. (1985). Baqueiro and Castagna (1988), Davy and Graham ( 1982), Philip- son (1989), Turner and Rose water (1938), In the Caribbean waters of Colombia, the dominant pearl oysters are Pteria colymhus and Pinctada imbricala and the fan shells are represented by Atrina serrata, Atrina seminiida and Pinna caruea. with the latter being the most dominant, Pteria colymlms lives attached to octocoral communities, Pinctada imhricata was found to be the dominant sessile species in sea grass (Thalassia testiidinum) communities (Urban 2000a) and Pinna carnea favors medium to coarse sandy substrates or mixed substrata (sand-rocks-corals). All three species are epifaunal species distributed in depths between i)5 m to 15 m. Except for Urban 20()()a and Urban 2000b on the aquaculture and population dynamics of Pinctada imbhcata. no further information on pearl oysters and fan shells from this region was found in the reviewed literature. The objectives of this paper are to study possible effects of oceanographic changes caused by El Niiio on reproduction and recruitment of three tropical bivalve species. Furthermore, the cy- clic reproductive patterns and temporal coupling of gonad produc- tion and spat variability are investigated. This includes duration of larval development under laboratory conditions and spat growth experiments. Reproductive strategies are discussed. MATERIALS AND METHODS Stud\' Area Spat and larval abundance and environmental factors, were monitored monthly between 1994 and 1998 in Chengue Bay, Tay- 1127 128 Urban rona National Nature Park (1 1°20'N, 74°10'W. Fig. la). Gameto- genic activity and condition cycles were studied monthly during one year, but during different years for each species: Pteiia c<>- lYinhiis (1.354 specimens) were sampled between March 1994 and February 1995 in lay rona Park, Piniui carnea between April 1995 and April 1996 ( 1469 specimens from Tayrona Park) and Pinctada imhiicara between March 1997 and March 1998 (782 specimens) in the •Cabo de la Vela", a Peninsula located in the Guajira prov- mce ( 12"10'N.72°20'W and 12°00'N. 72°I0'W). approx. 300 km from the Tayrona Park (for detailed description of the Cabo de la Vela study area see Urban 2000a. Fig. lb). Tayrona Park consists of several small bays offering a great variety of substrates (patches of sea grass, coral reefs, sandy-silty substrates and mixtures of coral and sand bordered by mangroves). Annual water tempera- tures vary between 26°C and 30'C with a mean salinity of 36%r and a tidal range o( less than 30 cm. Sampling Monthly zooplankton samples were taken between 9:00 and 1 1:00 a.m. with a Bongo net (0 35 cm. mesh size 150 (jini. fil- tering 25-30 m"" seawater). Samples were preserved in 10% alco- hol and all bivalve larvae of the genus Piiutada. Pteria or of the family Pinnidae were identified and counted. For larval identifi- cation, "morphotypes" from laboratory reared larvae, were com- pared with larvae obtained from zooplankton samples. Spat abun- dance was determined deploying collectors {= "traps") that con- sisted of three plastic mesh "onion bags" (80 x 30 cm. mesh size 0.8 cm), protected by a pi-opylene net-bag. Collectors were tied to a bottom long line at 5 and 10-m depths, and recovered 6 weeks later. Mean abundance collector"' was determined at monthly in- tervals. At each sampling date bottom temperature and salinity were recorded and seston samples were taken at 3-m water depth with a pump. Seston samples were filtered using Whatman Glass Microfibre filters, which were dried for 24 hours at 70°C. Particu- late organic matter (POM) was calculated after ignition of filters at 500' 'C for 4 hours. Gametogeiiic Aclirily. Cuiidilion and Giiiuid I'rodiicliiiii Maximum length (anterior-posterior axis) was recorded for all specimens with vernier calipers. Two sub-samples of 30-40 indi- viduals each were selected randomly covering the entire size range. One sub-sample was used for gametogenic analysis. Every month the reproductive stage of 30 individuals was determined Caribbean _ W^ S^ Tayrona Park Cabo dels Vela ^^ ^Tavrona Park ^-^^'W^^T based on microscopic observation on fresh gonad material (smear samples) using a semi quantitative scale (Urban 2000a: indifferent, developing 1. developing 2. ripe, spent). With the second sub- sample the condition (body weight cycle) was determined. Soft parts were removed and dried at 70°C to constant weight to de- termine shell free dry weight (SFDW). Ash free dry weight (AFDW) was obtained by ignition of dried soft parts at 550°C for five hours. Parameters "a" and "b" of the monthly relationships between shell length and shell-free dry weight (Eq. 1) were esti- mated with non-linear regression analysis. An annual shell free dry weight (i.e.. condition) cycle for a standard individual was calcu- lated as SFDW = a ■ SL" (1) where SFDW is the shell free dry weight |gl and SL the shell length [mm|. Gonad production was estimated using a new method (for details .see Urban & Riascos submitted) based on monthly quantitative samples, monthly length-gonad weight relationships, pooled annual length frequency data and observation of the game- togenic activity. First the gonad region was replaced by geometric abstractions. With linear measurements of the gonad region and geometric equations corresponding to geometric bodies, the gonad volume and thus the monthly gonad weights were estimated. As- suming the gametogenic stage "ripe" corresponds to the part of the population which spawns during a particular month, the monthly spawned gonad weight for each size interval was estimated. Fi- nally, the annual gonad production of the entire population is given by the sum of all months and length class intervals according to Eq. 2. W,. (2) annual gonad production of Ihe popidalion |g .AFDW nV^ yr'l gonad weight per length interval ■'i" durnig month "i' jg AFDWI pan of the population spawning per length inlcrval "i" during month "i" gonad volume per length interval ■•!" during month "\' I ml I total body volume per length interval "i" during month ' i" [mil total body weight per length interval "i" during month 'i' Ig AFDWI number of specimen In length interval "\ total number of specimen ol the pooled length frequency sample [yr"'] mean abundance from quantitative samples |n m"-| part of the population w ith gametogenic stages "ripe" during month j [9rl. Gonad production could not be estimated for PUwa canwci due to lack of quantitative samples. W„„„ , , 1 •' "P V. 1 V„„,j A ripe, Figure 1. Study area. Tayrona Park, Caribbean of Colombia. iMtval Divvlopment and Settlement Adults ol PiiHhuUi iinhriccita were adapted for four weeks to laboratory conditions (ultraviolet-irradiated, 45 |xm filtered seawa- ter, 23''C-25"C, 34%o. fed on a 1:1 mixture of Cluwtoceros gra- Reproductive Strategies in Tropical Bivalves 124 cilis and l\ear cycles of <;anietoj;enic activity and shell free dry weight cycle [Pterin colymhus. I'iiina cornea and Pinclada imhrieahi). l)-d) larval and spat abundance hcti\ecn March l'>'>4 and .\ugust 1998 (b: Pteria eohmhiis. c: Pineladu imbricata. d: Pinna earnea), e) temperature and salinity, 1) particulate organic matter. Rf.productivk Strategies in Tropical Bivalves 1131 1994 1 1995 1 1996 1997 1 199S temperature PI 1:1 Nino r {] A r. r 1 salinity ■ i|ii- T -m 1 1 n part, org matter n r ^ - . n . _p ^ M .1.^. "* "* O^C3s CP-<> OvOv t^'^ONO'C (Tf X ri H Ji. ' -'-!-■ ^ ' g 1 ^ i i 4^ C 3 aj ^ C ■ — 1 en P 3 t '^ ^ ?s «^ 0^ O^ O^ '-^ OS c p. o a E .2, Figure 4. Kl Niiio effects on temperature, salinity and particulate or- ganic matter. Decreasing or increasing tendencies are indicated by arrows. DISCUSSION In areas with a strong seasonality, reproduction (niattiration and spawning) is often toiind to be triggered directly by temperature (Alagarswami 1966. Urban & Campos 1994). Alternatively, many reports on temperature as an indirect trigger have been pubhshed. For example larvae abundance (Mytilus editlis, and Mcicoma hal- ihica larvae from the North Sea) are correlated to algal bloom in summer that depends on temperature (Niesel & Giinther 1999). In the current study under tropical conditions we would have ex- pected that reproduction depends on salinity and food availability rather than on temperature (Barnes 1957) because compared to temperate and Polar Regions in the tropics, annual temperatures are less variable. However, this prediction could not be confirmed, on the contrary spat abundance was best coirelated with tempera- ture. Recruitment (i.e.. spat abundance) that follows after reproduc- tive processes such as gonad production and spawning can be assumed to be temporally coupled. However, the literature gives numerous examples where reproduction and recruitment are highly variable and uncoupled. This is explained by factors such as patchy food supply for larvae (Langdon 1983), thus, causing an extension of the developmental period and a reduction of survival owing to larval food limitation (Olson & Olson 1989). Furthermore, oceano- graphic processes intluencing larval dispersal (Strathman 1974) and — especially in tropical regions — longer continuous reproduc- tive periods uncouple reproductive processes. This study con- firmed these results in the case of larval and spat abundance: Although in several cases larval and spat abundance is in line, the general picture is that both cycles are variable following different patterns. Apart from the possible explanations stated above, this could be due to the fact that zooplankton samples were taken only monthly. Thus, peaks could easily have been missed, whereas spat abundance that was recorded from collectors left for six weeks could be assumed to be more reliable. This aspect is important because at least for Pinna carnea and Pinctada imbricata, repro- duction (gametogenic activity or gonad production) correlate ex- tremely well with spat abundance while in Pteria colvmhiis these processes seem to be uncoupled. Owing to a lack of data for Pinna carnea. instead of the gonad production the gametogenic stage "ripe" is given. Urban and Riascos (submitted) found the nHuithly gonad \olunie cycle and the gametogenic stage "ripe" cycle of Pinna carnea from Tayrona Park to be positively corre- lated, explaining this finding as follows: The gametogenic stage "ripe" gives the percentage of specimens with ripe gonads (shortly before spawning). Gonads are most voluminous in this develop- ment stage as egg and sperm size is large. As the method to estimate gonad production is based upon the monthly gonad vol- ume, gametogenic stage "ripe" and gonad production are also cor- related. The apparent "direct" coupling of reproduction and recruitment in Pinna carnea and Pinctada iinbricata at first appears suspicious because a time lag would have been expected, i.e.. spat peaks and gonad production peaks should not be observed at the same time. Larval periods of bivalves have been the subject of numerous studies, and it is ob\ ious that their duration depend principally on temperature and food. For example, the larval development of Dona.x serra from the productive and temperate Benguela up- welling system lasts 3—1 weeks (Birkett & Cook 1987. Donn 1987). and for the tropical Pinctada maxima the larval develop- ment time is 28-35 days (Rose & Baker 1994). Besides this "time problem" it has to be taken into account that gonad production of Pinctada iinbricata was recorded for a population in Cabo de la Vela. 300 Km from Tayrona Park (where spat experiments were carried out). However. Pinctada imbricate/ spat from collectors after 6 weeks ranged from 1—1 mm shell length, which is in good agreement with larval growth experiments in which spat after 6 weeks had a mean length of 2.8 mm. Furthermore, laboratory- reared spat used for growth experiments after 6.7 weeks measured 2.9 mm (±0.675 S.D.). The question remains whether larvae could have been transported from Cabo de la Vela to Tayrona Park within this time. The Caribbean Cuirent (Thurman 1994) flows parallel to the Colombian coast into the Gulf of Mexico with a velocity of 0.5 m s"' (Tomczak & Godfrey 1994). thus spawned Pinctada imbricata larvae can be transported from the Cabo de la Vela to Tayrona Park within 6.9 days (Fig. lb). According to laboratory experiments, larval development until settlement lasts 23-28 days. Assuming a similar development time for Pinna car- nea. "collector periods" of six weeks would lead to such overlap- ping peaks of recruitment and reproduction as found in Figure 6, with no time lag between coupled gonad production and spat peaks being observed. At high latitudes seasonality is principally governed by the large annual temperature variability, and at low latitudes (i.e.. the tropics) with more constant temperatures, there is little seasonality. However, a clear seasonality of environinental factors other than temperature can be observed in tropical regions. In the study area a combination of precipitation and wind patterns seem to be the principal driving force. Annually, the rainy season (nor- mally between June and November) is followed by the dry sea- son (normally between December and May). Precipitation de- creases salinity and the run off of nutrients increases POM. On the other hand a local upwelling (Blanco 1988) towards the end of the rainy season leads to a drop of mean temperature (?4°C). This explains the apparent negative correlation between temperature and salinity. As mentioned above. El Nifio increases precipitation in the semi-arid region of the Colombian Caribbean close to Venezuela. Thus, decreasing salinity, increasing POM and negligible effect on temperature during El Nino can be postulated. This is an interest- ing parallel comparing El Nifio effects in the South American Pacific where temperature increases, and (in the pelagic regions), nutrients decrease. In the Caribbean, salinity variability owing to 132 Urban TABLE 1. Spearman correlation matrix of larval, .spat, temperature, salinity and particulate orj^anic matter (POM) cycles from March 1994 to .\ugust 1998. Values > = 0.5 are printed bold I^arvae Spat Pleria Pinclada Pinna Pleria Pinclada Pinna lohmhus imhricata cornea eolynihus iinhricala carnea Temperature Salinity P. iiuhncahi 0.6 Larvae P. carneu 0.4 0.7 P. calxnihiis 0.1 0.4 0.1 Spat P. inrhncithi 0.1 0.2 0.1 0.6 P. canieu -0.2 -0,1 -0.1 -0.2 0.1 Temperature -0.2 -0.3 -0.2 -0.5 -0.4 0.4 Salinity -0.1 0.0 0.0 0.2 0.3 0.0 -0.5 POM 0.2 0.1 0.1 0.0 0.0 -0.2 0.1 0.0 El Nino most likely has little inlliienee on the ecosystem, lii the Pacific how ever, the reduction of nutrients during El Nino, leads to a characteristic breakdown of food chains (Glynn 1988. Enfield 1988. Arntz & Fahrbach 1991 ). Considering that in the Caribbean during El Nifio 1997-1998 POM increased almost three fold, it 1994 1 1995 1 1996 1 1997 | 1998 | 3 - 4 ■ P, colymbus larvae ' 2 ■ n [In -| ^ S ■ 4 - 1 spat r 1 ' 2 - n .- ^ n 3 ■ A ■ 1 ■ larvae r P. imbricata 2 ■ 1 • r ri n- I7L ^ 3 ■ ■4 ■ 3 ■ spat i i - \ 1 2 ■ m ^ \L J r n n , 3 - 4 - 3 - larvae P. carnea 2 - 1 - n r 1 n np S - 4 - 3 - spat + X 2 1 _a — ^- -.XL-. n , ,r^ r A ^, Figure 5. Larval and spal ahundunce of Pleria cnlymhus. Pinclada iinhricala and Pinna earned. (Black arrov\s indicate a good agreement betv»'een larvae and spat peaks, i.e., a larvae peak shortly before a spat peak). could he assumed that suspension feeding secondary producers such as bivalves benefited from the increased POM pool, gaining more energy for production. However, except for Pinclada imhricata. no diffeiences in lar- val or spat abundance that could be attributed to El Niiio-induced changes of environmental factors were observed in this study. Moreover, Pinclada imhricata rather showed a leduction of spat abundance during El Nino. This finding might be explained by noting that filtration rales of filter feeding bivalves aie positively " 1994 199 5 1 1996 1 1997 I 1998 | ■ ' colymbus spat p - n - . P. mbricala r- ■ spat ■ 8 ■ T - -r -1 ^ F. arnei .pal T I - - £ in -a temperature [41 rf r ] ^ r.[\l " salinity n „ n n n r^n in n larl. org. matter „ r, n n . .I-.. Jl-a_- r... n n rnfTI [ ■ — ■ ^o T3 c - 5 o. y M s s .2, s! -3 e .2, "•Sea, Figure 6. Spat abundance of Pleria eolynihus. Pinclada imhricata and Pinna carnea compared with temperature, salinity and particulate or- ganic matter. (Periods of spat peaks are indicated with gray bars). a) b) § c) d) P. colymbus Reproductive Strategies in Tropical Bivalves Rcarnea R imbricata 1133 1W4 shell free dry weight gonad production spai 111 temperature □ salinity Ji LL 1995 _^J. rrn X garnet, stage "ripe' rt W:. XL Jll a IX 19971 i=i gonad production n EL Jl a Mm Apt Miy Jul Jul Aug Sep Oa Nov Dec MIS F* Api May Jim Jul. Auf. Stp Ocl Nov Dtc JllSd FA Mu Apr Mai Apt May lui M Aug, Sep Otl Nov Dec JaaNJ Fed Mar a) shell free dry weight of a standard individual [g] b) gonad production of the population [g AFDW m" yr '] (except P. carnea: garnet, stage "ripe" [%]) c) spat abundance [n collector'] d) temperature [°C] & salinity [%o] Figure 7. Temporal coiipiinn (black arrows) of yonad production and spat abundance of Fteriu colymbus (March 1994 to February 1995) Pimlada iiiihiicatci (April 1995 lo April I996| and Piiiiiii iiinnii (March 1997 to March 1998). related to particle concentratiod tmly up to a cectaiii optiiiiuin concentration (J0rgensen 1960. Foster-Sinith 1975). Should the POM concentration surpass this optimum no extra surplus energy would be assimilated. As to discuss reproductive strategies we have three epifaunal filter feeders distributed in the same study area. The principal differences in their niches .seem to be substrate preference (Pterin colymbus: octocorals, Pinctada iinhrictiur. sea grass coinniunity. Pinna carnea: buried semi-infaunal). They all display continuous reproductive cycles, typical tor tropical regions with spent gonads during at least seven months of the year, and larvae as well as spat throughout inost of the year. In contrast, for the clam Gari solida from Southern Chile (Urban & Campos 1994) and for the Antarc- tic bivalve Latermtia elliptica (Urban & Mercuri 1998) niorc dis- continuous reproductive cycles were observed with principally spent gonads present only during 2-3 months of the year. Apart from common characteristics, this study revealed the following differences: The two closely related pearl oysters {Pterin c(ilynil>n.\ and Pinctada imbricata) show several similarities such as high conelation between larval as well as spat cycles. In com- parison, spat cycles of Pinna carnea. are not coirelated with either of the spat cycles of both oysters. Furthermore, lor hcnh pearl oysters temperature is negatively correlated with spat, (and posi- tively correlated with Pinna carnea spat). On the other hand. Pte- ria colymbus had only half of the spat peaks, and thus less mean spat abundance along with uncoupling of gonad production and spat cycles that clearly displays differences from Pinctada imbri- cata. Also, gametogenic activity of Pteria colymbus is less con- tinuous than the latter two species, because at least during a two- month period (December 1994 to January 1993) little or no spent or ripe gonads were present. Thus, on a scale from discontinuous reproduction of polar species to continuous reproduction of tropi- cal species as the extremes, we would find Pteria colymbus in a more intermediate position inclined towards the "tropical group". El Nino S oi c o i: 1994 1995 I 1996 1997 |1998~ Figure 8, Kffects of Kl Nino on salinity, particulate organic matter and spat abundance o( Pteria colyitihiix. Pinctada imbricata and Pinna car- nea. Six months means of original data. 134 Urban ACKNOWLEDGMENTS The author would like to mention F. Marcos Ablanque. Juan Pablo Assnius. Claudia Castellanos. Carolina Garcia. Socorro Sanchez and Adriana Valero for their collaboration in the field, as well as working on the samples. My thanks to Dr. F. Zapata and Dr. T. Brey for supporti\e discussions on the analysis of time series data. LITERATURE CITED Alagarswami. K. 1966. Studies on Mime aspects of biology of the wedge clam Domix fcibu from Mandapam Coast in the Gulf of Mannar. ./. Mar.Biol. Ass. India 8:56-75. Arizpe. C. O. & R. Felix. 1986. Crecimiento de Pinna iiigosa (Sowerby. 1835) en la bahia de La Paz. Mexico. Univ Nal. Auton. Mexico. An. Inst. Cienc. del Mar y Limnol.. 1.^:167-172. Amtz. W. E. & E. Fahrbach. 1991. El Nino— Klimaexperiment der Natur. Berlin: Birkhauser Veriag. 264 pp. Baqueiro. E. & M. Castagna. 1988. Fishery and culture of selected bivahes in Mexico: past, present and future. ,/. of Shellfish Res. 7:433—143. Biirnes. H. 1957. Process of restoration and synchronization in marine ecology. The spring diatom increase and the spawning of the common barnacle Buhunis balanoides (L.) Ann. Biologiques 33:67-85. Blanco. J. 1988. Las variaciones ambientales eslacionales en las aguas costeras y su importancia para la pesca en la region de Santa Marta. Caribe colombiano. Tesis: Sc. Univ. Nac. de Colombia. 60 pp. Birkett. D. A. & P. Cook. 1987. Effect of the Benguela temperature anomaly, 1982-1983. on the breeding cycle of Dona.x seira Riiding. S. Afr. J. mar. Sci. 5:191-196. Cendejas. J. M.. M. G. Carvallo & L. M. Juarez. 1985. Experimental spat collection and early growth of the pen shell. Pinna riigosa (Pelecypoda: Pinnidae). from the Gulf of California. Aquaculture 48:331-336. Crisp, D. J. 1984. Energy flow measurements. In: N. A. Holme & A. D. Mclntyre. editors. Methods for the Study of Marine Benthos. London: Blackwell. 284-372 pp. Davy, F. & M. Graham. 1982. Bivalve culture in .Asia and the Pacific: Proceedings of a workshop held in Singapore. February 16-19. 90 pp. Donn. T. E. 1987. Long shore distribution oi Dona.x serra in two log-spiral bays in the Eastern Cape. South Africa. Mar. Ecol. Prog. Ser. 35:217- 222. Enfield. D. B. 1988. Is El Nino becoming more common? Oceanogr. Mag. 1:23-27 & 57. Foster-Smith. R. L. 1975. The effect of concentration of suspension on the filtration rates of pseudofaecal production for Mytihis ediilis (L.). Cerastoderma edule (L.) and Veneriipis pullastra (Montagu). J. Exp. Mar. Biol. Ecol. 17:1-22. Glynn. P. W. 1988. El Nino-Southern Oscillation in 1982 to 1983. Near- shore population, community and ecosystem responses. .Ann. Rev. Ecol. System. 19:309-345. Lalli. M. & T. R. Pearsons. 1993. Biological oceanography: An introduc- tion. Oxford: Pergamon Press. 301 pp. Langdon. C. J. 1983. Growth studies with bacteria-free oyster (Crassostrea gigas) larvae fed on semi-defined artificial diets. Biol. Bull. 164:227- 235. Jorgensen. C. B. I960. Efficiency of particle retention and rate of water transport in undisturbed lamellibranchs. / Coitc. Perm. Int. E.xphn: Mer. 26:94-116. Niesel. V. & C-P. Giinther. 1999. Spatial and temporal distribution patterns and their underlying causes. In: S. Dittmann. editor. The Wadden Sea Ecosystein. Stability Properties and Mechanisms. Berlin: Springer Ver- iag. pp. 77-94. Olson, R. R. & M. H. Olson. 1989. Food limitation of planktotrophic marine invertebrate larvae: does it control recruitment success? Ann. Rev. Ecol. Syst. 20:225-247. Philipson. P.W. 1989. The marketing of marine products from the South Pacific. Inst, of Pac. Stud, of the Univ. of the South Pacific. 307 pp. Rose. R. A. & S. B. Baker. 1994. Larval and spat culture of the Western Australian silver- or goldlip pearl oyster. Pinctada iiiaxinia (Jameson) (Mollusca: Pteriidae). Aqiiacidture 126:35-50. Strathman. R. 1974. The spread of sibling larvae of sedentary marine in\ei-tebrates. .4;n. Nat. 108:29-14. Thurman. H. 1994. Introductory oceanography. New York: Maximilian Publishing Company. 552 pp. Tomczak. M. & J. S. Godfrey. 1994. Regional oceanography: An intro- duction. Oxford: Pergamon Press. 422 pp. Turner, R. D. & J. Rosewater. 1958. The family Pinnidae in the western Atlantic. John.simia 3:285-326. Urban. H.-J. 1994. Upper Temperature Tolerance of Ten Bivalve Species off Peru and Chile related to El Nmo. Mar. Ecol. Prog. Ser. 107:139- 145. Urban. H.-J. & B. Campos. 1994. Population Dynamics ot the Bivalves Gari solida. Semele soUda and Protothaca thaca from a Small Bay in Chile at 36''S. Mar. Ecol. Prog. Ser. 115:93-102 Urban. H.-J. & J. Tarazona. 1996. Effects of EI Nino/Southern Oscillation on the Population Dynamics of a Gari .«)/i(/(i-Population in Peru (I4"S). .V/,jr Biol. 125:725-734. Urban. H.-J. & G. Mercuri. 1998. Population Dynamics of the Bivahe Ldlerniila elliptica from Potter Cove (King George Island. .Antarctic Peninsula). .4/it. 5c/. 10:153-160. Urban. H.-J. 2000a. Aquaculture potential of the Caribbean Pearl Oyster Pinctada imbricata I. Gametogenic Activity, Growth, Mortality and Production of a Natural Population, .\qiiacidture 189:361-373. Urban, H.-J. 2000b. Aquaculture Potential of the Caribbean Pearl Oyster Pinctada imbricata II. Larvae and Spat Collection, Juvenile Growth and Mortality in Culture Systems Related to Ambient Conditions. .\qitaculnu-e 189:375-388. Urban & Riascos, submitted. Estimating Gonad Index and Gonad Produc- tion in Tropical Bivalves with Fused Gonads and Continuous Repro- duction. .loiiriHil of Shellfish Resninh. Vol. 2(1, No. 3. I 135-1 144. 211(11. POPULATION STRUCTURE AND RECRUITMENT OF THE BIVALVE ARCTICA ISLANDICA (LINNAEUS, 1767) ON GEORGES BANK FROM 1980-1999 CRAIG V. W. LEWIS,' JAMES R. WEINBERG,- AND CABELL S. DAVIS' ^Dcpr. of liiici^rativc Bi(>l(>;i\. U.C. Bcrkclcw Berkeley. Californiii 9472U-3140: -National Murine Fisheries Service. Woods Hole, Massailnisells 02543: ^Wooils Hole Oeeaiiographie liisiitiiiion, Woods Hole. Massachusetts 02543 ABSTRACT .Airrica isliinilicii \s a commercially imponant bivalve from the North Atlantic Ocean. Due to its slow growth rate and longevity, long-term studies are needed to understand its population dynamics, particularly recruitment. This study describes the spatial distribution and population structure of A. isUiiulica on Georges Bank. USA from 1980-1999. Results are based on sainples collected with a hydraulic clam dredge during National Marine Fisheries Service ocean shellfish surveys. In all surveys from 1980-1999. individuals were abundant on the south tlank of Georges Bank between 60 and 75 m depth. In the 1980s, size distributions in tows were typically unimodal. and the population was dominated by large individuals. 75-90 mm in length. In contrast, bimodal size distributions were observed at many stations in the 1990s, and small uidividuals (<70 mm in lengthl often represented 20-50'f of the catch (by number). These small individuals were the result of spawning in the 1980s. Ages of clams were estimated from shell banding at one station in 1994. Most individuals had less than 40 bands, and individuals between 21 and 28 were rare. Among young individuals, modes were centered on 7 and 12 bands; assuming band formation is approximately annual, this suggests recruitment by cohorts spawned during 1986 or 1987 and 1981 or 1982. These results imply that annual recruitment of small A. islcmdicii on Georges Bank has been highly variable during the last 40 years and suggest an increase in recruitment to the fishery after 1990. KEY WORDS: Aniica i.\Uiiuhcti. Georges Bank, recruitment, growth, population dynamics INTRODUCTION The purpose of this stuiiy is to provide a comprehensive analy- sis of the NMFS research survey data collected on the ocean qua- hog. Arctica ( = Cyprinu) islandica {Linnaeus 1767). from Georges Bank. USA between 1980 and 1999. Specific goals were to describe the size and age structure, recruitment, individual growth rate and spatial distribution of .4. islandica. Previous re- search on A. islandica from Georges Bank includes stock assess- ments (e.g.. NEFSC 1990, 1996, 1998), relation.ships between age and growth (Ropes & Pyoas 1982). and population genetics (Dahl- gren et al. 2000). .Arctica islandica is unusual in having one of the slowest adult growth rates among tnarine bivalves (Turekian et al. 1975. Thotnp- son et al. 1980a. Thompson et al. 1980b. Murawski et al. 1982. Fritz 1991, Kraus et al. 1992). The species is widely distributed along the coasts of the North Atlantic Ocean (Nicol 19?1. Merrill & Ropes 1969. Brey et al. 1990. NEFSC 1998. Witbaaid et al. 1999). In the United States, a valuable comtiiercial fishery for this species has existed since the 1970s (Murawski & Serchuk 1989. Serchuk & Murawski 1997. NEFSC 1998). In this fishery, the clams are typically harvested between 80 and 100 mm in shell length (NEFSC 1998). A commercial fishery has also developed recently in Iceland (Thorarinsdottir 1997). Understanding the population dynamics of slow growing species, particularly recruit- ment, can provide the information needed to determine rates of sustainable exploitation. Inherent in the measurement of recruitment is the definition of the term recruitment. For purposes of this study, an individual is considered to have recruited when it is large enough to be mea- sured in the National Marine Fisheries Service (NMFS) stock as- sessment survey (this size is approximately coincident with the age of first reproduction (Thompson et al. 1980b) and the size at which individuals are vulnerable to commercial fishing equipment). This functional (and liberal) definition implies that a recruited indi- vidual has survived a host of planktonic. settlement and post- settlement hazards. For comparison, the term settlement will be used to describe the arrival of a meroplanktonic larvae onto suit- able substi'ate. In Arctica islandica the years between .settlement and recruitment involve tnyriad largely unknown risks and pro- cesses. Our interest focuses on developing an understanding of the temporal variation in recruittiient to the fishery using the NMFS data. The National Marine Fisheries Service has carried out standard research surveys of Arctica islandica off the US east coast since 1980. With the exception of an area south of Long Island. NY (Murawski et al. 1982), there was little evidence of small (50-70 mm) individuals throughout the 1980s (NEFSC 1998): however, the abundance of clams in this size class has always been uncertain because the clam dredge used to collect the survey samples does not retain all of the individuals below 70 mm in length (NEFSC 1998). The focal region of this study, Georges Bank, has recently become the focus of a tnajor interdisciplinary research project on interannual variability in biological and physical oceanographic processes (GLOBEC 1992). Georges Bank is unique for A. is- landica because a commercial ocean quahog fishery has not oc- curred there. This represents an opportunity to characterize popu- lation dynamics in the absence of direct fishing pressure. Fishing pressure on groundfish and sea scallops in this region has been intense (NMFS 1999), however, and is likely to have caused in- direct mortality to A. islandica and other infaunal species (Wit- baaid & Klein 1994, Bergman & Van Santbrink 2000). MATERIALS AND METHODS Sltidy Organism Arctica islandica is a filter-feeding bivalve connnon through- out the cold, coastal waters of the North Atlantic. It is the only extant species in its genus, which dates back to the Cretaceous (Nicol 1951, Abbott 1974). It is found along the North American coast from Cape Hatteras in the south, to Georges Bank, the coast of Maine, and Newfoundland in the north (Merrill and Ropes 1969. NEFSC 1998). No extant populations occur off the coast of Greenland (Jensen 1912. Ockelmann 1958). Along the coast of Europe it is found from the White Sea to the Bay of Cadiz, while 1 1 35 1136 Lewis et al. the Arkona Sea is the eastern limit in the Baltic (Nicol 1951, Merrill & Ropes 1969. Brey et al. 1990). The species has also been reported from the coasts of Iceland, the Faroes, the Shetlands. the British Isles and Norway (Nicol 1951. Thorarinsdottir 1997). Along the U.S. coast, A. islandica typically occurs at depths of 30 to 75 meters (NEFSC 1998). It is usually found in sand and sand- shell substrates (Fogarty 1981. Thouzeau et al. 1991a. b. Thorarins- dottir 1997). Arctica islimdica is one of the longest-lived bivalves on the continental shelf, ages of 50-100 years are common in natural populations, and the maximum repiirted age is approximately 200 years old (Thompson et al. 1980a. b. Murawski et al. 1982. Fritz 1989, Steingrimsson and Thorarinsdottir 1995). Rate of growth is relatively fast until the age of 7 years (Thompson et al. 1980a,b). The growth rate of older individuals is among the slowest known for bivalves, with the exception of cer- tain deep-sea species from soft sediments (Turekian et al. 1975, Thompson et al. 1980a, Murawski et al. 1982, Fritz 1991). Ropes and Pyoas (1982) showed that the growth rate of A. islandica is faster on Georges Bank than off the coasts of Long Island. NY. USA and Sable Island. Canada. Field studies have shown that among older individuals, females tend to be larger than males (Ropes et al. 1984. Fritz 1991 ). In the US Middle Atlantic Bight, age at maturity ranges from 4 to 14 years (Thompson et al. 1980b). Longevity and growth rate determinations have been based on tagging studies and length frequency modal progression (Muraw- ski et al. 1982), as well as on counts of microscopic growth lines within the shell. Numerous studies indicate that these lines are deposited annually and can be used to obtain a reliable estimate of individual age (Jones 1980. Thompson et al. 1980a. Murawski et al. 1982, Turekian et al. 1982, Ropes et al. 1983, Ropes 1984, Ropes & Jearld 1987, Weidman & Jones 1993, Weidman et al. 1994). The exact determination of age from band counts is not always possible because physical disturbances can cause ocean quahogs to produce additional bands (Turekian et al. 1982. Ropes et al. 1983. Ropes 1984, Kraus et al. 1992) and shell wear may result in loss of bands (Ropes 1984). Data Collection A stratified random sampling design was used by the National Marine Fisheries Service in surveys to assess stock biomass and size-structure of Arctica islandica on Georges Bank (Fig. 1; NEFSC (1998)). Surveys analyzed in this paper were earned out during the summer months of 1980, 1982, 1983. 1984, 1986, 1989. 1992. 1994, 1997, and 1999. A hydraulic clam dredge, 152 cm wide and weighing 3.2 t, was used to collect all samples. The openings in the body of the dredge are 2.5 x 5.1 cm. Samples of ocean quahogs were collected during surveys by towing the dredge for five minutes at 2.8 km h"' after it made contact with the bottom. The actual distance towed varied between stations due to differences in station depth and winch speed, which affected time to deploy and retrieve the dredge. Actual distance sampled at each station was determined from doppler readings for surveys conducted from 1980 to 1994, and from bottom contact sensors and inclinometers mounted on the dredge in 1997 and 1999. For analysis, catch at each station was standardized to a com- mon distance of 278 m (i.e., 0.15 nautical miles), assuming a linear relationship between catch per tow and distance sampled. Given the dredge width, this represents an area sampled of 423 m" per ^41 69 68 67 Longitude (°W) Figure 1. National Marine Fisheries Service ocean slicllfish survey strata on (Jeorges Hank. Boundaries of tile survey strata are at the 25, 30, 4(1 and 6(1 I'uthom isobaths ll fathom = 1.83 m). Dotted line indi- cates the boundary between L.S. and Canadian Haters. standard tow. This is a conservative estimate of the true density, as the dredge is likely to pass over and destroy .some clams (Smolo- witz & Nulk 1982). In addition, the dredge has only partial reten- tion of ocean quahogs that are <70 mm in shell length (NEFSC 1998. 2000); while specimens much smaller than the mesh size are frequently retained in the sample, their retention is likely to be very unreliable. Additional details about the clam dredge and sensors are given in Smolowitz and Nulk (1982). and NEFSC (1998). All live ocean quahogs at each station (i.e.. tow) were counted and shell lengths, defined as the maximum anterior to posterior distance in mm, were measured at sea to compute average size- frequency distributions per tow for each stratum. In 1994, a bimodal size structure was noted in several tows on the south tiank of Georges Bank (Lewis 1997). The existence of numerous small (<70 mm) clams implied a surge in recruitment to that area. To estimate when these ocean quahogs were spawned, a random subsample of 144 individuals (from a total of 883) was taken from Station 448 in Stratum 61 (41°9'N and 67°1 'W) for age estimation based on counts of internal bands. Shell lengths were also measured to estimate growth rate. The procedure for determining age involved several steps. Each clam was labeled and a valve was selected for sectioning. No preference was given to right or left valves. Shells were sectioned with a glass-cutting saw from the umbo to the most distant portion of the shell. The cut edge was polished on Buehler Polimet Pol- isher and glued to a microscope slide. A Buehler Isomet 1 1-1 180 Low Speed Saw was used to cut most of the shell away from the slide, leaving a 600-800 fjim thin section embedded in epoxy. Sections were ground to 250-300 \x.m on a Ingram Laboratories Model 305 Thin Section grinder and polished to 150 fjim. To determine clam age, shell bands in prepared sections were counted at least twice per individual using a microscope at x25 magnification. In the present study, counts of bands were used to indicate age-structure of the population, even though estimated age may not be exact for every individual in the sample. In most cases, an underestimate of age is likely. This is due to the timing and Population Structure and Recruitment of A. isl\ni>jca 1137 placement of the year I band which is deposited only 3 to 6 nioiitlis after settlement in a region of the shell that shows heavy abrasion. While not accounted for in our growth calculations, this adds further ambiguity to exact estimates of age. Age ami Leiif;tli Structure Analysis A major purpose of the present study was to detect recruitment into the population as indicated by the existence of multiple modes in si/e- and age-distributions in the NMFS survey data. We used a randomization procedure (Silverman 1981. 1983: Efron & Tib- sherani 1993) to statistically test each observed size- and age- distribution for the presence of multiple modes. The test statistic was the minimum value of a smoothing kernel width. /;, (referred to as "window size"), that would smooth the sampled frequency distribution until there was only one mode or local maximum. A large value of /i, is required to smooth a distribution with strong and widely separated peaks. The value of /;, from the observed data was compared to a probability density function generated by calculating h] for each of a large number of bootstrap samples from the original data. The null hypothesis; //,,: number of modes = 1. was rejected if the probability of finding /), greater than li\ was greater than or equal to 0.95. In cases where the null hspothesis of one mode was rejected, the procedure was repeated, calculating /n. a window size that generated 2 modes, with a new set of bootstrap samples based on that w indow size. In general, if the null hypothesis. W,,: number of modes = k. was rejected, then the hypothesis. H„: number of modes = k + 1 . was tested. The exact procedure for the bootstrap involved a number ot numerical steps. The minimum value of /; such that the function. ,/(/;//) = ±4'^) (II has only one mode or local maximum (c|)(rj is the standard normal density), was determined by successive approximation (values of .V, refer to length, to the nearest mm. or age. in years, of individual clams selected from among the n clams in a given tow). This minimum value was the test statistic. /?,. Each bootstrap sample contained /; individuals drawn at ran- dom from the smoothed distribution of the original data following the methods of Efron and Tibshirani (1993). In all tests conducted here, the distribution oi h\ was based upon 200 bootstrap samples. First, n individual values. y\ . . . y'„. were sampled, with replace- ment, from the original data set. .v, . . . .v„. The values of the ii members of the bootstrap sample were calculated from the y' values usinsi the formula; accepted. Otherwise, ii PUi\ > h,)- U'*?' the- null hypothesis that there was only one mode was accepted. The statistical test for multiple modes in a frequency distribu- tion was carried out on the observed size-frequency data at each station from Georges Bank. In addition, the test was applied to the observed age-frequency distribution at Station 448 in Stratum 61. defining the .v values as the number of shell bands per individual. RESULTS Anticii ishiiiJiia were concentrated around the periphery ot Georges Bank, with greatest concentrations on the south tlank between 60 and 75 meters depth (Fig. 2 and Fig. 3). Moderate densities of ocean quahogs were found on the steep north tlank between 50 and 1 00 meters depth. Relatively few ocean quahogs were collected from shallower stations located near the center of the Bank (Fig. 3). On the south tlank. .4. islaiulica were present at almost all stations between 50 and 80 meters depth. The deep water limit for the species was not completely resolved by these surveys, although samples taken at depths greater than 90 meters contained very few individuals (Fig. 2 and Fig. 3). From the 1980s to the 1990s there were changes in size struc- ture of the population on Georges Bank that were suggestive of an increase in recruitment. In the 1980s, the size structure within most strata was unimodal and dominated by large. 75-90-mm. individu- als. In contrast, the size-frequency distributions from 1992 to 1999 contained large individuals as well as numerous (i.e.. tens to hun- dreds) small clams. 40-70 mm in length. The dredge limitations make it difficult to assess the magnitude of the recruitment, but the consistency of the observations over several regions and years provide strong support for the hypothesis that the observations are more than a mere sampling artifact. The data from 1982-1984 and 1994 (Fig. 4) demonstrate typical size frequency distributions from surveys in the 1980s and 1990s, respectively. The appearance of small individuals in the 1990s was wide- spread along the south Hank of Georges Bank, including Strata 57. 59. 60. and 61 (Fig. 1. Fig. 5. and Fig. 6). This increase in small clams was unlikely to be a sampling artifact because the same dredge, bar spacing, and mesh liner were used throughout the 1000 XI = v + 1 -I-- ;(v/ );i=l,2. (2) (£ 750 7. 500 where v' was the mean of the y' values, a" was the variance of the original sample, .v. and the e, were standard normal random vari- ables. In practice, rather than determine the exact value of /j, needed to generate a smoothed bootstrap sample with one mode, it was only necessary to know whether li\ was greater than or less than /(, for each bootstrap sample. \i Pih', > h,)> 0.05 (10 out of 200 bootstrap samples), then the null hypothesis was rejected and the alternative hypothesis, that there was more than one mode, was Figure 2. Mean (bar) and 95'7f confidence interval for number of Arctica islandha per standard lo«. in five meter depth hitervals for all representative tows on (ieorges Bank from 1980-1999. Number of tows taken within depth intervals is shown at bott early in its life span (Fig. 10 and Table 1 ). Individuals grow faster on Georges Bank than in other regions for which growth curves exist. In the 1994 age saniple. clams pos.sessing 1 1-14 bands had an average length of 69.2 mm. The three clams with over 70 bands were smaller than sonie with a few as 29 bands, implying high variability in growth rate among individuals. An abrupt transition in band structure, from broad (."i-IO mni) to narrow (~l mm) spacing between successive bands. occuiTed between 7 and 10 bands, implying a shift from relatively rapid juvenile giMwlli to slow adult growth. DISCUSSION Several niechanisms niay explain the distribution of Antica IshiiidUa on Georges Bank between 60 and 90 )iiete)s. At these KIIK) HK. 10 UHH) 10 Siraium 37 S6 . M KHN) KM) HI J\ 1011(1 HKI 1(1 97 moo IIM) 10 99 30 40 50 60 70 90 100110 121 30 40 50 60 70 80 90 100110120 HHK) .^iiLiiiJin DU , ^m^ ^^r^^ xiL ^^M\m^ jgum^ jMMh^ 10 lUOO HKI 10 S 1000 3. MHl o "* ■^ lOlM) UJO .^^itPT^. 86 rfTTTh-i ,^ 94 ■iifcr^l^i i^.MU^i 99 30 40 50 60 70 80 90 1001)0)20 K) 40 50 60 70 SO 90 )00 1)0)20 S)ie)) )engi)i (mm) Figure 5. Size structure of .4r(7(((/ islandiia in four strata from the south side of Cleorges Hank, 19S0-I999. (Jraphs show mean (bars) and upper 95% confidence interval (line) for indicated years. Size classes <70 mm are shown in black to highlight the presence of small indi- viduals. depths, stratification is ra)ely bj'oken down by tidal tiiixing. Hence, bottom water tetnperature is typically cooler and there is lower seasonal variation than in the shallow tidally mixed region within the 60 meter isobath (Flagg 1987). On the south flank, the highest survey catches were recorded near the 60 meter isobath, roughly corresponding to the location of a convergent front that defines the boundaries of the stratified Shelf Water with the tidally mixed Georges Bank Water. This front may serve to accurnulate upward swimming larvae at the surface (Mann & Wolf 1983. Franks 1992). Adult ocean quahogs have low sur\ i\al at water temperatures greater than 20°C (Turner 1949. Nicol 1951. Landers 1976). Oc- casional high temperatures in the well-niixed central bank water ma> he sufficient to exclude Arctica iskindica from this region. Furthertnore. tidal currents over the central bank consistentiv re- suspend sediments which tnay ovowhehn the feeding capability of this filter-feeding bivalve. Finally, cotnpetition between A. is- Icimlica and the Atlantic surfclam. S/ii.'uda soUdissima. tnay favor ]hc latter in the shallow waters. The distributions of these two species are completnentary throughout the Middle Atlantic Bight, with A. ishindica found in coolei". deeper water (NEFSC 1998). The differing thermal loleiances of the two species are a likely factor in their segregation. The results suggested that Arctica ishindica recruit on Georges Bank on an irregular basis. A. ishindica populations may survive for long pericids of tiine even if recruitment occurs infrequently because of this species' longevity and iteroparous reproductive 1140 Lewis et al. 69 68 67 ^(Shell Lenglli<7Umm)= 66 69 68 67 Longitude ("W) lu <= 20 o 30 o 66 69 40 O 50 O Figure 6. Percent of catch that was smaller than 70 min in each t(iH from the 1980-1999 survey cruises. Radius of circle is proportional to percentage (the very large circle in 1994 reflects the presence of a single small clam at the stationl. Isobaths are in meters. mode. It is dilTicult Id identity tlie exact tiniini: and nature of large recruitment events because the ocean quahogs on Georges Bi\nk are not fully retained by the dredge until they are about 70 mm in shell length or about 10 years old. At the station where age- structure was determined, there were apparently 2 major cohorts that settled during the 1980s. In contrast, the absence of individu- als between the ages of 21 and 28 in 1994 suggests that little recruitment took place from 1967 to 1974. The high variability in the age structure sample (Fig. 1) also suggests highly intermittent recruitment at this site. Compared to Arcrica islandica. shorter-lived and faster- growing bivalves appear to recruit more frequently and with less variation in recruitment strength (Tremblay & Sinclair 1990. Tremblay & Sinclair 1992). Research on two large bivalves from Georges Bank. Spisiila solidissinui and sea scallop (Placopecten mugellaniciis). provides an interesting contrast with A. islandica. 69 68 67 66 69 68 67 Longitude (°W) Figure 7. Location of all tows in which two or more modes Here detected in the size structure (triangles) for each year, the lowest mode was centered helow 70 mm, and density of all individuals less than 70 mm was greater than 20 per standard to\\. Isobaths are in meters. Population Structurb and Recruitment oe A. islandica 1141 TABLE 1. Equations for growth curves shown in Figurt 10. Label Equation Source GB94 L = 97.57( I GBS2 L = 42.4(<;)"'"'' MAB L = 97.28(1 _ /"'"i«.+ mc,67,) COM L = 59.470(1 _(,-""??'"*"-"., Current .study. SA.S NLIN least squares fit Ropes and Pyoas (1982) NEFSC (1990. 1996) Kraus. et al. ( 1992) L = length; u = age S. soliilissiiihi has a )iia\iiiii))ii lilcspaitd (it'appro.xidialcly .-^.^ years (Weinberg 1999) and P. itnigclUinUiis appears to li\c for 111-15 years (Thouzeau et al. 1991a, Thouzean et al. I99lbl, which makes both species short-lived compared to A. islandica. S. so- lidissima recruited successfully along the US Atlantic coast in most years froni 1978 to 1997 (Weinberg 1999). Several studies have detected distinct peaks in the size structuie of P. ma^iclkuu- cus surveyed on Georges Bank, indicative of good annual reci'uit- nient (Serchuk et al. 1979, Serchuk and Wigley 1986. Thouzeau et al. 1991a, McGarvey et al. 199,^). Given low mortality levels, a long-lived species such as A. isUiudicn can maintain viable popu- lation levels without recruitment for many yeais. w hei'eas shorter- lived species require recruitment to prevent extinction. Many studies report that long-lived species have slow indi- vidual growth rates, and exhibit low resilience to and slow recov- ery from overexploitation (Musick 1999). Although there has been a significant commercial fishery for Aictica islandiai in the U.S. Middle Atlantic Btght since 1980 (NEFSC 1998). the population on Georges Bank has not been exploited yet commerically. Nev- ertheless, several major fisheries (e.g.. scallops, ground fish) have taken place on Georges Bank throughout the 1980s and 1990s (NMFS 1999). Fish trawls impact the benthic conimunity, causing substantial niojtalily in bivahes such as Spisitla sp. and A. is- landica in the North Sea (Bergman & Van Santbrink 2000). Thus, the changes in population structure of .4. islandica over linie re- ported in the present study are probably representative of popula- tion dynamics under conditions of low to moderate hunian distur- bance. We have not determined which biological and physical factors control recruitment of Arcrica islandica. We ha\'e examined sea bottom teniperature records from Georges Bank (Holzwarth & Mountain 1990), and there is no obvious relationship between observed recruitment and temperature at time of settlement. Other factors which are potentially important include predation, spawn- ing success, availability of larvae, and habitat quality on the sea TABLE 2. X^ tests comparing proportion of tows with 2 modes, a lower mode centered below 70 mm and at least 20 clams smaller than 70 mm in the 1980.S and 1990s. Whole Bank South Flank Decade "„,„, Decade 19S0 14(1 1990 258 X' = 13.49 p = 0.00024 1 980 4(1 1990 88 X= = 67.99 />< 0.00001 27 Figure 8. Age structure of ocean quahogs from Station 448 (Stratum 61. 1994 survey), from which a sample of 144 individuals were sec- tioned for age determination. Dashed line shows approximate mini- mum age that may to be undersampled by the dredge because of their small size. Seven individuals (not shown) were over 45 years old. Hoor. Potential predators tin newly settled clams include various groundfish. such as winter flounder (Steimle et al. 1994). sea stars (Franz et al. 1981). cancer crabs, and the gastropods Lunatin and Busycim. Mann (1982. 1985. 1986) provided a complete leview of 4;i- //<.'(( islaiidica'f, larval ecology in .southern New England and on Georges Bank. Although the spawning period is prolonged, there may only be nan'ow windows in time when oceanographic condi- tions allow for high larval survival and settlement. .Mann (1982) hypothesized that phytoplankton becomes available as a major food for A. islandica lar\ ae only after wind and storm events mix the stratified water column. It was further suggested that theie was 110 Figure 9. Size structure (for station #448, 1994 clam survey) for all individuals f>-IO (gray). 11-15 (white) and over 16 (black) years old. The shaded bars indicate the size structure of the subsample analyzed for age structure; the dark line shows the measured size distribution of the entire catch (88.^ individuals) from that tow (smoothed with a 5-mm moving average). 1142 Lewis et al. l_u 100 ^ _ _ - GB82 , - " " GB^M 80 ^^^Tfit '^^ ^ ' ' ' __■_-- MAB 60 - ; " ' X COM ■ 40 1 / y 1 - 20 i ■ ■ 0 10 20 30 40 50 60 Age (years) 70 80 yo 100 Figuri" 10. Relationships beHveen age and shell length in Arclica /,«- laiidica. Solid line ((;B94): a Von Bertalanffv curve fit to the data ("\") collected in 1^94 for the present study of Georges Bank. Dashed line (GB82): a power exponential cur\e from Ropes and I'voas (1982) for George Bank. Dash-dot line (MAB) a von Bertalanffv cur\e from the coast of New York, Middle Atlantic Bight, from NEFSC (IWd. 19%). Dotted line (GOMc a curve from a sample in the Gulf of Maine from Kraus et al. (1992). See Table 1 for equations of the growth curves. no single spawning stimulus; rather temperature, oxygen satura- tion, pH. and food availability all probably have an effect on spawning time (Mann 1982). Bottom characteristics vary greatly among different parts of Georges Bank. The south tlank. where Arctica isUiiulicci are plen- tiful, is composed of poorly sorted gravel, sand and mud. Larger rocks and drifting sand waves, common on the center of the Bank. are relatively rare on the south tlank. From a physical standpoint, the south flank is a stable habitat for A. isUmdica. The population structure oi Arctica islandica on Georges Bank differs from other regions. For example, compared to populations in the US Middle Atlantic Bight (New Jersey to Virginia), the population on Georges Bank in the 1990s had a higher ratio of small individuals (NEFSC 1998). This is consistent with the results showing that recruitment, as indicated by increased numbers of small individuals in samples, increased on Georges Bank in the 1990s. Comparisons between growth curves indicate that Arctica is- landica grows faster on Georges Bank than off the coasts of New York (Murawski et al. 1982). Maine (Kraus et al. 1992). and Iceland (Steingri'msson & Thorarinsdottir 1995). It is likely that the faster growth of Arctica islandica on Georges Bank reflects the region's high productivity (O'Reilly et al. 1987. Backus & Bourne 1987). The experimental v\ork of Kraus et al. (1992) in Maine demonstrated high phenotypic plasticity in the growth rate of this species. In addition. Dahlgren et al. (2000) determined the se- quence of the mitochondrial cytochrome b gene of Arctica is- landica and found little genetic subdivision between Nova Scotia. Canada, and Georges Bank and Virginia. USA. Growth rate in bivalves is affected by many environmental factors, and species may respond differently to these factors due to interspecific dif- ferences in physiology and feeding ecology. For example. Wein- berg and Helser ( 1996) compared the growth rate of Atlantic surf- clams on Georges Bank with that in other regions along the coast of the L'.S. In contrast with A. islandica. Spisiila solidissinia on Georges Bank were smaller at any given age than in warmer re- gions to the south, closer to the center of 5. solidissinia' s range (Weinbeig & Helser 1996). ACKNOWLEDGMENTS We appreciate the hard work on the part of the captain, crew and scientists aboard the NMFS Shellfish Assessment cruises while carrying out the NMFS clam surveys, with special thanks to the RA' Delaware II. We are grateful to Andy Solow for assisting with the statistical analyses and to S. Murawski, R. Mann, and F. Serchuk for reviewing the manuscript. Portions of this work were sponsored by a WHOI Coastal Research Center Grant and U.S. GLOBEC funding to C. Lewis. This is U.S. GLOBEC Contribu- tion Number 184. LITERATURE CITED New York: Van Nostrand Rein- Abhcm. R. T. 1974. American Seashells huld Co. 663 pp. Backus. R. H. & D. W. Bourne. 1987. Georges Bank. Cambridge. Massa- chusetts: MIT Press. Bergman. M. & J. Van Santbrink. 2000. Fishing morlality of populations of megafaima in sandy sediments. In: M. Kaiser & S. de Gioot. editors. Effects of fishing on nontargel species and habitats. Oxford: Blackwell Scientific, pp. 49-69. Brey. T.. W. Arntz. D. Pauly & H. Runiohr. 1990. Arclica (Cyprimi) islandica in Kiel Bay (Western Baltic): Growth, production and eco- logical significance. Jdurual of E.\pcrimenlal Marine BioUiii\ and Ecology 136:217-235. Dahlgren. T. J. Weinberg & K. Halanych. 2000. Phylogeography of the Ocean Quahog [Arctica islandica): influences of paleoclinia(e on ge- netic diversity and species range. Marine Biology 137:487^95. Efron. B. & R. J. Tibshirani. 1993. An introduction to the bootstrap. New York: Chapman Hall. 436 pp. Flagg. C. 1987. Hydrographic structure and variability. In: R. H. Backus & D. W. Bourne, editors. Georges Bank. Cambridge. Massachusetts: MIT Press, pp. 108-124. Fogarty, M. 1981. Distribution and relative abundance of the ocean quahog .^rcriea islandica in Rhode Island South and off Martha's Vineyard. Massachusetts. J. Shellfish Res. 1:33-39. Franks. P. J. 1992. Sink or swim: accumulation of biomass at fronts. Ma- rine Ecology Progress Series 82:1-12. Franz. D. R.. E. K. Worley & A. S. Merrill. 1981. Distribution patterns of common seastars of the middle Atlantic continental shelf of the North- west Atlantic (Gulf of Maine to Cape Hatteras). Biological Bulletin of the Marine Biological Lahoratoiy 160:394^18. Fritz. L. 1989. Seasonal condition of Arctica islandica in the Mid-Atlantic Bight. J. Shellfish Res. 8:462-163. Fritz. L. W. 1991. Seasonal condition change, morphometries, growth, and sex ratio of the Ocean Quahog. .Arctica islandica (Linnaeus. 1767) off New Jersey, U.S.A. / Shellfish Res. 10:79-88. GLOBEC. 1992. GLobal Ocean ECosystem Dynamics. Northwest Atlan- tic: Implementation Plan. Technical Report 6. Washington. D.C.: Joint Oceangraphic Institutions. Inc. Holzwarth. T. & D. Mountain. 1990. Surface and bottom temperature distributions from the Northeast Fisheries Center Spring and Full Bot- tom Trawl Survey program: 1963-1987. Northeast Fisheries Science Center Reference Doc. 90-03: 62 pp. POPLIl.ATION STRUrXURi-; AND RhCRUITMENT OF A. ISIANDICA 1143 Jensen. A. S. 1912. LanicllihranchlatLi. Pari 1. The Danish Ingulf- Expedition 2:1-119. Jones, D. 1980. Annual cycle of shell growth increment formation in two continental shelf bivalves and its paleoecologic significance. Piilehiol- «.i;.v 6:331-340. Kraus. M. G.. B. F. Beal. S. R. Chapman & L. McMartin. 1992. A com- parison of growth rates in Arnica islaiuliai (Linnaeus. 1767) between tleld and laboratory populations. J. Slwlljlsh Res. 1 1 :289-294. Lander. W. S. 1976. Reproduction and early development of the Ocean Quahog. Arclicci iskiiidicu. in the laboratory. Tlw Nautilus 90:88-92. Lewis, C. V. W. 1997. Biological-physical interactions on Georges Bank: plankton transport and population dynamics of the Ocean Quahog. Airrica islaiulica. PhD Thesis, Woods Hole Oceanographic Institution Joint Program with the Massachusetts Institute of Technology. Mann. R. 1982. The seasonal cycle of gonadal development in Antica islaiulica from the .Southern Nev\ England Shelf. Fishery Bulletin 80: 315-325. Mann. R. 1985. Seasonal changes in the depth-distribution of bivalve lar- vae on the southern New England shelf. J. Shellfish Res. 5:57-64. Mann, R. 1986. Arciica islaiulica larvae: active depth regulators or passive particles. American Malacological Bulletin 3:51-57. Mann. R. & C. Wolf. 1983. Swimming behavior of larvae of the ocean quahog Arctica islandica in response to pressure and temperature. Ma- rine Ecology Progress Series 13:2-3. McGarvey. R., F. M. Serchuk & I. A. McLaren. 1993. Spatial and parent- age analysis of stock-recruitment in the Georges Bank sea scallop {Placapccicn iikii;cllcinicus) population. Can. J. Fish. Aquar. Sci. 50: 564-574. Merrill, A. S. & J. W. Ropes. 1969. The general distribution of the surf clam and ocean quahog. ProceeJings of the National Shellfisheries Association 59:40—15. Murawski. S.. J. W. Ropes & F. Serchuk. 1982. Growth of the ocean quahog. .Arctica islaiulica. in the Middle Atlantic Bight. Fishery Bul- letin. U.S. 80:21-34. Murawski, S. & F. Serchuk. 1989. Mechanized shellfish harvesting and its management: the offshore clam fishery of the eastern United States. In: J. F. Caddy, editor. Marine invertebrate fisheries: their assessment and management. New York: John Wiley and Son.s. pp. 479-506. Musick. J. A. 1999. Life in the slow lane: ecology and conservation of long-lived marine animals. Belhesda: American Fisheries Society Sym- posium 23. 265 pp. Nicol, D. 1 95 1 . Recent species of the veneroid pelecypod .Arctica. Jinirnal of the Washington Academy of Sciences 41:102-106. NMFS. 1999. Our living oceans. Report on the status of the U.S. living marine resources, 1999. NOAA Tech. Memo. NMFS-F/SPO-41. U.S. Department of Commerce. 301 pp. Northeast Fisheries Science Center. 1990. Report of the Spring 1990 NEFC Stock Assessment Workshop. Northeast Fisheries Science Center Ref- erence Document 90-07. NOAA/National Marine Fisheries Service. Northeast Fisheries Science Center. 1996. 22nd Northeast Regional Stock Assessment Workshop. Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. Northeast Fisheries Science Center Reference Document 96-13. NOAA/Nalional Marine Fisheries Service. Northeast Fisheries Science Center. 1998. 27th Northeast Regional Stock Assessment Workshop. Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. Northeast Fisheries Science Center Reference Document 9H-I5. NOAA/Nalional Marine Fisheries Service. Northea.st Fisheries Science Center. 2000. Report of the 3()th Northeast Regional Stock Assessment Workshop. Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. Northeast Fisheries Science Center Reference Document (In Press). NO.AA/ National Marine Fisheries Service. O'Reilly. J.. C. Evans-Zetlin & D. Busch. 1987. Primary production. In: R. H. Backus & D. W. Bourne, editors. Georges Bank, Cambridge, Massachusetts: MIT Press, pp. 220-233. Ockelmann, K. l9,-)8. The zoology of East Greenland, marine Laniellihran- chiata. Meddr. Gronland 22:57-59. Ropes. J. W. & A. Jearld. 1987. Age determination of ocean bivalves. In: R. Summerfelt & G. Hall, editors. Age and growth in fish. Unknown, pp. 517-526. Ropes, J., D. Jones. S. Murawski. F. Serchuk & .A. Jearld. 1983. Docu- mentation of annual growth lines in the ocean quahog .Arcticti islandica Linne. J. .Shellfish Res. 3:100. Ropes. J., D. Jones. S. Murawski, F. Serchuk & A, Jearld, 1984, Docu- mentation of annual growth lines in the ocean quahog Arctica islandica Linne, Fislieiy Bulletin 82:1-19. Ropes, J, & D. Pyoas. 1982. Preliminary age and growth observations of ocean quahogs, Arctica islandica Linne, from Georges Bank. ICES CM. 1982/K:I5 Shellfish Committee. 6 pp. Ropes, J. W. 1984. Procedures for preparing acetate peels and evidence validating the annual periodicity of growth lines formed in the shells of Ocean Quahogs. Arctica islandica. Marine Fisheries Review 46:27-35. Serchuk. F. M. & S. Murawski. 1997. The offshore molluscan resources of the Northeastern Coast of the United States: surfclams, ocean quahogs. and sea scallops. In: C, L, MacKen/ie Jr., V. G. Burrell Jr., A Rosen- field & W. L. Hobart, editors. The history, present condition, and future of the molluscan fi.sheries of North and Central America and Europe. Volume I , Atlantic and Gulf Coasts. LI.S. Department of Commerce, NOAA Tech. Rep. 127. pp. 45-62. Serchuk, F, M. & S. E, Wigley. 1986. Evaluation of USA and Canadian research vessel surveys for sea scallops {Placopecten magclhinicus) on Georges Bank. / Northw. Atl. Fi.sh. Sci. 7:1-13. Serchuk. F. M., P. W. Wood, J. A. Posgay & B. E. Brown. 1979. Assess- ment and status of sea scallop (Placopecten magellanicus) populations off the northeast coast of the United States. Proc. Natl. Shellfish. Assoc. 69:161-191. Silverman, B. W. 1981. Using kernel density estimates to investigate niul- timodality. Journal of the Royal Statistical Society B 43:97-99. Silverman. B. W. 1983. Some properties of a test for multimodality based on kernel density estimates. In: J. F. C. Kingman & G. E. H. Reuter. editors. Probability, statistics, and analysis, Cambridge, UK: Cam- bridge Uni\. Press, pp. 248-260. Smolowitz, R. & V. Nulk. 1982. The design of an electrohydraulic dredge for clam surveys. Marine Fishery Review 44:1-18. Steimle. F. W., D. Jeffress, S. A. Fromm, R. N, Reid, J. J. Vitaliano & A. Frame. 1994. Predator-prey relationships of winter flounder, Pleu- ronectes americanus. in the New York Bight ape.x. Fishery Bulletin 92:608-619. Steingri'msson, S. A. & G. Thorarinsdottir. 1995. .Age structure, growth and si/e at sexual maturity in ocean quahog. .Arctica islandica iMollusca: Bivalvia). off NW-Iceland. ICES CM/K:54. Thompson. 1.. D. Jones & D. Dreibelbis. 1980a. Annual internal growth banding and life history of the ocean quahog Arctica islandica (Mol- lu.sca: Bivalvia). Marine Biology 57:25-34. Thompson. I., D. Jones & J. Ropes. 1980b. Advanced age for sexual maturity in the ocean quahog Arctica islandica (Mollusca: Bivalvia). Marine Biology 57:35-39. Thorarinsdottir. G. G. 1997. Distribution and abundance of juvenile ocean quahog {Arctica islandica) in Eyjatjordur Iceland. ICES CM/BB:17. Thorarinsdottir. G. G. & S. T. Einarsson. 1996. Distribution, abundance, population structure and meat yield of the ocean quahog. Arctica is- landica. in Icelandic waters. Journal of the Marine Biological Asso- ciation of the U.K. 76:1 107-1 1 14. Thouzeau. G.. G. Robert & S. Smith. 1991a. Spatial variability in distri- bution and growth of juvenile and adult sea scallops Placopecten ma- gellanicus (Gmelin) on eastern Georges Bank (Northwest Atlantic). Marine Ecology Progress Series 74:205-218. Thouzeau. G.. G. Robert & R. Ugarte. 1991b. Faunal assemblages of benthic megainvertebrates inhabiting sea scallop grounds from eastern Georges Bank, in relation to environmental factors. Marine Ecology Progress Series 74:61-82. Treniblav. M. J.. J. W. Loder. F. E. Werner. C. E. Naimie, F. H. Pane & 1144 Lewis et al. M. M. Sinclair. 1994. Drift of sea scallop larvae Placopeclen imti^el- luniciis on Georges Bank: li model study of the roles of mean advection. larval beha\ior and lar\al origin. Deep Sea Research II 41:7^9. Treinblay. M.J. & M. M. Sinclair. 1990. Sea scallop larvae Placopeclen magelkmiciis on Georges Bank; vertical distribution in relation to water column stratification and food. Murine Ecology Progress Series 61:1- Tremblay, M.J. & M. M. Sinclair. 1992. Planktonic sea scallop larvae {Placopeclen nuigellaiiicns) in the Georges Bank region: broadscale distribution in relation to physical oceanography. Canadian Journal of Fisheries and Aijualic Sciences 49:1597-1615. Turekian. K. K.. J. K. Cochran. D. P. Kharkar. R. M. Cerrato, J. R. Vais- nys, H. L. Sanders. J. F. Grassle & J. A. Allen. 1975. Slow growth rate of a deep-sea clam determined by """Ra chronology. Proceedings of ihe National Academy of Sciences 72:2829-2832. Turekian. K.. J. Cochran & Y. Nozaki. 1982. Determination of shell de- position rates of Arctica isUmdica from the New York Bight using natural ""''Ra and ""^Th and bomb-produced "C. Limnology and Oceanography 27:7.17-741 . Turner. H.J. 1949. The Mahogany Quahog resources of Massachusetts: Report on investigation of methods of improving the shellfish resources of Massachusetts. Technical Repori 12-16. Mass. Dept. of Conserva- tion. Di\ . of Mar. Fish. Weidman, C. R. & G. A. Jones. 1993. A shell derived time history of bomb '""C on Georges Bank and its Labrador Sea implications. Jonrnal of Geophysical Research 98(C8): 14.577-14.588. Weidman, C. R., G. A. Jones & K. C. Lohmann. 1994. The long-lived mollusc Arctica islamlica: a new paleoceanographic tool for the recon- struction of bottom temperatures for the continental shelves of the North Atlantic Ocean. Journal of Geophysical Research 99(C9): 18305-18314. Weinberg. J. R. & T. E. Helser. 1996. Growth of the Atlantic surfclam. Spisula solidi.ssima, from Georges Bank to the Delmarxa Peninsula. USA. Marine Biology 126:663-674. Weinberg, J. R. 1999. Age-structure, recruitment, and adult mortality in populations of the Atlantic surfclam. Spisula solidissiiua, from 1978 to 1997. Marine Biology 134:113-125. Witbaard. R. & R. Klein. 1994. Long-term trends on the effects of the southern North Sea beamtrawl fishery on the bivalve mollusc Arctica islandica L. (Mollusca, bivalvia). ICES Journal of Marine Science 51:99-105. Witbaard, R.. G. Duimexeld & P. de Wilde. 1999. Geographical ditler- ences in growth rates of Arctica islandica L. (Mollusca. bivalvia) from the North Sea and adjacent water. Journal of the Marine Biological Association of the UK. 79:907-915. Jaunuil ofShcllthli Rcscanh. Vol. 20. No. .^. 1 143-1 157. 2001. SHORT-TERM EFFECTS OF COMMERCIAL CLAM {MYA ARENARIA L.» AND WORM (GLYCERA DIBRANCHIATA EHLERS) HARVESTING ON SURVIVAL AND GROWTH OF JUVENILES OF THE SOFT-SHELL CLAM BRIAN F. BEAL'* AND KENNETH W. VENCILE" 'Universin- of Maine at Machias. 9 O'Brien Avenue, Macluas. Maine 04654: 'Beats Island Rei^ional Slwllfisli Hatchety, P.O. Bo.x S3 Beats. Maine 0461 1 ABSTRACT In Maine. USA. commeri.-ial fisheries for soft-shell clams. Mya arenaria L.. and blood worms. Glycera dihranclmmi Ehlers. occur simultaneously on muddy intertidal flats. Local and state clam managers frequently close flats to shellfishing for conservation purposes, but have no jurisdiction over wormers who are legally permitted to harvest C. dihranchiata on any intertidal flat. This sometimes causes conflicts, especially when wormers dig in clam conservation areas where clammers have enhanced stocks with wild or cultured "seed" clam,s «l cm shell length. SL). Clammers believe wormers kill or injure small clams directly or indirectly while commercially harvesting C. JihraiuhUiw. To help resolve these long-standing conflicts, we worked collaboratively with clammers and wormers and used an experimental approach to test the short-term interactive effects of clam and worm harvesting, harvesting intensity, time of harvest after seeding, and predator exclusion on the fate of small wild and cultured M. arenaria at an mtertidal mud flat m Brunswick. ME. We added .30 cultured juveniles of M. arenaria (SL = 12.5 mm) to 120 l-m" plots. 40 of which were undisturbed controls (20 protected with plastic netting— 6.4 mm aperture; 20 unprotected) from May to August 1996. The remaining 80 plots were assigned to one of 16 treatments. One half of the plots were protected from predators with the same plastic netting used in the undisturbed control plots. One half of the plots were harvested by a professional wormer or clammer who searched each plot for commercial size blood worms and soft-shell clams, respectively. Plots were harvested either once (after two weeks or four weeks) or twice (two weeks -i- two weeks, or four weeks + four weeks). Any effect due to clamming or worming on cultured clams or wild individuals of similar size was masked by clam losses exceeding 95% in the unprotected control plots. Intense predation by horseshoe crabs. Limulus polyphemus L. and the nemertean worm. Cerelyrauiliis lacteiis Leidy. are blamed for the high mortalities anions; clams. Only in protected plots was any effect detected and this depended on clam origin. Compared to the fate of cultured clams in protected controls, wormmg had no effect, but clamming contributed an additional 15% loss. Both types of commercial harvesting reduced wild clam numbers significantly compared to controls, but effects due to worming were more benign than effects due to clanimmg probably because wormers excavate less volume of sediments than clammers do as commercial size C. dihranchiaui are shallow burrowers compared to commercial size M. arenaria. Unless clam managers actively take steps to deter predators by using neltinn or other means, blood wormers should continue to harvest commercially from areas closed to shellfishing without reprisal or fear that they are causing damage to populations of juvenile soft-shell clams. KEY WORDS: blood v\orm. commercial harvesting, field experiment, fisheries contlicts. Glycera ddyranehiuia. intertidal. Maine. Mya arenaria. soft-shell clam INTRODUCTION Soft-bottom, unvegetated intertidal flats throughout the world provide habitat for commercially important infauna such as bi- valves and polychaetes (Jackson & James 1979. Emerson et al. 1990. Olive 1992. Brown & Wilson 1997). Harvesting these groups of organisms using hand-held implements such as fotks. hoes, rakes, or shovels (Creaser et al. 1983. Wallace 1997) or mechanical devices operated during tidal inundation (van den Heiligenberg 1987. Spencer et al. 1998) erodes or letiioves sedi- ments to some depth below which the fauna reside. Clatn digging, for example, causes sediments to become coarser and to lose sig- nificant amounts of organic matter compared with undisturbed intertidal areas (Anderson & Meyer 1986). Effects of turning over sediments on mudtlat residents may depend on size and geo- graphic extent of the fishery, sediment grain size, depth to which sediments are excavated, time of year, ability of organisms to evade harvesting or reburrow. or combinations of these factors. When the commercial species are shallow burrowers (e.g.. venerid clams such as Mercenaria mercenuria |L.] or Tapes pliilippinariiin [Adams & Reeve]), local impacts on dynamics of the target fishery such as individual growth and recruitment may not be very long lasting; however, changes in habitat may affect negatively nontar- geted infaunal or epifaunal populations (Peterson et al. 1987. Hall "Corresponding author: E-mail; bbeaKnmaine.edu & Harding 1997. Spencer et al. 1998). Harvesting deeper burrow- ing organisms (e.g.. polychaetes such as lugworms. Arenicola ma- rina [L.] [Cryer et al. 1987. Beukema 1995). and sandworms. Nereis virens Sars. (Brown & Wilson 1997]. or bivalves such as geoducks. Panopea alynipta [Conrad] [Campbell et al. 1998]) can have profound effects not only on the unharvested portion of the population, but also on benthic community structure. In Maine, two intertidal fisheries occur simultaneously. Soft- shell clams. Mya arenaria L.. are ubiquitous in soft sediments ranging from gravel and sand to soft mud (B. Beal. pers. obs.). M. arenaria is harvested for human consumption and is either pro- cessed (shucked) for chowder and fried clam markets or sold live to be eaten after steaming (Ellis & Waterman 1998). In 1999. 1.03.3 metric tons (t) worth $10.3 million were harvested commer- cially from Maine flats (NMFS 2001. http;//w ww.st.ntnfs.gov/stl/ commercial/annualjandings.html). Clams >50.8 mm (2 inches) in shell length (SL) are harvested using shon-handled hoes with 4 to 6 curved tines that are 30 to 38 cm in length (Robinson & Rowell 1990). The depth that commercial-size animals burrow to is typically sl3 ctii. but this depends on SL (Zwarts & Wanink 1989, Zaklan & Ydenberg 1997) and sediment type (Glude 1954). Blood worms. Gtycera dibrancl^iata Ehlers. are typically found in muddy sediments where they are excavated for bait using hoes similar in appearance to those used by clammers. except the tines are shorter, ranging from 20 to 25 cm (Dow & Creaser 1970, Creaser et al. 1983, B. Beal. pers. obs). In 1999. 233.5 t worth 1145 1146 Beal and Vencile $2.9 million were harvested commercially in Maine (NMFS 2001 ). Although both clammers and wormers ply their trade on nuiddy flats, digging style differs between the two groups. Clammers usu- ally seek areas of flats where relatively large, oval-shaped holes (that form when animals extend their siphons above the sediment- water interface to feed during tidal inundation) are visible and numerous (B. Beal. pers. obs.). Adults of M. arenaria frequently are distributed contagiously (either due to post-settlement pro- cesses that tend to move animals toward the upper intertidal [Mat- thiessen 1963, Emerson 1990] or to harvesting activity that redis- tributes sub-legal size animals). Clammers. therefore, tend to mo\e from place to place and dig where they notice an abundance of siphon holes. Thus, clam digging is more intense in some portions of the flats than in others. Generally, there is no regular pattern of sediment excavation in the harvested areas (B. Beal. pers. obs.). Conversely, although G. dihrwicliiara can prey on small fauna at or near the sediment surface (Ainbrose 1984) and may leave char- acteristic surface marks or holes (B. Beal, pers. obs.). these are largely ignored by bait worm diggers who collect the worms as they turn over sediments systematically leaving row after row of disturbed mud (Creaser et al. 1983, Brown 1993). Depending on worm density, some flats may be turned over three or more times during the summer months, whereas most clam flats may be turned over twice, but usually only once per year (Brown and Wilson 1997, B. Beal. pers. obs.). Historically, flats were harvested for worms or clams, not both. Rarely did participants in the two fisheries dig side by side. Flats harvested for worms contained too few commercial size clams or sediments were considered too soft to maneuver within easily for most clammers (D. E. Wallace, pers. comm.). However, as wild clam stocks in Maine fell during the 1980s and price per bushel (approximately 22.7 kg) increased from 6.4(93.6) 223.4(52.2) 33.0(15.8) 213.4(133.8) 7.4(4.7) 90.0(80.2) 3.8(2.5) 162.4(32.8) 3.6(2.1) 4.0 (3.9) 5.6(4.1) 14.8(9.7) 13.2(10.5) 28.8(20.3) 11.6(9.4) 52.4(16.9) L4(l.6) 3.8 (2.5) 5.0(3.6) 15.0(13.6) 7.8(8.1) 14.0(4.0) 15.4(4.1) 33.6(19.2) 0.0(0.0) 14.0(6.3) 0.0 (0.0) 7.2 (7.9) 0.8(2.2) 67.6 (9.5) 0.0 (0.0) 26.0 (47.3) 2.0(3.5) 77.2 (9.7) 1.2(3.3) 33.6 (35.6) 0.0(0.0) 86.8 (9.2) 0.0(0.0) 15.2(15.3) 0.0(0.0) - 0.2 (0.6) - 1.8(1.9) - 12.8(6.7) - 5.4(8.1) - 3.8(3.1) - 0.0 (0.0) - 0.0 (0.0) - 1130 Beal and Vencile 28, P < 0.001 ). clams grew, on average, only 4.8 ± 1.2 mm {n = 7 plots) over the nearly 3-mo test, attaining a final mean SL on August 1 of 17.3 ± 1.2 mm. The majority of growth occurred during the final month of the experiment (Fig. 2a). Control Plots: Wild Clams The mean number of wild clams varied widely through time (Fig. 3). Both main (sampling date and protection) factors and the interaction term were highly significant (P < 0.001 ). Single-factor ANOVAs for each lexel of protection were performed to facilitate interpretation of clam losses. Clams in both protected and unpro- tected treatments suffered significant losses through time (P < 0.01; Table 2b). SNK tests revealed that in unprotected plots, a significant number of clams disappeared between the first and second sampling date clams, but no significant additional losses occurred after July 1. In plots protected with netting, no significant differences in clam numbers were detected between May 20 and June 3. Significant losses occurred between June 3 and July 1, but 6 May N = 100 20 May N = 168 3 June N = 86 3 June N =638 ..1 1 July N = 102 ll. 1 July N = 503 1 August N = 163 Shell Length (mm) 6 B ID 12 14 13 18 20 22 24 26 2S 30 Figure 2. Size frequency distribution of cultured (a) and wild (b) soft- shell clams within 1-nr control plots (;; = 5) at Maquoit Bay (Bruns- wick, MF.I on each sampling date (except that wild clams on May 6 were sampled from a total of 10 benlhic cores |.\rea = 0.(I1S2 m-jl. Numbers (Nl associated with each jjraph represent total number of clams within phtts or cores (except on .\ugusl I. where N for the wild clams represents a random sample of taken from 4,386 individuals within 9(1 plots). ,\NOVA demonstrated significant increases in mean SI, through time for cultured, but not wild, clams. T 1 \ 6 IVIay 20 May 3 June 1 July 1 August Figure 3. Mean number (±95 '7r confidence interval) of wild individu- als of iV/. areiiaria within l-ni" control plots at Maquoit Bay (Bruns- wick. ME) on each sampling date (May 2(1 to .August 1. 1996). Initial density was determined from 10 benthic cores taken haphazardly within the experimental area on May 6, 1996. ANOVA performed on data from the 1-nr plots indicated both main factors (protection and time) and the interaction of these factors were highly significant [P < O.OOI). ;i = 5. then average numbers appeared to increase to pre-July levels. This variability in mean number may be related to the fact that two horseshoe crabs were found within a single protected plot sampled on June 3 (see above). That replicate contained 104 wild clams, whereas 393, 182, 179, and 209 were sampled from the other four protected plots. In addition, a group of small clams (<3 mm SL) appeared on the final sampling date in the protected plots that may have been too small to be sampled efficiently earlier. These were not 0-y class individuals, as each contained a disturbance ring (winter check) on both valves at 1 to 2 mm SL. By August 1, approximately 40 times the number of wild clams was found in protected plots (162.4 ± 32.83 individuals/nr) compared with un- protected plots (3.80 ± 2.89 individuals/m"). No significant difference in mean SL of wild clams was de- tected between protected and unprotected plots (F = 0.67. df = I, 32, P = 0.4191 ) nor was there any significant change in mean SL through time (F = 2.04. df = 3, 32, P = (J. 1280: Fig. 2b). The apparent lack of growth also may be related to the appearance of the small clams in samples on August 1. Mo\ement of wild clams within the l-ni" plots was measured indirectly. Since each plot was subdivided into four equal sampling areas for each of the first three sampling dates, we asked whether the distribution of clams observed in the four areas varied through time. We assumed that clam movement occuired if the distribution changed through time. Distribution did not vary significantly in protected (G = 9.27. df = 6. P = 0.1387} or unprotected (G = 8.66, df = 6, P = 0.1943) plots. 2'' Experiment Hatchery-Reared Clams Of the 1 1 sources of variation (hypotheses), the most important (i.e.. explaining the largest percent of total variation) associated with percentage of clam survival was that due to protection (meshed vs. unmeshed plots; 31.3%). but this effect depended on the rate plots were dug (P < 0.0001; Table 3). Mean survival was higher in protected treatments dug once vs. twice (23.3 ± 3.3% vs. Commercial Clam and Worm Harvlsting 1151 TABLK 3. Analysis ol' variance mi llic aiisine-transformed percentage sur>ival of hatchery-reared soft-shell clam juveniles within each dug plot after 12 wk (AuKiist I, l'W6). ;; =5. Source of Variation df SS MS Pr>F Type: Clammed vs. Wormed Protect: Mesh vs. L'limcsh Type X Protect Rate: Dug Once vs. Twice Type X Rate Protect X Rule Type X Protect x Rate Date (Rate! Type X Date (Rate) Protect X Date (Rate) Once: (Protect x Date) Twice: (Protect x Date) Type X Protect x Date (Rate) EiTor Total 2 2 2 1 1 2 64 79 4.513.95 23.57 0.38 0.50 700.82 39.51 47.45 31.45 663.18 73.87 589.31 146.61 1 .807.68 S.757.(i7 782.57 4.513.95 23.57 0.38 ()..5() 700.82 -39.5 I 23.72 15.73 331.59 73.87 589.31 73.3 1 28.24 27.71 159.81 0.83 D.Ol 0.02 24.81 1.40 0.84 0.56 11.74 2.62 20.86 2.W) 0.0001 0.0001 0.3644 0.9075 0.8945 0.0001 0.2413 0.4364 0.5758 0.0001 0.1104 O.OOOI 0.0825 3 ?■; to c: (U Vll (1) 13. IS r nj (I) ^ 10 a) Clammed ro 30 > 1 b) Wormed ^^ Unprotected CZS) Protected i 1 ^^M Unprotected rTVS Protected Dug Once Dug Once Dug Twice Dug Twice (2 weeks) (4 weeks) (2+2 wk) (4 + 4 wk) Figure 4. .Mean percentage of survival (±957f eontidence interval) of cultured soft-shell clams in the eight clammed and eight wormed treat- ments (Table I) at .Maquoit l$a> (Brunswick, MK) on .August I. 1996. Fifty cultured clams (mean SI. = 12.5 mm) were placed in the 1/4-m' center of// = 5 l-m" plots (m May 6, 1996. Data represent the per- centage of those clams sampled in the entire l-nr plot. .\N()V.\ was used to compare survi\al between type of tlsher. Ie\el of protection, intensity of digging, and length of time w hen digging occurred after the experiment was initiated (see Table 3). n = 5. 13.7 ± 4.44'7r. /? = 20). This trend was not observed in unprotected treatments where 3.0 ± 1 .709!- survived in plots duj: once vs. 6.9 ± 2.74% in those dug twice. The type of harvesting method, inde- pendent of other main or interactive effects, was also highly sig- nificant (P < 0.0001; Table 3; Fig. 4). Mean survival in clammed treatments was nearly one-half that observed in wormed plots (8.7 ± 2.51% vs. 15.75 ± 3.69%. /; = 40). The only other significant source of variation was that due to Date x Protection (Rate). De- composing this source of variation into its two orthogonal com- ponents revealed that for plots dug twice, the effect of protecting clams with netting, was more pronounced when they were dug on a 2 -I- 2 w k schedule than when dug on a 4 -I- 4 wk schedule (Table 3; Fig. 4). This result is not too surprising because unprotected plots dug twice on a 4 + 4 wk basis were actually protected with netting a month longer than animals in the 2-1-2 wk treatments. It also reflects the temporal difference in clam loss observed in the controls between June and July vs. July and August (Table 2a). TABLE 4. Analysis of variance on the untransformed relative growth data for living hatchery-reared soft-shell clams within each dug plot after 12 wk (.August I, 1996). Not all plots contained live clams at the end of the experiment, making the data unbalanced. Type III sums of squares were used in all hypothesis tests (Shaw & Mitchell-Olds 1993). Source of Variation df SS MS F Pr>F Type: Clammed vs. Wormed 0.0167 0.0167 1.06 0.3086 Protect: Mesh vs. Unmesh 0.1597 0.1597 10.14 0.0026 Type X Protect 0.0031 0.(J031 0.02 0.6587 Rate: Dug Once vs. Twice 0.0203 0.0203 1.29 0.2626 Type X Rate 0.0348 0.0.348 2.21 0.1439 Protect X Rate 0. 1 240 0. 1 240 7.88 0.0073 Type X Protect x Rate 0.0361 0.036 1 2.29 0.1370 Date (Rate) 0.0376 0.0188 LI9 0.3122 Type X Date (Rate) 0.0009 0.0005 0.03 0.9720 Protect X Date (Rate) 0.0570 0.0285 1.81 0.1750 Type X Protect x Date (Rate) 0.0522 0.0522 1.66 0.2015 Error 46 0.7244 0.0157 Total 61 1.2668 1132 Beal and Vencile To determine whether digging and type of digging generally had an effect on clam survival compared to the undisturhed con- trols, we conducted two additional, single-factor ANOVAs. The first analysis tested whether mean percentage of survival differed between unprotected controls (x = 1.6 ± 2.08%. n = 5) and unprotected, but dug. plots from the larger e.xperiment (.x^.,.,,,,,,,^.,; = 1.3i<7f. « = 20: x„ = 7.0 ±2.' 20). The second analysis tested whether mean percentage of survival differed be- tween protected controls (x = 32.4 ± I2.45'7r. ii = 5) and pro- tected, but dug, plots from the larger experiment (\,,„„,„^.j ± 3.24%. ;; = 20: x,. 24.5 ± 4.00. n = 20). Because these tests were not independent of those conducted previously, we con- servatively lowered the type I error rate for each according to Winer et al. { 1991 ). Data in both tests was used twice: therefore, we let /! = 2 in the following equation: c(_jj|„^,^.j = 1 - (0.95)"". This yielded an adjusted type I eiTor rate, a = 0.0253, No differ- ence in mean clam survival was detected between the unprotected controls and dug plots (F = 3.64. df = 2. 42. P = 0.0349). but a significant effect due to fisher type was observed among pro- tected treatments (F = 12.21. df = 2. 42. P < 0.0001). An a pasteiioh SNK test showed that mean survival in controls and plots dug by the wcirmer was similar and that both were signifi- cantly higher than mean sur\'ival of clams in plots dug by the clammer. Mean relative growth rates were unaffected by type of harvest- ing, but they did vary interactively with level of protection and harvesting rate (Table 4; Fig. 5). Clams grew approximately 10% slower in unprotected treatments dug once vs. twice (20.97 ± 10.12%, /! = 11 vs. 30.26 ± 9.857r, /; = 13). Conversely, clams grew nearly 6% faster in protected treatments dug once vs. twice (37.29 ± 3.66%, n = 20 vs. 31.38 ± 6.06%. n = 18). As with the survival analyses, we asked whether growth rates of animals in the controls differed from those in the larger, 2* experiment, and we divided this into separate analyses for unprotected and protected treatments. We used the conservative a = 0.0253. as above. Only two of five unprotected control plots contained live clams (RG = 18.43 ± 78.95%) vs. 10 of 20 unprotected clammed (RG = 24.92 ± 14.61%) and 14 of 20 unprotected wormed (RG = 26.77 ± 7.51%) plots. No significant difference in growth rate due to type of fisher was noted (F = 0.24. df = 2. 23. P = 0.7890). but the power of this test was very low (< 0.3 ). Similarly, for the protected plots, no effect due to type of fisher on relative growth was de- tected in protected plots (x^.^,„,„,| = 42.38 ± 10.06%, n = 5: Xcian,med = 33.47 ± 5.98%, « = I8;x„„„,,^ = 35.40 ± 4.18%. « = 20; F = 1.47, df = 2, 40. P = 0.2426). but again, the power to detect differences was low (0.6). Wild Clams Wild clams behaved similarly to the hatchery-reared animals (Table 5). Approximately 557f of the tola! variation in mean num- ber of live individuals was explained by the presence or absence of netting: however, this effect depended on both type (P = 0.0007) and rate of harvesting (P = 0.0001 ). Protecting clams with netting resulted in more than twice the number of wild clams in plots harvested by the wormer compared with those harvested by the clammer. For example, mean difference in number of wild clams between protected and unprotected treatments was 79.2 ± 21.3 individuals/nr in plots harvested by the wormer compared with 35.1 ± 11.69 individuals/nr in clammed plots (/; = 20: Fig. 6). Application of plastic netting to plots resulted in 2.5 times more individuals from treatments dug once vs. twice. The mean differ- ence in numbers of wild clams between netted and unnetted plots 100 90 80 70 60 50 40 30 20 10 0 a) Clammed o 5 (1) > cr Dug Once (2 weeks) Dug Once (4 weeks) Dug Twice (2 + 2 wk) Dug Twice (4 + 4 wk) Figure 5. Mean relative growth (±95% confidence interval) of cul- tured soft-shell clams in the eight clammed and eight wormed treat- ments (Table 1) at Maquoil Bay (Brunswick. MK) on August 1, 1W6. .\N()V.\ was used to compare relative growth between tvpe of fisher, level of protectiim, intensity of digging, and length of time when dig- ging occurred after the experiment was initiated (see Table A), n = 5. dug once vs. twice was 81.9 ± 22.22 vs. 32.4 ± 13.35 individuals/ nr (n = 20). In addition, the interaction of type and date of harvesting nested within harvesting rate was significant (P = 0.0389: Table 5). As with hatchery-reared clams, effects due to protecting wild clams with netting were more pronounced (both for clammed and wormed plots - Fig. 6) for the shorter (2-1-2 wk) compared to the longer (4 4-4 wk) harvesting intervals. We tested whether an overall effect due to digging and digging type existed by comparing, as above, mean number of clams in unprotected and protected controls, separately, to plots treated similarly in the larger. 2'* experiment. No significant difference iP > 0.0253) in mean number was observed in unprotected treatments (Xcnt, 3.8 ±2.34 I ,, = 17.7 ± 7.05 5] vs. x,|,„ = 14. :4.92 20] 20]: P = 0.0258). Significantly more wild clams were found in protected controls vs. protected but dug treatments (.x,.„„,,„ = 162.4 ± 32.50 |/; = 5] vs. x,,.,,,,,,,,,, = 49.2 ± 11.74 [ii = 20] vs. .\„„,,„,j = 96.9 ± 21.37 [n = 20|: P < 0.0001 ). An SNK test indicated signiticant differ- ences between all three means. Commercial Clams and Worms Harvested from I-m' Plots During the Kxperinient Experimental plots were dug on three occasions (May 20 = 40: June 3 = 60; July 1 = 20). We asked whether the square root- Commercial Clam and Worm Harvesting 1153 TABLK 5. Analysis of Miriance on the square root-transformtd nunilnr of «il(l soft-shell clams alJM' wllliin each (lii« plot after 12 «k (Aufjust 1, 19%). It = 5. Source of Variation df SS MS Pr>F Type: Clammed \s. Wormed Protect: Mesh vs. Unmesh Type X Protect Rate: Dug Once vs. Twice Type X Rate Protect X Rate Type X Protect x Rate Date (Rate) Type X Date (Rate) Protect X Date ( Rate ) Once: (Protect x Date) Twice: (Protect x Date) Type X Protect x Date (Rate) Error Total trutisformed tnean tiumber atid utitrunsfomied mean size of ani- mals harvested from each plot varied among sampling dates (Fig. 7). ANOVA detected no significant differences either in mean number (F = 0.78. df = 2. 57. P = 0.4630) or mean SL (F = 1.63. df = 2. 28. P = 0.2133) of clams through time. Interest- ingly, mean number of clams/nr adjacent to the experimental site on July I (0.'^) ± 0.41. ;; = 10) was nearly identical to numbers harvested initially frotii experimental plots on May 20 (0.95 ± 0.58. n = 20). Mean SL of cotiimercial clatiis within all plots was 64.5 ± 2.16 mtii (;i = 39). No differences were detected i)i mean number or size of worms per plot from May 20 to July 1 (F,„,,„hc, = 0.88. df = 2. 57. P = 0.4205; F,„^, = 0.06. df = 2,34. P = 0.9385; Fig. 7). Mean number of bloodworms per 1-m^ varied little through time. On May 20. 0.95 ± 0.44 indi\ iduals (/; = 20) weie harvested from the plots, whereas 1.2 ± 0.74 animals (n = 10) were harvested in plots outside the experimental site on July 1. Mean size (greatest diameter) of all womis collected was 5.4 ± 0.32 mm (;; = 37). Other Large Infauna Occurring in the Plots at the End of the 2'' Experiment The baltic clam, Macoma hulthica. was the most numerous large infauna within the plots: however, these animals were neither enumerated nor measured. Although horseshoe crabs were found in 50% of the control plots on August 1 (see above), only 15 of the 80 dug plots (18.75%) sampled on that date contained L. pohphc- niiis. Horseshoe crabs were found in six of the clammed and nine of the wormed plots. Mean CW of these 24 individuals was 45.5 ± 3.9 mm (range = 10.5-79.8 mm). C. tucleus was also found in most protected plots, but again, worms generally were small (ap- proximately < 40 mm TL) and were not enumerated because they tended to fragment easily. DISCUSSION This work was intended to investigate the short-term effects of clam and bloodworm harvesting within the soft bottom inteilidal on the fate of cultured and wild individuals of the soft-shell clam. Had we not included adec^uatc controls to monitor changes in mean 200 •^P 175 o •O 125 48.26 48.26 2 1 .58 O.OOOI o 395.28 395.28 176.77 0.00(11 28.67 2X.67 12,82 0.0007 b 3 6.73 6.73 .vol 0.0S76 Z 0.08 0.08 0.04 0.8475 0) 59.86 59.86 26.77 0.0001 2 0.76 0.76 ()..M 0.56 1 6 1 10.93 5.47 2.44 0.0948 T 2.95 1.48 0.66 0.5205 T 15.28 7.64 3.42 0.03.S9 1 0.86 0.86 0.38 0.5 1 m 1 14.42 14.42 6.44 0.0136 "f T -^..^7 2.78 1.25 0.2948 ■ft e ro 64 143.12 2.24 79 717.49 O a) Clammed Unprotected Protected 1 a> E 3 b) Wormed 1 ^^B Unprotected riT^^ Protected i Dug Once (2 weel^s) Dug Once (4 wee)(s) Dug Twice (2 + 2 wk) Dug Twice (4 + 4 wl<) Figure 6. Mean number of wild soft -shell clams (±95% confidence interval) in the eight clammed and eight wormed treatments (Tahle 1) at Maquoil I5a.\ (Brunswick. ME) on August 1. 1996. Initial density at the beginning of the experiment on May 6. 1996 was 229 ± 152 indi- viduals/nr. .\N()V.A was used to compare mean number between type of fisher, level of protection, intensity of digging, and length of time when digging occurred after the experiment was initiated (see Table 5). clam survival through time, we likely would have reached a dif- ferent conclusion about our results. The controls demonstrated clearly that natural losses due to processes that removed clams from the plots (predation and/or emigration) were more important than any effect observed either due to clam or bloodworm har- vesting. Each cotnparison of means between unprotected controls and unprotected, but dug. treatments in the 2"* experiment was not significant. Had our ability to detect differences between means not been reduced by the necessity to adjust the type I error rate of these tests, we would have concluded that survival of cultured clams and final mean number of wild clams was actually higher in the dug vs. control plots. That is, there might be some benefit to turning over the sediments. Only when we examined the fate of clams protected by netting could we discern any effects due to type of harvesting. The ANOVAs demonstrated that at least for hatch- ery-reared clams, effects due to worming on mean growth and survival were negligible compared to contiols. Clamming, on the other hand, contributed an additional 15% loss of juvenile clams compared to controls. We did detect a negative effect due to blood- worm harvesting on numbers of wild clams in protected plots, but the effect was not as severe as that from clamining. Approximately 40% and 70% fewer wild clams wei'e found in plots harvested by 1154 Beal and Vencile Clams Worms o c (D O" 0 (D C 40 0.05). the bloodwornier and ckiniiiier. respectively, compared lo pi'o- tected controls. Results from the 2"* experiment alone suggest (hat clamming had the gi'eater negative impact on clam survival than bloodvvomi- ing (Tables 3 and 5; Figs. 4 and 6); however, we did not examine the mechanism. Observed differences in survival of cultured clams and numbers of wild clams between clammed and wormed treat- ments may be related to the depth that harvesters must dig to find commercial size animals. For example, cominercial size M. arcihirici btiiTow > 15 cm (Zwarts & Wanink 1989) compared to 5 to 10 cm for Glyceia (Creaser et al. 1983). The tools each type of fisher uses also reflects these life history traits. That is. the tines of clam hoes are significantly longer than tines on hoes used by wormers: therefore, the volume of sediment excavated when searching for clams vs. worms is greater (B. Beal. pers. obs.). Since clam survival is inversely proportional to burial depth (Glude 1954), we infer that clam harvesting adds synergisticall) to the already large losses of small clams due to natural causes. Hatchery-reared clams used in this study were in the range of sizes of relatively mobile individuals (Baptist 1955, Beal. 1994). One question might be that since no bairiers or enclosure walls {sensii Peterson & Beal 1989) were used to restrict clam move- ments, how is it possible to distinguish the diffei'ence between emigration from unpiotected plots and clam mortality? We seeded small, cultured clams into the middle 1/4-m- of each 1-m- plot, while the remaining area around it was used as a buffer zone for emigrants. We assume that were emigration important in explain- ing the relatively high percentage of clams missing from the un- protected conti'ols (Table 2a). most movement fiom the plots would have occurred within the first 2 wk (May 6 to 20) after plots were seeded. On the first sampling date (May 20), 14% of cultured clams were missing within the five unprotected control plots. Of the remaining 83% of living hatchery-reared clams. 98% were sampled from the middle 1/4-m^ of each plot. In addition, although clam movement in protected controls was restricted to the 1-m" area, we found 94% of live animals to be within the seeded middle area on the first sampling date. These data suggest that if clams did emigrate from the l-nr plots, the rate was very low. In addition, comparable data showed that wild clams had moved very little. The severe losses of hatchery-reared clams in the controls (pro- tected = 68%; unprotected = 98% ) miiTored large losses of small wild clams in the same plots (protected = 27%; unprotected = 97%). Numbers of wild individuals in protected controls appeared to increase from July 1 to August 1. but this was likely due to a combination of two events. The first was a group of small clams (<5 mm) that apparently had grown into the size range sampled more efficiently by the 3.2-mm mesh. The second was predation. It is reasonable to conclude that the majority of clams lost from control plots was due to intense predation. Large predators such as horseshoe crabs were noted on each sampling date. The best evi- dence that predation occurred and was an important factor in ex- plaining the high percentage of clam losses came from a single protected plot on June 3. In four of five protected plots, the mean number of live hatchery-reared clams recovered out of 50 planted was 36.5. or IV/c. and the mean number of wild clams was 241. Two juveniles of L. polypliemus were found in the fifth plot that contained only 2 of 50 (4%) living, cultured clams (the rest were missing) and 104 (or 57% fewer) wild clams. Another line of evidence that predation was important was the number of nemertean worms observed in the control samples, as well as from the larger experiment. Rowell and Woo (1990) pro- vided the first experimental evidence from both field and labora- tory trials that C. Uutciis consumes individuals of M. arenaria. These worms leave no apparent damage to the valves such as crushing, chipping, or drilled holes commonly associated with crustaceans and naticid gastropods, respectively. C. lacteus gains entry to prey such as M. arenaria by wriggling its whole body or proboscis through either the siphons or the anterior pedal hole in the mantle (Kalin 1984). Since cytolytic and neurotoxic proteins are localized in the integumentary tissues of the bod> wall and proboscis of this predator (Kem 1985. Barnham et al. 1997), it is able lo kill its prey without damage to the vahes. In this study, numbers of dead, undamaged valves of both hatchery-reared and wild clams in control plots were high (Table 2a and b) and may have been related to nemertean worm predation. Stress due to mishandling cultured clams and/or disease of w ild or cultured individuals are alternative hypotheses for explaining the high proportion of dead, undamaged valves; however, these seem unlikely. Clams (8-20 mm SL) obtained from BIRSH were produced in 1995 and held over the winter (see B. Beal et al. 1995). Survival varied between 95%r and 98%c (B. Beal, pers. obs.). Hatchery-reared individuals within this range of sizes also were used in a manipulative field expeiiment initiated in April 1996 on Commercial Clam and Worm Harvesting 1 1 55 an interlidal mud Hat in eastern Maine (Jonesport; latitude 44°37'N, 67°34"W), In that study clams were planted at densities of 666. 1.332. and 2.664/m" in protected and unprotected experimen- tal units at each of three distinct tidal heights. Units (/( = 180) sampled on June 5 and August 3, 1996 contained 94.2% and 86.4% live clams, respectively (Beal et al. 2001). No C. lacteus or L. potyphemus occured on this flat (B. Beal. pers. obs.). Since clams used in both studies were of the same origin and were handled similarly prior to experimental manipulation, it is unlikely that stress due to handling was a significant factor in soft-shell clam mortality in the present study. In addition, although not specifi- cally tested here, since 1992. cultured soft-shell clam juveniles from BIR.SH have been routinely (annually) examined histologi- cally for gonadal (B. Barber. University of Maine. Orono. pers. comm.) and hemocytic (S. McLaughlin. NOAA. Cooperative Ox- ford Laboratory. Oxford. MD. pers. comm.) neoplasia prior to shipping seed out of state. No prevalence oi either type of neo- plastic cell has been reported. The level of protection was the most important source of varia- tion in each of the three ANOVAs associated with the 2'' experi- ment (Tables 3-5). but in each case, the Protect x Rate interaction term was significant. Wild and cultured clams were found in greater abundance within protected vs. unprotected plots regard- less of whether plots were dug once or twice; however, the effect due to netting was greater when plots were dug once \s. twice. This fact, and the failure of ANOVA to detect an effect of rate of digging alone, was partially a result of the experimental design, as unprotected plots dug twice were covered with netting during the time between harvests. In the case of treatments 15 through 18 (4 -I- 4 wk; Table 1). there is little difference in means between pro- tected and unprotected treatments (Figs. 4 and 6). In protected plots, cultured clam survival was 87f higher, and 43% more wild clams were found in plots dug once vs. twice. When we re- analyzed data from Tables 3 and 5 by ignoring all information from unprotected plots, these percentages were highly significant (Fsurv.vai = 19.44. df= 1 . 32. P < 0.000 1 ; F„„,„(,„'"= 17.50. df = 1. 32. P < 0.0001) and suggested that repeated digging of flats results in significant additional clam losses. The presence of net- ling enhanced growth of cultured clams by nearly 17% in treat- ments dug once, but this same effect was not observed in treat- ments dug twice. Again, this may be related to the extra time that unprotected plots were covered with netting. Clock and Chew (1979) found similar growth enhancement in Japanese littlenecks, Venerupis japonlca (DeShayes). at an intertidal flat in Washington state; however, previous field trials in eastern Maine using cultured seed of M. arenaria (Beal 1993) yielded no differences in clam growth between netted and unnetted experimental units. Except for Ambrose et al. ( 1998), previous studies on impacts of digging soft-bottom sediments on the fate of soft-shell clams have tocused directly or indirectly on effects caused by clammers (Glude 1954. Medcof & MacPhail 1964, Emerson et al. 1990, Robinson & Rowell 1990). Depending on sediment type, harvest- ing can be 60% to 85% efficient and can also damage or kill animals missed or are left behind intentionally. Unharvested clams may be exposed on the sediment, increasing their risk to predation by birds, fish, and crustaceans, or may be buried at depths that cause them to suffocate. Glude (1954) conducted field experiments in Maine to assess the effects of burial on soft-shell clams (9.5- 50.8 mm SL) in muddy sediments. With siphons oriented in three positions (upright, horizontal, and upside down), clams were bur- ied at depths from 2.5 to 23 cm. Survival was related directly to clam size, burial position, and depth. The range of estimates of direct mortality of clams from hoes or forks varies between 2% and 50% (Dtm et al. 1954. Medcof &. MacPhail 1964. Hruby 1982. Robinson & Rowell 1990). Both Dow et al. (1954) and Robinson and Rowell (1990) determined that breakage (chipped or cracked valves) of clams (13-57 mm) by hoes in commercially productive areas averaged between 17.6% and 19.6%. These means increased with increasing clam density and sediment com- paction. Dow et al. (1954) showed that breakage increases \silh increasing clam si/e. but Robinson and Rowell ( 1990) found no clear trend between mortality rate and clam size. Robinson and Rowell (1990) concluded that incidental mortality of clams due to harvesting could significantly affect clam stocks by decreasing yields. The work by Ambrose et al. ( 1998) represents the first exami- nation of bait worm digging on soft-shell clam populations. They concluded that commercial harvesting of G. dibianchiata has negative effects on clam abundance and that these effects should be considered in state and local clam management programs. Ex- perimental evidence presented here supports their recommenda- tion. We conclude that clamming and worming both contribute negatively to the loss of small soft-shell clams. However, these losses are considered unimportant during the summer months when natural mortality is high due to predation and other sources, especially in areas where netting is not applied to intertidal flats to protect clams from predators or to encourage recruitment of wild seed clams (Beal 1993). Our study and that of Ambrose et al. (1998) was a short-term trial conducted at a single-study site. Although it is dangerous to extrapolate results without spatial or temporal replication, our collective experience since the mid- 1980s working with coastal communities in Maine to enhance flats with hatchery-reared juveniles of M. aienuriu is that the rate of natural predation observed in this study is not unusual for intertidal areas along the southern Maine coast. In the absence of direct measures to enhance juvenile clam survivorship during summer months when natural predation rates are high (i.e.. applying pro- tective netting to portions of the intertidal or modifying sediments using gravel or shell (e.g.. Toba et al.. 1992]). we recommend that local and state managers permit bloodwormers to harvest G. 30 mm SL), we recommend that managers review the work of Glude (1954) and Ambrose et al. (1998). Furthermore, limiting clamming in shellfish conservation areas where natural predation rates are high will like!) not result in enhanced densities of small clams. Conservation closures for shellfish are more likely to benefit clams larger than 30 to 35 mm SL; that is. those individuals that have escaped the majority of threats by predators due to their relatively large size that enables them to burrow deeper in the sediments (Commito 1982). Conflicts between various fisher groups who harvest intertidal organisms in Maine and elsewhere are likely to increase in future as prices for dwindling supplies of natural stocks of commercial marine species continue to increase against a backdrop of condi- tions that restrict or limit access to fisheries (e.g.. management and/or rule-making decisions, local, state, and federal zoning laws, property rights issues, and aquaculture). Recently, it has become popular for fisheries managers to move toward decentralizing de- cision making and to spread out this process among communities, fishermen, and marine scientists (Wilson et al. 1994). Experimen- tal approaches can be useful tools in helping to resohe ambiguities 1156 Beal and Vencile and innuendos surrounding gear and other fisheries conflicts (Peterson at al. 1987. McAllister & Peterman 1992. Lenihan & Micheli 2000). This manipulative experiment demonstrated that local clam stewardship committees have a tool available to them to enhance survivorship of juvenile soft-shell clams. If adopted, this approach would necessitate short-term restriction of all types of digging t)f mud Oats in fenced or netted areas during that time of year when clam predation and other sources of mortality are great- est. Conversely, if no attempts are made to deter predation on small clams, no changes in current rules and regulations are war- ranted. ACKNOWLEDGMENTS We thank the municipal marine wardens from the town of Brunswick. A. Houston and S. Catlin. for their assistance in the field and for the use of the town's airboat. D. Caton (Wiscasset) and P. Hohnan (Brunswick) represented local professional worm- ers and clammers. respectively. J. Hawkes. L. Larkin. D. Kleman- ski, J. MacLeod. D. Michaud. T. Murphy. K. Nakazawa. J. Rodzen, and T. Simeone also assisted in the field. J. Aspinall and J. Rodzen helped measure clams and worms and transcribe data. Conversations about experimental design and statistical analyses with S. Fegley were helpful. The paper was improved by critical comments from D. Wallace and an anonymous reviewer. We also thank D. Wallace for providing accommodations. This study was funded primarily by the town of Brunswick. Maine and by the Maine Department of Marine Resources (Augusta). Other financial contributors include the Maine Aquaculture Innovation Center (Orono. Maine. University of Maine at Machias. and the Gulf of Maine Council on the Marine Environment. The manuscript was prepared while B. Beal was a visiting Fulbright Professor at the National University of Ireland. GaKvay and he thanks the Zoology Department and staff at Arus de Briin for pro\ iding office space and supplies during his stay. LITERATURE CITED Ambrose. W. C. Jr. 1984. Influences of predatory polychaetes and epibenthic predators on the structure of a soft-bottom commiinitv in a Maine estuary. J. Exp. Mar. Biol. Ecol. 81:115-145. Ambrose. W. G.. Jr.. M. Dawson. C. Gailey. P. Ledko\sky. S. O'Leary. B.Tassinari. H. Vogel & C. Wilson. 1998. Effects of baitworm dig- ging on the soft-shelled clam. Mya cireiuiria. in Maine: shell dam- age and exposure on the sediment surface. J. Shellfish Res. 17:1043- 1049. Anderson. F. E. & L. M. Meyer. 1986. The interaction of tidal currents on a disturbed intertidal bottom with a resulting change in particulate matter quantity, texture, and I'imilI quality Esiiiar. Const. Shelf Sci. 22:19-29. Baptist. J. P. 1955. Burrowing ability of juvenile clams. U.S. Fish WiUI. Sen: Spec. Sci. Rep. 140:1-1.^. Barnham. K. J.. T. R. Dyke. W. R. Kem & R. S. Norton. 1997. Structure of neurotoxin B-IV from the marine worm Cerehratulus lacteus: a helical hairpin cross-linked by disulphide bonding. J. Mol. Biol. 268: 886-902. Beal. B. F. 1993. Effects of initial clam size and type of protective mesh netting on the survival and growth of hatchery-reared individuals ot Mya arenaria in eastern Maine. J. Shellfish Res. 12:138-134. Beal. B. F. 1994. Biotic and abiotic factors influencing growth and survival of wild and cultured soft-shell clams, Mya arenaria L.. in eastern Maine. Ph.D. dissertation. University of Maine. Orono. 499 pp. Beal. B. F.. C. D. Lithgow. D. P. Shaw. S. Renshaw & D. Ouellette. 1995. Overwintering hatchery-reared individuals of the soft-.shell clam. Mya arenaria L.: a field test of site, clam size, and intraspecific density. Aquaculture 130:145-158. Beal. B. F.. R. Bayer. M. G. Kraus & S. R. Chapman, 1999. A uniLiue shell marker in juvenile, hatchery-reared indiv iduals of Ihe soltshell clam. Mya arenaria L. Fish. Bull. 97:380-386. Beal, B. F., M. Parker & K. W. Vencile. 2001. Seasonal effects of in- traspecific density and predator exclusion along a shore-level gradient on survival and growth of juveniles of the soft-shell clam. Mya arenaria L.. in Maine. USA. / Exp. Mar. Biol. Ecol. 264:133-169. Beukema. J. J. 1995. Long-term effects of mechanical harvesting of \ug- vjormsArenirola marina on the zoobenthic community of a lidal flat m the Wadden Sea. Nelh. J. Sea Res. 33:219-227. Botsfoid. L. W.. J. C. Castilla & C. H. Peterson. 1997. The management of fisheries and marine eco.systems. Science 277:509-515. Brown. B. 1993. Maine's baitworm fisheries: resources at risk? .4/». Zool. 33:568-577. Brown, B. & W. H. Wilson. Jr. 1997. The role of commercial digging of mudflats as an agent for change of infaunal intertidal populations. J. E.\p. Mar. Biol. Ecol. 218:49-61. Campbell, A., R. M. Harbo & C. M. Hand. 1998. Harvesting and distri- bution of Pacific geoduck clams. Panopea ahrupla. in British Colum- bia. In: G. S. Jamieson & A. Campbell, editors. Proceedings of the North acific Symposium on Invertebrate Stock Assessment and anage- ment. pp. 349-358. Can. Spec. Publ. Fish. Aquat. Sci. no. 125. Commito. J. A. 1982. Effects of Lunatia heros predation on the population dynamics o( Mya arenaria and Macoma balthica in Maine, USA. Mar. Biol. 69:187-193. Creaser. E. P.. Jr.. D. A. Clifford. M. J. Hogan & D. B. Sampson. 1983. A commercial sampling program for sandworms. Nereis virens Sars, and bloodworms. Glycera dibranchiuta Ehlers. harvested along the Maine coast. NOAA Tech. Rep. NMFS SSRF-767. Cryer. M.. G. N. Whittle & R. Williams. 1987. The impact of bait collec- tion by anglers on marine intertidal invertebrates. Biol. Consen: 42: 83-93. Dow. R. L.. D. E. Wallace & L. N. Taxiarchis. 1954. Clam (Mya arenaria) breakage in Maine. Maine Depl. Sea Shore Fish. Res. Bull. No. 15. Dow. R. L. & E. P. Creaser, Jr. 1970. Marine bait worms: a valuable inshore resource. Arl. Slates Mar. Fish. Coinm. Leafi. 12. Ellis, K. L. & M. Waterman. 1988. Maine clam handbook: a community guide for improving shellfish management. ME/NH Sea Grant Coll. Prog. Publ. 98-05. Orono, ME. Emerson. C. W. 1990. Influence of sediment disturbance and water flow on the growth of the soft-shell clam, Mya arenaria L. Can. J. Fish. Aquat. Sci. 47:1655-1663. Emerson, C. W., J. Grant & T. W. Rowell. 1990. indirect effects of clam digging on the viability of soft-shell clams. Mya arenaria L. Neth. J. Sea Res. 27:109-118. Clock. J. W. & K. K. Chew. 1979. Growth, recovery, and movement of manila clams. Venerupis japonica (Deshayes) at Squaxin Island, Wash- ington. Proc. Natl. Shellfish Assoc. 69:15-20. Glude, J. B. 1954. Survival of soft-shell clams. Mya arenaria. buried at various depths. Maine Dept. Sea Shore Fish. Res. Bull. No. 22. Hall. S. J. & M. J. C. Harding. 1997. Physical disturbance and marine benthic communities: the effects of mechanical harvesting of cockles on non-target benthic infauna. / Appl. Ecol. 34:497-517. Heinig, C. & D. Campbell. 1992. The environmental context of a Gyrod- iniiiin aureolum bloom and shellfish kill in Maquoit Bay. Maine. Sep- tember. 1988. J. Shellfish Res. 11:11 1-122. Hruby. T. 1982. The effect of digging on clam numbers. Mass. Audubon Soc. Gloucester. Mass. 12 February 1982. Jackson, M. J. & R. James. 1979. The influence of bait digging on cockle Commercial Clam and Worm Harvhsting 1157 Ccraslddcinui ctlulc popiikitioiis in Ndith Norfolk. ./. ■■\ppl. Ecol. Id: 671-679. Kalin. R. J. 19X4. Ob.sersalions ot a feeding method of the .Allantic nhhon worm. Ci'ivhnitiiliis Icureiis. E\li«irii's 7:170-180. Kern, W. R. 19f^.'^. Structure and actions of neniertine lo\in.s. Am. Zonl. 25:99-111. Lenihan. H. S. & F. Miclieli. 1999. Biological effects of shelllish harvest- ing on oyster reefs: resolving a fishery conllict by ecological experi- mentation. Fish. Bull. 98:86-95. Matthiessen. G. C. 1963. Intertidal zonation in populations of Myu arenaria. Lininol. Ocecmogr. 5:381-388. McAllister. M. K. & R. M. Peterman. 1992. Experimental design in the management of fisheries: a review. N. Am. J. Fish. MimuaemeM 12:1-1S, Medcof, J. C. & J. S. MacPhail. 1964. Fishing efficiency of clam hacks and mortalities incidental to fishing. Proc. Null. .Shellfish Assoc. 55:53-72. Olive. P. J. W. 1992. Management of the exploitation of lugworm Arcni- coUi marina and the ragworm Nereis virens (Polychaeta) in conserva- tion areas. Aquat. Consen: Mar. Freshwat. Ecosyst. 3:1-24, Peterson. C. H.. H. C. Summerson & S. R. Fegley. 1987. Ecological con- sequences of mechanical harvesting of clams. Fish. Bull. 85:281-298. Peterson. C. H. & B. F. Beak 1989. Bivalve growth and higher order interactions: importance of density, site, and time. Ecology 70:1390-1404. Robinson. S. M. C. & T. W. Rowelk 1990. A re-examination of the inci- dental fishing mortality of the traditional clam hack on the soft-shell clam, Myu urenaria Linneaus, 1758. J. Shellfish Res. 9:283-290. Rowelk T. W. & P. Woo. 1990. Predation by the nemertean worm. Cer- ehrutuhis lucleus Verill, on the soft-shell clam, Mya areiwriu Linneaus. 1758, and its apparent role in the destruction of a clam flat. / ShellfisI: fei. 9:291-298. Shaw. R. G. & T. Mitchell-Olds. 1993. ANOVA for unbalanced data: an overview. Eciihifty 74:1638-1645. Spencer. B. E.. M. J. Kaiser & D. B. Edwards. 1998. Intertidal clam harvesting: benthic community change and recovery. Ac/uueull. Res. 29:429-137. Steel. R. G. D. & J. H. Torrie. 1980. Principles and Procedures of Statistics. 2nd ed. New York: McGraw-Hill. 633 pp. Toba, D. R.. D. Thompson & K. K. Chew. 1992. Effects of substrate modification on natural recruitment, growth and survival of hardshell clams in Washington State. J. Shellfish Res. 1 1:207-208. \an den Heiligenberg. T. 1987. Effects of mechanical and manual harvest- ing of lugworms Arenicola marinu L. on the benthic fauna of tidal flats in the Dutch Wadden Sea. Biol. Conserw 39:165-177. Wallace. D. E. 1997. The molluscan fisheries of Maine. In: C. L. Mac- kenzie. Jr.. V. G. Burrell. Jr.. A. Rosenfield & W. L. Hobart. editors. The History. Present Condition, and Future of the Molluscan Fisheries of North and Central America and Europe. Vol. 1. Atlantic and Gulf Coast. U.S. Dept. Commerce. NCAA Tech. Rep. 127. pp. 63-86. Wilson. J. A.. J. M. Acheson. M. Metcalfe & P. Kleban. 1994. Chaos. complexity and community management of fisheries. Mur. Policy 18: 291-.W5. Winer. B. J.. D. R. Brown & K. M. Michels. 1991. Statistical Principles in Experimental Design. 3rd ed. New York: McGraw-Hill. 1057 pp. Zaklan. S. D. & R. Ydenberg. 1997. The body-si/e relationship in the infaunal clam Myu arenuriu. J. Exp. Mur. Biol. Ecol. 215:1-17. Zwarts. L. & J. Wanink. 1989. Siphon size and burying depth in deposit- and suspension-feeding benthic bivalves. Mur. Biol. 100:227-240. Joiinuil i>t Slulljhh Rcscanli. Vol. 20. No. 3. 1134-116?. 2(101. GENETIC VARIATION IN POPULATIONS OF THE COMMON CHINESE CUTTLEFISH SEPIELLA MAINDRONI (MOLLUSCA: CEPHALOPODA) USING ALLOZYMES AND MITOCHONDRIAL DNA SEQUENCE ANALYSIS XIAODONG ZHENG,' * RUCAI WANG,' XIAOFENG WANG,' SHU XIAO,' AND BING CHEN' 'Mariciilriirc Research Lab. Fisheries College. Qingdao Ocean llniversity. Qingclao. .Shandong 266(>0.\ P. R. China: -Molecular Virology Lab. School of Medicine. Qingdao University. Qingdao. Shandong 266071. P. R. China ABSTRACT Allo/.ymes and sequence of cytochrome o.\i(Jase subunil I (COI) gene were used to investigate the genetic variation in five populations of the common Chinese cuttlefish Sepiella maindroni in waters of China and Japan. Samples were examined at 14 enzymes comprising 23 putative allozyme loci. The results revealed moderate levels of genetic variability: the proportion of poly- morphic loci = 0.226og>j. mean observed heterozygosities per locus = 0.038, mean expected heterozygosity per locus = 0.03y. and the average effective number of alleles = 1.26. Nei's mean genetic distance ranged from 0.0001 to 0.0018. Part of the cytochrome oxidase subunit I gene was amplified with the polymerase chain reaction (PCRl and sequenced for 27 individuals: five to six from each of the five populations. Sequence data showed that there were 1 1 variable nucleotide positions In the 661 base-pair segments of the gene, and the 27 sequences could be grouped into nine haplotypes (A-I). No remarkable genetic difference was observed among those populations. KEY WORDS: genetic variation. Si-piclla maimlntiii. allo/ynie. cytochrome oxidase 1 gene. DNA sequencing INTRODUCTION MATERIALS AND METHODS The cuttlefish Sepiella maindroni de Rochebrune. 1884. is widely distributed in Asiu. It is an important and valuable fishery resource in China. Japan and South Korea (Dong 1991. Nesis 1987, Okutani 1995). Its output accounted up to about 609?- of the total cephalopod production in 1980s in China, and occupies an important position in Chinese marine products (Qi 1998). But its yield has declined severely since then, due to over-fishing. One of the main reasons for the decline of population size is that aspects of its population biology are poorly documented, especially the population structure and genetic diversity. Allozyme electrophoresis can provide direct information on the distribution of genotypes in accordance with the Hardy -Weinberg paradigm. It remains a valuable approach to describe breeding structure of a species (Carvalho & Nigmatullin 1998). Few genetic studies have been carried out on cuttlefishes, and all showed low levels of polymorphism (Perez-Losada et al. 1996. Perez-Losada et al. 1999). For providing useful information for resource manage- ment and protection, we examined four populations in the coastal waters of China, and one sample from Nagasaki port in Japan using allozyme electrophoresis. Intraspecific set^uence variation within mtDNA has also proven a powerful tool for examining population structure in marine organisms (Carvalho & Pitcher 1994). The combination of mitochondrial DNA and allozyme analysis may be helpful in understanding the evolutionary dynam- ics of these interacting populations of marine species (Hilbish 1996). We further investigated genetic constitution of those popu- lations through examining nucleotide secquence of part of the cy- tochrome oxidase subunit I (COI). *Correspondnig author. XiaoDong Zheng, Mariculture Research Lab, Fish- eries College, Qingdao Ocean University, Qingdao. Shandong 266003, P, R. China. Tel.: -H86-532-203-2873; Fax: -H86-532-289-4024; E-mail: xdzhengls'ouqd.edu.cn Samples Fresh samples were randomly collected from five locations in the waters around China and Japan (Fig. 1. Table!), and kept frozen at -20°C until dissection. Mantle length of all the speci- mens was measured and sex were recorded (Table 2). Mantle muscle tissue, liver, eyes and buccal muscle were taken from each individual in ice as quickly as possible, then transfened for storage at -80°C allozymic analyses. Muscle samples were removed and kept in liquid nitrogen for DNA analysis. Allozymic Analysis Electrophoresis was carried out using standard horizontal starch gel techniques (Harris & Hopkinson 1976). essentially as described by Zheng et al. (2()(llb). Forty-six putative enzyme- 120" 122' 124* 126- 128' 130" Figure 1. Sampling sites of Sepiella maindroni collected in the coastal waters of China and Nagasaki port in Japan. 1159 1160 Zheng et al. TABLE 1. Populations and samples used for electrophoresis and sequencing analysis in Sepiella inaindroni No. of Specimens Allozvme Sequencinj; Location Abbreviation of Location Date 21 fi4 50 IS 49 Nagasaki Putian Nan'ao Shenzhen Zhanjiang NG PT NA SZ ZJ 19W.4. 2000.4. 2000.5. 2000.11. 2000.10. coding loci and 4 different kinds of tissues of S. maindioni (e.g. liver, eyes, mantle and buccal muscle) were screened using several enzyme-staining recipes (Shaw & Prasad 1970. Shaw & Prasad 1990. Morizot & Schmidt 1992) in a range of buffer systems. After preliminary experiments for some specimens from five populations. 14 enzymes and optimal buffer systems and tissue (Table 3) were selected for further research. Allele frequencies were estimated for each sample, together with several measures of genetic diversity. The proportion of poly- morphic loci (P). values of the observed heterozygosity (H„), the expected heterozygosity (H^J. and the effectixe number of alleles (N.,) were calculated by direct census of the population data. Two criteria of polymorphism were used: ?„„„ and Pogj. Nei"s ( 1972) genetic distance (D^^.,) was calculated to quantify the genetic di- vergence between cuttlefish samples. Dendrogram to illustrate the genetic relationship among five populations was constructed from genetic distance using the unweighted pair-group arithmetic mean (UPGMA) cluster-analysis algorithm (Sneath & Sokal 1973). DNA Exiractiuii, Amplification and Sequencing Total genomic DNA was extracted using a CTAB method modified from Winneponninckx et al. (1993). The target DNA segments were amplified by PCR. The primers used for the am- plification of partial COI gene were: HC02198 (5'-TAA ACT TGA GGG TGA CCA AAA AAT-3') and LC01491 (5'-GGT CAA CAA ATC ATA AAG ATA TTG-3') from Folmer et al. (1994). PCR amplifications were performed under the following conditions: I2()s at 94°C. then 30 cycles of 40s at 94°C. I min at 50°C. and I min at 72°C. A total volume of 25 |jiL reactions consists of 0.3 units of Taij (TaKaRa), 0.3 pM each primer. 0.2 \s.M each dNTP. 2.3-3.3 niM MgCU. 2.3 |xl of 10 x buffer sup- plied with Ttiij and 4p.l (30-50 ng) of template DNA. The amplified fragments were purified by the PCR fragment recovery Kit (TaKaRa). Purified products were sequenced directly using the ABl PRISM BigDye Terminator cycle sequencing Ready Reaction Kit and AmplicTaq DN.'X polymerase with ABl PRISM 377XL DNA sequencer (Applied Biosystem Inc.). All final se- quences were obtained from both DNA strands for verification. Pliylogenclic and Sequence Difference Analyses The sequences were aligned with other gene sequences (out- group species) using a computer-assisted procedure. Clustal W (Version 1.80) (Thompson et al. 1994). No gaps were found in the COI sequences. The extent of sequence difference between indi- viduals was calculated by averaging pair-wise comparisons of se- quence difference across all individuals. One thousand bootstrap replicate data sets were produced using SEQBOOT. Under the Maximum Likelihood model, the sequence similarity and genetic distance were calculated with PHYLIP 3.5c (Felsenstein 1995). Distance-matrix trees were then constructed using the Fitch and Margoliash (1967) least-squares (LS) method and the neighbor- joining (NJ) method (Saitou & Nei 1987). Sepici offwinaUs was taken as the distant outgroup species (Carlini & Graves 1999. Accession no. AF000062). RESULTS Descripliiin and Interpretation of Allozynies Of the 14 enzymes assayed routinely. 15 loci (AAT-2*. ALP*. CK*. EST-2* FBP*. G3PDH-2*. GPI-I*. GPI-2*. GRS-I*. GRS- 2*. IDHP-I*. MPI-I*. MPI-2*. PGM-1*. SOD*) were monomor- phic in all samples, while the others (AAT-1*. EST-1*. G3PDH- I*. GPl-3*. IDHP-2*. NP*. PGDH*. PGM-2*) showed the most common allele frequencies were less than 0.99 in at least one sample. Allozymes that were consistently and clearly scored in all samples were considered useful genetic markers for population analysis. Genetic \ariability was estimated by the proportion of polymorphic loci and average heterozygosity. P. the proportion of polymorphic loci ranged from 0.043 to 0.217(<0.95) and from 0.217 to 0.261 (<0.99). respectively. When five samples were TABLE 2. Mantle length and sex composition of Sepiella maindroni from five populations No. of Specimens Mantle Length (mm) Sex Composition Location Range Mean Male Female 21 64 50 18 49 NG PT NA SZ ZJ 70-170 76-258 69.1-170 ,^7-97.4 3VI50.3 123.50 1 50.(1 127.4 68.71 54.97 12 35 18 9 31 32 Genetic Variation in Chinese Clttlefish 1161 TABLE 3. Enzymes examined, abbreviations. E. C. number, optimal buffer systems and tissue of the study Enzyme E. C. No Abbrev. Buffer Tissue Aspartate aniinotranst'erase Alkaline phosphatase Creatine kinase Esterase Glucosephosphate isomerase Fructose biphosphatase Glucose-3-phosphale dehydrogenase Isocitrate dehydrogenase Glutathione reductase Mannosephosphate isomerase Nucleoside phosphorylase Phosphogluconate dehydrogenase Phosphoglucomutase Superoxide disniutase 2.6.1.1 3.1.3.1 2.7.3.2 3.1. 1. 1 5.3.1.9 3.1.3.11 1 . 1 . 1 .49 1.1.1.42 1.6.4.2 5.3.1.8 2.4.2.1 1.1.1.44 5.4.2.2 1.15. 1, 1 AAT ALP CK EST GPI FBP G3PDH IDHP GRS MPI NP PGDH PGM •SOD CAMPV.n T-C 8.0 T-C 8.0 T-C 8.0 CAMP7.0 T-C 8.0 TVB-LB8.5 CAMP7.0 T-C 8.0 T-C 8.0 T-C 8.0 CAMP7.0 CAMP7.0 T-C 8.0 Mantle Liver Mantle Liver Mantle Mantle Mantle Mantle Liver Eye Eye Mantle Mantle Liver TABLE 4. Allele frequencies at 8 polymorphic loci IP), sample sizes (N( and indices of genetic variability within 5 samples n( Sepiella waindrnm. H,,. mean obser^ed heterozygosities: H,.. mean expected heterozygosity; Na, mean effective number of alleles. allele Samples Locus Nagasaki Putian Nan'ao Shenzhen Zhanjiang AAT-I* a 0.024 0.023 0.010 0.028 0.010 b 0.976 0.969 0.970 0.944 0.929 c 0.000 0.008 0.020 0.028 0.061 N 21 64 50 18 49 EST-1* a b 0.024 0.976 0 1 0 1 0 1 0 1 N 21 64 50 18 49 G3PDH-1* a 0.024 0.024 0.041 0.028 0.094 b 0.952 0.944 0.929 0.944 0.854 c 0.024 0.032 0.031 0.028 0.052 N 21 62 49 18 48 GPI-3* a 0.952 1 1 1 1 b 0.048 0 0 0 0 N 21 64 50 18 49 lDHP-2* a 1 0.984 0.989 0.944 0.888 b 0 0.016 0.0 1 1 0.056 0.112 N 21 63 47 18 49 NP* a 0.31 0.359 0.345 0.361 0.367 b 0.69 0.641 0.655 0.639 0.633 N 21 64 42 18 49 PGDH* a 0 0.095 0.064 0.059 0. 1 09 b 1 0.905 0.936 0.941 0.891 N 21 63 47 17 46 PGM-2* a 1 1 1 1 0.99 b 0 0 0 0 0.01 N 21 64 50 18 49 "0,95 0.043 0.130 0.130 0.217 0.217 "0,99 0.217 0.217 0.217 0.217 0.261 H„(S.E) 0.031 (0.008) 0.034(0.009) 0.034(0.0101 0.041 (0.012) 0.050(0.013) H,(S.E) 0,031 (0.008) 0.036(0.010) 0.034(0.010) 0.039(0.010) 0.055(0.014) Na 1.212 1.235 1.240 1.302 1.294 For 5 samples pooled: Po,^ =0.147; P,,,^ =0.226 Mean heterozygosities per locus = 0.038(H„); 0.039(H,.) The average effective number of alleles = 1 .26 1162 Zheng et al. TABLE 5. Genetic similarity (I) and genetic distance (D^iJ|) among five samples in the cuttlefish Sepiella maiitdroiii Samples Nagasaki Putian Nan'ao Shenzhen Zhanjiang Naga^aki Putian Nan'ao ShenZhen Zhanjiang 0.0007 0.0004 0.0006 0.00 IS 0.9993 0.0001 0.0002 0.0008 0.9996 0.9999 0.0001 0.0008 0.9994 0.9998 0.9999 0.0005 0.9982 0.9992 0.9992 0.9994 locus (observed and expected), and the effective number of alleles per locus were presented for all of the five populations pooled (Table 4). Genetic similarity (I) and genetic distance (D^^^,,) between all samples based on the 23 loci were given in Table 5. Nei's D of S.maindnmi populations varied from 0.0001 to 0.0018 (Table 5). Comparison of observed genotypic frequencies with Hardy- Weinberg expectation did not show significant deviation {P > 0.05) at NP* loci of five populations. The expected values of the genotype frequency were smaller than fi\e in other polymorphic loci, so a X" test for the Hardy-Weinberg equilibrium could not be performed. The dendrogram was constructed from Nei's D value (Fig. 2.) Sequence Variation of the COI Gene A base-pair ( bp ) ( 66 1 ) segment of the COI gene was sequenced for 27 individuals from five locations (NG. PT. NA. SZ and ZJ). Target segments were given by base pairing. The A + T percentage of 5 samples ranged from 6S.38'J to 68.99% (Table 6). which is a little higher than S. offhiiuilis's (63.47%) (Carlini & Graves 1999) Among these individuals, eleven nucleotide positions (position 146. 197. 263. 317. 362. 497. 543 and 617) were found to be variable (Fig. 3). As expected, ten variation sites were at third position of codon triplets, and only one variation site (position 621) was at P' position of codon triplets (Fig. 3). The base of positions 621 and 623 occuned to transverse from A to T at the same time in four individuals checked, and finally resulted in compatible amino acid change, in which Cys was in place of Ser. The other base substitutions were silent (.synonymous) changes that do not result in amino acid change. Analysis of Relationships and Occurrence Frequency of Haplolypes Nucleotide sequence difference between 27 individuals from Nagasaki in Japan to Zhanjiang in South China Sea ranged from 0.00 to 1.21%. Among these individuals, nine unique haplotypes (A-I) were found (Table 7). The haplotype D. with a frequency of 48% (13/27). was most commonly observed in all samples, though it was found only in the 4 populations living in the coastal waters of China. The haplotype A followed, and its highest frequency (4/6) was found in NG. The haplotype E was found in PT and ZJ. The other haplotypes were found each in only one individual. According to the phylogenetic tree of Neighbor-Joining (Fig. 4) and UPGMA (Fig. 5), there were two main branches: most of NG (67-83%) occupied one position from one branch, and the indi- viduals (81-86% ) living in the coastal waters of China took up the other branch. But genetic distance of 27 individuals as revealed by T--L--Y --F--I--F- -G--I--W-- S--G--L--L --G--T--S- GAACATTATA TTTTATTTTT GGTATTTGAT CAGGTTTATT AGGTACTTCA 50 -L--S--L-- M--I--R--S --E--L--G- -K--P--G-- T--L--L--N TTAAGTTTAA TAATTCGAAG AGAATTAGGA AAACCAGGTA CTCTATTAAA 100 — D--D — Q- -L — y — N — V — V — V--T --A--H — G- -F--I--M — TGATGATCAA TTATATAATG TTGTAGTAAC CGCCCACGGT TTTATCATAA 150 I--F — F — L — V--M--P- -I--M--I-- G--G — F--G --N--W--L- TTTTCTTTTT AGTTATACCT ATTATAATTG GAGGTTTTGG TAATTGATTA 200 -V--P--L-- M--L--G--A --P--D--M- -A — F--P-- R — M--N — N GTTCCCTTAA TATTAGGGGC ACCAGACATA GCCTTCCCTC GAATAAATAA 250 --M — S — F- -W — L — L — P — P — S — L — T — L — L- -L — S — S — TATAAGTTTT TGGTTATTAC CTCCATCTTT AACTCTTTTA TTATCATCCT 300 S--A--V--E — S — G--A- -G — T--G-- W--T — V — Y — P--P — L- CAGCTGTAGR AAGAGGT6CT GGAACTGGAT GAACAGTATA TCCTCCCTTA -S — S — N — L--S — H — A --G--P--S- -V — D — L-- A — I — F — S TCTAGTAATC TATCTCATGC TGGCCCATCT GTAGATTTAG CTATTTTTTC 4 00 --L--H--L- -A--G--V-- S — S--I--L --G--A — I- -N — F — I — TTTACATTTA GCTGGTGTTT CCTCAATCTT AGGTGCTATT AATTTTATTA T--T--I--L --N--M--R- -W--E--G-- L--Q--M--E — R--L--P- CAACTATTTT AAATATACGG TGAGAGGGTT TACAAATAGA ACGACTTCCT 500 NG PT NA SZ ZJ 10 10 10 10 Figure 2. Dendrogram derived from Nei's genetic distance (Nei.1972 using UPGMA cluster analysis. -L--F — V — W — S — V — F — I — T — A- -I--L — L-- L--L--S--L TTATTTGTTT GATCCGTATT TATTACAGCT ATTTTACTAC TATTATCCTT --P — V--L- -A--G — A-- I — T--M--L --L--T — D- -R--N--F-- ACCAGTTTTA GCTGGAGCCA TTACTATATT ATTAACCGAT CGAAATTTTA N — T — T--F --F--D--P- -S--G — G — G — D — P — I — L — Y — Q- ATACAACATT TTTTGATCCT AGAGGAGGAG GTGACCCTAT TTTATATCAA -H--L — F-- CATTTATTTT - 219 G 661 Figure 3. Nucleotide and deduced amino acid sequences of partial COI gene fragment in S. maindroni. Asterisks above the sequence show the variation nucleotide positions. Genetic Variation in Chinese Cuttlefish 1163 Sepiella Maindroni TABLE 6. Comparison in base composition of COI gene A^f C% G% T% Accession No. Nagasaki Putian Nan'ao ShenZlien Zhanjiang Outgroup Sepia officiiiali: 0.286 0.283 0.284 0.284 0.280 0.265 0.159 0.163 0.163 0.159 0.163 0.192 0.151 0.153 0.153 0.153 0. 1 53 0.174 0,404 0401 0.399 0.404 0.404 0.37(1 AF346S53 AF340032 AF361360 AF361359 AF361361 AF000(162 A: adenine; C; cytosine: G: guanine; T: thymine the COI sequence ranged IVotii 0.(1(1(1(1 to 0.(1(176 iisint; PHYLIP .3.3c. Geographical pro.xiinity had tio remarkable relevance to ge- netic-phylogenetic relationships among these haplotypes. DISCUSSION The study of enzyme loci in five populations of 5. iniiiiulroiu showed that this species had low levels of variability. The mean H,,{0.038) falls below the average for invertebrate and molluscs (0.120 and 0.145. respectivelyl (Ward et al. 1992). and also for marine molluscs (0.147) (Fujio 1997), but within the range for cephalopods (0.03 ± 0.03) (Zheng et al. 2001a). Low levels of heterozygosity have been found in other species of Cephalopoda, such as Loligo gahi, llle.x argentimis (Carvalho & Pitcher 1989, Carvalho et al. 1992), Sepioteuthis lessoniana (Izuka et al. 1994), Sepia officinalis. S.orhignyana. S.elegans (Perez-Losada et al. 1996). Very low levels of genetic variability are apparently a com- mon facet of cephalopods except Berryteuthis magister (Hf, = 0.131) (Katugin 19931. a feature probably unique over any major group of inveilebrates. In the five populations, the proportion of polymorphic loci of NG was the lowest (P0.95 = 0.043), and ZT was the highest. The sample of NG was obviously different from other samples at loci of AAT-1*, EST-1*. GPI-3*, IDHP-2* and PGDH*. The fi.xed allelic differences at AAT-1* between NG and other samples in coastal waters of China could serve as valuable markers to monitor geographic variation in stock distribution, though the genetic dis- tance was low (maximum value of D = 0.0018 being found be- tween the NG and ZJ samples |. Populations undergoing severe bottlenecks with rapid reduc- tions in population size may lose rare alleles through genetic drift (Wright 19781. Severe fishing pressure could bring about such population declines, especially if populations were founded each year by relatively few individuals. Population crashes are more likely in commercially harvested species (Nelson & Soule 1987). particularly when production models over-estimate escapement rates. In annual species such as S. maindroni, the genetic impli- cations of bottlenecks are particularly significant since there are no overlapping year classes to serve as a repository for genetic diver- sity. For instance, in China, the cuttlefish S. inaimlioni had once become one of the four biggest fisheries in the East China Sea. But now, its position has been replaced by other cephalopods. Clearly, it is facing the danger of extinction. A bottleneck effect may be one of the reasons for the low diversity of S. maindroni populations. The present COI gene analysis suggests that there was no re- markable genetic differentiation in the five samples, though there was evidence to indicate that NG samples differ slightly from the other samples in the coastal waters of China, especially in the number and type of haplotypes and the position in UPGMA and NJ trees. The result also supported the conclusion of biochemical research. Pelagic organisms in the open ocean were generally regarded as having low levels of population differentiation, resulting from probably ample opportunities of dispersal in various life history stages and the lack of physical barriers in the environment. The study on the population genetic structure of large pelagic fishes and some cephalopods in the open oceans, such as swordfish and TABLE 7. Variable nucleotide positions In part of the COI gene region of 11 haplotypes, and number of individual of each haplotype found in each locality. Site Number of Individuals in Each Locality Haplotype 146 197 263 317 362 497 543 617 620 621 623 NG PT NA SZ ZJ Total A C A G T A T T T T A A 4 0 0 1 0 B G C 1 0 0 0 0 C G C C I 0 0 0 0 D G T C C 0 3 4 3 3 13 E G T C C T T 0 1 0 0 ~t F G T c C A T T 0 1 0 0 0 G G C C C 0 0 1 0 0 H G 0 0 0 1 0 I T G A G c 0 0 0 0 1 Dots(-) indicate Identities. 1164 Zheng et al. Hh NA(4#) NG(5#) 0.0015 SZ(8#) 0,00154 I SZ(2#) I NG( 1,2,7,1 #) aooi5 0.00457 0.1 NG(8#) PT(1,2,3,26#) NA(5,6,25,26#) SZ(4,7,9#) ZJ{1,2,6,7,8#) PT(36#) ZJ(3#) Sepia officinalis Figure 4. NcighliDr-Joining tree ba.sed on COI gene sequence.s. Sepia officinalis was considered as distant outgroup species. diamond-shaped squid, showed very low levels of genetic diversity (Rosel & Block 1W6. Kitaura et al. 1998). The cuttlefish S. inaindroni is distributed widely in Asia — in the sea of Honshu (Japan) to the north, Malaysia and the Philippine Islands to the south, and Indian Ocean to the west. The present study suggests that for 5. inaindroni. much like a pelagic species with considerable mobility, behaves so that all the populations covered in this study may have migrated into each other and no physical barriers existed in the environment, especially regarding the samples from PT to ZJ. Genetic diversity in populations of the cuttlefish needs to be further examined with more extensive data from different geo- Figure 5. UPGMA phylogenetic tree based on COI gene Sepia officinalis was used as the distant outgroup species. - SZ(8#) I SZ(2«) I NG(I.2.7,II#) - PT(36«) I NG(5#) 1 N.A(4«) NG(8») PT(1,2,3.26()) NA(5,6,25.26») SZ(4.7.9#) ZJ(1.2.6,7,8#) ZJ(3«) Sepia officinalis sequences. graphical populations and also from more DNA markers such as highly variable microsatellite DNA markers. ACKNOWLEDGMENTS We are grateful to Prof. Yutaka Natsukari for generously pro- viding reagents and lab space for allozyme analysis. We thank Dr. T.X. Gao. Z.G. Chen, for their valuable suggestions and help on data analysis. We also thank X.D. Du. X.Z. Lin, M.Z. Zeng. W.Z. Chen, J.M. Zhao, for assistance in the collection of samples. This work was supported by the National High-tech Development Proj- ect arants 863-819-01-01. LITERATURE CITED Carlini. D. B. & J. E. Graves. 1999. Phylogenetic analysis of cytochrome c oxidase I sequences to determine higher-level relalionships within the coleoid cephalopods. Bull. Mar. Sci. 64:57-76 Carvalho. G. R. & Ch. M. Nigmatullin. 1998. Stock structure analysis and species identification. In: Rodhouse P. G. et al., editors. Squid recruit- ment dynamics. Rome: FAO Fisheries Technical paper No.376. pp. 199-232. Carvalho, G. R. & T. J. Pitcher. 1989. Biochemical genetic studies on the Patagonian squid. Loligo galii d 'Orbigny II. Population structure in Falkland waters using isozymes, morphometries, and life history date. J. E.xp. Mar.Biol. Ecol. 126:243-258. Carvalho. G. R. & T. J. Pitcher. 1994, Molecular Genetics in Fisheries. London: Chapman & Hall. 141 pp. Carvalho. G. R.. A. Thompson. & A. L. Stoner. 1992. Genetic diversity and population differentiation of the shortfin squid llle\ angenliims in the southwest Atlantic. J.Exp.Mur.Biol.Ecol. 158:105-121. Dong. Z. Z. 1991. Biology of the economic species of cephalopods in the world oceans. Jinan: Shandong Science and Technology Press, pp. 216-227. Felsenstein J. 1995. PHYLIP: Phylogeny Inference Package. Version 3.57c. Seattle: Department of Genetics. University of Washington. Fujio, Y. 1997. Genetic variability and heterotic effects within population of aquatic organisms. Fish. Genet. Breed. Sci. 24:43-52. Fitch, W. M. & E. Margoliash. 1967. Construction of phylogenetic tree. Science. 155:276-284. Folmer, O.. M. Black. W. Hoeh. R. Lut/, & R. Vnjenhoek. 1994. DNA primers for implication of mitochondrial cytochrome c oxida.se suhunit I from diverse metazoan in\ertebrate. Mol. Mar. Biol. Biotech. 3:294- 299. Harris, H. & D. A. Hopkmson. 1976. Handbook of Enzyme Electropho- resis in Human Genetics. North-Holland Publishing Company. Amster- dam. Hilbish. T. J. 1996. Population genetics of marine species: the interaction of natural selection and historically differentiated populations. / Exp. Mar.Biol. Ecol. 200:67-83. Izuka T., S. Segawa. T. Okutoni et al. 1994. Evidence on the existence of three species in the oval squid Sepioteiilhis lessoniana complex in Ishigaki Island. Okinawa, southwestern Japan, by isozyme analyses. Vra».v 53:217-228. Katugin. O. N. 1993. Genetic variation in the squid Berryteuthis magister (Breey. I9I3) (Oegopsida: Gonatidae). In: T Okutani. R. K. O'dor & T. Kubodera, editors. Recent advances in fisheries biology. Tokyo: Tokai University Press, pp. 201-213. Kitaura. J., G. J. Yamamoto & M. Nashida. 1998. Genetic variation in populations of the diamond-shaped squid Tliysanoteuthis rhombus as examined by mitochondrial DNA sequence analysis. Fisheri. Sci. 64: 538-542. May, B. 1992. Starch gel electrophoresis of isozymes. In: A. R. Hoelzel, editor. Molecular Genetic Analysis of Populations: A Practical Ap- proach. London: Oxford Univ. Press, pp. 1-27 and 271-280. Morizot. D. C. & M. E. Schmidt. 1990. Starch gel electrophoresis and histochemical visual lization of protein. In: D. H.Whitmore. editor. "Electrophoretic and Electro Focusing Technique in Fisheries Manage- ment". Florida: CRC Press, pp. 23-84. Nei. M. 1972. Genetic distance between populations. .\mer. Nat. 106:283- 292. Nelson. K. & M. Soule. 1987. Genetical conservation of exploited fishes. In: N. Ryman & T. Utter, editors. Population genetics and fishery management. Washington Sea Grant Program, Seattle: University of Washington Press, pp. 345-368. Genetic Variation in Chinese Cuttlefish 1165 Nesis. K. N.. I9S7. Cephalopods of the world. T.FH PuhliLUtions, Inc.. Ltd. USA. 9 pp. Okutani. T. 1995. Cuttletlsh and squids of the world in color. Tokyo: National Cooperative Association of Squid Processors. 4.^ pp. Perez-Losada. M.. A. Guerra & A. Sanjuan. 1996. Allozyme electropho- retic technique and phylogenetic relationships in three species of Sei>iu (Cephalopoda: Sepiidae). Comp. Biochem. Physiol. B. 114:11-18. Perez-Losada. M.. .\. Guerra & A. Sanjuan. 1999. Allozynie differentiation in the cuttlefish Sepia officinalis (MoUusca: Cephalopoda) from the NE Atlantic and Mediterranean. Heredity. 83:280-289. Qi , Z. Y. 1998. Economic Mollusca of China. Beijing: China Agriculture Press. 297 pp. Rosel. P. E. & B. A. Block 1996. Mitochondrial control region variabilily and global population structure in the swordfish. Xiphias i;laJiii\. Mar. Biol. 125:11-22. Saitou. N. & M. Nei. 1987. The neighhor-joining method: A new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406—125. Shaw. C. R. & R. Prasad. 1970. Starch gel electrophoresis of enzymes — A complication of recipes. Biochemical Genetics 4:297-320. Sncath. P H. .A. & R. R. Sokal. 1973. Numerical taxonomy-the principles and practice ot luimberical classification. W. H. Freeman c'i Co., San Francisco. Thompson. J. D.. D. G. Higgins & T. J. Gibson. 1994. CLUSTAL W: improv ing the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weigh matrix choice. Nude. Acids Res. 22:4673—1680. Ward. R. D., D. O. F. Skibinski & M. Woodwork. 1992. Protein heterozy- gosity, protein structure and taxonomic differentiation. Evol. Biol. 26: 7.VI59. Winnepenninckx. B.. T. Backeljau & R. De Wachter. 1993. Extraction of high molecular weight DNA from molluscs. Trends Genet. 9:407. Wright. S. 1978. Evolution and genetics of populations. Vol. 4: Variability within and among natural populations. Chicago: University of Chicago Press. Zheng. X. D.. R. C. Wang & Z. P. Wang. 200 la. Advance of studies on Cephalopod genetic variation. / Fisheri. China. 25:84-89. Zheng. X. D.. Y. Natsukari. R. C. Wang, et al. 2001b. Biochemical genetic studies on cuttlefish Sepiella nmindroni (Cephalopoda: Sepiidae)- Active loci screening of isozyme in S.maindroni. Oceanologiu et lim- nolo^iu sinica. 19:345-349. Journal of Slwllfish Kcscurch. Vol. 20. No. 3. 1 167-1 171, 21101. SEASONAL VARIATION IN PHYSIOLOGICAL CONDITION 0¥ AMBLEMA PLICATA IN THE UPPER MISSISSIPPI RIVER EMY M. MONROE AND TERESA J. NEWTON* t U.S. Gt'olof^iciil Survey. Biclciiical RcMnirce.s Division. Upper Midwe.st Eiivironincntul Science.s Center. 2630 Ftiiihi Reed Road. La Crosse. Wiseonsiii 54603 ABSTRACT Measures of physiological condition are being used as sub-lethal endpoints in studies with unionids exposed to a variety of stressors, yet the natural seasonal variation in these measures are largely undocumented. We measured concentrations of glycogen in foot and mantle ti.ssue and a tissue condition index (TCI) in Amhlema plicata (Say 1817). about monthly, for 2 years in mussels that were: (1) obtained directly from the Upper Mississippi River (riverine group); and (2) relocated from the river into an artificial pond (relocated group). In both groups, we observed significant seasonal variation in all physiological indicators. Seasonal variation in glycogen was 729c in mantle and ,')2'7r in foot tissue and paralleled reproductive activity in this short-term breeder. In the relocated group, most of the variation in glycogen occurred dunng the first six months after relocation, suggesting that handling stress may have been a contributing factor. The significant seasonal variation in the TCI paralleled glycogen in riverine mussels. We observed tissue-specitlc differences in glycogen in the riverine group, but not in the relocated group. These data suggest that an interaction of environmental and biological factors influence the energetic status of mussels in natural populations. A better understanding of this variation is needed to interpret changes in physiological condition due to stressors such as relocation. KEY WORDS: glycogen, condition index, seasonal variation. Amhlcmci pliccini. relocation INTRODUCTION Unionids are the most imperiled group of animals in North America with 72% of the 297 species classified as extinct, endan- gered, threatened, or Hsted as a species of special concern (Wil- liams et al. 1993). Declines in species density and diversity have been attributed to habitat degradation, changes in fish host distri- butions, commercial exploitation, and introduced species (Began 1993). Thus, manv agencies are examining approaches to conserve this fatma. Relocation is being used extensively as both a conser- vation and management tool (Cope & Waller 1995), however, it is difficult to draw conclusions on the relative success of relocations because of inconsistencies in the methods used and in the selection of endpoints. Often, survival and growth are the only endpoints measured and they are usually monitored for <1 yr (Cope & Waller 1995). However, one-year survival estimates may underestimate the mortality that occurs during longer relocations (Newton et al. 2001 ). and short-term measures of growth may be of limited value for these long-lived animals (Naimo et al. 1998). Research suggests that changes in certain physiological mea- sures often precede changes in survival (Haag et al. 1993). How- ever. Newton et al. (2001) were unable to predict mortality in AmbU'inu plieata based on physiological indicators measured an- nually over a three- to four-year period. This lack of predictive power may have been related to the frequency of measuretnent or because the measures were not sensitive enough. It is ditticull to draw conclusions about the physiological condition of unionids from measurements made at a few points in time without tnider- standing the seasonal variation in such measures. Various sub-lethal indicators have been examined m unmnids. TCI has been used to assess the condition of freshwater bivalves exposed to zebra mussels (Baker & Hornbach 1997) and contami- nants (Naimo et al. 1990), or infected with parasites (Jokela ct al. 1993). Because the measurement of condition indices requires sac- rificing the individual, their application is litnited. More recent techniques allow tissues to be biopsied for physiological and bio- chemical constituents (Berg et al. 1995. Byrne & Vesk 1996. Naimo et al. 1998). Glycogen is the primary storage form of car- bohydrates in bivalves; consequently, it is being used as a physi- ological indicator in unionids subjected to emersion stress (Greseth 1998). quarantine and relocation (Patterson et al. 1997. Naimo et al. 1998), zebra mussel infestation (Haag et al. 1993. Hallac & Marsden 2000) and parasitic infestations (Jokela ct al. 1993). The few studies measuring seasonal variation in physiological indicators in natural populations suggest that much of this varia- tion is related to environmental conditions or reproduction (Dietz & Stern 1977. Holopainen 1987, Pekkarinen 1993). For example, monthly sampling of Spluicritiiii transversum for total protein and carbohydrates showed a positive conelation with reproductive ac- tivity during the fall and winter and a decline towards the end of the reproductive season (Dietz & Stern 1977). Pekkarinen (1993) found the periods of gravidity for Uiiio piclonim and U. uimidiis were correlated with rising and maximum water temperatures and condition indices rose concomitant with the rise in water tempera- tures. Seasonal changes in the physiological condition of unionids is a complex process involving, at a minimum, environmental conditions and the energetic demands of reproduction. Most studies have measured physiological condition at one point in time in response to a given stressor. In addition, most of these studies measured seasonal variation in physiological indica- tors on bivalves with substantially different reproductive cycles than unionids. Information on the seasonal variability in physi- ological measures in individuals is largely undocumented. Thus, our objectives were to ( 1 ) examine the seasonal variation in gly- cogen concentrations and a TCI in two groups of A. plicata, an undisturbed riverine group and a relocated group; and (2) compare seasonal patterns in glycogen between foot and mantle tissue. MATERIALS AND METHODS We examined the seasonal variation in glycogen concentrations ♦Corresponding author. E-mail: Teresa_Newton@usgs.gov tPublished and in a TCI in two groups of /I. plieata: a riverine group and a formerly as Teresa J. Naimo relocated group. For the ri\erine group, we obtained 10 zebra- 1167 1168 Monroe and Newton mussel-free individuals (80-90 mm shell length) about monthly from May 1997 to October 1998 from Lake Onalaska. a riverine lake in the Upper Mississippi River, near La Crosse, WI. For the relocated group, we relocated 200 zebra-mussel-free individuals (80-90 mm shell length) from the Upper Mississippi River into an 0.04 ha earthen pond supplied with well-water at the Upper Mid- west Environmental Sciences Center in La Crosse, WI. in May 1995. Ten individuals were obtained from the pond for analysis about monthly until November 1996. Although no supplemental food was added, the pond produced extensive blooms of green algae each summer. Because these two groups were sainpled dur- ing different years, we have made no attempt to compare results between groups; rather, we are examining seasonal patterns in physiological measures within a group and between tissues within a group. To obtain tissues for analysis. 30-100 mg pieces of foot and mantle tissue were removed from sacrificed animals, placed on dry ice. and stored at -84°C until analyzed for glycogen. Tissue samples were consistently removed from the same area (Naimo & Monroe 1999). The remainder of the soft tissues and shells were dried to a constant weight at I05°C for determination of the TCI, calculated by dividing the tissue dry mass by the shell dry mass and multiplying the quotient by 100. For the relocated group. TCI analysis was conducted only in 1996. Because A. plkata are not sexually dimorphic, the sex of individuals was not determined. Concentrations of glycogen were determined by the alkaline di- gestion and phenol-sulfuric acid spectrophotometric method (Naimo et al. 1998). The accuracy of glycogen determinations was quantified by the use of procedural blanks, replicates of an in-house reference ma- terial (Naimo et al. 1998). triplicate analysis of 4 aqueous calibra- tion standards, and triplicate analyses of 3 known additions or 3 matrix standards. Glycogen in the tissue samples averaged 0. 1 7 mg (range. 0.03-0.97) in foot tissue and 0.29 mg (range. 0.04-1.86) in mantle tissue. As a measure of precision, we estimated the relative standard deviation of triplicate analyses of known additions ac- cording to APHA et al. ( 1995). Relative standard deviations aver- aged 8% in foot tissue and 12% in mantle tissue. Bias associated with our glycogen determinations was estimated by recovery of matrix standards and known additions (APHA et al. 1995). mean percent recovery of glycogen averaged 106% in foot tissue and 99% in mantle tissue. We assumed glycogen concentrations (mg/g dry weight) fol- lowed a gamma distribution. The gamma distribution is a plausible model for random variables, like glycogen concentration, that can- not take values less than zero and for which the variance is ap- proximately proportional to the squared mean (McCullagh & Nelder 1989). We used the model for mean glycogen concentration in tissue j from group /. p.,,, given by: sinu.soidal seasonal effect at a frequency of one cycle per year. This model allows us to detect long-term trends in glycogen, in- cluding seasonal effects that can be described by a sinusoidal response, and possible differences between tissues in each group. We modeled Var(p,) by; Vaiip.)- M^ *' C-) where tj) is a scale parameter. Log(|jL) can be viewed as a gener- alization of analysis of covariance that accommodates a gamma- distributed response, one categorical main effect (tissue), a sinu- soidal seasonal effect, and a logarithmic trend that may differ between tissues. We fitted the model given by equations ( I ) and (2) using maximum likelihood estimation. We modeled the mean TCI similarly to the model given above for glycogen, except the parameters for tissue type were removed. RESULTS In the riverine group, glycogen and the TCI varied significantly on a seasonal basis (glycogen P = 0.02, TCI P < 0.0001). and showed a similar pattern in both years with highest values eariy in the season (usually June) and lowest values late in the season (usually October; Fig. I ). Although a significant seasonal pattern was apparent in all indicators, it was most pronounced in glycogen in mantle tissue. For example, glycogen concentrations varied 72% [((annual maximum - annual minimum) / annual maximum) x 100], during 1997 and 46% during 1998 (Fig. I ). In contrast, the seasonal change in TCI varied only 20% during 1997 and 38% during 1998. Although seasonal patterns in glycogen and TCI were similar in both years, the annual maximum of glycogen in foot and mantle tissue was greater in 1997 than in 1998, while TCI maxi- mums were similar between years. In the relocated group, glycogen and the TCI also varied sea- sonally (glycogen P = 0.01. TCI P = 0.002). Like the riverine group, highest values were observed eariy in the season (usually May) with lower values each fall (usually November). Most of the seasonal variation in physiological measures in this group were likely driven by the large decline in glycogen the first 6 months after the relocation (Fig. I ). Glycogen values declined by 72% in mantle tissue and 65% in foot tissue from May to November 1995. During 1996. glycogen concentrations continued to decline, but onlv by 35% in mantle tissue and 49% in foot tissue. Glycogen concentrations varied between foot and mantle tissue in the riverine mussels iP < 0.0001), but not in the relocated mussels (P = 0.09). In the riverine group, glycogen concentrations in mantle tissue ranged from 1.9 to 3.6 times those in foot tissue over both years. In the relocated mussels, glycogen concentrations in foot and mantle tissue were similar during both years (Fig. 1). M-„ = M-n exp T^-i-piog(r)-i- p, log(;) 1-nt + 7, sin^ — j-h72Cos(^ — 277/ DISCUSSION The significant seasonal variation in glycogen was anticipated, (1) given that glycogen is the primary reserve energy store and is re-allocated from the body towards the gonads during gametoge- where p.,, is the baseline (mantle tissue trom mussels m the poiitl) nesis (Gabbott 1983). It is difficult to compare the amount of mean glycogen concentration, t, is the effect of tissue type (foot or annual variation in glycogen concentrations in our riverine group mantle), t is time measured ininonths since the start of the year (range. 32-72%) with other studies because few have measured during which the trial began. P is a trend coefficient, p^ is a trend glycogen with the frequency that we used and several have used coefficient due to tissue / and y, and y, are coefficients in a species with substantially different reproductive traits. However, Physiological Condition oh Amblema 1169 800 600 -a B c (U o (J I I I I I ] I I I I I I I I I I I I 0 May Jun Jul Aug Sep fXi N(i\ F-ch Apr May Jun Jul Aug Sep Oct Nov 1997 1998 soo 600 400 200 -• — Glycogeu in foot tissue -■ — Glycogen in mantle tissue Relocated mussels -T — Tissue condition index ^ ■ Water temperature H en C O o 3 a. o 3 I— ( 3 CL CD X p 3 H 3 4 ^ 3 fti o o '- X May Jun Jul Aug Sep Oct Nov Feb Apr May Jun Jul Aug Sep Oct Nov 1995 1996 Figure 1. Mean (±1 SE) concentrations of glycogen and the tissue condition index in Amblema plicata (solid lines) and mean water temperature (dotted lines). Riverine mussels were sampled from the Upper Mississippi River ahout monthly from May 1997 to October 1998. Relocated mussels were remo\ed from the I pper Mississippi River and placed into an earthen pond (May 1995), then sampled about monthly until November 1996. total body glycogen content varied 39% between July and Octt>ber in Anodonta pisciiuilis. during the time of glochidial development (Jokela et al. 1993). Similariy. the total carbohydrate content of fingernail clams (5. tnmsversiim) varied 809r when sampled monthly in a given year (Dietz & Stem 1997). Thus, it seems that significant seasonal variation exists in glycogen concentrations in a variety of freshwater bivalves, even those with substantially different reproductive strategies. Seasonal variation in glycogen concentration is likely driven by environmental factors such as water temperature, which subse- quently influences reproductive condition. A. plicata is a short- term breeder with maturing gametes and glochidia present from May to early August in the Upper Mississippi River (Holland- Bartels & Kammer 1989). By late May, 100% of the females and 80% of the males contain mature gametes, indicating a progression towards full reproductive maturation as water temperatures rise from 13° to I9°C (Fig. 2). Seasonal patterns in glycogen concen- tration in riverine mussels in this study parallel reproductive ac- tivity. As water temperatures warm in the early spring, male and female mussels begin producing mature gametes and glycogen is either being stored or is re-allocated from other tissues to meet the energetic demands of reproduction. As temperatures begin to sta- bilize in late June, gametes are fertilized, mature glochidia are released, and glycogen concentrations are depleted (Fig. 2). Condition indices are frequently used as a measure of nutritive status in marine bivalves and represent the total energetic status of an individual, including stores of protein, lipids, and carbohydrates (Bayne et al. 1985). Thus, seasonal patterns in conditiini indices should follow food availability and reproductive status. In the riverine mussels in this study, seasonal patterns in the TCI paral- leled reproductive activity. Prior to reproduction (April to June), the condition of mussels increased (they were adding soft tissue 1170 Monroe and Newton 100 Glycogen Temperature Apr May Jul Monlh Sep Figure 2. Mean (±1 SE) concentrations of glycogen in mantle tissue in Ambleiiia plicala from the Upper Mississippi River (luring IWS (solid line), and water temperature at the time of mussel collection (dotted linel. Pie charts estimate the reproductive stage in A. plicala with the black area representing the percentage of male and female mussels on each date containing mature gametes (data from Holland-Bartels & Kammer 1989). weight) as they prepared for reproduction. After fertdization and glochidia release (June to July), energetic stores were depleted and the TCI decreased. This coupling of the condition inde.\ and the reproductive cycle was also seen in the short-term breeders [/. tumidiis and U. pictonim (Pekkarinen 1993). The similarity in the seasonal patterns between glycogen and the TCI in riverine A. pUcata suggests that in natural populations, glycogen represents a large proportion of the total energy stored. In another short-term breeder, Quadnda pusndosa. carbohydrates av- eraged 76'J of the total energy stores (Greseth 1998). Conversely, in relocated mussels, patterns in glycogen and the TCI become uncoupled, suggesting that energetic stores besides glycogen are being used. In natural populations, we hypothesize that measures of glycogen and the TCI provide comparable data on the relative condition of an indisidtial. but when unionids are stressed, such as during relocation, these two measures provide different data, re- flecting varying energy sources during periods of stress. If this can be shown for additional species, resource managers may have another tool for use in studies with threatened and endangered species. The substantial difference in glycogen between tissues in riv- erine mussels, but not in relocated mussels, suggests that energy stores may be re-allocated during periods of stress. These data support our earlier conclusions that glycogen may be preferentially stored in mantle tissue in A. pUcata when environmental condi- tions are favorable (Naimo & Monroe 1999). Jokela (1996) found that energy was allocated differently among i-eproduction, somatic growth, and biochemical storage, depending on the time of the year when A. pisciiudis were transplanted. Thus, because the re- located mussels were moved during an active reproductive period, the glycogen that had accumulated in mantle tissue may have been re-allocated to maintenance metabolism. The seasonal variation in physiological measures in native populations indicates that numerous biological and environmental parameters interact to influence the relative condition of unionids. Researchei's measuring physiological responses of unionids to a particular stressor (zebra mussels, contaminants, relocation, etc.) at a given time need to understand where these measures fall on the natural response curves to better inteipret how a given stressor affects the physiological energetics of the species under consider- ation. Finally, this research was conducted on a single species; studies with other species are critically needed before resource managers can make sound management decisions that will con- serve this declining faunal group. ACKNOWLEDGMENTS We thank staff from the Wisconsin Department of Natural Re- sources. La Crosse. WI. for obtaining mussels for the relocation portion of this study. Steve Gutreuter from the Upper Midwest Environmental Sciences Center provided statistical guidance. LITERATURE CITED Americiin Puhlic Health Association. American Water Works Association. & Water Environment Federation ( APHA). 199.^. Standard methods for the examination of water and wastewater. 19th edition. American Pub- lic Health Association. Washington DC. 993 pp. Baker S. M. & D. J. Honihach. 1997. Acute physiological effects of zebra mussel (Dreissemi palvDiorplui) infestation on two unionid mussels, Acrinonaia.s Ugamenuiui and Amhicma pUcata. Can. J. Fish. Aquat. Sci. .S4:.'il2-.519. Bayne. B. L.. D. A. Brown. K. Burns. D. R. Dixon. A. Ivanovici. D. R. Livingstone. D. M. Lowe, M. N. Moore. A. R. D. Stehbing & J. Widdows. 1985. The effects of stress and pollution on marine animals. New York: Praeger Publishers. .^84 pp. Berg. D. J.. W. R, Haag, S. I. Guttman & J. B. Sickel. 1995. Mantle biopsy: a technique for nondestructive tissue-sampling of freshwater mussels. ./. N. Am. Bcntliol. Sac. 14:577-581. Bogan, A. E. 1993. Freshwater bivalve extinctions (Mollusca: Unionoidal: a search for causes. Amer. Zool. 33:599-609. Byrne. M. & P. A. Vesk. 1996. Microanalysis of elements in granules in Hyriclella depresui (Bivalvia): multivariate analysis and biomonitoring potential. Aiisrnilcis. .1. Ecotoxicoi 2:91-97. Cope W. G. & D. L. Waller. 1995. Evaluation of freshwater mussel relo- cation as a conservation and management strategy. RegidaleJ Rivers: Reseanh and Manage. 1 1:147-155. Dietz. T. H. & E. M. Stern. 1977. Seasonal changes In reproductive activity and biochemical composition of the tlngernail clam. Spluwiiiim trans- versum. Naulilus 91:136-140. Gabbott. P. A. 1993. Developmental and seasonal metabolic activities in marine molluscs. In: P. W. Hochachka, editor. The Mollusca, Vol 2, Environmental Biochemistry and Physiology. New York: Academic Press, pp. 165-217. Greseth, S. L. 1998. Sub-lethal effects of emersion on three species of unionid mussels. Master's Thesis. University of Wisconsin-La Crosse. 31 pp. Haag. W. R.. D. J. Berg. D. W. Garton & J. L. Farris. 1993. Reduced survival and fitness in native bivalves in response to fouling by the introduced zebra mussel (Dreissena pnlymorplia) in Western Lake Erie. Can. J. Fi.di. Aquat. Sci. 50:13-19. Hallac. D. E. & J. E. Marsden. 2000. Differences in tolerance to and recovery from zebra mussel (Dreissena potymorpha) fouling by Ellip- lii) camplanata and Lampsilis radiata. Can. J. Zool. 78:161-166. Holland-Bartels. L. E. & T. W. Kammer. 1989. Seasonal reproductive development of Lampsilis cardiwn. .\mhtema pUcata plicala. and Pota- milas alatiis (Pelecypoda: Unionidae) in the Upper Mississippi River ,/. Freslm. Ecol. 5:87-92. Holopainen. I. J. 1987. Seasonal variation of survival time in anoxic water and the glycogen content of Sphaerimn corneiim and Pisidium amni- cum (Bivalvia. Pisidiidae). Am. Malacol. Bull. 5:41-48. Jokela, J., L. Uotila & J. Taskinen. 1993. Effect of the castrating trematode Ph^ siOLOGicAL Condition of Amulema 1 171 parasite Rhipulocolyle feimicu on energy allocation ot fresh-v^ater clam Anoilonta piscinalis. Fund. Ecol. 7:332-338. Jokela. J. 19%. Within-season reproductive and somatic energy allocation in a freshwater clam. Anoilonhi pisciiuilis. Oecologki 105:167-174. McCullagh. P. &. J. A. Nelder. 1989. Generalized linear models. 2"'' edi- tion. London: Chapman and Hall. 511 pp. Naimo, T. J.. G. J. Atchison & L. E. Holland-Bartels. 1990. Sublethal effects of cadmium on physiological responses in the pocketbook mus- sel. Lampsilis veiilricosa. Environ. Toxicol. Chem. 11:1013-1021. Naimo. T. J.. E. D. Damschen, R. G. Rada & E. M. Monroe. 1998. Nonlethal evaluation of the physiological health of unionid mussels: methods for biopsy and glycogen analysis. ,/. Nortli .\m. Bcnthol. Soc. 17:121-128. Naimo. T. J. & E. M. Monroe. 1999. Variation in i>lvcot;en concentrations withni mantle and foot tissue in .\nililcnia pluaia plimki: Implicalums for tissue biopsy sampling. Am. Makicol. Bull. 15:51-56. Newton. T. J.. E. M. Monroe, S. J. Gutreuter. R. Kenyon. P. A. Thiel & K. I. Welke. 2001. Evaluation of relocation of unionid mussels into artitR-ial ponds. J. North .Am. Benlhol. Soc. 20:468^85. Patterson. M. A.. B. C. Parker & R. J. Neves. 1997. Effects of quarantme times on glycogen levels of native freshwater mussels (Bivalvia: Unionidae) previously infested with zebra mussels. Am. MiiUicol. Bull. 14:75-79. Pekkarinen, M. 1993. Reproduction and condition of unionid mussels in the Vantaa River. South Finland. Arch. Hydrobiol. 127:357-375. Williams. J. D.. M. L. Warren. K. S. Cummings. J. L. Harris & R. J. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Fisheries 1 8:6-22. Journal of Shellfish Resciinh. Veil. 20, No, 3, I 173-1 176, 2001 . EFFECTS OF STOCKING DENSITY ON GROWTH AND SURVIVAL OF ATLANTIC SURFCLAMS IN BOTTOM CAGES VERSUS MESH BAGS RANDAL L. WALKER* Shellfish Aquucultiire Lnhoratorx. University of Gcuriiid Marine Extension Service. 20 Ocean Science Circle. Savannah. Ceorfiia M4I I -llll I ABSTRACT Atlantic suiiVUims, Sivsiila .soliili.wmia. were reared in lield eullure to determine the elTeets of: (a) stocking densities on clams cultured in bottom cages and mesh bags; (b) bottom cage versus mesh bag culture; and (c) bag mesh diameter on growth and survival of clams in coastal Georgia. With the exception of the 100 stocking density, mean clam size decreased with increases in stocking densities in cages, i.e.. 200 clams/m- (x= 56.5 mm) = 300 (x = 57.0 mm) > 100 (x = 52.8 mm) > 400 (x = 49.1 mm) = 500 (X = 51.0 mm) > 600 (x = 46.5 mm). No significant difference (p = 0.1994) in clam survival occurred among any cage treatments; however, mean survival of clams ranged from 52% in cages stocked at 300 clams/m- to ST7r in cages stocked at 500 clams/m-. Significant differences (p = 0.0258) in clam survival occurred among the mesh bag stocking density treatments with survival ranging from 22% in densities of 100 clams/m' to 73.5% at 500 clams/m". In the mesh bags, clam size was significantly greater in the 200 clams/m- (x = 32.9 mm) = 400 clams/m- (x = 32.8 mm) = 500 clams/m" (x = 34.6 mm) treatments than 300 clams/nr (x = 29.4 mm) which were greater than the 100 clams/m" (x = 26.6 mm) bag treatment. Atlantic surfclams had significantly (p = 0.0100) higher survival in 3-mm mesh (62%) versus 6-mm mesh bags (18.5%-) with low survival (0.3%') in 12-mm mesh bags. Clam size was similar (p = 0.0623) in all three mesh size treatments. Atlantic surfclams had greater growth and survival rates with greater production in bottom cages over mesh bags for the field culture of clams in Georgia. A culture strategy using stocking density data for maximi/mg the harvest, hence marketmg window of the Atlantic surfclam, is discussed. KEY WORDS: SpimUi. growth, mortality, stocking density, cage. bag. mesh INTRODUCTION The Atlantic surtclatii. .Spisulci sDlitlissiiini (Dillwyn, 1817). natural fishery was valued at ,S3(I.4 million in the United States in 1999 (O'Bannon 2000). Atlantic surfclams occur from shore to 40 m depth offshore, from the maritime provinces in Canada to .South Carolina (Abbott 1974). However, the fishery is centered primarily between New York and Virginia. Large adult clams greater than 12 cm in shell length (L) are harvested and used in the fried clam strip and chowder markets (Goldberg 1989). The Atlantic surfclam does not naturally occur in Georgia; however, a subspecies Spisiila Miliilisxinui siiuili.s does occur (Walker & Heffeman 1994). The southern surfclam, SpisiiUi so- lidissima similis is now classified as Spi.'iiila raveueh (Conrad. 1831) (American Fisheries Society 1998). Insufficient stocks of Spisula raveneU occur in coastal Georgia to support a fishery (Walker & Heffeman 1994). The Atlantic surfclam cannot survive the summer water temperatures in coastal Georgia, whereas the southern suifclam survives the first summer, but generally dies during the second summer (Walker & Heffeman 1994). Research into the development of an Atlantic surfclam aqua- culture fishery in Georgia has shown promise (Goldberg & Walker 1990: Walker & Heffernan 1990a. Walker & Heffeman 1990b. Walker & Heffernan 1990c). Earlier growth studies examining the feasibility of culturing the northern Atlantic surfclam in Georgia showed that surfclams grow well over winter, but die by summer (Goldberg & Walker 1990. Walker & Heffeman 1990a. Walker & Heffeman 1990b. Walker & Heffeman 1990c). Atlantic surfclams (10 mm in L) when field planted in October after water tempera- tures drop belov\ lethal levels (28 C) in Georgia will grow 50 to 70 mm L by the following spring. Thus, marketable size Atlantic surfclams can be grown in 6 months as compared to 2 to 2.5 years *Corresponding author. E-mail: walkerfe'arches. uga.edu of growth required for culturing the native northern quahog. Mer- cenaria merceiiaria (Linnaeus. 1758) (Walker 1997). The potential for aquaculture for Atlantic surfclams is to pro- duce clams (approximately 50 mm L) for the lucrative fried, raw and steamer markets. The rapid growth rate of the clam in Georgia allows a marketable product in six months from field planting. Extension efforts in developing an aquaculture industry for surf- clams has also shown promise. Taste tests at local seafood festivals in Georgia indicated that the clam has potential for development into an aquaculture industry (unpublished data). Taste testers rated the clam as an excellent product with a slight sweet and mildly nutty flavor; however, clams must be placed in floating cages for a brief period to allow clams to purge sand from inside their shell prior to marketing. Marketing attempts with local seafood dealers showed that they could sell the clam at an equivalent price to that of the northern quahog. The only drawback of the marketing at- tempts was that the Atlantic surfclam had a short shelf life. A preliminary attempt at flash freezing the clams and placing them in shrink packing, resulted in the cracking of the shells. The shells v\ere weak presumably due to the rapid growth rate of clam. Clams were increasing their length and width at the margins of the shell rather than producing a thick layer of shell. Another problem in developing an aquacultural industry for surfclams in Georgia is that clams need to be harvested, sold, and consumed in a very short period (May). One possible means of increasing the marketing window and increasing shelf life, may be to harvest clams earlier in the season prior to the warmer waters adding an increased physiological stress level on the animal. By manipulating field stocking densities of surfclams in culture units, a clam farmer may optimize growth conditions producing a harveslable product ear- lier in the season. This study examines the effects of cage and mesh-bag stocking densities on the growth and survival of cultured surfclams in the coastal waters of Georgia in an effort to optimizing the field har- vesting window of opportunity. In addition, the study determines 1173 1174 Walker the effects of bag mesh size on growth and survival of the Atlantic surfclam. MATERIALS AND METHODS Cage Stocking Density Study To determine the effects of cage stocking density on growth and survival of Atlantic surfclams, cages were stocked with At- lantic suifclanis of 13.3 ±0.17 (S.E.) mm L at treatment dL-nsities of 100. 200. .^00. 400. 300. and 600 clams per cage. Each treatment had three replicate cages. Eighteen 6-mm mesh vinyl-coated-wire cages ( 1 X 1 X 0.6 m) were partially buried (0.3 m) in a sandy-mud substrate in a straight line at the spring-low-water mark on an intertidal flat at the mouth of House Creek. Little Tybee Island. Georgia. Two l-m poles (to anchor cages) were driven into the sediment at opposite corners of each cage and attached to the cage with cable ties. Cages were randomly stocked according to a ran- dom number table in October 5, 1995. Cages were harvested 222 days later in May 14. 1996 with all clams per cage counted and 50 randomly selected clams measured for shell length (anterior- posterior) to the nearest 0.5 mm with Vernier calipers.. To determine the effects of stocking density on growth utilizing the mesh-bag-linc culture system designed for the culture of the northern quahog (Walker & Hurley 1995), surfclams were stocked at various densities in 6-mm mesh bags. Five mesh bags (0.5 m" each) were attached to a long line according to Walker and Hurley (1993). Atlantic surfclams at 15.5 ± 0.17 mm were stocked at densities of 50. 100, 150, 200, and 250 clams per bag with three replicate bags per treatment. These stocking densities are equiva- lent to densities of 100. 200, 300, 400, and 500 clams/nr, respec- tively. Results are reported in terms of clams/m" to compare with the same density as in the cage experiment. Bags were randomly assigned to long lines and were deployed in October 5, 1995. Bag lines were secured with stakes adjacent to the cages. Clams were harvested in May 14, 1996 with all clams per bag counted and 50. if available, randomly selected clams measured for shell length (anterior-posterior) to the nearest 0.5 mm with Vernier calipers. Mesh Size Study To determine the effects of bag mesh diameter on growth and survival of Atlantic surfclams, 3-mm (N = 4). 6-mm (N = 3). and 12-mm (N = 3) mesh bags were stocked with 200 Atlantic sur- fclams at a mean size of 15.5 ± 0.17 mm. Five bags each were randomly attached to two long lines (Walker & Hurley 1995) and staked at 0.5 m below the spring-low-water mark on an intertidal flat at the mouth of House Creek, Little Tybee Island, Georgia. Bags were harvested in May 14. 1996 with all clams per bag counted and 50 indi\iduals measured for shell length. Statistical Analysis One-way Analysis of Variance (ANOVA) (a = 0.05) and Tukey's Studentized Range Test (SRT) (a = 0.05) using SAS for PC (SAS Institute Inc. 1989) were carried out in each study. All proportion survival data was arcsine square root transformed prior to statistical analysis. RESULTS Cage and Mesh Bag Stacking Density Study Greater Atlantic surfclam survival (p = 0.0040) occmred in cages (77%) than in mesh bags (53%). No significant diftcrcnce (p = 0.1994) in clam survival occurred among any cage treatments (Table I); however, mean survival of clams ranged from 529c in cages stocked at 300 clams/m~ to 87% in cages stocked at 500 clams/m~. Significant differences (p = 0.0258) in clam survival occurred among the mesh bag treatments with survival ranging from 22%- in densities of 100 to 73.3%r at 500 clams/ni". Although survival of clams in bags increased with increases in stocking densities (Table 1), clam survival at the 100 clams/m"^ stocking density was only significantly lower than that at 300 clams/nr. Significant differences in clam size occun-ed among stocking densities in both cages (p < 0.0001) and mesh bags (p < 0.0001; Table 1). In general and with the exception of 100 stocking den- sity, mean clam size decreased with increases in stocking densities in cages, i.e.. 200 clams/ni" (x = 56.5 mm) = 300 clams/m" (x = 57.0 mm) > 100 clams/m" (x = 32.8 mm) > 400 clams/m- (x = 49.1 mm) = 500 clams/m- (x = 51.0 mm) > 600 clams/m" (x = 46.5 mm). In the mesh bags, clam size was significantly lower in the 100 clams/m" bags (x = 26.6 mm) which was greater than 300 (X = 29.4 mm) > 200 clams/nr (x = 32.9 mm) = 400 clams/nr (X = 32.8 mm) = 500 clams/m" (x = 34.6 mm). Overall in the mesh bags, clam size increased with increases in stocking density. Surfclams achieved a greater size in cages versus mesh bags re- gardless of stocking density. In cages, mean Atlantic surfclam size ranged from 46.5 mm at 600 clams/m" to 37 mm at 300 clams/m". whereas, in mesh bags mean clam size ranged from 26.6 mm in 100 clams/nr to 34.6 mm in 5(X) clams/m" treatments (Table 1). Mesh Size Study Growth and survival data of the Atlantic surfclam cultured in various mesh bag diameters are given in Table 2. ANOVA showed that no significant difference (p = 0.0623) in clam size occurred among treatments: however, Tukey's SRT showed that clams from the 12-mm mesh bags were smaller (26.1 mm) than those from the 3-mm (33.1 mm) and 6-mm (32.2 mm) mesh bags. The results of the 12-mm mesh bags are moot, as only two animals survived (0.3%-). ANOVA of survival data showed significantly greater sur- vival occurred for clams stocked in the 3-mm mesh bags (62%) versus the 6-mm mesh bags (19%). DISCUSSION Surfclams demonstrated greater growth and survival when cul- tured in bottom cages as compared to the mesh-bag culture tech- nique. Surfclams grown in cages achieved approximately twice the size of those grown in mesh bags at equivalent stocking densities (Table 1 ). The lowest size of clams in cages (46.5 mm in 600 clams/nr treatments) were 34% larger than the greatest size of clams in the mesh bag (34.6 mm in 500 clams/m") treatment. Atlantic surfclams had 1.5 times higher survival rate in cages than in mesh bags (Table 1 ). Surfclams grow slower In mesh bags than in cages. The reason for this is unknown, but restricted water flow within the bags and increased exposure to turbulence are two possible explanations. Because the bags are lying on the sediment, water flow does not pass through the bags as it does for a cage. Instead water flows over the top of the bags. The cage sides cause a battling effect thereby allowing food particles to fall out of the water column and consequently being more readily available to the clams. The physi- cal pressure of the bag lying on the clam may have prevented efficient siphoning by the clam and hence, resulted in reduced feeding capabilities. In addition, cuiTents appear to effect the bags more than the cages in this study. Although a few cages exhibited SuRFCLAM Culture in Coastal Ghorgia 1175 TABLE 1. Mean clam size in shell length and sur\i\al ol Atlantic suifclanis cultured al various stockin;; densities I clams per m"l in bans and cages on an Intertidal Hat at House Creek, Little Tybee Island, (leorgia. AN()\ A and Tukey's SRT results are given for overall means where similar letters indicate means that are not signitlcantlj different. Cages Bags Mean ± SE Mean ± SE Density Number .Survival {%) Size in mm Number Survival (% ) Size in mm Clams/m- Surviving (p = 0.1994) (p< 0.0001) Surviving (p = 0.0258) {p < 0.0001) 100 76 76.0 54.8 ± 0.89 16 32.1) 23.9 ± 0.96 100 57 57.0 51.5 + 0.77 17 33.0 29.2 ±0.87 100 53 53.0 52.2 ±0.81 0 0 overall 186 62.0a 52.8 ± 0.49a 33 22.011 26.6 ± 0.80a 200 182 91.0 52.7 + 0.63 0 0 200 154 77.0 59.1 ±0.59 56 56.U 30.5 ± 0.53 200 156 78.0 57.6 ±0.78 55 55.0 35.2 ± 0.68 overall 492 82.0a 56.5 ± 0.45b 111 37.0ab 32.9 ± 0.49b 300 251 83.7 59.0 ± 0.58 85 .S6.7 27.1 ±0.49 300 218 72.7 .S4.9 ± 0.73 94 62.7 31.8 ±0.52 300 0 0 0 0 overall 469 52.1a 57.0 ±0.5 lb 179 39.8ab 29.4 ± 0.43c 400 356 89.0 49.7 ± 0.58 154 77.0 33.2 ± 0.63 400 305 76.3 53.9 ± 0.78 159 79.5 31.9 ±0.57 400 339 84.8 43.9 ± 0.79 126 63.0 33.5 ±0.63 overall 1000 83.3a 49. 1 ± 0.53c 439 73.2ab 32.8 ± 0.41b 500 454 90.8 50.3 ± 0.87 210 84.0 32.1 ±0.49 500 454 90.8 52.4 ±0.78 169 67.6 35.1 ±0.55 500 396 79.2 53.5 ± 0.79 172 68.8 36.7 ±0.58 overall 1304 86.9a 51.0 ± 0.48c 551 73.5b 34.6 ± 0.35b 600 467 77.8 46.3 ± 0.60 600 468 78.0 46.7 ± 0.64 600 530 88.3 46.8 ± 0.54 overall 1465 8 1. 4a 46.5 ± 0.34d some wash-out of sediment, most mesh bag-lines were I'oimd twisted and in some cases completely relocated. Several times the lines had to be untangled and relaid. The physical movement of clams within bags by currents combined with the effects of in- creased densities resulting in a concentration of the clams into one end of the bag could retard growth. The effects of current on the bags may also explain why the higher bag stocking densities pro- duced better clam growth and survival than that occurring in the lower density bags. The weight of the additional numbers of clams may provided a more stable environment within the baa. In a similar study, northern quahogs. Mciceinnia ineneimria. grew better in cages and in gravel filled trays than in the mesh- bag-line system (Walker 1997). Although growth was lower for northern quahogs cultured in mesh bags as compared to that for clams in cages or trays, survival of animals was found significantly higher in the mesh-bag treatment (Walker 1997). This was not the case for surfclams in this study. Atlantic surfclams in bags had 53% survival compared to 77% for those in cages. In most cases, greater growth and survival of surfclams occurred in bags with higher stocking densities. The increased bioinass associated with higher densities may have provided a somewhat more stable en- vironment consequently, the bags were not disturbed as much as less densely filled bags. During the northern quahog experiment at this same site (Walker 1997). bags were found filled with sediment at the end of that experiment; whereas, in this studv. most bags were devoid of sediment. The effects of stocking density on the growth of various bivalve species is well documented, whereby growth tends to be reduced TABLE 2. Mean clam size in shell length and survival of Atlantic surfclams cultured in 3-mm, 6-mm, and 12-mm mesh bags at the spring-lovv-vvater mark on an intertidal Hat at House Creek, Little Tybee Island, Georgia. Mesh Size Number Survival {''/r 1 Mean ± SE Size (mm) Surviving ip = 0.0100) (mm) (p = 0.06231 3 mm Bag 1 146 73.0 32.1 ±0.51 Bag 2 125 62.5 32.6 ± 0.42 Bag 3 138 69.0 35.2 ± 0.54 Bag 4 85 42.5 32.5 ± 0.55 overall 494 61.8a 33.1 ± 0.27a 6 mm Bag 1 56 28.0 30.5 ± 0.53 Bag 2 S? 27.5 35.2 ± 0.68 Bag 3 0 0 overall 111 is..';b 32.2 ± 0.49a 1 2 mm Bag 1 2 1.0 26.1 ±0.14 Bag 2 0 0 Bag 3 0 0 overall T 0.3b 26.1 ± 0.14b ANOVA (mean shell length; p = 0.0623 and survival; p = 0.0100) and Tukey's SRT results are given for overall means where similar letters indicate means that are not significantly different. 1176 Walker with increases in stocking density (Adams et al. 1994. Eldridge et al. 1979, Walker 1984, Goldberg 1989. Walker & Hurley 199.^). Goldberg ( 1 989) planted 1 3.7 mm Atlantic surfclams at 500. 1000, and 2000 clam,s/m" in bottom cages and after approximately t'oin' months, clams had grown to a mean size of 47.3 mm. 40.8 mm. and 32.0 mm L, respectively. In our study, 15.5 mm Atlantic surfclams planted in bottom cages at densities from 100 to 600 clams/m" showed significant decreases in growth between 200- 300 and 400-500 clams/m" with slowest growth occurring at 600 clams/m- (Table 1 1. Interestingly, the lowest stocking density in bottom cages in our study (100 clams/m") yielded slower growth rates than the 200 and 300 clams/m" stocking densities, but had higher growth rates than did the 400 to 600 clams/nr stocking densities. Furthermore, the lowest stocking density of 100 clams/ m" regardless of growout protocol yielded lower growth rates across both culture treatments compared to the 200 and 300 stock- ing densities (Table 1 ). The optimum stocking density for culturing .Atlantic surfclams in Georgia to achieve ma.ximum size is 300 clams/m''; however, as discussed below, optimum stocking density for obtaining marketable size animals efficiently is 500 clams/m-. Aquaciilture Recommendalion Aquaculture of .Atlantic surfclams has excellent potential in coastal Georgia. Atlantic surfclams are easily spawned and reared under hatchery conditions. Seed can be grown in the hatchery to a field-planting size by mid-October, once ambient water tempera- tures decrease below 28°C. Ambient river temperatures in Georgia which may reach 31°C are lethal to the Atlantic surfclam. so seed must be reared in the hatchery at lower water temperatures throughout the summer months (Walker et al.. 1997). Clams must be harvested by May to early June before ambient water tempera- tures reach 28"C. Atlantic surfclams are marketable within six months from planting. The results of this work show that cages can be seeded at different densities to produce an optimum sized clam (50 mm) for the lucrative steam or fried clam markets via a stag- gered crop method. Cages seeded at densities of 200 to 300 clams/ m- can reach the targeted market size of 50 mm by March to April v\'ith cages seeded at 500 clams/m" achieving the same market size in May to June. Thus, by seeding at the optimum density for producing a maximum sized clam, the culturist can produce mar- ketable clams early in the season. By seeding at 500 clams/m- for maximum production, clams will attain market size toward the end of the growing season (May to early June). ACKNOWLEDGMENTS The author thanks Capt. J. Whitted of the R/V SEA DAWG for his efforts in transporting cages and animals from the field. Dr. F. O'Biern. Mr. D. Hurley and Ms. D. Moroney are thanked for their help in deploying cages, bags and animals and in terminating the experiment. LITERATURE CITED Abbou, R. T. 1974. American seasliells. .second edition. New York: Van Nostrand Reinhold Co. Adams, M. P., R. L. Walker & P. B. Heffernan. 1994. Effects of stocking density, bag mesh size, and bottom sediment on the growth and sur- vival of the eastern oyster. Cnissostrea virginica. with emphasis on controlling oyster spat fouling. J. Appl. Aquaculture 4:25—14. American Fisheries Society. 1998. Common and Scientific Names of Aquatic Invertebrates from the United States and Canada, Bethesda: American Fisheries Society Special Publication 26. Eldridge, P. J., A. B. Eversole & J. M. Whetstone. 1979. Comparative survival and growth rates of hard clams, Mercenaria mercenaria. planted in trays subtidally and intertidally at varying densities in a South Carolina estuary. Proc. Natl. Shellfish. Assoc. 69:30-39. Goldberg. R. 1989. Biology and culture of surf clams. In: J. J. Manzi & M. A. Castagna. editors Clam mariculture in North America. New York: Elsevier, pp. 263-276 Goldberg. R., & R. L. Walker. 1990. Cage culture of yearling surf clam. Spisulii vo/u/mvi/ik/. in coastal Georgia. J. Shellfish Res. 9:187- 194. 0"Bannon. B. K. 2000. Current fisheries statistics No. 9900. Fisheries of the United States. 1999. Washington. D.C.: Department of Commerce. National Marine Fisheries Service. SAS Institute Inc. 1989. SAS User's Guide: Statistics, version 6. Gary. North Carolina: SAS Institute Inc. Walker. R. L. 1997. Comparison of field grow out methods for the northern quahog in the coastal waters of Georgia. University of Georgia. Marine Extension Bulletin 17. Athens, Walker. R. L. 1984. Effects of density and sampling tmie on the growth of the hard clam, Mercenaria mercenaria. planted in predator-free cages in coastal Georgia. The Nautilus 98:1 14—1 19. Walker, R. L. & P. B. Heffernan. 1990a. The effects of cage mesh size and tidal level cage placement on the growth and survival of clams. .Mer- cenaria mercenaria (L.) and Spisula solidissiiiia (Dillwyn). in the coastal waters of Georgia. NE. Gulf Sci. 11:29-38. Walker. R. L. & P. B. Heffernan. 1990b. Plastic mesh covers for field growing of clams. Mercenaria mercenaria (L.). Mya arenaria (L.). and Spisula solidissima (Dillwyn). in the coastal waters of Georgia. Ga. J. Sci. 48:88-95. Walker. R. L. & P. B. Heffernan. 1990c. Intertidal growth and survival of northern quahogs, Mercenaria mercenaria (Linnaeus, 1758), and At- lantic surf clams, Spisula solidissima (Dillwyn. 1817). in Georgia. J. Worid .Aquaculture Soc. 21:307-313. Walker, R. L. & P. B. Heffernan. 1994. Age. growth rate and size of the southern surf clam. Spisula solidissima siinilis (Say. 1822). (Bivalvia: Mactridae). J. Shellfish Re.s. 13:433-+4I. Walker. R. L. & D. H. Hurley. 1995. Biological feasibility of mesh bag culture of the northern quahog Mercenaria mercenaria (L.) in soft- bottom sediments in coastal waters of Georgia. University of Georgia. Marine Extension Bulletin 16. Athens. Walker. R. L.. D. H. Hurley. F. X. O'Beirn & D. A. Moroney. 1997. Culture of juvenile Atlantic surfclams. Spisula solidissima solidissima and Spisula solidissima similis, in forced-flow upwellers in a bivalve hatchery in coastal Georgia. J. World Aquaculture Soc. 28: 28-33. JiHinuil ol Shclllhh Research. Veil. 20, No, 3, 1 177-1 I S6. :()()l. COMPARISON OF RECRUITMENT FREQUENCY AND GROWTH OF SURFCLAMS, SPISULA SOUDISSIMA (DILLVVYN, 1817), IN DIFFERENT INNER-SHELF HABITATS OF NEW JERSEY MARNITA M. CHINTALA' " AND JUDITH P. GRASSLE' Instiuitc (if Marine and Coastal Sciences. Ri(ii;ers, The State University of New Jersey. New Brunswick. New Jersey: 'U.S. Enviroiiineiital Protection Afiency. NHEERL. Atlantic Ecoloi^y Division. Nanv.i>ansett. Rhode Island ABSTRACT S/n.snl the slallons where surl'clanis were sampled in 1993. Lines paralleling the coast delineate the 1-mile zones for NJ DEP sampling. Numbers indicate the zone delineations. Cities and towns indicate the origins of U.S. EPA weekly transects measuring surface and bottom temperature and oxygen concentration in May to September, 1979-1991. Clams were collected by a commercial hydraulic clam dredge equipped with a 72-iiich knife. The dredge tloor was lined with Springfield fencing with 3 1 -mm square mesh. Ring diameter in the metal bag at the cod end of the dredge was 72 mm (Giust 1993). The smallest clam that was caught with this dredge and measured was 33 mm in length. We followed the same assumptions as Wein- berg ( 1999) in that clams smaller than this were completely missed during sampling and that other clams that could pass the ring size of the dredge were not sampled quantitatively. One tow was taken per station, and was assumed to randomly sample surfclams fully retained by the dredge in that area. Clams were haphazardly se- lected from a one-bushel subsample at each station (sample size ranged from 6-69). These surfclams were shucked on board and the shells were returned to the laboratory for age and growth determination. The data from the different stations were analyzed by zones as described below . Recruitment and Crnwth Curve Determinations The shells were cleaned and height was measured. Because the clams were large, they were not embedded prior to cutting and counting the growth lines (Jones, pers. comm.). For consistency, w henever possible, only the right valve was cut for age and growth determinations. On some occasions it was necessary to use the left valve when the right valve broke during sectioning. A single cut was made with a diamond saw from the umbo to the ventral margin along the axis of maximum height (Jones et al. 1978). The shell edge was polished on a lapidary wheel for 2 to 3 min until the edge was smooth. The dark internal growth lines were counted and used to back-calculate the year of settlement for each of the clams since Jones et al. (1978) have shown that the growth lines are annual. The height at each age was measured directly from the dark. internal growth lines, which were traced from the chondrophore to the outer shell surface. The shell was examined under a dissecting microscope at 60x or 12()x (objective and eyepiece magnification) and the exit point of the growth line was marked with a pencil. The straight-line distance from the umbo to the point of exit of each growth line at the shell's outer surface (Jones et al. 1978) was measured to the nearest 0.1 tnm with calipers. In this way. a series o'i height-at-age measurements were obtained from each indi- vidual. Analysis of Data Data were analyzed as in Cerrato and Keith (1992). A von Berlalanffy curve was fit to the shell height vs. age data by non- linear least-squares analysis using SAS for all clams measured in each of the zones (Cerrato 1990). The algebraic form of the von Bertalanffy curve used in these analyses was h: H,)(l -.v'-'). (I) where h is the shell height at age t. H, is the expected mean height at age 1. H,,,.,^ is the asymptotic maximum size, and .v is the expected fractional reduction in annual increment growth (Cerrato & Keith 1992). Parameter estimates were determined for H,. H„,^,, and X. The growth parameter, k, which is a measure of how fast a population reaches its maximum size, was calculated as follows: .v = e-^ (2) The 1988 year class was the predominant year class in most of the samples. Therefore, we ran a separate analysis on the repre- sentatives of that year class. Because individuals in this year class had not yet approached their asymptotic growth stage, the alge- braic form of the von Bertalanffy equation used for this analysis was (Cerrato & Keith 1992) h = H, -KH, -H|K1 -.v'"')/(l -V^». (3) where H^ is the height at age 5 and the other parameters are the same as in equation ( 1 ). Parameter estimates for H,, H,, and .v were determined and analyzed. Analyses of the parameter estimates from the nonlinear regres- sion were done with likelihood ratio tests, and 95'7f confidence intervals were generated by projecting the likelihood regions onto the appropriate axes (Cerrato 1990. Cerrato & Keith 1992). Data were analyzed by zone, with each zone containing two to five stations. The experiment-wise eiTor rate was 0.0.3. and the per comparison rate was 0.05/p. with p = 36. the number of compari- sons. Temperature and Dissolved Oxygen Data Temperature and dissolved oxygen data, taken from 1 and 3 miles off the New Jersey coast for years 1979 to 1991, were Slirfclam Recruitment and Growth in Coastal NJ 1179 obtained tioni iIil- U.S. EP.A ( 1944). SluIjcc ( I in below surface) and bottom ( 1 m abo\e boiloni) data were eolleeted as part of the U.S. EPA helicopter water i|uality sample collection for the New York Bight area. Sampling occurred weekly during the summer each year from appro.ximately mid-May to late Septem- ber. The weekly values were averaged to get a mean surface and bottom temperature for each station and station values within a zone were averaged. The sampling entailed nine transects e.xtend- ing east from the NJ coast, with five stations in each transect. The only transects used were these from Bayhead (corresponding to zones 1 and 3), Seaside (zones 1 and 3), Barnegat (zones 4 and 6l. Beach Haven (zones 4 and 6). Atlantic City (between zones 4 and 7, and 6 and 9:. and Strathmere (zones 7 and 9) (Fig. 1; Chintala 1997). Most of the other zones (4-8). were dominated by the 1988 year class (Fig. 2, D-H ). This year class accounted for 45% to 95% of the clams within these zones. Although the 1988 year class was the most prevalent in zone 9 {i37( ). there were also a large number of 1976 and 1977 year class clams in these samples (29.6% and 18.5%. respectively; Fig. 21). Recruitment across mile zones was similar since the same age-class pattern was seen in zones 1 through 3. zones 4 through 6. and zones 7 through 9 within each mile; however, towards the south and farther offshore, the number of clams in the post-anoxia year classes (1976-1978) became more prominent (Fig. 21). The northern stations appeared to be repre- sented by a wider range of year classes than farther south (Fig. 2). South of Barnegat Bay. the samples appeared to be overwhclm- inulv dominated bv the 1988 vear class. RESULTS Growth Ciir\e\ of S. solidissinia Recruitment History of S. solidissinia The 1988 year class was predominant in the samples (53.5% of all clams sampled), followed by the 1986 year class (8.3%), and the 1983 year class (5.0%). Zones I, 2, and 3 had surfclams from many different year classes (Fig. 2, A-C). Zone 1, the zone located furthest north and within 1 mile of the shore, was the only zone where the 1983 year class was the most abundant (Fig. 2A). The 1986, 1987, and 1989 year classes were also well represented (>10%) in this sample. Zone 2, between miles 1 and 2, was domi- nated by five year classes, 1982, 1983, 1986, 1987, and 1988 (Fig. 2B). The northern zone farthest offshore, zone 3, was more clearly dominated by the 1986 year class (Fig. 2C), but several other year classes were represented in the sample. The likelihood ratio tests indicated that the von Bertalanffy growth curves were the same for zones 1 and 4 and for zones 8 and 9. For all the other pairwise zone combinations, the growth curves were significantly different from one another (Table I). The growth in the first 3 to 4 y of life was similar for the nine zones (Fig. 3). The growth curves up to age 5 for zones 5 through 8 were lower and more variable than those from other zones, probably due to the dominance of the 1988 year class (Fig. 3). The expected mean height at age 1 (H, ) was the same for zones 1 and 4 through 9 (Fig. 4A). Clams in zones 2 and 3 were signifi- cantly smaller at age I than those in the other zones. The expected maximum height (H^,,^) varied among zones and had smaller con- fidence intervals than H, (Fig. 4B). Again, H,„,^ in zones 2 and 3 were not significantly different from one another, but were greater Z0NE1 N = 82 --■-- bI-.H!! 72 73 74 75 76 77 76 79 80 ei 82 83 54 65 65 87 t YEAR 1 30 Z 60 UJ 40 d. 2C B ZONE 2 N = 79 .... II .III 7: 73 74 75 75 77 76 73 80 61 82 63 &4 65 86 87 88 89 90 YEAR ZONE 3 N = 73 jLj 72 73 74 75 76 77 78 79 60 81 82 93 B4 85 66 87 BB 89 90 YEAR ZONE 4 N = 125 72 73 74 75 76 77 78 79 80 8t 82 83 S4 85 86 87 68 89 90 YEAR 100 E ZONE 5 ec W= 156 o □: HI 40 a. 20 0 - . ■ > - - ■ , 72 73 74 75 76 77 78 79 80 81 &: 83 84 8S 8 6 87 88 89 90 YEAR t- z UJ UJ 0. 100 80 60 40 20 0 F ZONE 6 N= 143 _ _ 1. 72 73 74 75 76 77 78 79 eO 81 82 83 84 as 86 87 88 89 M YEAR PERCENT . y g s s s H ZONE 8 N= 167 72 73 7 A 75 76 77 78 79 80 81 82 63 84 85 86 67 66 59 90 YEAR 100 1 ZONE 9 60 N = 81 2 fO UJ U Q. 1 :-o .h . 1 72 73 74 75 76 77 78 79 SO 81 82 S3 84 85 66 97 88 89 90 YEAR Figure 2. .\-I. Percentage of surt'clam sample in each .vear class for nine inshore /.ones off New Jersey. N indicates sample size for thai zone. 1180 Chintala and Grassle TABLE I. Likelihood ratio tests comparing the von Bertalanffy growth parameters among zones (1-9) off the Neu Jersey coast. The test statistic is -2log(A) where A is the likehhood ratio. Significant outcomes are underhned. The maximum hkelihood estimates under the null hypotheses were calculated hy iteratively reweighted least squares due to heterosccdasticity in the error variances between stations. The experiment-wise error rate for each hypothesis was 0.05. and the per comparison error rate was a = 0.05/p. where p is the total number of pairwise comparisons (based on the Bonferroni inequality). A paired comparison was significant if-2log( A) > »log| 1 + (q/i')F",| ,,|. where q is the number of parameters being compared, n is the total sample size, v = n - 6, and F",,.,. is the upper a percentile point of the F distribution with q and i' degrees of freedom. (Cerrato 1990. Cerrato & Keith 1992). Stn I 2 3 4 5 6 7 8 9 H„: H|, H„,„^, .V equal, all year classes 1 2 311.18 _ 3 182.11 19,56 - 4 4.40 .341.16 206.71 - 5 54.9(1 387.29 246.74 45.10 - 6 289.7U 130.98 1 30.60 303,88 271,39 - 7 229.49 581.80 425.90 187.93 92,68 396.81 - 8 75.63 170.78 104.51 110.27 273.30 322.66 503.23 9 93.66 209.03 118.58 132.22 215.06 323.21 463.11 H„: H, equal, all year classes 2 14.15 _ 3 17.99 0.53 - 4 0.296 12.18 16.00 - 5 2.49 23.84 27.43 5.08 - 6 2.03 2 1 .35 25.13 3.94 0.003 - 7 5.64 29.12 32.41 9.63 0.72 0.45 - 8 0.92 21.04 24.69 2.79 0.70 0.56 3.19 0 4.79 25.35 29.10 7.33 1.03 0.77 0.17 H„: H,,,,^ equal, all year classes t 172.95 _ 3 92.43 8.79 - 4 1.48 198.38 110.83 - 5 7.83 117.76 53.45 15.69 - 6 283.62 71.40 107.43 301.10 241.02 - 7 2.73 158.40 92.88 0.44 15.04 282.23 - 8 0.15 113.00 58.75 1.68 3.06 240.74 2.83 9 11.57 142.84 64.38 22.70 0.008 255.36 19.03 H„: k equal, all year classes 2 8.88 _ 3 2.47 1.27 - 4 0.08 10,71 - 5 31.36 2.26 7.51 - 6 212.71 111.17 124.86 225.73 122.19 - 7 37.62 5.04 11.50 43.20 1.08 95.96 - 8 8.52 28.11 15.05 7.41 67.90 264.29 73.27 9 1.55 15.44 6.41 1.00 43.64 228.69 49.82 H„: 1 H,, H5, X equal. 1988 year class 1 31.36 _ 3 21.08 2.88 - 4 0.17 49.91 21.97 - 5 13.36 80.04 29.89 83.42 - 6 24.81 97.80 34.14 161.01 35.18 _ 7 16.13 83.08 30.84 104.23 8.98 29.12 - 8 30.23 33.53 18.49 235,18 446.30 599.90 465.58 9 39.51 7.89 8.73 124,96 175.68 212.12 179.30 6.59 2.78 3.94 2.71 39.82 continued on next page SuRFCLAM Recruitment and Growth in Coastal NJ 1181 Stn TABLE 1. continued H„: H, equal. 1988 year class 2 0. 1 1 8 - 3 0.07? 0.003 - 4 0.003 0.103 0.048 - 5 1.480 2.448 0.786 9.3 1 2 - 6 0.303 0.90(1 0.319 2.275 2.998 - 7 0.232 0.704 0.329 1 .304 5.851 0.416 - 8 ().l,?l 0.557 0.264 0.922 8.621 0.896 0.300 9 0.004 0.119 0.066 -0.006 3.823 0.880 0.523 i»: H; equal, 1988 year class 2 24.63 - 3 19.73 2.77 - 4 0.022 39.61 20.48 _ 5 1 1 .69 75.78 29.30 68. 1 5 - 6 24.4.S 95.04 33.79 155.86 30.66 - 7 15.09 81,25 30.62 93.87 2.40 18.79 - 8 3.61 26.67 16.52 17.10 225,41 389.72 289.83 9 21.68 1.48 5.92 62.37 147.70 192.40 161,16 0.339 39.215 than all but zone 6. Zones 1.4. 7. and 8 were not significantly different. Likewise, zones 5. 8. and 9 were not significantly dif- ferent from each other, and zones 1 and 5 were the same (Table 1 ). The largest H„,^^ was in zone 6. one of the offshore zones. The clams with the smallest H„,^^ were found in /ones 1. 4. and 7 within I mile of the New Jersey coast (Fig. 4B). The growth parameter k ranged from 0.164 to (J. 396 (Fig. 5). Zone 6. which had the highest H^^^^, also had the lowest k. w hile zone 8 had the highest k. The statistical analysis showed that the k \alue for zones 1. 3. 4. and 9 were not significantly different from one another, nor were zones 2. 3. and 5 different from each other. Growth Curves of 1988 Year Class ofS. solidLssima Testing if all the von Bertalanffy parameters were equal for all zones showed the following homogeneous groups: zones 1 and 4; zones 2, 3. and 9; zones 1 and 5; and zones 5 and 7 (Fig. 6; Table 1 1. The analysis of H, showed that there were no significant dif- ferences among the zones (Fig. 7A; Table I ). The largest clams at age one were in zones 5 and 6. The greatest variability in H, was 10 15 AGE (YEARS) Ficure 3. Mean hel};hl (mlllimelers) at each ajje fur all clams collected at each of the stations in the nine zones off the coast of Ne« .lersey. in zones 1 through 3. zones that also had the lowest number of 1988 year class clams (Fig. 7A). The estimate of the height at age 5 showed that clams from zones 2. 3. and 9 were similar and were also the largest (Fig. 7B). The smallest clams were in zones 5, 6. and 7. with only zones 5 and 7 not significantly different from each other (Table I ). Zones 1 and 4 were the same as were zones 1 and 8 (Fig. 78). Temperature and Dissolved Oxygen Data Only the summer means (from mid-May to late September) for the surface and bottom temperatures and dissolved oxygen mea- surements for the stations that corresponded to clam sampling stations were examined. The data for stations in each of the inshore zones (1, 4. and 7) and the offshore zones (3, 6, and 9) were averaged. In zones 1 and 3, mean summer surface and bottom temperatures tracked each other closely from 1979 to 1991. with surface temperatures being 4°C to 5°C above bottom temperatures in every year (Chintala 1997). hi zones 1. 3. 4. and 6 there were lower average temperatures in 1988 than in all other years (Fig. 8). At the southern border of zones 4 and 6. and in zones 7 and 9. temperatures were low in 1988. but not uniquely so. during this 13-year period (Fig. 8). Dissolved oxygen data showed very little variation at any of the sites, varying by less than 3 mg/L through- out the 13 y. and there was no evidence of widespread, long-lasting hypoxia (see Chintala 1997 for complete temperature and dis- solved oxygen data). DISCUSSION S. solidissima commonly experiences episodic recruitment (Loesch & Ropes 1977) and age-class dominance in dense popu- lations (Murawski & Serchuk 1989). Since the anoxic event off the New Jersey coast in 1976 that killed 62% of the New Jersey surfclam resource in one area (Carlo 1982). it had been thought that the fishery was supported predominantly by the 1976 and 1182 Chintala and Grassle 100 123456789 ZONE I O 160 120 80 1 40 - ♦ ♦ ♦ 123456789 ZONE Figure 4. A. Exptctcd luinlit at age 1 lor all ilams sampled in eaeli of the nine zones as determined bv von Bertalanffv analysis. Values were determined by nonlinear regression. Bars indicate 95'7f confidence intervals for each estimate. B. Expected maxinunn height for all clams sampled in each of the nine zones determined by von BertalanlT\ analysis. \ alues were determined by n(mlinear regression. Bars indi- cate 95% confidence intervals for each estimate. 1977 year classes (Murawski & Sercliiik 19S9). These year classes survived, most likely because of the anoxia-related mortality of the main predators of juvenile S. soliclissinui. including the crab. Oni- lipes oci'llcitiis, and the starfish. Aslerias forhesi (Garlo 19X2. MacKenzie et al. 1985). Cenato and Keith (1992) found that in 1988. the surfclam populations in estuarine waters off Long Island Sound, which were not affected by the hypoxia further south, were instead dominated by the 1983 and 1984 year classes. Previously, it had been thought that there was little or no re- cruitment to the fishery inshore or offshore subsequent to the 1976 05 04 0 3 02 0 1 *-i- i 123456789 ZONE Figure 5. The time each clam population takes to reach maximum size (k) for all clams collected in each zone. Values were determined by Eq. 2 in the text. Bars indicate the 95% confidence intervals for each estimate. X o UJ I AGE Figure 6. Mean height (millimeters) at each age for the 1988 year class collected from each of the nine zones off the coast of New Jersey. and 1977 cohorts (NEFSC 1989. Murawski & Serchuk 1989). In the area designated as Northern New Jersey (area from the New York/New Jersey border south to 39" 10' or 30°30' latitude, de- pending on distance offshore (Serchuk et al. 1979)). NMFS deter- mined in a 1992 EEZ as.sessment study that there was no evidence of any substantial recruitment since 1976-1977. However, in southern New Jersey, they determined that there was some recruit- ment into the fishery between the years 1986 and 1992 (Weinberg lU z o UJ < I o UJ X Q O Ul D. X LU UJ > U- UJ O < X o Ul I a LU Q. X 30 25 20 15 10 0 120 100 80 60 40 A ■ '"(It , 1* * ^-'-'i 4 5 6 ZONE 20 1 3 4 5 6 7 8 9 ZONE Figure 7. A. Expected height at age 1 calculated by von Bertalanffy analysis for the 1988 year class clams sampled in each of the nine zones. Values were determined by nonlinear regression. Bars indicate 95 9r confidence intervals for each estimate. B. Expected height at age 5 calculated by v(m Bertalanffy analysis for the 1988 year class clams sampled in each of the nine zones. Values were determined by nonlin- ear regression. Bars indicate 95% confidence intervals for each esti- mate. SURFCLAM RhCRL'ITMENT AND GROWTH IN COASTAL NJ 1183 Figure 8. Average summer temperature for surface (I ni belowl and bottom (1 ni above bottom l waters off the eoast of New Jersey from 1979 to 1991 for selected transect locations. Data courtesy of the I .S. EPA (1994). 1993). The 1988 year class, which dominated the cential and southern regions of our sampling regime, tails within this time frame. These results suggest that a significant settlement e\ent occurred off southern New Jersey in both the inshore and EEZ populations during this time. Recent data have shown that the EEZ surfclam fishery has indeed been supported by multiple year classes throughout the 1980s and 1990s (Weinberg 1999). In an area offshore from our sampling stations in NJ. it appeared that after the 1976 hypoxia event, there were several years when surf- clam recruitment was relatively high, followed by several years of relatively low recruitment. A 1997 assessment showed that the 1976 and 1977 cohorts were then only a small component of the population, which was represented hy at least 19 year classes (Weinberg 1999). The results from our inshore collections suggest that a new year class ( 19S8) had recruited to the fishery, at least off the central and southern New Jersey coasts. Six of the nine zones sampled were dominated by this year class, and in some zones, such as 5. 7. and 8. it was the only major year class present. The only substantial remnant of the post-anoxia population (1976 and 1977 year classes) was seen in zone 9. The NJ DEP has reported that the inshore surfclam population structuic in 1992 was still dominated by the 1976 and pre-1976 year classes that survived the anoxic conditions (Giust 1993). the average shell length of the surfclanis being 94.1 mm. To compare our results with those from the NJ DEP reports, we converted the surfclam shell heights to lengths using a height:length ratio for each station. The mean length of the dominant 1988 year class in our samples at age 4 (corresponding to 1992) varied between 60 and 80 mm, depending on the station, but it was difficult to com- pare these values with the overall mean in the NJ DEP study since their clams were not aged. The issue of when the dark growth bands used in age determi- nations are laid down in the shell has been extensively discussed by Jones et al. ( 1983). who used oxygen and carbon stable isotopic analyses in a single surfclam shell taken from 9 to 10 m depth off the Bainegat Light (at the southern border of zone 1 in the present study) to examine the correlation between shell increments and monthly average sea surface temperatures. He concluded that in this inshore environment, shell growth is most rapid in spring and early summer, slows in late summer and fall, and is extremely slow or nonexistent in the winter. This agrees well with our own ob- servations at the LEO- 15 site on Beach Haven Ridge (zone .'i in this study: Weissberger 1998. Ma 1999. Grassle unpubl. data). Past research has suggested that the age bands were laid down annually in surfclanis (Jones et al. 1978, Jones et al. 1983). One notable feature of Jones (1983) study was the large shell length (-56 mm) at age I of the individual surfclam that was studied. Jones (1983) noted that the values of 8 "^O predicted froin the monthly sea surface temperatures and the measured oxygen iso- tope values in the first growth increment agreed well except during the months of January through March, and that it indeed repre- sented only 1 year's growth. If settlement of this individual had occuned in the early summer, as is likely off Barnegat Light, and the first clearly visible growth line was laid down in the winter of the following year, the first growth increment may, in fact, have represented growth over 1 6 to 1 8 mo, making the length at "age 1 " less anomalous. In the present study, the expected mean shell heights at age 1 (H,) were considerably lower (ranging from 8.0 mm in /one 3 to 17.5 mm in /one 9. corresponding to shell lengths 1184 Chintala and Grassle of -10.3 and 22.0 mm, respectively), than the previous the shell height at age one (56 mm) of the individual in Jones ( 1983) study. The first hand is often incomplete and faint in surfclams, and these lines may have consistently been missed or ignored in earlier studies, resulting in apparently large shell heights at age one (Jones 1983, Ropes 1985). Some error could have been introduced in the determination of the first age band, but the most likely factor to affect H| would be the time of settlement of the larvae. Larvae that settle later in the year would be expected to have a smaller H,. One of the most striking characteristics of the 1988 year class is the fact that H, was the same in all the zones. This suggests a single, widespread settlement event within 3 miles of the coast in 1988. It has also been suggested previously that growth lines might be laid down at the time of spawning (Jones 1980). There does not seem to be any positive evidence for this in surfclams living off New Jersey, and contrary to previous suggestions that surfclams first reach sexual maturity at age 2 (Ropes 1979), inshore surf- clams as young as 3 mo of age and 5 mm in shell length were found to be sexually mature (Chintala & Grassle 1995). It is likely that inshore and offshore surfclams spawn at different times each year, with the inshore clams spawning as soon as bottom tempera- tures increase above 12-"'C to 13°C in the late spring or early summer. The offshore clams probably do not spawn until bottom temperatures reach this level as a consequence of the breakdown of the thermocline in the early fall, or because of storm mixing (Haskin et al. 1979. Weissberger 1998). Benthic community stud- ies at LEO- 15 indicate that surfclam settlement occurs not only in the early summer, but at low levels in the fall (e.g., October and December of 1993; Weissberger 1998). and molecular iden- tifications of bivalve larvae indicate the presence of surfclam lar- vae in the water column in winter (December and February of 1995/1996; Gregg. Tucker & Grassle, unpubl. data). Thus, it seems likely that the larvae produced by spring/early summer spawning of inshore surfclams settle in early summer, while those produced in the fall may settle in the winter and grow very little prior to the following summer. As a result, the first winter growth line may therefore only be visible in the shells of clams that settle in the summer. A number of hypotheses to explain slower growth (and there- fore a decrease in von Bertalanffy parameters) from New Jersey and Delmarva surfclam populations in the EEZ were put forward by Weinberg and Helser (1996) and Weinberg (1999). Among these were the rapid population increase following the 1976 anoxia event, which might ha\ e increased intraspecific competition; com- mercial size-selective harvesting; and the use of hydraulic dredges that damage clams, expose them to predators, resuspend bottom material, and potentially decrease food quality. Although these hypotheses were for surfclam growth differences from offshore regions that do not overlap with our sampling area, each of these hypotheses could also explain the differences in growth curves in the inshore New Jersey regions that we sampled, since the 1976 anoxic event also affected many inshore regions of New Jersey, and hydraulic clam dredges are used to harvest clams within the 3-mile New Jersey coastal zone. Results suggest uniformly slower growth along the northern coast of New Jersey within I mile of shore, and off southern New Jersey 2 to 3 miles offshore. This slower growth could be a lin- gering effect of competition from the 1976 and 1977 year classes, even though those year classes are now essentially gone from the populations except in zone 9. The explanation for this growth differential is complex since there is a spatial and temporal com- ponent to the variability. In addition, the smallest maximum heights were from surtclatns within 1 mile of shore (zones 1, 4, and 7), while the largest maximum heights were in populations 2 to 3 miles offshore (zones 2, 3. and 6). Also, the lowest k (how fast the population reaches maximum size) was in zone 6, suggesting a less rapid reduction in annual growth with age, while the highest k was in zones 1 . 4. 8. and 9. suggesting a rapid reduction in annual growth with age. These findings do not support previous results that inshore clams always grow more slowly and reach a smaller maximum size than those farther offshore (Jones et al. 1978, Am- brose et al. 1980. Jones 1980. Wagner 1984. Cerrato & Keith 1992). There were also many differences in the expected mean height at age 5 (H^) among the zones, suggesting that environmental or growth conditions are different among zones. The fastest growth for the 1988 year class was in zones 2, 3, and 9, while the slowest growth appeared to be in zones 5, 6, and 7. The Hj for this 1988 year class in zones 1 and 7 were less than in zones 3 and 9. suggesting that in the northern and southern zones the surf- clams furthest offshore had grown faster and reached a larger size than those inshore. This pattern was not supported in the cen- tral zones (zones 4-6). These H^ differences could be related to density, since Cerrato and Keith (1992) found that H^ decreased significantly with surfclam density in the waters off Long Island. Density had no effect on the H, (0-300 clams per m"). but a significant decrease in H.^ was found with densities >100 per m~. In the present study, some of the largest clams were found in zones 2 and 3. and likewise, the highest H, for the 1988 year class clams was in these zones. In a study off the Delmarva Pen- insula in an area closed to commercial harvesting from 1980 to 1991, Weinberg (1998) found that the intensity of apparent competition among surfclams was greatest in areas of the highest density. This competition was evident in a strong nega- tive relationship between mean surfclam length and density of clams aged 5 to 12 y. Unfortunately, in the inshore New Jersey areas where we sampled, intense harvesting could have ob- scured the relationship between density and growth rate by distur- bance from the hydraulic clam dredges and removal of large num- bers of clams from an area in a short period of liine (Weinberg 1998). Reasons for the success of the 1988 year class are likely to be multiple and complex. The stations were located within 3 miles of the New Jersey coast. There did not appear to be a pattern of settlement within the mile zones. There did appear to be a latitu- dinal gradient, however, with the northern regions being domi- nated by numerous year classes and the more southern zones being donnnated o\erwhelmingly by the 1988 year class. Gradients in age-structure have been identified for other surfclam populations along the east coast. Cerrato and Keith (1992) reported that east of Fire Island Inlet, New York, on the south shore of Long Island, the surfclam population was characterized by a wide range of ages with no evidence of age class dominance, but in Long Island Sound and at Long Beach, one or two age classes dominated and there were few or no clams >8 y old. Temperature records from the U.S. EPA (1994) indicate that the mean summer surface temperatures in 1988 were the lowest in 13 y over much of the study area (ranging from 17°C to 18 "C), and mean bottom temperatures were 12°C to 15°C in zones 1 through 6. These low mean summer bottom and surface tempera- tures likely reflect the frequency and persistence of near-coastal SuRFCLAM Recruitment and Growth in Coasi al NJ 1185 Mimnier upwcllinj: e\ents. The differences among the /ones along- shore in different years are not surprising considering the way in which the upweUiiig nodes along the New Jersey coast are distributed (Glenn et al. 1996). Downwelling that follows the upwelling e\ents in the \ icinity of the LEO- 15 site in zone 5 have been shown to be associated with relatively brief pulses of very high concentrations of S. soUdisslina lar\ae (Ma 1997). and subsequent larva! settlement (Ma unpubl. data). Thus, it is not surprising to find a correlation between a measure of up- welling (low mean summer surface or bottom temperature) and surfclam year class strength. Good growth years and good recruitment years for S. si>litlissiiiui have previously been associated with cool rather than warm years (Jones 1980). Generally, cool summers and frequent inner-shelf upwelling events may be associated not only with high larval settlement, but also with reduced mortality due to changes in predator abundance or predator actisity (Weissberger 1998). ACKNOWLEDGMENTS We would like to thank Lori Guist Mayer. Jetf Normand, and the rest of the field crew for their assistance in collecting surf- clams. Doug Jones and Michael Kennish provided tremendous assistance in cutting and aging the surfclam shells. Robert Cerrato was invaluable in sharing statistical analyses. Lastly, we thank Steve Schimmel, Satn Wainright. Peter Morin, Gary Taghon. Rob- ert Cerrato. Jim Weinberg, and two anonymous reviewers for their editorial comments. This is contribution No. 2001-11 from the Institute of Marine and Coastal Sciences at Rutgers Liniversity and Contribution No. AED-OI-017 from the United States Environ- mental Protection Agency. Although the writing of this manuscript was partially funded by the U.S. Environmental Protection Agency, it does not necessarily reflect the views of the Agency. A portion of this project was supported under Rutgers/CMER project 9.V()5 tilled "Surfclam Recruitment." LITERATURE CITED Abbutt. R. T. 1474. American Seashells, 2nd ed. New \iirk: Van Nostrund. 663 pp. Ambrose, W. G.. Jr.. D. S. Jones & I. Thompson. 1980. Distance from shore and growth rate of the suspension feeding bivalve. Spisula vo- lUlissimii. Pioc. Natl. Shellfish Assoc. 70:207-2 1.^. Botton. M. L. & H. H. Haskin. 1984. Distribution and feeding of the horseshoe crab. Linmlus polypheimis. on the continental shelf off New Jersey. US Natl. Mar. Fish. Sen: Fish. Bull. 82:383-389. Cerrato, R. M. 1990. [nterpretable statistical tests for growth comparfsons using the von Bertalanffy equation. Can. J. Fish. Aqtuit. Sci. 47(7): 1416-1426. Cerrato, R. M. & D. L. Keith. 1992. Age structure, growth, and morpho- metric variations in the Atlantic surf clam, Spisuhi solidissima. from estuarine and inshore waters. Mar. Biol. 1 14:.'i8 1-593. Chintala, M. M. 1997. Population biology of surfclams iSpisuUi solitlis- siiiia) in inshore New Jersey waters. M.S. Thesis. Rutgers University. New Brunswick. NJ. 109 pp. Chintala, M. M. & J. P. Grassle. 1995. Early gametogenesis and spawning in "juvenile" Atlantic surfclams. Spisula solidissiina (Dillwyn. 1819). J. Shellfish Res. 14(2):301-306. Carlo, E. V. 1982. Increa.se in a surfclam population after hypoxic water conditions off Little Egg Inlet, New Jersey. J. Shellfish Res. 2( 1 ):59- 64. Giust. L. 1993. Inventory of New Jersey's surf clam tSp:siila Milulissniia) resource. NOAA, NMFS report. Project 3-IJ-I5. Glenn. S. M.. M. F. Crowley. D. B. Haidvogel & Y. T. Song. 1996. Underwater observatory captures coastal upwelling e\cnts off New Jersey. Eos. Am. Geophys. Union 77:233-236. Haskin, H. H. 1978. The onshore surf clam resource along the southern New Jersey coast. Proc. Natl. Shellfish Assoc. 68:79-80. Ha.skin, H. H.. R. R. Schneider & M. Tarnowski. 1979. Management studies of surf clam resources off New Jersey. April 1. 1977-Oclober 31, 1977. Unpublished report to the Fi.sheries Management Branch (FNE3), NMFS. Gloucester. MA. Jones. D. S. 1980. Annual cycle of shell growth increment Ibrmation in two continental shelf bivalves and its paleoecologic significance. Paleobi- ology 6:i?,\-?.-W. Jones. D. S. 1981a. Annual growth increments in shells of Spisula solidis- sima record marine temperature variability. Science 21 1:165-167. Jones. D, S. 1981b. Reproductive cycles of the Atlantic surfclam. Spisula .solidissiina. and the ocean quahog Arclica islaiidica off New Jersey. J. Shellfish Res. l(l):23-32. Jones, D. S.. I. Thompson & W. Ambrose. 1978. Age and growth rate determinations for the Atlantic surf clam, based on internal growth lines in shell cross-sections. Mar. Biol. 47:63-70. Jones. D. S.. D. F. Wdliams & M. A. Anhur. 1983. Growth history and ecology of the Atlantic surf clam. Spisula solidissimu (Dillwyn). as revealed by stable isotopes and annual shell increments. J. E.\p. Mar. Biol. Ecol. 73:225-242. Loesch. J. G. & J. W. Ropes. 1977. Assessment of surf clam slocks in nearshore waters along the Delmarva Peninsula and in the fishery south of Cape Henry. Proc. Nad. Shellfish As.soc. 67:29-34. Ma. H. 1997, Time series analyses of meroplankton in moored pump samples at LEO- 1 5: the relationship between the abundance of surf- clam larvae and nearshore upwelling events. M.S. Thesis. Rutgers Uni- versity, New Brunswick, N.J. 1 10 pp. Mackenzie. C. L.. Jr.. D. J. Radosh & R. N. Reid. 1985. Densities, growth and mortalities of juveniles of the surf clam (Spisula solidissima) (Dillwyn) in the New York Bight. J. Shellfish Res. 5(2):81-84. Merrill. A. S. & J. W. Ropes. 1969. The general distribution of the surf clam and ocean quahog. Proc. Natl. Shellfisli Assoc. 59:40-45. Murawski. S. A. & F. M. Serchuk. 1989. Mechanized shellfish harvesting and its management: the offshore clam fishery of the eastern United States. In: J.F. Caddy, editor. Marine Invertebrate Fisheries: Their As- sessment and Management. New York: J. Wiley and Sons. pp. 479- 506. NEFSC (Northeast Fisheries Science Center). 1989. Report of the yth Northeast Regional Stock Assessment Workshop (9th SAW). Assess- ment updates lor Middle Atlantic. Southern New England, and Georges Bank surfclam populations. NEFC Ref Doc. 89-09. Woods Hole. MA. pp. 9-15. NEFSC (Northeast Fisheries Science Center). 1998. Report of the 26th Northeast Regional Stock A.ssessment Workshop. B. Surfclam assess- ment. NEFSC Ref Doc. 98-04, Woods Hole. MA. pp. 13-23. NEFSC (Northeast Fisheries Science Center). 2000. Report of the 30th Regional Stock Assessment Workshop. E. Surfclam advisory report. NEFSC Ref. Doc. 00-04, Woods Hole. MA, pp. 39-43. Ropes. J. W. 1979. Shell length at sexual maturity of surfclams. Spisuki solidissijna. from an inshore habitat. Proc. Natl. Shellfish .Assoc. 69: 85-91. Ropes. J. W. 1985. Modern methods used to age oceanic bivalves. Nautilus 99(2-3):53-57. Ropes. J. W.. A. S. Merrill. S. A. Murawski. S. Cheng & C. L. MacKenzie. Jr. 1979. Impact on clams and scallops. Part I. Field survey assess- ments. In: R. L. Swanson & C. J. Sindermann. editors. Oxygen Deple- tion and Associated Benthic Mortalities in New York Bight. 1976. Washington. D.C.: U.S. National Oceanic and Atmospheric Adminis- tration Professional Paper 11:263-275. Serchuk. F. M.. S. A. Murawski. E. M. Henderson & B. E. Brown. 1979. The population dynamics basis for management of offshore surf clam 186 Chintala and Grassle populations in tlie Middle Atlantic. Proc. Nonheast Clam Industries: Management for the Future. Univ. of Mass-MIT, Sea Grant. SP-1 12. 157 pp. U.S. EPA (United States Environmental Protection Agency). 1994. STORET Database. Office of Water. Washington, D.C. Wagner, E. S. 1984. Growth rate and annual shell structure patterns in a single year class of surf clams SpisiiUi solidissima off Atlantic City. New Jersey. M.S. Thesis. Rutgers University, New Brunswick, NJ. 160 pp, Weinberg, J. R. 1993. Surf clam populations of the Middle Atlantic, South- em New England, and Georges Bank in 1992. NEFSC Lab. Ref. Doc. 93-01. 14 pp. Weinberg. J. R. 199S. Density-dependent growth in the Atlantic surfclam. Spisiila soliJissinui. off the coast of the Delmarva Peninsula, USA. Mar. Biol. 130:621-630. Weinberg, J. R, 1999. Age-structure, recruitment, and adult mortality in populations of the Atlantic surfclam, Spisiila .- phorae Guilding. the hard clam Mercenariix inercenaria Linnaeus, and the dwaif surfclam Mulinia lateralis Say. Here we report the presence of the vertebrate sequence in telomeric regions of all chromosomes in all four species. Our results along with previous studies suggest that the vertebrate telomere sequence, (TTAGGG)n, may be the telomeric sequence for all bivalve mol- luscs, which at present are the only invertebrates showing the vertebrate lineage in telomeric DNA. MATERIALS AND METHODS The eastern oysters used for this study were from a Rutgers University strain, NEH, which has been selected for MSX- resistance for about 40 years and for Dermo-resistance for about 10 years. MSX and Dermo are two diseases caused by the parasites Haplosporidiiim nelsoiii and Perkinsus inarinus. respectively. The mangrove oysters were from a Fl Caribbean population produced and maintained at the Harbor Branch Oceanographic Institute, Florida. The hard clams were obtained from a clam culture com- pany. Biosphere Inc., New Jersey. The dwarf surfclams were col- lected from beaches of Rhode Island in the fall 2000. Cliromosiime I'reparaliaii Metaphase chromosomes were obtained from cuhiued embryos in all four species, according to the methods described b\ Guo and Allen (1997). For oysters and the dwarf surfclam. gametes were 1 187 1188 Wang and Guo collected by dissecting gonads. For the hard clam, females and males were induced to spawn by thermal simulation. Eggs were collected by passing through a 60-(jLm nytex screen to remove large tissue debris and rinsed on a 25-|jLm screen. Sperm suspension was filtered through a 25-p.m screen. Eggs from sev- eral (3-6) females and sperm from several males were collected and pooled. Eggs were resuspended in 23-25°C seawater (at about l()()(l/ml) and fertilized by adding sperm suspension to a final density of about 5-10 sperm per egg. Embryos were cultured in filtered and UV-treated seawater at 23°C. At 6-8 hours post fer- tilization (about 64-cell stage), embryo suspension was concen- trated on a 25-|jim screen and harvested into a 15-ml tube: 0.2 ml of embryos per 10 ml seawater. Embryos were treated with 0.005% colchicine in seawater for 10-15 min. Embryos were pel- leted by centrifugation at the end of the treatment. Colchicine was replaced by 0.075 M KCI. and the hypotonic treatment lasted for 10-15 minutes. Embryos were then fixed v\ith freshly made Car- noy's fixative. l:3(v:v) acetic acid and methanol. Fixative was changed twice before storing at 4-7°C. Chromosome spreads were made by dropping embryo/cell sus- pension onto clean glass slides and flooding the slides with 2-3 drops of 1:1 methanol and acetic acid. Slides were air-dried at an angle of 45°C. Slides were stored at -20°C until FISH analysis. Ill \ilii Hybridization FISH was conducted according to a protocol from Oncor Inc. modified and described in Guo and Allen ( 1997). Prior to hybrid- ization, slides were pretreated in x2 SSC (0.3 M sodium chloride and 0.03 M sodium citrate. pH = 7.0) at 37°C for 30 minutes. Slides were then dehydrated in 70%. 80% and 95% ethanol for 2 min each and air-dried. Chromosomes on the slides were denatured in 70% formamide in 2x SSC at 72°C for 2 min and dehydrated in a series (70%. 80% and 95%) of ice-cold ethanol. Slides were air-dried, and areas containing chromosome spreads were marked. A commercially available all-human telomere probe. (TTAGGG)n labeled with digoxigenin and supplied in hybridiza- tion solution (Oncor). was used for this study. A 10-p.l aliquot of probe was denatured at 72°C for 5 min. immediately cooled on ice and applied to the marked area on each denatured slide. A cover glass was applied and sealed with rubber cement. Slides were then incubated overnight in a humidified chamber at 37°C for hybrid- ization. After hybridization, slides were washed for 5 min in 2x SSC at 72°C. and for three times. 2 min each, in Ix phosphate buffered detergent (PBD. provided by Oncor) at room temperature. The stringency of the post-hybridization wash was increased to ().5x SSC to test hybridization strength. Negative treatment con- trols were included and made by replacing the probe with equal amounts of hybridization solution. Hybridization was detected with fluorescein-labeled anti- digoxigenin antibodies (Ventana). Sixty |j.l of detection reagent were applied to each slide, covered with a plastic coverslip and incubated at 37"C for 5 inin. Detection reagent was washed three times with Ix PBD. Slides were counterstained with 18 p.1 of propidium iodide/antifade and analyzed under a Nikon epi- fluorescence microscope with PI and PI/FITC dual-pass filters. Metaphase images and FISH signals were captured with a 3CCD camera and analyzed using the Image-Pro Plus software. RESULTS Preparations from early embryos provided sufficient mctaphases for FISH analysis, and most contained elongated chro- mosomes. All mctaphases analyzed in the eastern and mangrove oysters had 20 chromosomes as expected (Longwell et al. 1967, Rodriguez-Romero et al. 1979). Mctaphases from the two clam J species had 38 chromosomes as previously reported (Wada et al. " 1990, Menzel & Menzel 1965). All chromosomes of dwarf-surf clam are acrocentric. FISH with human telomeric probe (TTAGGG)n produced strong hybridization signals on interpha,se nuclei and metaphase chromosomes in all four species examined. Analysis of metaphase chromosomes showed that FISH signals were located at telomeric regions of all chromosomes in all four species: the eastern oyster (Fig. I A), the mangrove oyster (Fig. IB), the hard clam (Fig. IC) and the dwarf surfclam (Fig. ID). The signals were clear and strong in all four species even under high wash stringency (0.5x SSC). Most images were collected after high stringency wash, although results from low stringency (2x SSC) wash were almost identical, or occasionally stronger. All signals were limited to telo- meric regions, and no minor or weak interstitial sites were delected even under low wash stringency. FISH signals were present at telomeres of two sister- chromatids of each chromosome, although occasionally the signal intensity varied. On some chromosomes, the signal on one chro- matid was stronger than that on the other chromatid. In some cases, signal intensity \aried somewhat among chromosomes. From .sev- eral mctaphases analyzed, there was no apparent difference in the size or intensity of signals among the four species. No signals were observed in the treatment controls. DISCUSSION Results o\ this study clearly demonstrate that the vertebrate telomeric sequence (TTAGGG)n is present in telomeric regions of all chromosomes in all four bivalve species examined. The strong FISH signals obtained under high stringencies are an indication of complete homology between the probe and targeted sequence, rather than cross-hybridization to related sequences. In a Cnrhii- liini clam and the bay scallop, the (TTAGGG)n sequence hybrid- ized to genomic DNA in Southern blot, while other telomeric sequences, such as the insect (TTAGG)n, the nematode (TTAG- GC)n and the protozoan (TTGGGG)n. did not (Okazaki et al. 1993, Estabrooks 1999). The lack of cross-hybridization among different (although similar) telomeric sequences has also been demonstrated in many other species (Okazaki et al. 1993. Meyne et al. 1995). Other telomeric sequences were not included in this study, because they are not comparable with the vertebrate se- quence in length and labeling. Nevertheless, it would be helpful to test all telomeric sequences in molluscs by FISH in the future. Chromosomal location of the vertebrate sequence has not been studied by FISH in the Corhiciilliui clam and bay scallop. In the Pacific oyster and a Ddiulx clam, the vertebrate sequence has been located to telomeric regions of all chromosomes (Guo & Allen 1997, Gonzales-Tizon et al. 1998). Including the four species ex- amined here, all eight bivalve molluscs studied so far shared the vertebrate telomeric sequence. Those findings lead us to believe that the vertebrate .sequence. (TTAGGG)n. may be the telomeric sequence for most, if not. all bivalve molluscs. This finding is interesting, because at present, bivalve molluscs are the only in- vertebrate group showing the vertebrate lineage in telomeric DNA. Nematodes and most insects have different telomeric sequences (Zakian 1995). However, telomeric sequences in most inverte- brates have not been studied, and it is possible that the vertebrate sequence is shared by other invertebrates. Vfrtebrate Telomeric Sequence in Molluscs 1189 F^H ->S^ ^^ ^ ' m^ ^^^ ^z^ ^ Fijjure I. Fluorescence (';; silu hyliridizatioii i)l the \ertebrate telomeric sequence (TTAGGG)n to telomeres of all clironu)somes in four bivalve molluscs: A, the eastern oyster, Crassoslrea virgiiiica: B, the mangrove oyster, Crassostrea rhizophorae; C, the hard clam, Mercenaria mercenaria: and D, the dv\;nf surl'clam, Miiliniu laleralis. Bar = 5 //m. It should be pointed out that telomeric DNA in most organisms is more complex than just repeats of one simple sequence. Besides the dominant simple sequence, minor simple repeats and moderate repetitive sequences may occur in telomeric regions (Zakian 1QQ5, Meyne et al. 1995). The complete understanding of hi\al\e telo- meric DNA requires cloning and sequencing of telomeric regions. The vertebrate sequence used in this study can be useful as a probe for cloning studies in bivalves. Telomeric sequences have been found at interstitial sites in many organisms (Zakian 199S). In this study, all FISH signals were restricted to telomeric regions of chromosomes, and no in- terstitial sites were detected even under low stringency conditions. No interstitial sites were observed in the Pacific ovster and Doiuix clam (Guo & Allen 1997, Gonzales-Tizon et al. 1998). The lack of interstitial sites for telomeric sequence may be an indication of liniiled karyotypic rearrangement through evolution. Karyotypes of Crassostrea oysters are highly conserved, and all species stud- ied so far had a diploid number of 20 chromosomes, which are similar in morphology (Nakamura 1985, Xu et al. 2001). It is possible that interstitial sequences are present, but too short to be detected by FISH. The occasionally different signal intensities among some chromosomes and between chromatids mav be due to random variations in hybridization condition, or may reflect true differences in target sequence size. In conclusion, this study provides further evidence that the vertebrate telomeric sequence. (TTAGGG)n, is the telomeric se- quence for bivalve molluscs. Studies on additional species are needed to test this hypothesis. It would be interesting to determine if other molluscan groups such as gastropods and cephalopods, also share the vertebrate telomeric sequence. The vertebrate telo- meric probe and FISH analysis can be useful in studies on telo- meric changes in molluscan development and aging (Harley et al. 1990. Estabrooks 1999). ACKNOWLEDGMENT The authors thank Dr. John Scarpa for pro\ iding mangrov e oysters and Dr. Tinwlhy Scott for providing dwarf surfclams. This study was conducted at Rutgers University and supported by grants from National Sea Grant's Marine Biotechnology Program (Grant B/T-98()l ). USDA NRI (No. 96-35205-3854) and New Jer- sey Commission of Science and Technology (No. 00-2042-007- 20). Yongping Wang is a visiting student from Institute of Ocean- ology. Chinese Academy of Science, and partly supported by grants from China's National Science Foundation (No, 39825121 ), the 863 program (819-01-07). and the Chinese Academy of Sci- ences. This is publication No. 2001-20 of IMCS/NJAES and Sea Grant No. NJ.SG-0 1-472. 1 1 90 Wang and Guo LITERATURE CITED Bluckburn. E. H. 1491. Structure and function ot telemeres. Naluiv 330: 569-573. Estabrooks, S. T. 19W. The telomeres of the bay scallop. Argopccu^fi iiradians (Lamarck). ./. Shellfish Res. 18:401-404. Gonzales-Tizon. A.. A. Marti'nez-Lage, L. Marinas. R. Feire, L. Comudella & Y. Mendez. 1998. Cytogenetic characterization of Doiua irwiciilu.s {Mollusca. Bivalvia). Cytogeiiet. Cell Genet. 81:109. (abstract only) Guo. X. & S. K. Allen Jr. 1997. Fluorescence in situ hybridization of the vertebrate telomere sequence to chromosome ends of the Pacific oyster. Crassoslrea gi.i^iis Thunberg. J. Shellfi.sh Re.s. 16:87-89. Harley. C. B.. A. B. Futcher & C. W. Greider. 1990. Telomeres shorten during aging of human fibroblasts. Nature (Loiulon) 345:458—160. Hastie, N. D.. M. Dempster, M. G. Dunlop & A. M. Thompson. 1990. Telomere reduction in human colorectal carcinoma and with aging. Nature (UmdonI 346:866-868. Lichter. P. & T. Cremer. 1992. Chromosome analysis by non-isotopic in situ hybridization. In: D. E. Rooney & B. H. Czepulkowski. editors. Human Cytogenetics. A Practical Approach. New York: IRL Press, pp. 157-192. Longwell, A. C. S. S. Stiles & D. G. Smith. 1967. Chromosome comple- ment of the American oyster Crassostrea virgiiiica. as seen in meiotic and cleaving eggs. Can. J. Genet. Cytol. 9:845-858. Menzel. R. W. & M. Y. Menzel. 1965 . Studies on chromosomes to two species of quahog clams and their hybrids. Biological Bulletin 129: 181-188. Meyne. J.. H. Hirai & H. T. Imai. 1995. FLSH analysis of the telomere sequences of bulldog anls [Myrmccia: Formicidae), Chromosoimi 104: 14-18. Meyne, J., R. L. Ratliff & R. K. Moyzis. 1989. Conservation of the human telomere sequence (TTAGGG)n. among vertebrates. Proc. Natl. Acad. Sci. USA 86:7049-7053. Nakamura. H. K. 1985. A review of molluscan cytogenetic information based on the ClSMOCH-computerized index system for molluscan chromosomes. Bivalvia. Polyplacophora and Cephalopoda. Veiuis 44: 193-226. Okazaki. S., K. Tsuchida. H. Maekawa, H. Ishikawa & H. Fujiwara. 1993. Identification of a pentanucleotide lelonieric sequence (TTAGGG) in the silkworm Boinhyx iiiori and in other insects. Mol. Cell Biol. 13: 1424-1432. Rodriguez-Romero, F., M. Uribe-Alcocer, A. Laguarda-Figueras & M. E. Diupotex-Chong. 1979. The karyotype of Crassostrea rhizophorae. Venus 38:135-140. Wada. K. T.. J. Scarpa & S. K. Allen Jr. 1990. Karyotype of the dwarf-surf clam Mulinia lateralis (Say 1822) (Mactridae, Bivalvia). J. Shellfish Res. 9:279-28 1 . Xu, Z., X. Guo. P. M. Gaffney & J. C. Pierce. 20111 . Chromosomal location of the major ribosomal RNA genes in Crassoslrea virginica and Cras- sostrea gigas. The Veliger 44:79-83. Zakian, V. A. 1995. Telomeres: beginning to understand the end. Science 270:1601-1607. Joiinuil of Slicllfish Rescuinh. Vol. 20. No. .^, 1141-1 IMS. 20(11. ABSTRACTS OF PRESENTATIONS Presented at 54th ANNUAL MEETING OF THE NATIONAL SHELLFISHERIES ASSOCIATION PACIFIC COAST SECTION & PACIFIC COAST SHELLFISH GROWERS ASSOCIATION Kah-Nee-Tah Rest)rt Warm Springs, Oregon September 27-29. 2000 1191 National SIicIUisIilmics Associalion. Warm Springs. Oieyini Abstracts. September 201)0 I 19."? CONTENTS Amilee Caffey, Brady Blake, and Wall Cooke Eiihaneenient efforts on state ticjelands by the WDFW Intertidal Shellfish Enhancement Project 11 95 H. Calik, M. T. Morrissey. P. Reno, R. Adams, and H. An The use of high hydrostatic pressure for the reduction of Vihrin piinilnicmohlicti\ in shell oysters 1 1 93 Daniel Cheney, Ralph Ehlon, Brian MaeDonald. Kendra Kinnan, and Andy Snhrhier The roles of environmental stressors and culture methods on the summer mortalU) ol tlie Pacific oyster Cmssostrea gigas I 1 9.3 Brett Dumhauld. David Armstrong, Cnrtis Roegner. Kristine Feldman, Lihhy Loggerwell, and Steven Rumrill Implementing a study to determuic the value of molluscan shelllish culture areas as fish habitat in West Coast estuaries 1 1 96 Benoit Eudeline Comparison of the field performances of the "lOO'/r or natural" triploid Pacific oyster Cnissoslrcii iilgcis with chemical triploids and diploids I 1 96 Ronald Figlar-Barnes, Andrea Randall, and Brett Dumhauld Monitoring the status ol the Kuropean green crah unasion in Washington State coastal estuaries i 196 Graham E. Gillespie, Mia Parker, and Bill Merilees Biology and fisheries potential of the varnish clam. NiittiiUiii nhsctircitd. in British Columbia 1196 Laura Hauck. Sylvia Belirens Yaniada, and Sahre Mahaffy Where does the green crab fit into the hieraichv of nativ e crab predators? 1 1 97 Chris luingdon. Dave Jacohson, Sean Matson, and Eord Evans Imprinemenis of \ ields of Pacific in sters through genetic selection 1 197 Timothy Loher and D. A. Armstrong Spatial stock structure m Bristol Bay (Alaska) red king crab and its intluence on long-term recruitment trends 1197 Daniel E. Penttila Intertidal spawning ecology of three species of marine forage fishes in Washington State 1 198 Donald Velasquez and Steve Burton Puget Sound Dungeness crab ( Cancer niagisler) molting patterns 1198 National Slielltisheries Association. Warm Springs. Oregon Abstracts. September 2()()0 I 193 ENHANCEMENT EFFORTS ON STATE TIDELANDS BY THE \VDF\V INTERTn)AL SHELLFISH ENHANCEMENT PROJECT. Aniilee Caffev, Brady Blake, and Walt Cooke. Washington Department of Fish and Wildlife. Point Whitney Shellfish Laboratory. Brinnon WA 9S32U. Due to an increasing demand for clam and oyster resources on public tidelands in Washington, the Washington Department of Fish and Wildlife (WDFW) developed the hitertidal Shellfish En- hancement Project in 1988. The goal of this project is to increase recreational opportunities by u a\ of planting clam and oyster seed, as well as harvestable size oysters, onto public tidelands. Species targeted in the past have been Pacific oysters, geoduck and Manila clams. In the last few years, research has been conducted on cul- turing and enhancing native species such as butter, native littleneck clams and cockles. Enhancement efforts occur throughout Puget Sound. Creel surveys and population assessments of targeted clam species are conducted each year by WDI^V and tribal staff. This infonnation is passed to the Intertidal Enhancement Project and beaches are seeded according to Intertidal Management plans and needs. The Enhancement Project purchases Pacific oyster seed and harvestable size oysters to be planted by commercial growers as directed by WDFW staff. The Point Whitney Shellfish Hatchery provides geoduck and Manila clam seed. Research performed by the WDFW Shellfish Hatchery on native species of slams such as butter, littleneck clams and cockles, have provided the Enhance- ment Project with small groups of seed for test plots, which will, in turn, lead to large scale planting of native species in the future. THE USE OF HIGH HYDROSTATIC PRESSURE FOR THE REDUCTION OF VIBRIO PARAHAEMOLYTICUS IN SHELL OYSTERS. H. Calik. M. T. Morrissey, P. Reno. R. Adams, and H. An, Oregon State University Seafood Laboratory. 2001 Marine Drive, Astoria, OR. High Hydrostatic Pressure (HHP) Technology has shown good potential pathogen reduction. The effect of HHP treatment on pure Vp culture was tested. Clinical and environmental strains of Vp were acquired from FDA was tested. Both clinical and environ- mental Vp strains were treated with HHP at different settings ( 1-10 min at 35K psi: 1-5 min at 40K psi: 30-120 s at 45K psi: 10-50 s at 50K psi). Total Vp was enumerated before and after HHP treatment to determine survival. Additional tests were performed to determine if Vp in shell- oy.sters responds to HHP treatment differently than Vp in pure culture. After the inoculation of the same clinical and environmen- tal Vp strains, oysters were treated with HHP at the same time and pressure settings. Results showed that Vp is susceptible to HHP m both pure culture and shell-oyster treatments. The optimum conditions for reducing Vp counts from 109 cfu/ml to 101 cfu/ml was achieved at 50K psi m 30 seconds. At 35K psi. >10 min were required to reduce the count to 101 cfu/ml Vp survival. HHP treatment was successful in reducing Vp counts. Vp in oysters and pure Vp culture respimded similarly to HHP treat- ments. No resistance variation to HHP uas detected between clini- cal and environmental Vp strains. HHP can be a viable means to reduce the Vp counts in oysters. THE ROLES OF ENVIRONMENTAL STRESSORS AND CULTURE METHODS ON THE SUMMER MORTALITY OF THE PACIFIC OYSTER CRASSOSTREA GIGAS. Daniel Cheney, Ralph Elston, Brian MacDonald, Kendra Kinnan, and Andy Suhrbier, Pacific Shellfish Institute. 120 State Ave NE #142. Olympia. WA 98501. During the late summer to early fall period, Pacific oysters cultured on the west coast of the United States and elsewhere may e.\perience high levels of mortality. In the 1960's to 80' s, this condition was subject to intensive investigation focusing on broad areas of disease pathology, genetics, physiology and the environ- ment. Results of these studies were largely inconclusive, or pointed to a poorly defined etiology. Although several factors, such as a bacterial and herpes-like virus infections could be linked to certain mortality events, no clear picture emerged. Recent studies in Puget Sound. Washington USA and Tomales Bay, California USA center on the influence of multiple stressors and their affects on oyster survival, physiology and pathology. The goal of this research is to identify possible modifications in culture practices, brood stock selection or grow-out location to increase survival of Pacific oysters. Field observations indicate oysters are subject to extreme varia- tions in a number of parameters during intertidal cycles. Annual or seasonal variations in those parameters and culture practices ap- pear to play a major role in oyster survival. An increased rate of oyster mortality and modified physiological response appears to be strongly correlated with both elevated temperatures and extended periods of depressed DO. A long period of neap tides with low and slack water during the evening was ob.served to result in daily and successive reductions in DO to levels ranging from 0.5 and 2 mg/L. The DO reductions are sometimes coupled with heavy mac- roalgae blooms and high phytoplankton densities. This and other work indicate oyster summer mortality rates are also strongly in- fluenced by ploidy and broodstock origin/stock selection. These observations have renewed interest in testing selected stocks with reduced rates of summer mortality, which still retaining desirable characteristics of good growth and meat vield. 196 Abstracts, September 2()()() National Slicllfisheries Association. Warm Springs. Oregon IMPLEMENTING A STUDY TO DETERMINE THE VALUE OF MOLLUSCAN SHELLFISH CULTURE AREAS AS FISH HABITAT IN WEST COAST ESTUARIES. Brett Dumbauld, Washington State Department of Fish and Wildlife, P.O. Box IQO. Ocean Park. WA. 4S64(). David Armstrong. Cur- ti.s Roegner, Kristine Feldnian. Libbj Loggerwell. School of Fisheries, Box 355020, University of Washington. Seattle, Wash- ington 98195, and Steven Rumrill. South Slough National Estua- rine Research Reserve. P.O. Box 5417. Charleston, Oregon 97420. The ecological role of bivalve molluscs in estuarine systems has recently been recognized in other areas including Europe and eastern North America where studies have been completed and in some cases shellfish restoration efforts initiated. Comparable stud- ies are lacking however from estuaries on the West Coast, where bivalve (particularly oyster and clam) aquaculture often dominates the intertidal landscape. Due to increasing interest, especially with regard to the potential impacts of oyster aquaculture activities on juvenile salnionids (driven by the listing of several stocks under the Endangered Species Act), we are initiating a study designed to quantify both adverse, but also beneficial impacts of shellfish farming on selected estuarine fauna and tlora. We will focus our initial efforts on oyster ground culture and on eelgrass as benthic habitats given the extent and previously documented value of these habitats respectively. Field and laboratory objectives include: 1 ) utilizing remote sensing and ground-trulhing to document anntial variability in eelgrass cover in oyster culture and eelgrass mead- ows in Willapa and Coos Bay estuaries; 2) compare species di- versity, density and biomass in culture areas as well as eelgrass meadows; 3) conduct field experiments to examine the impacts of various culture activities on eelgrass and associated infaunal and epifaunal communities; and 4) conduct surveys of fish utilization in oyster beds and eelgrass meadows. Finally, we hope to prepare guidelines to assist both shellfish farmers and estuarine managers in avoiding and/or reducing adverse impacts on estuarine habitat while maximizing the potential beneficial impacts of aquaculture activities. COMPARISON OF THE FIELD PERFORMANCES OF THE "100% OR NATURAL" TRIPLOID PACIFIC OYSTER CRASSOSTREA GIGAS WITH CHEMICAL TRIPLOIDS AND DIPLOIDS. Benoit Eudeline, Taylor Resources INC. Whiskey Creek Shellfish Hatchery, 2975 Netarts Bay Road W.. Tillamook OR 97141. For about 3 years, the Whiskey Creek Shellfish Hatchery has been producing commercially a new kind of triploid Pacific Oys- ters {Cnissosliea fiigas) called "Natural" or "lOO'/r" triploid. made from the cross between regular diploids and tetraploids. The first advantage is that these "natural" triploids are always 100 percent triploid as verified by flow cytometry, insuring a high quality product for the customers. Hatchery seed was deployed in grow- out bags at two different locations for a 26-month study to examine the growth performance (length, whole v\ eight, meat weight) and gonadal development of the natural triploids. chemical triploids and diploids. The "natural" triploid outperformed both diploids and chemical triploids on most of the characteristics studied. The final length was increased by 7 to 9% in the natural triploids compared to the diploids, the whole weight improved by 30 to 40% and the meat weight by 19% in the most productive area. The "natural" triploids outperformed as well the chemical triploids. with an increase of 7% in length, 15% in whole weight and 1S% in meat weight in the most productive area. No histological dif- ferences were detected between "natural" and chemical triploids. but their gametogenesis was significantly reduced when compared to the diploids for the females and in a lower extend for the males. MONITORING THE STATUS OF THE EUROPEAN GREEN CRAB INVASION IN WASHINGTON STATE COASTAL ESTUARIES. Ronald Figlar-Barnes. Andrea Ran- dall, and Brett Dumbauld. Washington State Department of Fish and Wildlife, RO. Box 190, Ocean Park, WA, 98640. The European green crab (Carciiuis inaeiuis) was discovered on the West Coast of North America in 1989. A series of recruit- ment events led to expansion northward from San Francisco Bay including a particularly strong recruitment pulse which resulted in numerous crab being found in coastal estuaries from Oregon to British Columbia in 1998 and 1999. Washington State responded by establishing a statewide monitoring and control program for this invader in 1998. Results from the monitoring effort to date in Willapa Ba\ and Grays Harbor. Washington suggest that the population in these estuaries is declining, despite evidence that local reproduction is potentially occurring. Average catch per unit trapping effort has declined markedly at a station located near the mouth of Willapa Bay. Mating couples and females with viable eggs were found during winter months and recruitment was documented in 1999, however larger crab representing the 1997/98 year class still domi- nate the catch. A volunteer trapping effort was successfully insti- gated along the western shore of Willapa Bay in 2000 and trapping methods continue to he refined as the primary control technique. BIOLOGY AND FISHERIES POTENTIAL OF THE VAR- NISH CLAM, NVTTAU.IA OBSCIRATA. IN BRITISH CO- LUMBIA. Graham E. Gillespie, Mia Parker, and Bill Merilees, Fisheries and Oceans Canada. Pacific Biological Station, Nan- aimo, BC. V9R 5K6. Varnish clams, Nutlatlia nhsciirata. have recently become es- tablished in British Columbia (BC), likely from introduction m ballast water. They are currently distributed from Cameleon Har- bour (50''22'N) to at least Port Townsend (48''()rN) and on the west coast of Vancouver Island in Barklcy and Clayoquot Sounds. National Shellt'isheries Associatiini. Warm Spiinys. Oregon Abstracts. September 2(K)() 197 They have been reported trom Oregon, but not from Grays Har- bour or Willapa Bay. WA, Varnish clams tend to be distributed at higher tidal elevations than Mamla clams, prefer sand or gravel substrates and are often found in areas of freshwater outtlows or seepage. When found w ith Manila clams, varnish clams tend to bury deeper in the substrate. Varnish clams use pedal feeding to augment suspension tecding and this, as well as their tendency to live deep in the substrate, may allow them to live at higher tidal elevations. Varnish clams spawn in May in the western Pacific and histo- logical e,\amination of a BC sample taken late in March showed evidence that spawning had occurred. Varnish clams appear to grow at similar rates to Manila clams, reaching .-(iS lum in length in four years. Varnish clams have potential as a commercial and lecreational resources, as attractive clams with good meat-to-shell ratio and good shelf life. They are generally gritty and can recycle purged substrate from tanks. The soft-shelled nature of varnish clams present a problem due to breakage during harvest and transport. Work is currently underway in BC to explore market potential and develop policy leading to commercial exploitation of varnish clams. IMPROVEMENTS OF YIELDS OF PACIFIC OYSTERS THROl (;H genetic selection. Chris Lanjjdon. Dave Jacobson, Si-an Matson, and Ford Evans, Coastal Oregon marine Experimental Station and Department of Fisheries and Wildlife. Oregon State University. Newport. Oregon. 97.^6.S. Crosses were made in 1998 among nine families showing the highest yields (wet weight per bag) of 48 families planted at an intertidal site in Tomales Bay. California. The resulting full-sib families were plated at an intertidal site in Totten Inlet. Puget Sound. Washington, in 1998. In addition, full-sib families derived from crossing non-selected '"wild" oysters were planted at this site as well as groups of oyster seed from various commercial sources. After about two years of growth, the families were harvested and the average yield (kg live weight) per bag determined for each family. The estimated reali,^ed heritability for yield was 0.42. Fam- ily yields varied greatly depending on the composition of the pa- rental cross, with the average yield of inbred families being sig- nificantly less (p = 0.0028; Fisher's PLSD) than that of out-bred families from the selected broodstock; however, the average yield of the lop nine families from the selected broodstock was IS'/c greater than that of industry oysters. Based on these results, it should be possible for industry to significantly improve oyster production by crossing specific selected families as part of a long- term breeding program, hibreeding should be avoided as it results in signincantly reduced yields. WHERE DOES THE GREEN CRAB FIT INTO THE HIER- ARCHY OF NATIVE CRAB PREDATORS? Laura Hauck, Sylvia Behrens Yamada, and Sabre Maliaffy. Zoology Depart- ment. Oregon State University. Corvallis. Oregon 97.^31-2914. We compared prey size selection, prey consumption rates and mechanical advantage of the claw lever systems in the European green crab. Cairiiui.s inaenas. and five species of native crabs. Individual crabs were housed inside perforated plastic boxes in open seawater tanks at Hatfield Marine Science Center. We of- fei-ed each crab 5 small (5-8 mm) 5 medium (8-1 I mm) and 5 large (11-14 mm) snail (Liiidhim shkaiui) per day and recorded their consumption rate. Green crabs eat significantly more snails than the shore crabs Hemigrapsus niidis and H. oregonesis of similar size. While all .3 crabs prefer smaller snails, only green crabs eat the largest. They simply insert the slender tips of their minor claw into the snail's aperture and pull out the soft tissue. Green crabs exhibit the same feeding rates as young Dungeness (Cwiccr mugistei) and red rock crabs (C. prothictiis) of similar size. The mechanical advantage of claw lever systems are 0.25 for Dungeness. 0.28 for Hemigrapsus nudis and H. oregonesis, 0.36 for the master claws of male green crabs and 0.39 for red rock crabs. These comparisons suggest that the impact of green crabs on hard-shelled prey will be significantly higher than that of shore crabs and of the same order of magnitude as that of red rock crabs of similar size. This research was supported by Oregon Sea Grant. SPATIAL STOCK STRUCTURE IN BRISTOL BAY (ALASKA) RED KING CRAB AND ITS INFLUENCE ON LONG-TERM RECRUITMENT TRENDS. Timothy Loher and D. A. Armstron!;. University of Washington. School of Fish- eries. Box 355020. Seattle. WA 98195. The Bristol Bay red king crab {Paniliiltodes vaiiiisclniticus) stock once supported the most lucrative fishery in the world, but low catches over the last -20 yrs have prompted fishery closures and a host of restrictive regulatory measures. However, these ac- tions have had little effect on stock rebuilding, suggesting that factors outside the fishery may exert strong influence on popula- tion abundance. We hypothesize that the population is regulated by the survi- vorship of early post-settlement stages that require complex sub- strates as nurseries, and that low recruitment levels are in part the result of climate forcing that has altered larval .source-sink dynam- ics. We present evidence of alterations in broodstock distribution that occurred during the early I980"s, and changes in spatial re- cruitment patterns that roughly mimic the shifts in the adult popu- lation: the possible effects of the 1978 "climate regime shift" in initiating these changes will be discussed. 1 198 Abstracts, September 2(K)() Natuiiuil Shellfisheries Association. Warm Springs, Oregon INTERTIDAL SPAWNING ECOLOGY OF THREE SPE- CIES OF MARINE FORAGE FISHES IN WASHINGTON STATE. Daniel E. Penttila. Washington Department of Fish and Wildlife. Marine Resources Div.. LaConner. WA 98257. The Pacific heiring iChipca). surf smelt {Hypomesus). and Pa- cific sand lance (Ammadyles) are common schooling forage fishes in the Puget Sound basin and coastal estuaries on Washington State. In spite of their pelaigic habitats, they are intimately asso- ciated with nearshore/interlidal benthic environments during their annual spawning activity. Herring use nearshore beds of eelgrass and marine algae for spawn deposition. Surf smelt and sand lance use upper intertidal sand-gravel beaches for spawning. Marine forage fish spawning sites and seasons are unpredict- able from year to year. Spawning commonly occurs for several months each year at any given spawning site. Preservation of all known spawning sites is considered critical for these species' long- term conservation. In many areas, forage fish spawning sites and seasons overlap and coincide with shellfish aquaculture areas and commercial har- vest activities. Means must be found to minimi/e disruption of forage fish spawning sites and seasons. Cunent environmental regulations consider all known marine forage fish spawning sites to be "habitats of special concern." and they are afforded "no net loss" protection from impacts oi in-vvater human activities. Efforts are being made to inventory all Washington State shorelines for evidence of forage fish spawning activity. PUGET SOUND DUNGENESS CRAB [CANCER MAGIS- TER) MOLTING PATTERNS. Donald Velasquez and Steve Burton, Washington Department of Fish and Wildlife. Ifil)18 Mill Creek Blvd., Mill Creek, WA 98012-1296. The Washington State Department of Fish and Wildlife and Treaty Tribes currently regulate the harvestable surplus of Dunge- ness crab in Puget Sound based on the following criteria: 1 ) male crab onlv. 2) a minimum size of male crab, and 3) avoiding pot harvest on soft male crab coming out of a major molt. With these conservation tools in place, the harvestable surplus is then allo- cated between the State and Treaty Tribes for a 50/50 split. State/ tribal management plans set pot fishery seasons around the major molts and any deviation from these dates must be accompanied by tests to determine that SO'/r of the legal-sized, male crabs are in hardshell condition. Data from these shell condition tests indicates that the molting time period once considered appropriate for all of Puget Sound is not applicable for all of it's subareas. The results also suggest a trend in the molt timing from the northern regions to the southern regions. JoKiiial of Shellfish Rcuanh. Vol. 20. Nci. .■?, 1 IW-l.MS. ^(lOI. Proceedings of The 3rd International Conference on Molluscan Shellfish Safety Southampton, New York June 19-24,2000 Guest Editor: STEVEN JONES JACKSON ESTUARINE LABORATORY University of New Hampshire 85 POINT ROAD DURHAM, NEW HAMPSHIRE 03824-3406 1199 3rd International Conference on MoUuscan Shellfish Safety Conference Chair Sandra E. Shumway (USA) Organizing Committee Stephen Jones (USA) Gary Roddrick (USA) Dorothy Leonard (USA) International Advisory Committee Rhodora Azanza (Philippines) Ken Buckle (Australia) Henrik Enevoldsen (Denmark) Yosuwo Fukyo (Japan) Susan Gallacher (United Kingdom) Joanne Jellett (Canada) Jennifer Martin (Canada) Beatriz Reguera (Spain) Brian Roughan (New Zealand) Benjamin Suarez-lsia (Chile) Local Committee Gregg Rivara (Chair; Cornell Cooperative Extension) John Aldred (East Hampton Shellfish Hatchery) Nancy Boland (Southampton) Stuart Buckner (Town of Islip Dept of Environmental Conservation) Maureen Davidson (NY State Dept of Environmental Conservation) Ken Gall (New York Sea Grant) Jeffrey Kassner (Brookhaven Town Dept of Environmental Robert Nuzzi (Suffolk County Dept of Public Health Protection) Lisa Tettelbach (NY State Dept of Environmental Conservation) * International Conference on Molluscan i Shellfish Safety JUNE 19-24, 2000 SOUTHAMPTON COLLEGE LONG ISLAND UNIVERSITY Southampton, New York USA A VERY SPECIAL THANKS TO OUR SPONSORS LONGlSIAND NATIONAL SHKII FISHKRIKS ASSOC I AH O.N OCM«0s^ UNESCO National Sea Grant New York Connecticut Maine/New Hampshire nt Sea Rhode Island Florida S.C. Sea Grant Consortium Louisiana SeaUfant Virginia MARINE INSTITUTE Journal of Shellfish Research. Vol. 20. No. .^. i:().V1214, 2()()1. MONITORING FOR TOXIC CONTAMINANTS IN MYTILUS EDULIS FROM NEW HAMPSHIRE AND THE GULF OF MAINE S. H. JONES,' * M. CHASE,' J. SOVVLES," P. HENNIGAR,' N. LANDRY,^ P. G. VVELLS,'^ G. C. H. HARDING," C. KRAHFORST.^ AND G. L. BRUN'* 'j(((/v.sy);( Estiiarine Laboratory. University of New Hampshire. Durham. New Hampshire 03824: 'Maine Department of Murine Resourees, Boothlniy Harlwr. Maine; Environment Canada. Dartmouth. Nova Scotia B2Y 2N6. Canada; '^New Hampshire Department of Environmental SeiTices. Concord. New Hampshire 03301; ^School for Resource and Environmental Studies. Dalhousie University. Halifax. Nova Scotia B3H 3E2. Canada; '^Department of Fisheries ami Oceans. Bedford In.'ititute of Oceanography. Dartmouth. Nova Scotia B2Y 4A2, Canada; ^Marine Monitoring and Research. Massachusetts Coastal Zone Management. Boston. Massachusetts 02114; ^Environmental Conservation Branch. CWS Moncton. New Brunswick. Canada ABSTRACT Gulfwatch is an international monitoring program that uses Myiilus cdiilis as the sentinel species for habitat exposure to to.xic contaminants in the Gulf of Maine. Since 1991. the Gulfwatch program has measured the concentrations of 10 trace metals. 17 chlorinated pesticides. 24 polynuclear aromatic hydrocarbons (PAHs). and 24 polychlorinated biphenyl (PCB) congeners in mussel tissue from over W) sites throughout the Gulf of Maine. In 1998. the Gulfwatch program included a more intensive survey of sites in New Hampshire. Of the 21 sites sampled in the Gulf in 1998. six sites were sampled in New Hampshire and the other 15 sites were in Nova Scotia. New Brunswick. Maine, and Massachusetts. Results for the Gulfwide program showed a southward trend of increasing concentrations for organic contaminants and .Ag. Cr. and Pb. reflecting major local and regional pollution sources. The distribution of the remaining trace metals was more uniform, without obvious impacts from the more heavily urbanized areas of the southern gulf. At some sites there appears to be localized hot spots for a variety of contaminants. Gulfwide geometric mean (GM) concentrations were lower than GM +8.'i<7f confidence level (CL) concentration for the NOAA National Status and Trends (NS&T) Mussel Watch program, except for Hg. Two sites had Pb concentrations that exceeded U.S. FDA guideline levels. While Hg concentrations did not exceed the U.S. FDA action concentration at any site, the NS&T GM +85'7f CL was exceeded at all 1998 Gulfwatch sites. The spatially intensive New Hampshire program has been useful in identifying local sources of contaminants and determining the significance of mussel exposure to oil from chronic discharges and larger spills. Similar to comparisons between Gulfwatch and NS&T data, interpretation of New Hampshire results benefit from comparisons to the greater Gulfwatch program. The Gulfwatch database provides a useful regional perspective to interpret the results from the New Hampshire program and for other localized studies. Gulfwatch provides unique. Gulfwide status and trend information and helps to focus efforts to reduce loading of contaminants such as Hg. KEY WORDS: Mxiiliis eihili<.. toxic contaminants. Gulf of Maine, spatial trends INTRODUCTION The Gulf of Maine (GOM) extends from Cape Sable. Nova Scotia, through New Brunswick. Maine, and New Hampshire to Cape Cod, Massachusetts, and includes the Bay of Fundy and Georges Bank. The ecosystem is one of the world's most produc- tive systems and supports a vast airay of species, including some of great commercial importance. Commercial fisheries, including aquaculture. is its principal income-generating enterprise. Tourism is also a significant source of income to GOM coastal communi- ties. Increases in coastal populatiinis and industrial and residential development have contributed to the deteriorating quality of sec- tions of the Gulfs coastal environment (Crawford & Sowles 1992. Dow & Braasch 1996). Despite efforts to improve pollution treat- ment, the result of increased human growth and activities is the steady input of toxic chemicals into estuarine and coastal en\ iron- ments. Many human-made chemicals can be bioaccumulated to concentrations significantly above ambient levels. Furthermore, some of these environmental contaminants may also be present at concentrations considered toxic to organisms, and thus induce ad- verse biological effects on productivity . reproduction, and survival of marine organisms and humans (Kawaguchi et al. 1999. Wells & Rolston 1991). *Corresponding author. E-mail: shj@hypatia.unh.edu Monitoring programs may use a variety of organisms as indi- cators of environmental contamination. One organistn that has been commonly employed and has proven useful for biomonitor- ing chemical contamination is Mytihts edulis L (Sericano et al. 1995. Tripp & Farrington 1985). M. edulis and other bivalves have been used throughout the world as a measure of the spatial and temporal trends in habitat exposure to chemical contatninants. In addition. M. edulis has been successfully used as an indicator organism in environmental monitoring programs throughout the world to identify variation in chemical contaminants between sites, contributing to the understanding of trends in coastal contamina- tion (NAS 1980. NOAA 1991. O'Connor 1992, O'Connor &i Be- liaeff 1995. 'Widdows et al. 1995. Cantillo 1998). The blue mussel is the indicator organism for the NOAA NS&T. Gulf of Maine Gulfwatch and New Hampshire Gulfwatch programs because of the follow ing favorable characteristics: mussels are indigenous and abundant in many coastal areas, and are easy to collect and pro- cess; much is known about mussel biology and physiology: they are a commercially important food source; mussel contamination is a public health concern; and they are sedentary suspension- feeders that concentrate chemicals in tissue, making measurement of chemicals in mussel tissue an assessment of local biologically available contamination. The iiverall objective of this paper is to introduce the Gulf- watch program to a wider audience of scientists concerned about the safe consumption of molluscan shellfish. The 1998 results are 1203 1204 Jones et al. presented to illustrate local, regional, and international implica- tions and to address issues related to toxic contaminant pollution in the marine environment. Between 1991 and 1997 the Gulfvvatch program measured the concentrations of trace metals, polynuclear aromatic hydrocarbons (PAHs). polychlorinated biphenyls (PCBsl. and chlorinated pesticides in mussel tissue at 56 sites throughout the GOM. In 1998. the Gult'watch program was in the sixth year of a nine-year monitoring plan. By design of the Gult- watch program, the 1998 sampling was the second samplmg of sites previously sampled in 1995. These sites will be resampled in 2001 (Jones et al. 1998). Native mussels collected at 21 sites throughout the Gulf of Maine were analyzed for organic and in- organic contaminants (Crawford & Sowles 1992) to build on the program's Gulfwide assessment of long-term exposure to bioavail- able contaminants. One location in each of the five jurisdictions served as a baseline station that has been resampled every year. The 1998 Gulfwatch program also included more spatially inten- sive sampling in New Brunsv\ick. MATERIALS AND METHODS The 1998 Gulfwatch sample collection and analysis is the sixth year of the program's nine-year sampling design (Sowles et al. 1997). The 1998 sampling represents the third year of the second three-year cycle. As such, a sub.set of the total Gulfwide stations sampled in 1998 were previously sampled in 1995. Some select sites have been sampled each year since 1993 for more intensive temporal analysis of contaminant concentrations. The New Hamp- shire Gulfwatch program added six new sites in addition to six sites previously sampled under the Gulfwide Gulfwatch program; 1998 is the first year of sampling at three of the new sites in New Hampshire. Locations for sites sampled as part of the Gulfwatch program prior to 1998 are shown in Figure 1 . The stations sampled in 1998 are presented in Table 1 with reference to site numbers in Figure 1. The site list includes two New Brunswick sites located in St. John Harbor that were new in 1998. The six New Hampshire sites sampled in 1998 are shown in Figure 2. including three sites that had been previously sampled. NHLH. NHDP. and MECC. and three sites that were sampled for the first time: NHGP. NHSS. and NHNM. Clark Cove (MECC) in Portsmouth Harbor is a bench- mark site that has been sampled each of the previous five years. Gulfwatch benchmark sites were established to enable trend analy- sis and were selected as sites considered to be important in each jurisdiction. Field and Laboralory Procedures Details regarding the mussel collection, measurement, and sample preparation are published in Sowles et al. (1997). All field sampling was conducted in the fall of 1998. No sampling occurred during or shortly after periods when storm water runoff and wave resuspension of bottom sediment could result in enhanced uptake and accumulation of sediment in the mussel gut, as previously reported (Chase et al. 1998). Mussels were collected from four discrete areas within a segment of the shoreline that is represen- tative of local water quality. Using a wooden gauge or a ruler. 45-50 mussels of 50-60 mm shell length were collected. The mussels were cleaned of all sediment, epibiota. and other accre- tions in clean seawater from the collection site, placed in clean containers, then transported to the lab in coolers with ice packs. Mussels were not depurated prior to processing. Prior to shucking. Julfwatch Stations 1991-1998 No 8ite NO 8la MfUM 1 »ll.SN z *UP1 ;iu •irttB a *UD\ 31 MflB 4 M4tO 32 Mcrp 5 »UNI 33 Mrpi 6 M41I 34 MCLR 7 M4ftl 35 mcm: 8 *uin 36 MC*H 9 •UWN 37 MrMR lO M4l»tt 36 MCCK 11 Misn 39 NB.-K 12 *i4Mn 40 NBNR ia •141 P 41 NBni 14 •u»tc 42 NBLN 15 Nnn* 43 NBUl 16 NHitn 44 NBMI 17 SHLM 45 ^»4R IB NtlSI 46 r*in 10 ISM DP 47 VS\C ZO Mrcc 4B vso* 21 MCBM 49 VMC 22 ►IC»4 50 V^DI 23 HCPn 51 !^BC 24 MCPR 52 N&.SC 25 MCBC 53 N»BC 26 MCRT 54 ISATR 27 MCkN 55 NS\Q 2S MC*n 56 P^BP Figure 1. Location of Gulfwatch stations in the Culf of Maine. mussels were thoroughly washed to minimize tissue contamination from an> remaining surface debris. In the laboratory, mussels were shucked directly into appropri- ately prepared Mason jars for metal and organic analysis, respec- ti\elv (for details see Sowles et al. 1997). Composite samples (20 mussels/composite; four composites/station) were capped, labeled, and stored at -15"'C for three to six months prior to analysis. Analytical Procedures Analytical procedures followed those reported for the previous years (Chase et al. 1998. Jones et al. 1998). Table 2 contains a summary of measured trace metal and organic compounds. Inor- ganic contaminants were analyzed at the State of Maine Health and Environmental Testing Laboratory (Augusta. ME). Analyses for mercury were done on a subsample of 1 to 2 g of wet tissue and measured by cold vapor atomic absorption on a Perkin Elmer Model 503 atomic absorption spectrometer. Analyses for all other metals were conducted on 5 to 10 g of wet tissue dried at 100°C. Zinc and iron were measured by tlame atomic absorption using a Perkin Elmer Model 1100 atomic absorption spectrometer. All remaining metals (Ag, Al. Cd, Cr, Cu, Ni, and Pb) were analyzed using Zeeman background-corrected graphite furnace atomic ab- sorption with a Varian Spectra AA 400. The analyte detection limits for the metals in |jig/g dry weight (DW) are as follows: Ag, 0. 1 ; Al, 4.0; Cd. 0. 1 ; Cr. 0.2; Cu. 0.4; Fe. 4.0; Hg, 0. 1 ; Ni, 0.4; Pb, 0.4; and Zn. 0.5. Organic contaminants in mussel samples were analyzed at the Environment Canada, ECB Laboratory in Moncton, NB. The ana- Toxic Contaminants in Mytilvs eduus 1 205 TABLi: I. (iull (il Miiiiii' (;ulf'\v:iti'h ^llltl^ silo Uiiations in IWS. Station Massachusetls Sandwich Boston Inner Harhor Pines River New Hampsliire Little Harhor Gypsiuii Plant North Mill Pond Schiller Station Dover Point Clark Cove Maine Kennebec River Damariscotta Boothbay Harbor New Brunswick Niger Reef Chanicook Letang Estuary Limekiln Bay Tin Can Beach Coast Guard Wharf Nova Scotia Cornwallis Digby Broad Cove Map # (Figure ll I 9 11 18 Portsmouth Harbor Portsmouth Harbor Portsmouth Harbor 20 21 28 32 30 41 42 43 44 John Harbor St St. John Harbor 50 52 53 Statitin Code MASN MAIH MAPR NHLH NHGP NHNM NHSS NHDP MECC MEKN MEDM MEBB NBNR NBCH NBLN NBLB NBTC NBCG NSCW NSDI NSBC samples. The Moiicton laboratory participated in the NIST Status and Trends Intercomparison Marine Sediment Exercise IV and Bivalve Hoinogenate Exercise. Internal quality control and method performance specifications are described in the Enxironment Canada Shellfish Surveillance Protocol (Sowles et al. 1997, Jones et al. I99S). The protocol includes mandatory QC measures with every sample batch including method blanks, spike matrix samples, duplicate samples. suiTogate addition, and certified ref- erence materials (SRM 1974a). The protiicol specifies the perfor- mance criteria relevant to method accuracy, precision, and detec- tion limits and data reporting requirements for the analysis of organic contaminants in shellfish samples. Dula Analysis Total PAH (IPAH,_,). total PCB (iPCB.jl. and total pesti- cides (XTPESTi^) values were calculated from the sum of all individual compounds or congeners with \alues greater than the detection limit for the compound. Total DDT (XDDT,,) is the sum of o.p'-DDT and p.p'-DDT and hotnologues (o.p'-DDE, p,p'- DDE. o.p'-DDD. and p.p'-DDD). Several tissue samples for met- als and organics were below the detection level. Variables in which all replicate measurements were below the detection limit were treated as /.ero and recorded as not detected (ND). However, if at least one of the replicates was greater than the detection limit, then the other replicates were recorded as 1/2 the detection limit. All metal data, with the exception of Ag and Ni. were log,,, transformed to conect for heterogeneity of variances, whereas all organic contaminant. Ag. and Ni data vxere log|,,(x + I) trans- lyte detection limits ranged from .3.0 to 10 ng/g for aromatic hy- drocarbons, from < 1 .0 to 2.0 ng/g for PCB congeners, and from <1.() to 2.0 ng/g for chlorinated pesticides. Eighteen of the PCB congeners identified and quantified correspond to congeners ana- lyzed by the National Oceanographic and Atmospheric Adminis- tration's (NOAA) National Status and Trends (NS&T) Program designated congeners. Other organic compounds selected for analysis are also generally consistent with NOAA National Status and Trends mussel monitoring (NOAA 1989). A description of the full analytical protocol and accompanying performance-based QA/QC procedures are found in Sowles et al. (1997). and more comprehensively in Jones et al. (1998). Tissue samples were extracted by homogenization with an organic solvent and a drying agent. Solvent extracts were obtained by vacuum filtration, and biomatrix interference was separated from target analytes in extracts by size exclusion chromatography. Purified extracts were subjected to silica gel liquid chromatography, which provided a nonpolar PCB/chlorinated pesticides fraction and a po- lar chlorinated pesticide fraction. PCBs and pesticides were ana- lyzed by high-resolution dual column gas chromatography/ electron capture detection (HRGC/ECD). Following PCB and pes- ticide analysis, the two fractions were combined and the resulting extract was analyzed for aromatic hydrocaibons by high-resolution gas chromatography/mass spectrometry (HRGC/MS). Standard laboratory proceduies for metals incorporated method blanks, spike inatrix samples, duplicate samples, surrogate addi- tion, and standard oyster tissue (SRM 1566A). The method blanks were inserted as follows: three at the beginning of the run. one at the end, and six at various intervals during the run. Duplicate samples and matrix spike reco\'eries were conducted on I.S'r of the 1 V,> -, /MAli^E ^ NEWHAMPSfUR€/^\.^ 2 " Great k S^S'-) Figure 2. Location of 1998 Gulfwatch stations in New Hampshire. 1206 Jones et al. formed. At each site, arithmetic means were used to suminarize the results of replicate samples and are used in all subsequent tables and figures. In addition, geometric means were calculated for each metal and organic contaminant for comparison with other data sets. The confidence levels (CL) around the geometric mean were cal- culated as: TABLE 2. Inorganic and organic contaminants analyzed in mus.sel tissues from the seacoast of New Hampshire in 1W8. L| =antilog[(mean of logK) - t „5|„_||SQRT(,yi;,,,/;))] (1) and Inorganic Contaminants Organic Contaminants Metals Aromatic hydrocarbons Ag Naphthalene Al 1-Methyhiaphlhalene Cd 2-Methynaphthalene Cr Biphenyl Cu 2,6-Dimethylnaphthalene Fe Acenaphthylene Hg Acenaphthene Ni 2.3,5-Triniethylnaphthalene Pb Fluorene Zn Phenanthrene .Anthracene 1 -Methylphcnanlhrene Flouranthene Pyrene Benzo((i]anthracene Chrysene Benzo|/)|noaranthrene BenzolAlllouranlhrene Benzohdpyrene Benzoic Ipyrene Perylene Indenol 1 ,2.3-rJ|pyrene Dibenzo[(i./!]anthracene Benzo[g./(./]perylene Chlorinated Pesticides Hexachlorobenzene (HCB) gamma-hexachlorocyclolie,\ane (HCHi Heptachlor Heptachlor epoxide Aldrin Mirex ci.s-Chlordane trans-Nonachlor Dieldrin Alpha-Endosulfan bela-Endusulfan DDT and homologues 2.4'-DDE 2.4'-DDD 2.4'-DDT 4.4'-DDE 4.4'-DDD 4.4'-DDT PCB congeners PCB 8 PCB IX PCB 2S PCB 29 PCB 44 PCB 50 PCB .^2 PCB 6ft PCB 77 PCB .S7 PCB 101 PCB 105 PCB IIS PCB 126 PCB I2S PCB 138 PCB 133 PCB 169 PCB 170 PCB 180 PCB 187 PCB 195 PCB 206 PCB 209 Z., = antilog[(mean of logF) + t ,|^j„_||SQRT(.sf„„,//))] (2) s" = sample variance RESULTS AND DISCUSSION Trace Metal Conlainiiianis Table 3 contains the metal concentrations (arithmetic mean ± SD. (JLg/g DW) for mussels from all site composite (;i = 4) samples in I99S. Overall metal concentrations for indigenous mus- sels are given as geometric means and CI range (Table 3). Trace metals were detected at all sites except for Ag, which was below the detection liinit (0.1 p,g/g DW) at 10 of the 21 sites. No site had extremely high concentrations of any trace metal, although some sites like MAIH and most of the New Brunswick sites had elevated concentrations of some of the metals. Table 3 includes values for geometric means (GM) and GM +S5% CL from the 1990 NS&T Mussel Watch data (O'Connor & Beliaeff 1995) for comparison. Using the NS&T GM -i-85''f CL value as a reference for compara- tive purposes, two sites exceeded the Ag value, four sites exceeded the Cr value, six sites exceeded the Cu value. 17 sites exceeded the Hg value, two sites exceeded the Ni value, and six sites exceeded the Pb value out of the total of 21 sites. This suggests localized sources of these contaminants at those sites. However, for Cr, Cu, Pb, and especially Hg, more widespread elevated levels suggest possible regional sources of these contaminants. The tissue analysis for Al and Fe is included to serve as an indication of the degree to which sediment contamination may be present in mussel tissue (i.e., the gut). The fact that four of the six New Brunswick sites had relatively high concentrations of Al and Fe suggests that the mussel tissue samples contained elevated lev- els of sediment. The presence of sediment in the mussels was suspected in samples having elevated concentrations of some met- als (iron, aluminum, and associated metals) (Lobel et al. 1991, Robinson et al. 1993). Sites in the Bay of Fundy are dominated by extensive intertidal mudflats that can be subject to considerable resuspension during windy storm events. The concentrations of most metals were relatively evenly dis- tributed around the Gulf of Maine (Table 3). There were no ap- parent spatial trends; however, an occasional hot spot of elevated concentrations was observed. Figure 3 shows the concentrations of Ag, Pb. and Hg measured in the tissue of M. ediilis at the 1 998 sampling stations. The order in which sites are presented is from Cape Cod Bay, up along the coasts of New Hampshire and Maine, into the Bay of Fundy, and along the western shore of Nova Scotia. Silver concentrations ranged from below the detection limit (0.1 jig/g DW) to 1.82 ± 0.20 (NBCG) and showed strong geo- graphical hot spots of elevated concentrations in areas in each jurisdiction except New Hampshire along the Gulf of Maine (Fig. 3). The highest concentrations were observed in Massachusetts from Boston Harbor south to Sandwich and in New Brunswick around St. John Harbor. Elevated silver exposure concentrations have been shown to coincide with regions receiving municipal sewage (Sanudo-Wlhelmy & Flegal 1992. ten Brink et al. 1996). Because of silver's use in the photographic and jewelry industries, the coastal waters of Massachusetts are up to 1,000 times more concentrated in Ag than in other Gulf of Maine waters (Krahforst & Wallace 1996). The high levels observed at MASN, which is in an area with no known significant source of municipal waste, may Toxic Contaminants in Mytilvs edulis 1207 TABIi; 3. Tissue metal concentrations (fig/g dry weight, mean ± SD) for (iuH'Hatcli mussels in IW8. Station As Al Cd Cr Cu Ff Hg Ni Pb Zn MASN 0.93 ± l).4U 70 ± 20 1.93 ±0.49 1.13 ±0 23 6.37 ± 1.52 217±47 0.36 + 0.06 0.97 3.23 ± 0.75 15,6 ±0.1 MAIH 0.15 + 0.13 155 ±34 2.65 ± 0.34 2.23 ± 0.80 17.75 ±7.68 502 ± 1 23 0.55 ±0.01 1.35 ±0.10 26 ± 13 14.6 ±0.5 MAPR ND 167 ±60 1.80 ±0.22 2.35 ±0.91 8.15 ± 1.67 322 ± 78 0.48 ± 0.07 1.43 ±0.21 5.65 ± 1.84 14.8 ±0.3 NHDP ND 202 ± 39 2.80 ±0.28 2.95 ± 0,06 6.06 ± 0.69 385 ± 50 0.97 ± 0.05 1.70 ±0.20 3.02 ± 0.3 1 130 ± 14 NHGP ND 175 ±49 1.92 ±0.52 2.08 ± 0.56 4.70 ±1.27 358 ± 103 0.86 ± 0.09 1.35 ±0.24 3.32 ± 0.56 1 1 1 ± 25 NHLH ND 162 ±31 2.42 ±0.10 2.75 ± 0.97 5.12±0..33 400 ± 45 1,00 ±0.05 1.72±().17 4.65 ± 0.37 105 ± 17 NHNM ND 260 ± 54 1 .98 ± 0.37 2.32 ±0.43 6.55 ± 0.60 482 ± 99 0.79 ± 0. 1 2 1.24 ±0.20 5. 1 8 ± 1 .45 135 ±21 NHSS ND 192 ±34 2.25 ±0.51 2,30 ±0,1 8 6.12 ±0.49 385 ± 38 1,08 ±0.10 1.45 ±0.24 3. 15 ±0.48 128+ 10 MECC ND 298 ± 64 2.08 ±0.13 3, 18 ±0.69 7.20 ±0.67 528 ± 80 0.82 ±0.1 1 2.32 ± 1,08 5,75 + 0,70 135 ±24 MEKN 0.1U±0.1>4 117 ±26 2.08 ± 0 43 1,27 ±0,23 5.33 ±0.62 225 ± 42 0 41 ±0 09 1 I)2±(),I2 1 ,58 ± 0,40 53 ± 9.9 MEDM 0.15 ±0.06 292 ± 94 1.23 ±0.10 1,25 ±0,21 5.23 ±0.51 345 ± 101 0.34 ±0.10 1,03 ±0,15 1,75 ±0,19 67 ± 3.5 MEBB 0.07 ± 0.03 258 ± 103 0.%±0.10 1.25 ±0.26 13.5 ±1.9 380 ±109 0.52 ± 0.04 0,90 ±0,14 15.8 ±2,1 1 1 1 ± 1 3 NBNR ND 285 ± 45 0.74 ± 0.05 0.82 ±0.11 4.58 ±0.39 358 ± 43 0.22 ± 0.03 0,96 ± 0,05 0,58 ±0,1 2 65 ± 3.4 NBCH ND 175 ±21 0.88 ±0.1 7 0.69 ± 0.06 5. 15 ±0.47 245 ± 24 0.22 ±0.05 0.87 ± 0.08 0,58 ±0,1 7 68 ±8 NBLN 0.05 ± 0.07 777 ± I4S 1.50 ±0.10 17.5 ±31.5 12±4 678 ±315 0. 1 1 ± 0.00 9.8 ± 17.4 1.6 ±0.4 82 ±8 NBLB 0.04 ± 0.03 835 ± 53 1.50 + 0.10 2.5 ± 0.5 13 ±2 609 ±61 0.14 ± 0 1)3 1.4 ±0,1 2.7 ±0.1 85 ± 1 1 NBTC 0.16 ±0.13 2925 ± 1870 2.50 ± 0.30 12.9 ±7.8 29 ±8 2131 ±764 0.33 ± 0.03 6,6 ±3,3 2.3 ±0.8 110 ±24 NBCG 1.82 ±0.20 793 ± 1 79 2.00 ± 0.30 4.0 + 1 .0 29 ± 10 696 ± : 60 0.29 ± 0.02 1,9 ±0.4 2.3 ± 0.3 139± 10 NSBC ND ISO ±35 1.70 ±0.00 1.33 ±0.12 6.73 ±0.21 3 1 3 ± 46 0.37 ± 0.03 1,20±0.10 1.67 ±0.1 5 84 ±8 NSCW 0.08 ± 0.05 388 ± 46 2.73 ±0.39 1.70±0.14 5.70 ± 1.15 522 ± 59 0.45 ± 0.07 1.88 ±0.26 3.40 ± 0.39 87 ±21 NSDI 0,08 ± 0.05 338 ±31 1.60 ±0.18 1 .43 ± 0.22 5.33 ± 1 .44 478 ± 39 0.46 ± 0.06 1.63 ±0.15 2.70 ± 0.22 94 ± 16 Geometric mean 0,11 270 1 63 1.78 7.61 416 0.42 1.27 2.63 96 LCL to UCL 0.03 to 0.41 141 io5l7 1.11 lo 2.4 0.96 to 3.29 4.62 10 1 2.6 271 to 638 0.24 to 0.76 0.84 to 1.91 1.31 Io5,29 69 to 134 NS&T Mu-,scl Wau h 1990 data (O'Connor & Be laelf 1995) Geonielrjc mean 0.17 none 2.8 1.7 8.9 none 0.09 none 1,9 130 CM +1 SD 0.58 none 5.7 3 11 none 0.24 3.3 4,3 190 The geometric mean and CL tor all mussels are ^i\en; u = A leplicuites pci ND = not detected. be a function of transport and iJeposition of sewage-iJerived par- ticles (Bothner et al. 1993) that are sequestered in Cape Cod Bay and accutnulated by mussels. The concentration of lead ranged from a value of 0.58 ± 0.12 (xg/g DW at two New Brunswick sites (NBNR and NBCH) to 26 ± 13 jxg/g DW at Boston Inner Harbor (MAIH) (Table 3. Fig. 3). Lead levels at Boothbay Harbor. ME (MEBB) were also much higher than most other sites. Sediment particles containing Pb may be transported to Boothbay Harbor from the Kennebec- Androscoggin watershed (Larsen & Gaudette 1995). Lead concen- trations were generally higher in the southwestern areas compared to the northern and eastern sites (Fig. 3). Mean concentrations of Pb in mussels from coastal regions generally range from I to 16 |j.g/g DW (Fowler 1990). All six of the NH sites sampled in 1998 exceed the Maine reference concentration (ME-RM; .Sowles 1993) of 2.6 ±1.1 p.g/g DW. but no site exceeded the Maine high \alue (ME-HV) (6.00 (jig/g DW). MAIH is located in a highly urbanized harbor and subjected lo heavy industry, marine transport activities. and municipal waste discharges. Elevated lead in the New Hamp- shire sites may be related to the close proximity of the sites to the Portsmouth Naval Shipyard where waste plating sludge and lead batteries, respectively, were disposed and stored (NCCOSC 1997). The potential for the shipyard to be a source of lead to estuarine biota was demonstrated in July 1999 when significant ann)unts of soil from a contaminated site at the Shipyard, containing as much as 14.2 mg Pb/g soil DW. was discovered to be eroding into the Piscataqua River (Cohen 2000). Mussels collected from sites in close proximity to the eroding soil contained Pb concentrations as high as 199 jjig/g DW. The concentration of nietcuiy in nutsscl tissue ranged from a value of 0.11 [xg/g DW at NBLN lo 1.08 ± 0.10 |a.g/g DW at NHSS (Table 3. Fig. 3). Mercury exceeded the NS&T CM 4-85% CL of 0.24 p.g/g DW at 17 of the 21 sites. The New Hampshire sites are markedly higher than sites in other jurisdictions. There are several known historical mercury sources in the New Hampshire Seacoast, including some that are suspected to be related to the Portsmouth Naval Shipyard (NCCOSC 1997) and, especially, the PSNH Schiller Station (NHSS) on the Piscataqua River, where mercury steam was used from 1950 to 1968 (Nelson 1986). Analysis of the mussel tissue concentrations of Hg revealed that there was a sig- nificant difference in Hg concentrations between NHSS and all other New Hampshire sites except NHLH. Mean values of Hg in Mytihis spp. from coastal regions worldwide range from 0. 1 to 0.4 p.g/g DW (Kennish 1997). but can be much higher in areas like the southwest Pacific, where sites average as much as 2.7 |j.g Hg/g DW (Fowler 1990). In a review of the first five years of the Gulfwatch program, tissue concentrations of Hg were discussed as being unusually high and a possible concern (Jones et al. 1998). Recent studies have shown that a mercury problem exists in freshwater systems of the northeast United States and maritime provinces of Canada (Welch 1994. DiFranco et al. 1995. Evers et al. 1996). About 479f of mercury deposition in the region origi- nates from sources within the region. 30"/? originates from U.S. sources outside the region, and 23% originates from the global attnospheric reservoir (NESCAUM 1998). On June 8. 1998, the New England governors and eastern Canadian premiers agreed to cut regional mercury emissions from power plants, incinerators, and other sources in half by the year 2003. However, until recently few large coastal systems have been known to be impacted hy Hg pollution. Atmospheric mercury deposition measurements made at New Castle. NH. at the mouth of Portsmouth Harbor, showed -8 ng/nr total mercury was deposited during 1998 (MDN. unpub- 1208 Jones et al. 45 ' 40 30 25 :o LEAD p n flnnni^nn^r. n n Fr m m ("I H ^^s;^^.?^^ ^v ^N. ^■: ,>>>V".<^^<<^V>%-VV^ Figure 3. Distribution of silver, lead, and mercury tissue concentra- tion.s (arithmetic mean ± SD, pg/g dry weiglit) in mussels at all Gulf- watch stations in 1998. lished). The New Castle site, along with two other Maine coastal sites in Casco Bay and Acadia National Park, showed somewhat elevated total mercury atmospheric deposition compared to nearby, upstream inland sites. Other areas in the Gulf of Maine have elevated (3-20 ng/g) sediment mercury concentrations (ten Brink et al. 1997). including the Penobscot River near Onington. where permitted and accidental discharges from the Holtra-Chem facility have resulted in sediments having much higher (>100 p,g/ g) Hg concentrations (MEDEP. unpublished). Thus, data on mus- sel tissue mercury levels may have added importance in assessing cun-ent contamination problems and the effects of discharge re- duction efforts related to Hg in the future. Organic Contaminants The total concentration of detectable polynuclear aromatic hy- drocarbons (iPAH^j). polychlorinated biphenyls (TPCB^j). and organochlorine pesticides (VTPEST.v) measured in mussel tissue samples of indigenous mussels are presented in Table 4. Indi\ idual analyte concentrations of each compound class are not shov\ n. In 1998. as in previous reports (Sowles et al. 1996. Chase et al. 1998. Jones et al. 1998). IDDT,, and its degenerative metabolites were the main contributors to total detectable pesticides (data not shown). Analytes within each category of organic contaminant were detected at most sites, except for ZPAH:,4 at MEDM and SPCB.4 at NBCH, NSBC. and NSCW. There were much wider ranges in concentrations of organic compared to trace metal con- taminants. None of the Gulfwatch sites had organic contaminant concentrations that exceeded any human health criteria. For example, the U.S. FDA Action Limit for PCB is 2 |xg/g DW (USFDA 1990). and the concentration for the highest Gulfwatch site was 0.74 |j,g/g DW at MAIH. which is well below this limit. A clear pattern of higher concentrations in the southwestern Gulf compared to the northeastern Gulf is apparent for all three types of organic contaminants (Table 4). This pattern can be seen in Figure 4. which shows the concentrations of XPAH^j and ZPCB,j measured in M. ediilis tissue from the 1998 sites, pre- sented from south to north. The mean of the XPAH-,4 concentra- tions ranged from not detected (ND) at MEDM to 3.3.^0 ng/g DW at MAIH. Mean concentrations of SPAH.j at all but a few of the Maine and New Brunswick sites were as high as those reported from areas mlluenced by oil spills and municipal sewage outfall (148 ng/g in Rainio et al. 1986. 63-1.060 ng/g in Kveseth et al. 1982). However, only samples from the Boston Harbor station (MAIH) had concentrations as high as in industrialized areas af- fected by coking operations, as documented in Sydney Harbor, NS (1,400-16,000 ng/g) (Kieley et al. 1988) or smelting operations in Saudafijord. Norway (5.1 1 1-225.163 ng/g) (Bjorseth et al. 1979). The wide range in ZPAH^j concentrations suggests that sites with elevated concentrations may be in close proximity to recent or ongoing sources. Table 4 includes values for overall geometric mcMs (GM) and GM -^85% CL from the 1990 NS&T Mussel TABLE 4. Tissue organic concentrations (ng/g dry weight, mean ± SD) lor Gulfwatch mussels in 1998. Chlorinated Station PAH PCB Pesticides MASN MAIH MAPR NHDP NHGP NHLH NHNM NHSS MECC MEKN MEDM MEBB NBNR NBCH NBLN NBI.B NBTC NBCG NSBC NSCW NSDI Gecimetnc mean LCL to UCL NS&T Mussel Watch 1990 data (O'Connor & Beliaeff 1995) Geometric mean 270 110 -^6* GM -^ 85% CI 1020 470 120* The geometric mean and 8,S<7r confidence levels (CL) for all mussels are given below. » = 4 replicates per site. ND = Not Detected. * Value for DDT only. 13.0 ±2.0 27.2 ±7.1 29 ± 3.0 3330 ± 223 740 ± 39 133± 14.5 553 ± 33 131 ±7.8 60 +11.8 231 ±29 32 ± 8.5 16.1 ±2.6 164 ± 12.6 25.5 ± 1.5 14.1 ± l.S 78 ± 11.6 12.4 ±1.8 102 ±0.83 645 ± 55 65 ±9.1 67 + 9.7 187 ±47 }Q ± 5.4 14.6 ± 1.9 1 99 ± 25 42 ±7.4 15.4 ±2.3 .S8+ 19.6 16.7 ±3.8 5.2 ± 0.49 ND 3.7 ± 0.34 4,6 ±0.5(1 11 20 ±58 44 ± 5.5 61 ±2.2 106± 11.3 1.6 ±0.18 7.0 ±0.97 24± 13.1 ND 6.5 + 2. 1 13.8 ±3.0 6.6 + 0.50 5.3 ± 1.6 16.2 ±4.8 6.7 ± 1.3 5.8 ± 2.5 164 ± 12.4 33 ± 5.2 13.8 ± 1.0 229 ±7.7 38 ± 2.3 37 ± 1.6 133 ± 17.8 ND 5.1 ± 1.6 138 ± 160 ND 2.5 ±0.42 106+ 13.9 3.5 ± 0.43 6.0 ± 1.2 127 23.5 13.6 40 to 401 8.8 to 62.6 5.9 to 31.6 Toxic Contaminants in Mytiu's kduus 1209 3500 3000 2500 2000 1500 1000 500 0 .nnn n n TOTAL PAHs , n „ - n n n, ,» n n v^ /:?=^W^j^ ^s'y/ ^f^^s^^V^" #V^ ,s^^>* /;<#,^y:/ ^/-y ///#V/ /^-^v* Pifjure 4. Distribution of total PAH and PCB tissue concentrations (aritlimttic mean, (ig/g dry weight) in mussels at all Gulfwatch stations in 1998. Watch data (O'Conimr & BeliaetT 1995) for comparison. Two sites, MEBB and MAIH. exceeded the N.S&T GM +85% CL for PAH (1,020 ng/g DW), and two other sites, MAPR and NHNM. were at concentrations over half of this value. MAIH. and to a lesser degree MAPR, has been subject to high levels of all types of contamination. Relatively uncontaniinaled mussels deployed in 1995 had -1.570 ng PAH/g DW after 60 days in cages at MAIH (Chase et al. 1996). MEBB (Boothbay Harbor, ME) had not been sampled since 1991 when organic analyses were not conducted. However, analysis of tissue samples showed mussels from MEBB to contain elevated levels of trace metals, especially Pb and Zn (Jones et al. 1998). Thus, it is a site that is apparently subject to a variety of contaminants, possibly because it is downstream from the Kennebec-Androscoggin watershed. The possible source of the PAHs at NHNM is not known. In contrast, mussels at NHDP. which was impacted by the 1996 Provence oil spill at the .Schiller Station oil terminal, and at NHSS. which is in clo.se proximity to the oil terminal, had much lower ZPAH^j concentrations than at NHNM in 1998. Examination of the individual PAHs detected at NHNM reveals a marked dominance of higher molecular weight and nonalkylated PAHs (Chase et al. 2001 ). This pattern was con- sistent for all 1998 New Hampshire sites and suggests that the PAHs may be from pyrogenic sources, as opposed to recent pe- troleum sources. The pattern also strongly suggests that the sources may be historical, or reflect past exposure. Lower molecular weight PAHs degrade faster (Shiaris 1989) and are more mobile in the envirt)niiicnt. and bivahes lend to metabolize and excrete higher molecular weight P.AHs at slower rates (Widdows & Donkin 1992). Sediments from sites in North Mill Pond, especially upstream sites, had XP-AH17 concentrations ranging from <690 to 23,600 ng/g DW (ANMP 1998). It is possible that PAH- contaminated sediments from upstream sources could be taken up and accumulated by mussels at the downstream NHNM site, es- pecially during high How or storm events at low tide. The concentrations of ZPCB24 ranged as widely as the ZPAH,4 concentrations. The mean of XPCB^j ranged from ND at NBCH, NSBC. and NSCW to 740 ng/g DW at MAIH (Table 4). Table 4 also shows the overall GM -H85'7r CL of iPCB^j concen- trations for all 1998 Gulfwatch sites. The same pattern of elevated concentrations in the southwest compared to the northeast sites can be seen. Two Massachusetts sites had the highest concentrations, MAIH (741 ng/g DW) and MAPR (131 ng/g DW). As described previously. MAIH is a site in Boston's Inner Harbor and has been subject lo high levels of all types of contamination. Relatively uncontaminated (-37 ng ZPCB-,4/g DW) mussels deployed in 1995 had -361 ng PCB/g DW after 60 days in cages at MAIH (Chase et al. 1996). In 1995. XPCB,4 concentrations for MAPR. a site north of Boston Harbor, were the highest (131 ng/g DW) of any Gulf of Maine mussels sampled (Chase et al. 1996). Most of the New Hampshire sites exhibited relatively uniform and some- what elevated concentrations relative to the Gulfwide geometric mean. The ZPCB24 concentration at NHNM was the third highe.st of the 1998 sites at 65 ng/g DW. As with PAHs. the source of the PCBs in North Mill Pond is not known. Analysis of sediments from North Mill Pond conducted on samples collected in 1997 showed no detectable PCBs (ANMP 1998), although detection limits 02.400 ng/g DW for seven Aroclors) were relatively high for that study. Sites in Portsmouth Harbor have had relatively high sediment PCB concentrations compared to other areas in the Gulf of Maine, except for Boston Harbor (ten Brink et al. 1997). The XPCB24 concentration in mussels at MEBB (44 ng/g DW) was also elevated compared to other 1998 sites, as was the case for other contaminants already mentioned. It is instructive to look at data from New Hampshire as an example of how a regional database can be useful for interpreting local study results. Organic contaminant concentrations at NHNM are higher than at any other New Hampshire site (Table 4. Fig. 5). and this is a previously undocumented concern. Local community groups have focused on restoring this site, and it has been sus- pected to be a relatively contaminated site (ANMP 1998). How- ever, in the context of the overall 1998 Gulfwatch database. NHNM has much lower XPAH-,4 and ZPCBi4 concentrations than at MAIH and either MAPR (IPCB.4) or MEBB (IPAH,4). An- other characteristic of the New Hampshire data is that concentra- tions of both SPAH-,4 and XPCB24 are uniformly elevated com- pared to Gulfwatch sites north and east of New Hampshire, sug- gesting that there may be common sources for these relatively localized sites. Thus, interpretation of the New Hampshire data benefits greatly from comparison with the overall Gulfwatch da- tabase. Coordination of sampling and use of common laboratories for tissue analyses is a critical aspect oi both programs that allows for added merit within the comparisons. Another use of the Gulfwatch program in New Hampshire has been to document the effects of an oil spill on tissue PAH con- centrations. On July I, 1996. there was an oil spill from the vessel Provence into the Piscataqua River. Approximately 1.000 gallons of #6 fuel oil was dispersed with water currents into nearby areas 1210 Jones et al. Tfin 1 600 500 TOTAL PAHs 400 300 200 K 1 1 1 i 100 ^iv.^ NHLH NHSS NHGP NHUP NHNM MECC 60 50 TOTAL PCBs 40 i 30 1 20 f 10 1 0 l! 1 NHLH NHSS NHGP NHDP NHNM MECC Figure 5. Distribution of total PAH and PCB tissue concentrations (arithmetic mean, ng/g dry weight) in mussels at New Hampshire (Julf- watch stations in 1998. of the Great Bay Estuary. Fuel oils are known to contain a variety of PAHs, especially 2- to 4-ring PAHs. although hundreds of or- ganic compounds, including larger PAHs. are present in all crude oils (Kennish 1997). The Gulfwatch station at Dover Point (NHDP). located at the continence of the Piscataqua River and Little Bay approximately 2.5 miles upstream of the source of the oil spill, was sampled in 1994. Sampling was again conducted in July and October 1996 to determine if contaminants from the spill were taken up by mussels and to find the degree of elimination of the contaminants over time. The 1994 data serve as useful back- ground information for assessing the degree of residual exposure of the 1996. 1997, and 1998 mussel tissue samples to the oil spill contaminants. ZPAH^j concentrations were determined in mussel tissue samples collected in 1994, in 1996 on July 16 and October 1.16 days and 3 mo, respectively, after the spill, and in 1997 and 1998 (Fig. 6). ZPAH^4 concenti'ations in the 1996/16-day sa)nples were much higher than in 1994, as well as in comparison to other sites in New Hampshire. Elevated levels of PAHs were also ob- served in oysters {Crassostrea virginka) collected at another site impacted by the spill (Chase et al. 1997). The average iPAH^j co)icentrations were 639. 298. 266. 238. and 187 ng/g DW for the 16-day, 3 mo. 1997. 1998. and 1994 samples, respectively. Thus, the total PAH concentration has decreased greatly from 1996. less drastically from 1997 to 1998. and levels have almost returned to prespill (1994) levels. Tempiiral Varialion in Contaminant Ciincvntratiuns Sampling has occurred in up to five previous years at the five "benchmark" Gulfwatch stations to provide data for temporal trend analysis (Jones et al. 1998). Data for MECC. a benchmark station sampled since 1993. are presented here as one illustration of tem- poral trends in the Gulf of Maine (Table 5). There were few obvious trends in concenti'ations of tiace metals and organic con- taminants. Trends at the other four benchmark sites also showed variable results, with most discernable trends being decreases in contaminant concentrations. There has been significant variation for between-year comparisons for some contaminants (Chase et al. 2001). Acci'plahlc Levels and Standards of Mnssel Contamination Despite the wealth of information on the effects of toxic Ciin- taminants on a variety of species, limited information is available on observed human health effects resulting from the consumption of chemically contaminated shellfish. While documented epide- miological effects may be limited, the results of laboratory assays and isolated occurrences of acute human poisonings are respon- sible for focusing attention on human health impacts from eating chemically contaminated marine fish and shellfish. For example, in New Hampshire there are currently human consumption advi- sories for Hg and PCBs (NHDES 1998. NHEP 2001). The advi- sory for Hg is based on elevated Hg levels in inland lakes and rivers and is for all freshwater fish. For marine waters, there is a consumption advisory for both lobsters and bluefish based on el- evated levels of PCBs. The PCB advisories for bluefish and lob- sters are based on studies done in 1987 and 1991. respectively. Published tolerance or action levels for PAHs in commercial marine species are not available in Canada or in the United States. Closure of commercial fisheries as a result of high contamination levels has been dealt with on a case-by-case basis in marine areas where PAH contamination may be a human health concern. In general, most concentrations reported in the literature are on a wet weight basis in contrast to Gulfwatch dry weight values. To fa- cilitate general comparisons with Gulfwatch values, an average moisture content of 85% has been applied to wet weight health values to derive dry weight equivalents. All Gulfwatch organic concentrations are within acceptable concentrations for those com- pounds that have established FDA Action Limits in fish and shell- fish. PCB concentrations found in Gulfwatch mussels (Table 4). including concentrations measured in MAIH mussels (0.74 ± 0.003 |xg/g DW) are less than the action level of 13 p.g/g dry weight (USFDA 1990. CSSP 1992). The action level for the pes- ticides dieldrin. aldrin. chlordane. heptachlor. and heptachlor ep- oxide is 2.0 |xg/g dry weight (USFDA 1990). Only dieldrin and chlordane were detected in the 1998 mussel survey, but at con- centrations barely above detection limits, which are orders of mag- nitude below the action levels. The total DDT concentrations found are several orders of magnitude below the action level of 33 (jLg/g dry weight (USFDA 1990. CSSP 1992). Canadian limits for agricultural chemicals exclusive of DDT are 0.67 |jig/g DW. As presented in Table 6. admissible levels of methyl mercury, expressed as |jig Hg. are less than 6.7 p,g/g dry weight, or I p.g/g wet weight in the United States (USFDA 1990). and less than 3.3 ixg/g dry weight, or 0.5 jjig/g wet weight in Canada (CSSP 1992). The highest concentration of mercury found in the 1998 Gulfwatch study was 1.20 |xg/g DW, in one replicate sample from the Schiller Toxic Contaminants in Mytilus f.dul/s 1211 700 Months Figure 6. l.oiig-lcrm effects of a July 1W6 oil spill on PAH contamination in mussels at Dover Point, NH. First PAH concentration sample collected in October 1994. Station. New Hampshire, which is well below both federal action concentrations. A series of FDA "Guidance Documents" (USFDA 1993) for cadmium, chromium, lead, and nickel was released in the United States to complement the FDA Mercury Action Level. These "alert" levels are guidelines and by themselves do not warrant the issuance of health advisories. In Table 6. guidance concentrations are reported on both wet weight and dry weight bases and are compared to the highest observed concentration in any single rep- licate analyzed in the 1998 Gulfwatch Project. All nickel, chro- mium, and cadmium concentrations in 1998 Gulfwatch mussels were well below the guideline values. However. Pb concentrations were above the FDA guideline alert level of 11.5 p.g/g DW at MAIH and MEBB. and are thus of regional and local concern. The highest observed concentrations from the 1998 Gulfwatch data for other trace metals for which there is no guideline or action limit are included in Table 6. This highlights hot spots of localized elevated contamination as well as sites where elevated levels may also be TABLE 5. Long-term tissue contaminant concentrations (arithmetic mean ± SD) for a Gulfwatch baseline station: Clark Cove, ME (MECC). Contaminant 1993 1994 1995 1996 1997 1998 Metals ( ng/g dry weight! Ag 0.10 ±(1.05 0.07 ±0.03 0,06 ± 0,03 0.05 ± 0.02 0.04 ± 0.02 0.03 ±0.01 Cd 2..18 ± (J.27 1,33 ± 1,50 1,37 ±0.87 1.26 ±0,27 0.94 ± 0.50 0.74 ± 0.44 Cr .\31 ± 1.28 2.29 ± 1 .44 2.08 ± 0.09 1 .69 ± 0.62 1 .33 ± 0.67 1.08 ±0.52 Cu 7.51 ±0.87 4.19 ±4,70 4.31 ±2.72 3.98 + 0.87 2.97 ± 1 .56 2.34+1.40 Pb 5.35 ±2.18 3.77 ± 2.24 3.39+1.50 2.72 ± 1.04 2.16 ±1.08 1.75 ±0.83 Hg 0.74 ± 0.06 0.40 ± 0.48 0.42 ± 0.28 0.40 ± 0.08 0.30 ±0.15 0.23 ±0.14 Ni 2.60 ± 0.2 1,40 ± 1.7 1 .47 ± 0.99 1.39 ±0.29 1.04 + 0.54 0.81 ±0.49 Zn 126 ± 17 71 ±77 73 ±45 66 ± 15 50 ±26 39 ±23 Al 187±81 134 ±75 119 ±52 95 ± 38 76 ± 38 62 ±29 Fe 535 ± 138 336 ±281 322 ± 1 64 276 ± 78 210± 110 169 ±91 Organics (ng/g dry weight) PAH 1 ,54 ± 47 137 ± 10 158 + 39 203 ± 22 147 ± 19 200 ± 26 PCB 70 ± 1 1 67 ±5 35 ±10 38 + 2 37 ±8 42 ±8 Pesticides 11.1 ±5,3 12,5 ± 1,3 13.8 ± 1.0 7.2 ± 1.5 1 5 ± 5 16±2 1212 Jones et al. TABLE 6. Comparisons of U.S. Food and Drug Administration federal guideline and action concentrations for trace metals with the (iulfwatch results, and the highest observed 1998 Gulfwatch concentrations. Highest Observed 1998 (iulfwatch \ alue Metal Wet Weight Dry Weight (l)r> Weight 1 Guideline concentration Cd 3.7 |xg/g 25 |xg/g 3.1 |j,g/g Ci* 13|JLg/g 87 [JLg/g 19.7|xg/g 4.2 M-g/g Pb 1 .7 |xg/g 1 1 .5 |ig/g 37 (jLg/gt 18M.g/gt Ni* 80 |j.g/g 533 jjLg/g 9.4 (JLg/g 2.1 (JLg/g Federal action concentration Hg (US) 1 .0 ^g/g 6.7 |xg/g 1.2M.g/g Hg iCA\ iy5 ixg/o 3.3 ji-g/g No federal guideline or action concentration Highest observed 1998 Gulfwatch value Ag* 2.0/1.3 Al* 5640/440 Cu* 43/26 Fe* 2950/610 Zn 310 Location Cornwallis, NS Tin Can Beach. NB Little Harbor. NH Boston Inner Harbor, MA Boothbay Harbor. ME Tin Can Beach. NB Cornwallis. NS Schiller Station. NH Coast Guard wharf, NB/Sandwich, MA Tin Can Beach, NB/Cornwallis. NS Coast Guard wharf. NB/Boston. Inner Harbor, MA Tin Can Beach. NB/Boston Inner Harbor, MA Boston Inner Harbor, MA * Two values for some metals are presented because the highest values are all from New Brunswick sites where tissue samples tended to contatn excessive inorganic sediment, as indicated by elevated Al and Fe concentrations. t Gulfwatch highest values e,\ceed U.S. FDA Guideline concentration. Four replicate sample concentrations ranged from 30-37 (jLg/g DW at MAIH and 13-18 jjLg/g DW at MEBB. associated with excessive sediment in tissue samples (New Bruns- wick sites). Many of the areas with ele\ated concentrations are currently under more detailed investigations by other agencies. ACKNOWLEDGMENTS The authors are grateful to the following individuals: Noel Carlson. Andrea Riley. Deb Lamson at UNH/JEL. Joanne McLaughlin at NHOSP. and Andrea Bowman. Amber Currier. Paul Currier. Rob Livineston, and Eric Williams at NHDES. All authors on previous Gulfwatch annual reports are gratefully ac- knowledged for their contributions to the conceptual framework and some of the program-consistent text for this report. Generous financial support for the program came from NOAA. the New Hampshire Department of Environmental Services, the Gulf of Maine Council on the Marine Environment, and Environment Canada. Community groups in Canada. ACAP St. John and ACAP Charlotte Waterways, are also acknowledged for their help with sampling and supporting the 1998 program. LITERATURE CITED ANMP (Advocates of North Mill Pond I. 199S. The state of the North Mill Pond, Portsmouth, NH. Portsmouth, NH: New Hampshire Estuaries Project. Bjorseth. A„ J. Knutzen & J. Skei. 1979. Determination of polycyclic aromatic hydrocarbons in sediments and mussels from Saudafijord, W. Norway, by gas capillary chromatography. Sci. Toial Environ. 13:71- 86. Bothner. M. H.. M. Bucholtz ten Brink, C. M. Paramenter. W. M. d'Angelo & M. W. Doughten. 199.^. The distribution of silver and other metals in sediments from Massachusetts and Cape Cod Bays. U.S. Geological Survey Open-File Report, pp. 93-725. Cantillo. A. Y. 1998. Comparison of results of mussel watch programs of the United States and France with worldwide mussel watch studies. Mar. Pollut. Bull. 36:712-717. Cha.se, M.. P. Hennigar, J, Sowles, S. Jones, R. Crawford, G. Harding, J. Pederson. C. Krahforst. D. Taylor & K. Coombs. 1998. Evaluation of Gulfwatch 1997 — Seventh year of the Gulf of Maine environmental monitoring plan. Augusta, ME: The Gulf of Maine Council on the Marine Environment. State Planning Office, Chase, M. E., S. H. Jones, P. Hennigar, J. Sowles, G. C. H. Harding. K. Freeman, P. Wells, C. Krahforst, K. Coombs, J. Pederson & D. Taylor. 2001. Gulfwatch: Monitoring spatial and temporal patterns of trace metal and organic contaminants in the Gulf of Maine ( 1991-1997) with the blue mussel Mytilus edulis L. Mar. Poll. Bull. 42:491-505. Chase, M.. S. Jones. P. Hennigar. J. Sowles. K. Coombs. R. Crawford. G. Harding, J. Pederson & D, Taylor. 1997. Evaluation of Gulfwatch 1996 — Sixth year of the Gulf of Maine environmental monitoring plan. Augusta, ME: The Gulf of Maine Council on the Marine Environment. State Planning Office. Cha.se, M., S. Jones, J. Sowles, P. Hennigar. G. Harding. R. Crawford, J. Pederson, K. Coombs, D, Taylor & W. Robinson. 1996. Evaluation of Gulfwatch 1995 — Fifth year of the Gulf of Maine environmental moni- toring plan. Augusta, ME: The Gulf of Maine Council on the Marine Environment. State Planning Office, Cohen, D. J. 2000, Interim offshore monitoring program Round I data package for Portsmouth Naval Shipyard, Kittery. Maine. Lester, PA: Northern Division, Naval Facilities Engineering Command. Crawford. R. & J. Sowles, 1992. Gulfwatch Project— Standard procedures Toxic Contaminants in Mytjlus edulis 1213 for field sampling, measurement and sample preparation. Gulfwatch Pilot Period 1991-19^2. Augusta. ME: The Gulf of Mame Council on the Marine Environment. State Planning Office. CSSP (Canadian Shellfish Sanitation Program). 1992. .Action levels and tolerance.s and other values for poisonous or deleterious substances in seafood. Appendix 111. Manual of operations. Ottawa. ON: Fisheries and Oceans and Environment Canada. DiFranco. J.. L. Bacon, B. Mower & D. Courtemanch. 1995. Fish tissue contamination in Maine Lakes — Data report. Augusta. ME: Maine De- partment of Environmental Protection. Dow. D. & E. Braasch 1996. The health of the Gulf of Maine ecosystem: Cumulative impacts of multiple stressors. In: D. Dow & E. Braasch. editors. Regional .Association for Research on the Gulf of Maine (RAR- GOM) Report 96-1. April .^0. 1996. Evers, D. C. P. Reaman. J. Kaplan & J. Paruk. 1996. North American Loon Biomonitoring Program: 1995 field season report — 1989-1995 comprehensive report. Paradise. MI: Biodiversity. Inc. Fowler. S. W, 1990. Critical review of selected heavy metal and chlori- nated hydrocarbon concentrations m the marine environment. Mar. Eiivinin. Rc'\. 29: 1 . Jones. S. H,, M. Chase, J. Sowles. P. Hennigar. W. Robinson, G. Harding. R. Crawford. D. Taylor. K. Freeman. J. Pederson. L. Mucklow & K. Coombs. 1998. Evaluation of Gulfwatch: The first five years. Augusta. ME: The Gulf of Maine Council on the Marine Environment, State Planning Office. Kawaguchi. T., D. Porter. D. Bushek & B. Jones. 1999. Mercury in the American oyster Crassoslrea virginica in South Carolina, LIS. A., and public health concerns. Mar. Poll. Bull. 38:.^24-.^27. Kennish, M. J. 1997. Practical handbook of estuarine and marine pollution. Boca Raton: CRC Press. Kieley. K. M.. P. A. Hennigar. R. A. F. Matheson & W. R. Ernst. 1988. Polynuclear aromatic hydrocarbons and heterocyclic aromatic com- pounds in Sydney Harbor. Nova Scotia. A 1986 survey. Environment Canada, Environmental Protection Service, Atlantic region Surveil- lance Report EPS-5AR-88-7. Dartmouth. NS: Environment Canada, Krahforst, C. F. & G. T. Wallace. 1996. Source estimates and the parti- tioning of silver and other trace metals in Massachusetts coastal waters. In: A. W. Anden & W. T. Bober. editors. 4th International Conference on Transport. Fate, and Effects of Silver in the Environment. Madison. WI: University of Wisconsin Sea Grant. Kveseth. K.. B. Sortland & T. Bokn. 1982. Polycyclic aromatic hydrocar- bons in sewage, mussels, and tap water. Chertiosphere 1 1 :623-639. Larsen. R. P. & H. E. Gaudette. 1995. Spatial and temporal aspects of sedimentary trace metal concentrations in mid-coast Maine. Mar. Poll. Bull. 30:347^44. Lobel, P. B., S. P. Belkhode. S. E. Jackson & H. P. Longerich. 1991. Sediment in the intestinal tract: A potentially serious source of error in aquatic biological monitoring programs. Mar. Environ. Res. 31:163- 174. NAS (National Academy of Sciences). 1980. The international mussel watch. Washington, D.C.: National Academy of Sciences. NCCOSC Naval Command (Control, and Ocean Surveillance Center). 1997. Estuarine ecological risk assessment for the Portsmouth Naval Shipyard. Kittery, ME, vol, I: Technical Report. Revised draft final. Lester, P.\: Northern Division, Naval Facilities Engineering Com- mand. Nelson, J. I.. Jr. 1986. The presence of mercury, chromium, lead, nickel, copper, and zinc in the Great Bay Estuarine System, New Hampshire. M.S. thesis. Dept. of Civil Engineering. Univ. of New Hampshire. Durham. NESCAUM. 1998. Northeast States/Eastern Canadian Provinces Mercury Study, February, 1998. NHDES (New Hampshire Department of Environmental Services). 1998. State of New Hampshire: 1996 Section 305(b) Water Quality Report. Concord. NH: NH Department of Environmental Services. NHEP (New Hampshire Estuaries Project). 2001. .A technical character- ization of estuarine and coastal New Hampshire. S. H. Jones, editor. Portsmouth. NH: New Hampshire Estuaries Project. NOA.'K (National Oceanic and Atmospheric Administration). 1989. A sum- mary of data on tissue contamination from the first three years (1986- 1988) of the mussel watch project. National Status and Trends Program for Marine Environmental Quality Progress Report. NOA.A Technical Memorandum NOS OMA 49. NOAA (National Oceanic and Atmospheric Administration). 1991. Mussel watch worldwide literature survey — 1991. Rockxille. MD: NO.A.A Technical Memorandum NOS ORCA 63. O'Connor. T. P. 1992. Recent trends in coastal en\ironmcnlal quality: Results from the first five years of the NO.AA Mussel Watch Project. Silver Springs, MD. NOAA/NOS. O'Connor. T. P. & B. Beliaeff 1995. Recent trends in coastal environ- mental quality: Results from the Mussel Watch Project. Silver Springs. MD. NOAA/NOS. Rainio, K.. R. R. Linko & L. Ruotsila. 1986. Polycyclic aromatic hydro- carbons in mussels and fish from the Finnish Archipelago Sea. Bull. Environ. Conlum. To.xicol. 37:337-343. Robinson, W. E.. D. K. Ryan & G. T. Wallace. 1993. Gut contents: A significant contaminant of Mytilus edulis whole body metal concentra- tions. .4/r/i. Environ. Conlam. To.xicol. 25:415—421. Sanudo-Wlhenily. S. A. & A. R, Flegal. 1992. Anthropogenic silver in the Southern California Bight: A new tracer of sewage in coastal waters. Environ. Sci. Technol. 26:2147-2151. Sericano. J. L.. T. L. Wade. T. J. Jackson. J. M. Brooks, B. W. Tripp, J, W. Farrington. L. D. Mee. J. W. Readman, J. P. Villeneuve & E. D. Goldberg. 1995. Trace organic contaminafion in the Americas: an over- view of the US National Status and Trends and the International "Mus- sel Watch" programs. Mar. Pollut. Bull. 31:214-225. Shiaris. M. 1989. Seasonal biotransformation of naphthalene, phenan- threne. and benzo(i(lpyrene in surficial estuarine sediments. Appl. En- viron. Microbiol. 55 : 1 39 1 - 1 399. Sowles. J. 1993. Maine mussel watch heavy metal baseline survey in blue mussels: 1988-1992. Augusta. ME: Maine Department of Environmen- tal Protection Tech. Report. Sowles, J., R, Crawford, P. Hennigar. G. Harding. S. Jones. M. E. Chase. W. Robinson. J. Pederson. K. Coombs. D. Taylor & K. Freeman. 1997. Gulfwatch project standard procedures: Field and laboratory. Gulf- watch implementation period 1993-2001. Augusta. ME: Gulf of Maine Council on the Manne Environment, State Planning Office, Sowles, J., R. Crawford. J. Machell, P. Hennigar, S, Jones, J. Pederson, K. Coombs. G. Atkinson. D. Taylor, G. Harding. M. Chase & W. Robin- son. 1996. Evaluation of Gulfwatch 1993— Third year of the Gulf of Maine environmental monitoring plan. Augusta. ME: The Gulf of Maine Council on the Marine En\ ironment. State Planning Office. ten Brink. B. M.. F. T. Manheim, J. C. Hathaway. S. H. Jones. L. G. Ward, P. F. Larsen. B. W. Tripp & G. T. Wallace. 1997. Gulf of Maine contaminated sediment database: Draft final report. Orono, ME: Re- gional Marine Research Program for the Gulf of Maine. ten Brink. B. M. R.. F. T. Manheim & M. H. Bothner. 1996. Contaminants in the Gulf of Maine: What's here and should we worry? In: D. Dow & E. Braasch. editors. The Health of the Gulf of Maine Ecosystem: Cumulative Impacts of Multiple Stressors. Regional Association for Research on the Gulf of Maine (RARGOM) Report 96-1. April 30, 1996. Tripp. B. W. & J. W. Farington. 1985. Using sentinel organisms to monitor chemical changes in the coastal /one: Progress or paralysis. In: Pro- ceedings of the Ninth .-Annual Conference of the Coastal Society. At- lantic City, NJ. USFDA (United States Food and Drug Administration). 1990. U.S. Food and Drug Administration Shellfish Sanitation Branch. Washington. D.C. 1214 Jones et al. USFDA (United States Food and Drug Administration). 1993. U.S. Food and Drug Administration Shellfish Sanitation Branch. Washington. D.C. Welch. L. 1994. Contaminant burdens and reproductive rates of bald eagles breeding in Maine. M.S. thesis. University of Maine. Orono. ME. Wells. P. G. & S. J. Rolston. 1991. Health of our oceans. A status report on Canadian Marine Environmental Quality. Conservation and Protec- tion. Ottawa. ON and Dartmouth. NS: Environment Canada. Widdows, J. & P. Donkin. 1992. Mussels and environmental contaminants: Bioaccumulation and physiological aspects. In: E. Gosling, editor. The mussel Mxtilus: Ecology, physiology, genetics and culture. New York: Elsevier Science Publishers, pp. 383—124. Widdows. J.. P. Donkin. M. D. Brinsley. S. V. Evans. P. N. Salkeld, A. Franklin. R. J. Law & M. J. Waldock. 199.'i. Scope for growth and contaminant levels in North Sea mussels Mytilus edulis. Mar. Ecot. Prog. Ser. 127:131-148. Jounuil of Shellfish Research. Veil. 20. No. 3. 1215-1221. 2001. DISTRIBUTION PATTERNS OF POLYCHLORINATED BIPHP]NYL CONGENERS IN MARINE SEDIMENTS AND WILD MUSSELS FROM GALICIA COAST (NORTHWESTERN SPAIN) N. CARRO.* V. SAAVEDRA. I. GARCIA, M. IGNACIO, AND J. MANEIRO Cciitni dc Control dc Calidad del Medio Marino. Peirao dc Vlla.xodn s/n. 3661 1. \'ilci;j,arciu dc Annisa. Poincvi'drii. Spidn ABSTRACT To know dislnbution p;itlerns ol |iolychloiinatctl hiplienyl congeners in ihe niaiine environment from Galicia. poly- clilorinated biphenyls (PCBi) no. 31. 28. 52. 101. I IS. 153. 105. 138. 156. and 180 were isolated by Soxhlet extraction in wild mussels and surface marine sediments; the quantification was performed using gas chromatography-mass spectrometry (GC-MS) and gas chromatography-electron capture detection (GC-ECD). PCBs 101. 138. 153. and 180 were the most abundant congeners in all samples analyzed. Distribution of PCB congeners was shifted in favor of higher chlorinated compounds. By means of multivariate techniques of data exploration, such as hierarchical cluster analysis and principal components analysis. mus,sel sainples from contaminated (Ferrol and A Coruna zones) areas were clearly separated from uncontaniinated (all the other samples) areas. Sediment samples were more heterogeneous. If tri- and tetrachlorinaled biphenyls were considered when carrymg out the multivariate analysis. Vilaboa samples formed a group differentiated from the other samples. KEY WORDS: polychlorinated biphenyls (PCBs). wild mussels, marine sediments, principal components analysis (PCA). hierar- chical cluster analysis (HCA) INTRODUCTION The increase of human activity and industrial development in Galicia is intensely disturbing the natural marine environment. Its progressive deterioration is mainly catised by contributions of vari- ous different industrial discharges that reach the coast withotit previous treatment or purification. Taking into account the high importance of shellfishing in Galicia. especially bivalve mollusks (the aquaculture of mussels is one of the main resources), it is necessary to carry out a baseline analysis of micropollutants in the estuarine hays ("Rias") from Galicia. Polychlorinated biphenyls (PCBs) are considered to be thei'- nially, chemically, and physically stable tnicropolltitants; this sta- bility has been responsible for PCB marine environmental con- tatninatioii problems (Erickson 1997). PCB compounds are vola- tile, lipophilic, and persistent. They tend to bioaccumulate and bioconcentrate in fatty tissues of biota and to as.sociate with benthic and suspended sediments (Berger & Brevik 1996. Beyer et al. 1996). Mussels, with low PCB metabolism, are a good coastal indicator for micropollutants in the water column and seditnents. Various instrumental techniques are available for the extraction and determination of PCBs in environmental samples. Techniques for the extraction of PCBs include Soxhlet extraction (Piiieiro et al. 1996). microwave-assisted extraction (MAE) (Carro et al. 20()(), Carro et al. 1999b), Supercritical fluid extraction (SEE) (Haw- thorne et al. 1993. Soren & Berit 1994). solid phase microextrac- tion (SPME) (Llompart et al. 1999). solid phase extraction (SPE). accelerated solvent extraction (ASE) (Bandh et al. 1998) and oth- ers; techniques for the determination include gas chromatography- electron capture detection (GC-ECD) (Chen & Ling 1992). gas chromatography-mass spectrometry (GC-MS) (Pavoni et al. 1991, Leonards et al. 1996. Shin & Oh-Shin 1999). gas chromatography- atomic emission detectiott (GC-AED). gas chrotiiatography- Eourier transfortii infrared (GC-FTIR) (Hembree et al. 1993), and others. Previous works on distribution of chlorinated biphenyls in mollusks from some Galicia estuarine bays were developed using different extraction methods (Fumega et al. 1984, Fianco et al. *Corresponding author. 1984. Fernandez Muiiio et al. 1991, Alvarez Piiieiro et al. 1994. Alvarez Piiieiro et al. 1995). The objective of this study was to describe the distribution of PCBs in wild mussels {Mytiliis galloprovincialis) and in surface marine sedinients collected from 19 sites in the Galicia coast. Ten PCB compounds (lUPAC No. 31, 28, 52. 101, 118, 153, 105, 138, 156, and 180) recommended by the International Council for Ex- ploration of the Seas (ICES) (Duinker et al. 1988) have been isolated using Soxhlet extraction and analyzed by GC-ECD and GC-MS. Multivariate techniques of data exploration as hierarchi- cal cluster analysis (HCA) and principal component analysis (PCA) have been applied to sediments and mussels data. MATERIALS AND METHODS Materials and Apparatus Dichloromethane. n-pentane. and isooctane for organic trace analysis were purchased from Merck (Damistadt. Gemiany); alu- niinum oxide, silica gel. and anhydrous sodium sulfate were also purchased from Merck. Copper turnings were supplied by Aldrich (Steinheitn. Germany). Analytical reagent grade PCB individual congener standards were purchased from Dr. Ehrenstorfer (Augs- burg. Germany). Wild mussels (Mytiliis galloprovincialis) and surface sediments were collected from 19 coastal subzones in the Estuarine Bays C'Ri'as") from Galicia (December 1999): Ribadeo (Ri'a de Rib- adeo). Foz (Ria de Foz). and O Vicedo (Ri'a do Barqueiro) frotn Lugo zone (samples labeled as I ); As Pfas Monton. As Pias Puntal. Barallobre, and Mugardos (Ri'a de Ferrol), Pontedeume (Ri'a de Ares), and Mirio (Ri'a de Betanzos) from Ferrol zone (samples labeled as 2); Santa Cruz and Pasaxe (Ri'a de A Coruna) from A Corutia zone (sample labeled as 3); Anilons (Ri'a de Corme-Laxe). Camarinas (Ria de Camarihas), and Corcubion (Ria de Corcubion) from Costa da Morte zone (samples labeled as 4); Raxo and Louri- zan (Ri'a de Pontevedra), Vilaboa, Arcade, and Baiona (Ri'a de Vigo) from Pontevedra zone (samples labeled as 5) (see Fig. 1 ). Two certified reference materials. HS-2 (Canadian marine sedi- ment) supplied by National Research Council of Canada and IAEA 142 (mussel tissue homogenate) supplied by International Atomic 1215 1216 Carro et al. 1 Iberian / PeninsulaC r 1 ^V^ , Figure 1. Map with •tampling points. Energy Agency (Vienna. Austria), were used tor quality control. Standard stock solutions were prepared by weighing an appropri- ate amount of each standard and diluting to 5 mL with isooctane. Working solutions were made by appropriate dilution of the stock solution. All solutions were stored at 4°C. For quantitative gas chromatographic determinations, calibration was carried out at four concentration levels for each congener spanning the range of 4—100 |j,L and using CB 155 (1 mg/L) as an internal standard. The concentrated extracts were analyzed by gas chromatogra- phy using a Perkin-Elmer Autosystem gas chromatograph equipped with an electron capture detector. A TRB-5 (Teknok- roma. Spain) y/i diphenyldimethyl silo.xane capillary coluinn (60 m X 0.20 mm i.d. x 0.4 p-m phase thickness) was used. The chromatographic conditions were as follows; the column tempera- ture program was 90°C (3 min) to 215°C (40 min) at 30°C/min and 275°C (30 min) at 5°C/min: the injector temperature (splitless mode. 1.8 min) was 270°C; the electron capture detector tempera- ture was 365°C; canier gas was hydrogen supplied by Air Liquid (Spain). The identification of extracted compounds was performed on a Varian Saturn 2000 gas chromatograph-ion trap detector mass spectrometer. The chromatographic conditions and capillary col- umn were identical to those described above: carrier gas was he- lium supplied by Air Liquid (Spain). The mass conditions were: the trap temperature 170'C: the scan range 100-400 m/z: the mul- tiplier delay 5 min: the emission current 10 |j.A; the multiplier voltage 1.450 V: the maximum ionization time 25.000 p.sec: and the manifold temperature 120 C. Data numerical analysis was carried out by means of the sta- tistical package (SPSS Inc.: 1989-1999 V.9.0). Sample Preparation Thirty individual mussels of similar sizes (40-80 mm) were used for each analysis. Homogenates of the mussel flesh were frozen (-30' C). freeze-dried and Soxhlet extracted (5 g of sample; 150 niL of dichloromethane-pentane. l-I: 8 h) by duplicate, and the extract was concentrated under vacuum evaporation to ap- proximately 1 mL. An aliquot of the extract was used to determine gravimetrically the lipid content. Lipids were removed from an appropriate portion of the concentrate by chromatography over alumina (6% deactivated) and the CB fraction was purified by chromatography on silica (I'/r deactivated). CB 155 was added as an internal standard prior to analysis by gas chromatography. Sediments were collected by hand using a metal scoop at 0-30 cm depth and immediately frozen (-30°C). The sediments were freeze-dried and sieved to obtain a particle size below 63 p,ni. Activated copper turnings were used to remove elemental sulfur during Soxhlet extraction (5 g of sample: 150 niL of dichlo- romethane-pentane. 1-1: 8 h) in duplicate. The sediment extracts were concentrated under vacuum evaporation to 1 mL and were treated in the same manner as the mussel samples. RESULTS AND DISCUSSION Validation of Analytical Procedure of PCB Compounds Calibration curves were obtained at four concentration levels using adequately diluted standards. Each concentration level was injected in duplicate, and peak heights were fitted by linear re- gression. The correlation coefficient was 0.999 for all the target compounds. Method reproducibility experiments were carried out in five replicate mussel and sediment samples, providing a mean relative standard deviation of S.SVr and 12.71%. respectively: the mean recoveries of PCBs were 92.09% and 101.31% from mussel and sediment samples, respectively. Individual Congener Concentrations: Cnivariate Analysis Concentrations of individual congeners of PCBs (lUPAC No. 31,28.52. 101. 118. 153. 105. 138. 156, and 180) expressed as the average of two analytical replicates in ng/g dry weight (dw) from mussels and marine sediments and lipid content of mussels in percentage of studied subzones are summarized in Tables 1 and 2. The contents of PCBs in wild mussels and marine sediments varied from 0.06 ng/g to 115 ng/g dw and from 0.01 to 19.4 ng/g dw. respectively. In both mussels and marine sediments, congeners making the largest individual contributions to total congener con- centration (sum of 10 congeners) were the penta-, hexa-, and hep- tachlorobiphenyls (PCBs 101, 138. 153. and 180): these com- pounds are less volatile, more resistant to metabolic and microbial degradation, adsorb readily to sediments, and are more soluble in lipids than lower chlorinated congeners (Shiu & Mackay 1986). The low concentrations of tri- and tetrachlorobiphenyls (PCBs 28, 31, and 52) in samples were due to their higher volatility, low partition coefficient between n-octanol and water (greater aqueous solubility), and higher tendency to biodegradation in marine .sedi- ments (Bright et al. 1995). However, the Vilaboa subzone showed very high levels of lower chlorinated compounds, mainly in the sediment sample: these results will be discussed below. In general, concentrations of PCBs in mussels coming from the same zone were very homogenous; levels of PCBs in mussels collected from Ferrol and A Corufia zones (labeled as 2 and 3, respectively), which possessed the greatest lipid contents, were higher than those found in the other zones: it is logical because these areas present the largest degree of urbanization and Indus- PCBs IN Marine Sediments and Mussels from Spain 1217 TABLK I. Mean of concentration of PCBs (nfj/gdw) and lipid content {'t I in mussels from (iaiicia coast. Zone Subzone Cb31 Cb28 Cb52 CblOl CbllS Cbl 53 Cbl()5 Cbl38 Cbl 56 Cbl 80 Lipid Content 1 Ribadeo 0.18 0.19 o.s 2.47 1.78 12.94 0.5 6.63 0.52 2.06 5.85 1 Foz (1.15 0.3S 0.24 1.32 1.12 7.33 0.28 3.98 0.28 1.43 5.72 1 O Vicedo 0.09 0.06 0.22 3.79 2.01 47.62 — 24.35 2.12 7.93 4.3 2 Pias Monlon 0.79 1.77 3.57 19.47 17.36 87.67 6.47 50.38 3 10.9 8.58 T Pias Puntal 0.7 1.44 3.27 21.15 23.53 106.21 4.73 64.12 3.95 13.91 8.65 1 Barallohrc 0-51 0.S8 4.07 22.16 19.79 87.38 3.21 47.28 3.11 7.49 7.55 2 Mugardos 0.(1 1 1.44 3.13 20.25 23.08 115.52 4.36 64. 1 6 3.47 12..35 10.95 2 Pontedeiime 0.2 0.28 0.82 4.29 3.54 26.38 0.79 14.76 0.81 2.29 9.4 2 Mifio 0.21 0. 1 1 0.57 3.87 4.18 33.78 0.86 19.-54 0.61 3.6 6.5 3 Sta. Cru/ 0.39 0.76 0.92 7.74 8.73 67.33 1.74 36.82 2.58 13.17 6.06 3 Pasaxe 0.36 0.6 1 46 10.2 7.75 79.06 1.81 5 1 .4 1 3.25 14 6.99 4 Anilons 0.46 0.59 1.04 5.15 1.26 9.08 0.63 4.73 0.85 1 .08 6.34 4 Camarillas 0.19 0.29 0.43 2.36 2.45 11.36 0.62 6.67 0.4 1.48 5.27 4 Corcubion 0.71 1.15 1.86 4.89 4.99 13.29 1.17 9.57 0.57 1.43 6.4 5 Raxo 0.28 0.55 1.22 5.03 5.83 17.49 2.55 11.76 0.77 2.58 8.48 5 Lourizaii 0.43 094 1.26 6.9 5.8 28.2 2.7 18.98 1.76 7.39 8.22 5 Vilaboa 2.26 3.5 2.6 6.49 6.06 33.85 1.23 20.03 1.7 5.48 6.83 5 Arcade 0. 1 9 0.37 0.58 4.76 4.54 33.6 0.94 18.63 1.14 4.16 7.8 5 Baiona 0.77 0.66 0.44 2..56 3.45 8.08 0.81 5.67 0.53 0.91 6.49 trialization in Galicia. PCB Lonceiitralions in sediment samples showed a greater intrazonal heterogeneity. To select the most significant variable to distinguish between the studied zones, data in Tables I and 2 were subjected to uni- variate analysis. A one-way analysis of variance (ANOVA) re- vealed that CBI53 (present in enhanced concentrations in indus- trial mixtures) was the most significant variable for mussel and sediment samples; Figure 2 shows the box-whisker plot corre- sponding to this variable. Congener Distrihutiini Patterns: Multivariate Analysis In view of the previous results, it was logical to use the set of variables at hand to achieve more precise distinction of the zones. Hence, the data (in the case of luussels, normalized to lipid con- tent) in Tables 1 and 2 were subjected to PCA and HCA. Initially, both types of multivariate analysis were applied separately to data for the mussel and sediment saiuples. Mussel Samples Cluster analysis of data (concentration of higher chlorinated biphenyls, CBs \0\. 118. I3.\ 105. 1.^8. 156. and \S0) for the mussel samples using the average linkage method led to two fairly robust clusters (Fig. 3A). The most compact cluster was formed by the samples from Ri'a de Ferrol (zone labeled as 2) and Ria de A Coruiia (zone labeled as 3). the most contaminated zones: at higher T.\BLE 2. Mean of concentration of PCBs (ng/g d« ) in sediments from Galicia coast. Zone Subzone Cb31 Cb28 Cb52 CblOl CB118 Cbl53 CB1(I5 Cbl38 Cbl56 Cbl 80 1 Ribadeo — — 0.04 0.2 — 1.35 — 0.57 0.05 0.61 1 Foz — — 0.12 — — 0.24 — — — — 1 0 Vicedo — — — 0.04 — 0.95 0.27 0.45 0.05 0.87 ") As pias Monton 0.09 0.19 0.42 1.61 0.95 5.69 0.2 3.59 0.32 7.45 1 As pias Puntal — — 0.04 0.16 — 0.97 — 0.39 0.01 0.35 -) Barallobre — — 0.83 6.38 1.35 16.22 0.17 8.12 O.S 8.35 -) Musardos 0.02 0.19 0.26 3.95 0.79 11.4 0.14 5.49 0.52 5.07 1 Pontedeiime — 0.52 0.88 0.4 0.96 0.1 0.65 — 0.25 T Mirio 0.07 0.14 0.21 0.26 — 0.82 — 0.38 0.01 0.28 3 Sta Cruz — — 0.03 0.17 — 0.72 — 0.36 — 0.38 3 Pasaxe — — 0.13 1.38 0.25 5.8 — 3.41 0.36 6.04 4 Anilons — — 0.03 — — 1.41 — — — — 4 Carmariiias 0.17 0.45 0.55 0.72 0.1 1.28 — 0.59 0.01 0.23 4 Corcubion — — 0.03 0.11 — 0.99 — 0.5 0.05 0.93 5 Raxo — — 0.01 — — 0.6 — 0.02 — 0.11 5 Lourizan 0.03 0.05 0.19 4.75 1.05 19.4 — 12.06 1.25 13.96 5 Vilaboa 3.44 5.28 2.15 1.7 0.57 5.73 0.33 3.6 0.27 5.19 5 Arcade — — 0.07 0.26 — 0.73 — 0.3 — 0.02 5 Baiona — — 0.03 0.01 — 0.61 — — — — 1218 Carro et al. m o Zone o 30' 20' Oe 10' ? :;:; r 0- in 2.00 3,00 Zone Figure 2. Bo\-\vhiskcr plot of CB153 for the five sampling zones (mus- sel and sediment samples). Data are divided into areas of equal fre- quency. Box: middle >i)'i frequency. Lower whisker: from the first quartile to the smallest data point, llpper whisker: from the third quartile to the largest data point. similarity, these samples were separated in two new clusters that grouped mussels from Ri'a de Ferrol and Ri'a de A Coruna. respec- tively. The other principal cluster (less compact) was formed by samples from slightly contaminated zones. Mussel samples from Pontedeume and Mifio subzones (Ri'a de Ares and Ria de Betanzos. respectively) belonging to Ferrol zone (labeled as 2) were placed in this cluster. If tri- and tetrachlorobiphenyls (CBs 31. 28. and 52) were used to carrv out the cluster analysis, mussels from Vilaboa subzone (Pontevedra zone, labeled as 5) were clearly differentiated from all the others (Fig. 38). The same conclusions are extracted using PCA; retaining the first two factors. 85.93% of the initial variance was explained ( PC 1 . 66.42% : PC2. 1 9.5 1 % ). Figure 4 shows the projection of the objects (sampling subzones) on the plane of the first two compo- nents. Ri'a de Ferrol and Ri'a de A Corufia samples were associated with the positive part of the first component that comprised the penta-. hexa-. and heplachlorobiphenyls. The distribution of the uncontaminated samples group was mainly influenced by the negative part of the first factor. The Vilaboa sample was associated with the second component that was constructed with the CBs 31. 28. and 52 contribution. In mussel samples. CB 153 and the other higher polychlori- nated biphenyls showed a clear differentiation between contami- nated and uncontaminated zones independently of the geographi- cal proximity. Pollution of industrial zones does not reach the nonindustrialized sites; this behavior can be mainly observed in Fenol zone. Pontedeume and Mifio subzones are slightly contami- nated and are considered as unspoiled and not industrialized areas, while the other subzones (P. Puntal. P. Monton. Mugaidos, and Barallobre) are highly contaminated industrial and urbanized ar- eas. Actually, there are no studies of water (lows in the Ferrol zone hut the distribution of PCB compounds suggest that the water circulation of the estuarine bays (Ares. Betanzos. and Femtl) is very slow in this area. Lourizan and Vilaboa subzones presented high contents of PCBs. Lourizan is placed near to Pontevedra city with an impor- tant pulp paper and chlorine bleaching mill: Vilaboa is sited near to the most industrialized city in Galicia (Vigo I and placed in a very closed /one of the estuarine bay. Sediment Samples Cluster analysis of data (sum of congeners) for the sediment samples using the average linkage method (Fig. 5) distinguished two groups. One of the groups included the most contaminated sediments. Barallobre subzone (sample labeled as 2) and Lourizan sub/one (sample labeled as 5). The other group consisted of two subgroups. The first subgroup included the slightly contaminated samples Mugardos and Monton subzones (samples labeled as 2) and Pasaxe subzone (sample labeled as 3) and Vilaboa subzone (sample labeled as 5). The second subgroup included the least contaminated samples. Analytical results for all of the sediment samples were more heterogeneous than for the mussel samples. Sediment samples from Lourizan and Vilaboa subzones with a high content of total organic carbon (TOC) (silty muds) presented a greater concentration of PCBs in comparison with expected val- ues of concentration from their mussel samples. This could be due to the particular conditions developed in these subzones that pro- duce a low PCBs bioavailability from sediments to mollusks: these sub/ones are sited in very stagnant areas in the estuarine bay where redissolution of organic compounds from sea bottom to water is improbable. Moreover, the high content of total organic carbon in sediments causes a high accumulation of PCBs (hydrophobic con- taminants) because they tend to be strongly associated with or- ganic-rich particles. These sediments can act as a temporary pol- lutant trap before accumulation in biota. In general, the bioavail- ability and toxicity of sediment-associated contaminants decrease with increasing TOC content (Lake et al. 1990). In further re- search, the TOC values of sediments (with different structures and compositions) will be considered to attain a good geographical comparison and a coirect data interpretation in contamination stud- ies. The sediment TOC and the lipid content of the organisms are key factors in the partitioning of organic contaminants. Mussels from Puntal (sample labeled as 2) and .Sta Cristina (sample labeled as 3 ) subzones had a higher concentration of PCBs than their corresponding sediment samples. This could be due to the low depth of sampled sediments (0-30 cm). In these subzones, the surface sediment, placed in a tidewater zone, is a "clean" sandy PCBs IN Marine Sediments and Mussels from Spain 1219 CASE 0 Label Num 4- Mussel samples Ribadeo 1 Baiona 5 Camarlnas 4 Anllon A Foz 1 Pontedeume 2 Corcubion 4 Mino 2 Raxo 5 Arcade 5 O vicedo 1 Vilaboa 5 Lourizan 5 sea Cruz 3 Paxase 3 P. PunCal 2 Mugardos 2 P. Monton 2 Barallobre „^ Rescaled Distance Cluster Combine 5 10 15 20 B CASE C Label Nun + Mussel samples Mino 2 Arcade 5 Camarinas 4 Foz 1 Pontedeume 2 Raxo 5 Bibadeo 1 Paxase 3 Vi cedo 1 .Anllons 4 Sta cruz 3 Corcubion 4 lourizan 5 BaiOna 5 ? . Monton 2 Mugardos 2 ?. PunCal 2 Baiallobre 2 '1 laboa S Rescaled Distance Cluster Combine 5 10 15 20 Figure 3. (A) Dendrogram slio«in); Ihe results of a hierarchical cluster analysis on higher chlorinated biphenyl distributions in Galicia mussels. (B) Dendrogram showing the results of a hierarchical cluster analysis on lower chlorinated biphenyl distributions in Galicia mussels. deposit with very low levels ofTOC and organic pollutants. Some- times, accumulation of conlaminanis in seseral species (including mussels) does not depend on pollutant concentration in sediments but it can be strongly affected by a selective uptake of contami- nated food. In these zones, the local (urban) point discharges are frequent. If only tri- and tetrachlorobiphenyls CBs 28. 31. and 52 were considered when carrying out the cluster analysis, the sample from Vilaboa was clearly differentiated from all the others: the behavior was similar to the mussel sample. Two principal components were found to account for about 91.25% of the initial variance. The most contaminated group is associated with the first component (constructed with penta-. hexa-, and heptachlorinated biphenyls). The Vilaboa sample is related to the positi\e part of the first and second components (the latter is constructed with tri- and tetrachlorinated hiphenyls). Correlation between Mussels and Sediments from the .Same Zone Due to the cycling of micropollutants in marine inlets, mussels and marine sediments must be significantly couelated in a unique manner that should predict the general distribution patterns of PCBs. In this way. the possible sources of organic micropollutants can be detected and actions initiated. To confirm this hypothesis, the data sets in Tables I and 2 were subjected to multivariate analysis. In general, cluster analysis in Figure 6A shows separation between sediments and mussels, but there are some slightly contaminated mussels that were placed in the sediment cluster and some highly contaminated sediments (Lourizan subzone) that were placed in the mussel cluster (Baiona. Camarinas, Anllon subzones). Mussels of Ri'a de Fen'ol were clearly separated from all samples. If only tri- and tetrachlorinated biphenyls are considered. Vila- q: Vilaboa a Corcubion Ria de Fer Am Ions / As Pias-Monton' J- ^ Raxo Loun. f 0 ViCedo a /'Santa Cruz nV. '.. ° Pasa,, ) ' ^" Ria de A Coruna ol Sub-zone ° 5.00 ^ 4.00 ° 3,00 2,00 1 ,00 2.0 BART factor score 1 for analysis 1 Fifiure 4. DistrihutMm o{ mussels in the plane ni the tirst and second component. Label Num + Sediment samples O vicedo Corcubion Ribadeo Anil 6ns Arcade Sta Cruz P. Puntal Mifio Raxd Baiona Foz Pontedeume Camarinas Mugardos Vilaboa P. Monton Pasaxe Barallobre Lourizan 5 Rescaled Distance Cluster Combine 5 10 15 20 Figure 5. Dendrogram of Galicia sediments. Sum of congeners was considered to carrv out cluster analvsis. 1220 Carro et al. CASE Label Num + +■ Mussel and sediment samples Sed.Foz Sed. Baiona Sed.Anilons Sed. Ai cade Sed. Baxo Sed.Camarinas Sed . Pontedeume Sed.P.Puntal Sed.Sta cruz Sed.Mino Sed.O Vicedo Sed. Ribadeo Sed.Corcubion Mus . Ribadeo Mus . Foz Mus .Anllons Mus -Camarinas Mus . Baona Mus . Lourizan Mus.Vilaboa Mus .Mino Mus. Arcade Mus . Pontedeume Hus . CoEcubion Mus. Raxo Sed.Mugardos Sed. Vilaboa Sed. P.Monton Sed. Pasaxe Sed. Barallobre Mus . Sta . Cruz Mus . Pasaxe Mus.O Vicedo Sed. Lourizan Mus. P.Puncal Mus .Mugardos Mus. P.Monton Mus . Baral lobre Bescaled Distance Cluster Combine 5 10 15 20 B Rescaled Distance Cluster Combine 5 10 15 20 Label Num + +- Mussel and sediment samples Sed. 3 Sed. 5 Sed. 4 Sed. 5 Sed. 5 Sed. 1 Sed. 1 Sed. 4 Sed. 1 Sed. 2 Sed. 2 Sed. 2 Sed. 2 Sed. 3 Mus. 1 Mus. 4 Mus. 1 Sed. 4 Sed. Mus. Mus. Mus. Mus. Sed. Mus . Mus. Mus. Mus . Mus. J Mus . Mus . Mus . Mus. Sed. Vilaboa Mus .Vilaboa ^ Figure 6. (A) Dendrogram shoHing the results of a hierarchical cluster analysis on chlorinated hiphenvl dislrihutions in (iaiicia mussels and sediments. (B) Dendrogram showing the results of a hierarchical cluster analysis on lower chlorinated biphenji distributions in Calicia mussels and sediments. boa samples (mussels and sediments) were closely grouped (Fig. 6B). The abnormal behavior of the Vilaboa subzone with regard to lower chlorinated biphenyls is probably related to several factors that can cause some special conditions in the aquatic system. The Vilaboa subzone can be considered as an undisturbed area because the presence of strong air and water flows is unlikely, so a major contribution of the highly volatile compounds (with a high vapor pressure) can be caused by atmospheric input. These compounds can be present in sufficient concentrations so that they remain in a free phase and are not dissolved in water. Sediments from Vilaboa subzone possessed high organic car- bon contents. This may influence the fate and bioavailability of some contaminants by affecting physicochemical characteristics, such as water solubility, sorption capacity, fugacity. and partition- ing (Kenaga & Goring 1980). It is also known that under anaerobic conditions, microorganisms can partially dechlorinate the more highly chlorinated congeners producing the lower chlorinated compounds (Erickson 1997). There are reasons to believe that PCB composition will change with time in environmental samples and that the resulting propor- tions of congeners will not resemble the pattern of industrial mix- tures. The lack of definitive assignments of congener composition ' ''**^S5:r;-^... Figure 7. Distribution of (iulicia mussels and sedime three principal components (Varimax rotated). ° Corcubion ° Carmarinas ° Anllons " Pasaxe ° Sla Ciuz ° Mino ' Pontedeume Mugardos Barallobre As pias Puntal "^ As pias Monton ° O Vicedo ' Foz ° Ribadeo nts in the first PCBs IN Marine Shdiments and Mussels from Spain 1221 of industrial mixtures and other products contributes to an igno- rance of the possible transformations of these compounds in the animal tissues and. in general, in all en\ ironmcnlal samples (water, sediment, soil, atmosphere, etc.). PCA followed the cluster analysis. With the first two factors. 83.4691- of initial variance is explained. A representation of the first three rotated (Varimax) components is shown in Figure 7. In the tridimensional space, three sample groups can be identified: the first is closely related to the positive part of the second axis (Vila- boa sediment). The distribution of the second group (mussels and sediments from Ferrol /.one. mussels from P. Montcin. P. Puiital. Barallobre. and Mugardos and sediments from P. Montoii. Pont- edeume. Barallobre. and Mugardos) is mainly infiuenced by the positive part of the first factor. Finally the third group (all the other samples) is related to the negative part of the second axis. By means of cluster analysis and principal component analysis. the most polluted samples (i.e.. the group of Ferrol zone samples of mussels and sediments) and the group of Vilaboa samples (mus- sels and sediments) were separated from the more heterogeneous group. This study has described the profile and base levels of the organochlorine pollution in estuarine bays froin Galicia. In this way. we can now detect abnormal concentrations and possible sources of these compounds that may damage marine water quality and shellfish culture destined for human consumption. ACKNOWLEDGMENT We would like to thank A. Mouteira for sample preparation. LITERATURE CITED Alvarez Pineiro. M. E.. J. Simal Lozano & M. A. Lage Yusty. 1994. GC determination of PCBs in mussels from Galicia. / AOAC Int. 77:985- 988. Alvarez Pineiro. M. E.. J. Simal Lozano & M. A. Lage Yusty. 1995. Organochlorine compounds in mussels of Estuarine Bays of Galicia (North-West Spain). Mai: Pollut. Bull. .^0:484-487. Bandh. C. E. Bjoerklund. L. Mathiasson, C. Noef & Y. Zebuehr. 1998. Determination of PCBs in Baltic Sea sediments using accelerated sol- vent extraction ASH. Oiganohalogen. Compd. 35:17-19. Berger. J. A. & E. M. Brevik. 1996. Uptake of metals and persistent organochlorines in crahs [Cancer pagiiiii.\) and flounder {Plutkhthys flesiis) from contaminated sediments: Mesocosm and field experi- ments. Mui: Pallia. Bull. .^.^:46-55. Beyer. J.. M. Sandvik. K. Hylland. E. Fjeld. E. Egaas. E. Aas. J. U. Skare & A. Goksoyr. 1996. Contaminated accumulation and biomarker re- sponses in flounder [Plalichlhys fle.ms L.) and Atlantic cod {Gaclus oiiuia L.j exposed by caging to polluted sediments in Sortjorden. Nor- way. Aquat. Toxicol. 36:75-98. Bright. D. A.. S. L. Grundy & K. J. Reimer. 1995. Differential bioaccu- mulation of non-ortho-substituted and other PCB congeners in coastal arctic invertebrates and fish. Environ. Sci. Tecbnol. 29:2504-2512. Carre. N., I. Garcia & M. Llompan. 2000. Closed-vessel assisted micro- wave extraction of polychlorinatod hiphenyls in marine mussels. Aiuiliisis. 28:720-724. Carro. N.. Y. Saavedra. I. Garcia & M. Llompart. 1999b. Optnnization of microwave assisted solvent extraction of polychlormated biphenyls from marine sediments. J. Microcolumn Sep. 1 1 :544-549. Chen. C. Y. & Y. C. Ling. 1992. Optimization electron capture detector performance for gas chromatographic analysis of polychlorinated bi- phenyls. Chromatographia 33:272-278. Duinker. J. C. S. E. Schulz & G. Patrick. 1988. Selection of chlorinated hiphenyl congeners for analysis in environmental samples. Mar. Poll. Bull. 19:19-25. Erickson. M. D. 1997. Analytical chennstry of PCBs. 2nd ed. Boca Raton and New York: CRC Press, Lewis Publishers. FemandezMuifio. M. A., J. De la Montaiia & J. Simal Lozano. 1991. A GC method for chlorinated pesticides and PCBs in mussels. Chro- inalographia 3 1 :453^56. Franco. J. M.. M. J. Fernandez & M. A. Murano. 1984. El impacto de la contaminacion por compuestos dorados en el literal gallege. Cuad. Area Cienc. Mar. 1:385-393. Fumega, J., J. M. Masso & F. Schultze. 1984. Distribucion de DDT deri- vades bifenilos peliclorados (PCBs) en la Ria de Ponte\edra. Cuad. Area Cienc. Mar. 1:379-384. Hawthorne. S. B.. D. J. Miller, J. J. Langenfeld & D. Burford. 1993. Effects of temperature, density, modifier identity and concentration on super- critical fluid extraction rates of PCBs and PAHs from environmental samples. In: Prec. 15th Intem. Symp. on Capillary Chromatogr. Rixa del Garda, Italy, May 24-27. pp. 1701. Hembree, D. M.. Jr.. N. R. Smyrl. W. E. Davis & D. M. Williams. 1993. Isomeric characterization of polychlorinated biphenyls using gas chro- mategraphy-Fourier transform infrared, gas chromatography-mass spectrometry. Analysi. 1 18:249-252. Kenaga. E. E. & C. A. I. Goring. 1980. Relationships between water solubility, soil sorption, octanol-water partitioning and concentration of chemicals in biota. In: Aquatic Toxicology: Proceedings of the 3rd Annual Symposium on Aquatic Toxicology, pp. 78-115. Lake. J. L.. H. L. Rubinstein. C. A. Lake. J. Heltshe & S. Pavignano. 1990. Equilibrium partioning and bioaccumulation of sediment-associated contaminants by infaunal organisms. Environ. To.xicol. Chein. 9:1095- 1106. Leonards. P. E. G.. U. A. T. Brinkman & W. P. Cofino. 1996. The use of gas chromatography with ion-trap MS-MS detection for the determi- nation of planar PCBs in biota and sediment. 1996. Chemospliere 32: 2381-2387. Llompart. M.. K. Li & M. Fingas. 1999. Headspace solid-phase microex- iraction for the determination of polychlorinated biphenyls in soils and sediments. J. Microcolumn Sep. 11:397-402. Pavoni. B.. A. Sfriso & S. Raccanelli. 1991. Quantification of PCBs in environmental samples: Comp;msen of results obtained with different analytical instruments (GC-ECD, GC-MS) and .standards. Inl. J. Envi- ron. Anal. Chem. 44:11-20. Piiieiro. M. E. A., J. S. Lozano. M. A. L. Yusty & S. T. C. Gon/alez- Barros. 1996. Comparison ef two extraction methods for determination of PCBs and PCTs in mussels from Galicia. Talaiira 43:487-491. Shin. H. S. & Y. Oh-Shin. 1999. A Study of quantitation of polychlorinated hiphenyls (PCBs) in sediments. Anal. Sci. Teclinol. 12:151-158. Shiu. W. Y. & D. Mackay. 1986. A critical review of aqueous solubilities, vapor pressures. Henry's law constants and octanol-water partition co- efficients of polychlorinated biphenyls. J. Pli\.\. Chem. Ref. Data 15: 911-929. Soren. B. & J. Berit. 1994. Analysis of PCBs in sulfur-containing sedi- ments by off-line supercritical fluid extraction and HRGC-ECD. .\nal. Chem. bii-.bbl-biy. .I.nmial »/ Shellfish Research. Veil. :(). Nci. 3. 1:23-1227, 2001. VALIDITY OF ESCHERICHIA COU, ENTEROVIRUS, AND F-SPECIFIC RNA BACTERIOPHAGES AS INDICATORS OF VIRAL SHELLFISH CONTAMINATION LAURENCE MIOSSEC,' * FRANCOISE LE GUYADER,' DOMINIQUE PELLETIER." LARISSA HAUGARREAU.' MARIE-PAULE CAPRAIS/ AND MONIQUE POMMEPUY' 'IFREMER. Lahoraunir de Microhiologie. HP 21105. 44311 Nantes Cede.x 3. France: 'IFREMER. Labonitoiic de Mad\einalique,s Applic/uecs a I' Exploitation des Ressources Halieutiques et Aqiiacoles. BP 211(15 44311. Nante.s Cede.x 3. France: 'iFREMER. Laboratoire de Microbiologie, BP 70, 29280 Ploiizane. France .ABSTRACT The sanitary classification of har\esiing areas tor bivalve mollusks in France is based on the level of E.wherichia eoli contamination detected in shellfish meat, as defined in EC Directive 91/492 EEC. However, outbreaks of gastroenteritis or hepatitis after consumption of shellfish meeting cunent bacteriological standards suggest that £. call is a poor indicator of viral contamination. The purpose of this study was to assess the adequacy of enterovirus and F-specific RNA bacteriophages as new indicators of human enteric viruses. Shellfish were sampled over a 37-mo period to characterize microbial contamination in two coastal areas subjected to different sewage contamination inputs. Contamination by E. eoli, F-specific RNA bacteriophages (F+ RNA) and human enteric viruses (enterovirus. EV; hepatitis A virus, HAV; Norwalk-like virus, NLV: astrovirus. AV; and rotavirus, RV) was measured in the same samples. E. eoli analysis was performed by conductance measurement, enteric viruses were detected by reverse-transcription poly- merase chain reaction (RT-PCR) and hybridization, and F-i- RNA was evaluated by culture according to the ISO 10705-1 method. Statistical analysis based on bootstrap methods was performed on 95 series of paired observations. The validity of E. eoli. enterovirus, and F-specific RNA bacteriophages as viral indicators was evaluated by measuring their sensitivity and specificity in the presence of enteric viruses. None of the tested indicators proved adequate to protect the public from viral shellfish contamination. The sensitivity of all indicators was better m the highly contaminated zone, and enteroviruses showed the highest specificity for both sites. KEY WORDS: indicator, Eseheyiehiu eoli. F-specific RNA bacteriophages, enteric viruses, viral contamination, shellfish INTRODUCTION Determinatiiin of the iiiicrdbioUigical saniluiy quality of shell- fish-growing waters is based on the use of fecal coliforms to evalu- ate the presence of fecal pollution. In Europe, sanitary criteria are cunently established by EC Directive 91/492 EEC. However, pre- vious studies have shown that there is very little relationship be- tween traditional fecal indicators and the presence of viral patho- gens (Le Guyader el al. 1994, Dore & Lees 1995. Tree et al. 1997, Lee et al. 1997. Le Guyader et al. 1998). This has been confirmed by outbreaks of viral gastroenteritis or hepatitis after consumption of shellfish meeting current bacteriological standards (Dovvell et al. 1995, Leeset al. 1995). In fact, fecal coliforms are rather poor indicators of viral con- tamination because they are different than viruses in their resis- tance to environmental conditions, sewage, or water treatment pro- cesses and marine water. They only survive in seawater for a few- days, whereas viruses can persist for several months. A few other microbial organisms, such as enteroviruses and bacteriophages, have been proposed as alternative sanitary indicator organisms. As viruses, their survival conditions in seawater should be similar to those of enteric viruses (Havelaar 1993, Metcalf et al. 1995. Mail- lard 1996, Sobsey 1997). Three main groups of bacteriophages are of particular interest as potential indicators; somatic phages, Bacteroide.'i fnigilis phages, and F-specific RNA phages. Analysis of existing studies! Lee etal. 1997, Tree et al. 1997, Havelaar et al. 1993) indicated that the last two groups were most suitable for tracing fecal contamination. However, the liinited number of B. fragilis phages in the environment, as well as methodological dif- ficulties in detecting them, have led some researchers to choose *Corresponding author. F-specific RNA phages as potential indicators of viral risk in shell- fish (Lee et al. 1997). The relation between these indicators and enteric viruses in environmental samples (wastewater, groundwa- ter, fresh and marine waters, and shellfish) is uncertain. Some authors have found a good con^elation between these candidate indicators and viruses (Gantzer et al. 1998. Havelaar et al. 1993). whereas others are more doubtful (Legnani et al. 1998. Ricca & Cooney 1999. Griffin et al. 1999. Leclerc et al. 2000). The purpose of this study was to compare the validity of £. co/i. enterovirus, and F-specific RNA bacteriophages as viral indicators by measuring the sensitivity and specificity of each candidate in the presence of enteric viruses. MATERIALS AND METHODS En virimmeiilal Sampling Shellfish samples were collected over a 3-y period (September 1995 to September 1998) in two coastal areas with different pol- lution levels. Forty-seven oyster ( Cra.v,tt«//-efl gigos: Thunberg. 1793) samples were collected at three sampling points in a shell- fish area temporarily classified as a category A site (site I ) and 48 mussel (Mytilits galloprovinciali.s: Lamarck. 1819) samples at two sampling points in a highly contaminated area not open to shellfish commercialization (site 2). Sample collection was uniform throughout the year: 459f during the autumn-w inter season (re- spectively. 47% in site 1 and 44% in site 2) and 55% during the spring-summer sea.son (respectively. 53% in site 1 and 56% in site 2). Contamination by Esclu-ricliia eoli. F-specific RNA bacterio- phages (F-i- RNA). and human enteric viruses (enterovirus. EV: hepatitis A virus. HAV: Norwalk-like virus. NLV; astrovirus. AV; and rotavirus. RV) was evaluated in the same samples. 1223 224 MiOSSEC ET AL. TABLE 1. Sensitivity and specificity measurements of an indicator. Viral Contamination ■^es No Po^itive level of indicatcir Negative level of indicator TP FP FN TN Sensitivity = TP/(TP + FN). Specificity = TN/(TN + FP). TP: true-positive: FN: false-negative; FP: false-positive; TN: true-negative. Microbiological Analyses E. coli Quantitative estimation of fecal indicators (£. coli) in shellfish was performed by conductance measurements (Dupont et al. 1996). The detection threshold of the conductance method is 30 E. coli per 100 g of shellfish (meat and liquor). Enteric \i ruses and Bacteriophages Upon reception in the laboratory. 10 to 12 shellfish were washed and shucked, and the stomach and digestive tissues were dissected, mixed, aliquoted. and frozen. For analysis, one aliquot was thawed on ice and crushed. Viruses, enteric viruses, and F-i- RNA were concentrated by PEG precipitation, as previously de- scribed (Atmar et al. 1995). For liuinan enteric virus detection, the PEG pellet was sus- pended in phtisphale buffer and nucleic acid and purified as pre- viously described (Atmar et al. 1495). All primers, probes, and reverse-tran.scription polymerase chain reaction (RT-PCR) condi- tions have already been reported and used in different studies (Le Guyader et al. 2000). For F-specific RNA bacteriophage enumeration, the PEG pellet was then suspended in 3 mL of peptone water, and F+ RNA phages were detected according to the ISO 10705 ( 1 ) 1995 method. The double agar overlay technique was used for preparation and de- termination of phage stocks and enumeration of the coliphage content of oysters. The F-i- RNA assay was performed with host Salmunella typhinuirium WG49. The detection thresholds of the assay were 2.5 PFU/1.5 g of digestive tissue for oysters and 5 PFU/1.5 2 of diszestive tissue for mussels. with no viral contamination was estimated by the specificity of the indicator. Sensitivity was defined as the ratio of the number of true- positive values (positive level of the candidate indicator organism in the presence of virus) to the number of true-positive values plus the number of false-negative values (negative level of the candi- date indicator organism in the presence of virus). Specificity was defined as the ratio of the number of true-negative values (negative level of the candidate indicator organism in the absence of virus) to the number of true-negative values plus the number of false- positive values (positive level of the candidate indicator organism in the absence of virus) (Table I). Sensitivity and specificity es- timates ranged between 0 and 1 (or 0% and 100%). Ideally, the best indicator should have a sensitivity and specificity of I (or I00'7f). An indicator was considered to be positive or negative accord- ing to a specific threshold. The negative level of the threshold was less than 230 E. coli per 1 00 g of shellfish meat for E. coli. the detection level for F-t- RNA and absence of detection for EV. The positive level of the threshold was equal to or greater than 230 E. coli per 100 g of shellfish flesh for E. coli. above the detection level for F+ RNA and positive detection for EV. The vaiiance estimators for sensitivity and specificity were evaluated using resampling techniques (Efron 1982). Resamples were generated from a nonparametric bootstrap procedure based on the initial samples, consisting of 47 pairs of data for site 1 and 48 pairs of data for site 2 for each organism. Each pair related to the concentration of the organism (£. coli. F+ RNA) or the pres- ence/absence of the organism (EV), or the presence/absence of the virus. One thousand replicate samples were generated for each organism based on these data. The mean and variance of both sensitivity and specificity were calculated for the three indicator organisms. Two conditions were tested for viral contamination. The first concerned all the enteric viruses (VI. i.e., HAV and/or NLV and/ or AV and/or RV) except EV, which was used here as an indicator. The other focused on viruses known to be involved in shellfish- borne viral disease (V2, i.e.. HAV and NLV). For each condition, a r-test at a 95"^ confidence level was used to compare differences in mean sensitivity and mean specificity between two indicators. Statistical analyses were performed using S-Plus 2000 (MathSoft. Inc.. Seattle. WA). Data Analysis Data were first summarized by the minimum, maximum, and geometric mean of the shellfish concentration for E. coli and F-i- RNA and by the presence or absence of detection for enteric v i- ruses (EV. HAV. NLV. AV. and RV). The validity of an indicator organism to discriminate shellfish samples contaminated by virus was evaluated using the sensitivity of each candidate indicator organism for virus detection. The accuracy of identifying samples RESULTS The results obtained throughout the study period for E. coli and F-i- RNA are shown in Table 2. At site 1, E. coli concentrations in shellfish ranged from <30 to 5.670 per 100 g of shellfish, with a geometric mean of 55. At site 2. the maximum v alue and geomet- ric mean were 20.200 and 283 E. coli per 100 g. The F+ RNA content in shellfish ranged between <2.5 and 87 phages per 1.5 g TABLE 2. Concentrations (minimum, maximum, and geometric mean) of £. coli and F+ RNA in shellfish collected in the study areas. H Analyzed Samples E. coli (/too S of Shelllish) F+ RNA (/L5 g of Digestive Tissue) Min. Max. (Jeometric Mean Min. Max. Geometric Mean Site 1 Site 2 47 4H <3() <3(.) .^.(i7() 20. 200 55 283 <2.5 <5 87 4.12 3.y 46 Indicators of Viral Shellfish Contamination 1225 of digestive tissue with a geometric mean of 3.9 pliages per 1.5 g at site 1 . Bacteriophage contamination was greater at site 2. w ith a niininuim and maximum of <5 and 4.125. phages per 1.5 g and a geometric mean of 46 phages per 1.5 g. The minimum values obtained at sites I and 2 corresponded to the detection limits of the analytical methodology. All enteric viruses except hepatitis A vi- rus at site 1 were found in sampling areas (Table 3). Shellfish were contaminated nwre often by all types of \iruses at site 2 than at site I. All /-tests comparing mean sensiti\ity and mean specificity were significant except, at site 1 . for the sensitivity results between E. coli and EV for V2 (viruses involved in shellfish-borne \'iral disease, i.e., HAV and VNL). At site 1 (Table 4), F-i- RNA had low sensitivity measurements of 0.23 for VI (all enteric viruses) and 0.50 for V2. This potential viral indicator was associated with viral contamination in only one of five samples for VI and one of two for V2. The results obtained for E. coli and EV sensitivity were lower (0.09 for E. coli and 0.13 for EV for VI. and both 0.20 for V2). For V2, the difference in sensitivity between F-i- RNA and the other indicators was particularly striking. For specificity, enterovi- ruses gave the highest estimates (0.92 for VI and V2) compared to E coli (0.80 for VI. 0.86 for V2) and F-h RNA (0.84 for VI: 0.89 for V2). The adequacy of the viral indicator relative to viral contami- nation was greater at site 2 (Table 5) than site I. Sensitivity esti- mates ranged between 0.57 (EV) and 0.80 (F-l- RNA) for VI and between 0.73 (EV) and 1.00 (F-i- RNA) for V2. The sensitivity of E. coli was intermediate, (i.e.. 0.66 for VI and 0.80 for V2). Enterovirus specificities were highest (1.00 for VI and 0.82 for V2). whereas F-i- RNA had the lowest capacity to identify non- contaminated shellfish samples (specificity estimates of 0.33 lor VI and 0.37 for V2). The specificity estimates of £. coli were between those calculated for EV and F-t- RNA (0.55 for V 1 and 0.52 for V2). The sensitivity and specificity estimates for a given candidate were statistically different from those obtained for the others, according to VI or V2 (f-test. 5% in Table 5). DISCUSSION The method applied here is currently being used in clinical experiments (Desenclos et al. 1997) to evaluate the accuracy (sen- sitivity and specificity) of a new screening procedure compared to a gold standard. Sensitivity represents the ratio of individuals clas- sified as "ill" by the method among those known to have the disease, and specificity represents the ratio of indiv iduals classified as "not ill" among those known not to have the disease. These criteria are calculated according to a threshold, and the relative frequency of true-positive (TP). false-negative (FN), false-positive (FP). and true-negative (TN) results depends on the level of this TABLE i. Detection of \iruses in shellllsli ciilleited in the sluil> area (a sample niav ha>e been contaminated l)v more than one virus). n .\nal\/.ed Samples in) HAV (») NLV (HI AV ill) RV («) Site 1 Sue : 47 4S 17 10 12 II ly EV: enterovirus; HAV: hepatitis A virus: ustovirus; RV: rotavirus. TABLE 4. Sensili>it> and specilleit) measurements lor each indicator organism in site l(;i = 471. VI V2 Site 1 Sensitivity Specificity Sensitiv ty Specificity E. coli EV F-l- RNA 0.09 0.13 0.23 0.80 0.92 0.84 0.20 0.20 0.50 0.86 0.92 0.89 VI represents the presence or absence of HAV and/or NLV and/or AV and/or RV: V2 represents the presence or absence of viruses involved in shellfish-borne viral disease (HAV and NLV I. All results were signifi- cantly different from others (Mest. 5'~'i). except for the two 0.20-obtained for V2. threshold. A low threshold increases sensitivity and reduces speci- ficity, and the opposite is true for a high threshold. The diagnostic strategy requires that sensitivity or specificity be concordant with the objective (avoiding false-negative or false-positive results). In environmental research, this method was used to compare the validity of £. coli. EV and F-i- RNA as viral indicators. The purpose of sanitary analysis of shellfish is to provide guarantees that protect consumers from public health hazards due to the con- sumption of polluted shellfish. In this context, the sensitivity of an indicator is the prime criterion. In other respects, the specificity of an indicator has commercial implications, limiting the period of temporary closure of a shellfish area because of public health concerns. Moreover, sensitivity measurements have greater public health implications in shellfish-growing areas classified as cat- egory A sites since these shellfish are consumed directly (without any purification treatment). Accordingly, the following comments are concerned mainly with sensitivity criteria and focused on re- sults obtained for site I . No indicator was highly sensitive to viral shellfish contamina- tion in samples from the slightly contaminated area (site I ). Re- sults for F-l- RNA sensitivity, though better than those for the other indicators, were not fully predictive of the viral risk related to shellfish consumption. Classification errors were numerous: 50% of the results were FN for viruses involved in shellfish-borne viral disease, V2 (80'7r FN for all enteric viruses, VI). F+ RNA sensi- tivity could not be improved because the tested threshold was the detection level of the F-l- RNA analytical method (2.5 PFU/1 .5 g of digestive tissue for oysters and 5 PFU/1 .5 g of digestive tissue for mussels). TABLE 5. Sensitivity and specificity measurements for each indicator organism in site 2 (;i = 48l. NLV: Norwalk-like virus: AV: VI V2 Site 2 Sensitivity Specificity Sensitivity Specificity E. coll EV F-l- RNA 0.66 0.57 0.80 0.55 1.00 0.33 0.80 0.73 1.00 0.52 0.82 0.37 V 1 represents the presence or absence of HAV and/or NLV and/or AV and/or RV: V2 represents the presence or absence of viruses involved in shellfish-bome viral disease (HAV and NLV). All results were signifi- cantly different from others (r-test, 5%). 1226 MiOSSEC ET AL. The inadequacy of£. coli standards for public tiealtli protection was confirmed in this study. £. ccli was the least sensitive and specific tool for identifying samples contaminated by enteric vi- ruses. Enterovirus was not belter than E. coli in predicting a public health risk, but showed a high degree of accuracy in predicting the absence of viruses, with a low error classification {S'7c FPl. For sporadic contamination, the tested indicators (£. coli. EV. and F+ RNA) were not sensitive to the targeted viruses, either because the epidemic cycles were different or because the behavior and sur- vival of the viruses in seawater were variable. Although E. coli is part of the bacterial flora of warm-blooded animal feces. F-l- RNA phages are an infrequent component of feces. However, these phages are abundant in sewage treatment plants where they pro- lifei-ate at low temperatures. As they are resistant to various sew- age treatment processes and to chlorination, they are more suitable for tracing a sewage effluent than indicating fecal pollution (Have- laar & Pot-Hogeboom I9S8. Leclerc et al. 2000). The presence of enteric viruses in sewage water and eventually in coastal water and shellfish is correlated with viral diseases in the general population. Epidemics of viral diarrhea with NLV. RV. and AV etiologies usually follow a seasonal pattern, with a peak in winter (Kapikian et al. 1996. Vinje et al. 1997. Desselberger 1998. Evans et al. 1998. Otsu 1999). Outbreaks of hepatitis A can occur throughout the year, depending on the different exposure conditions (i.e., con- taminated food and water, contact with sewage, travels in endemic countries, direct human transmission) (Hollinger & Ticehurst 1996). Pathologies caused by enteroviruses occur most frequently in summer (Melnick 1996). These epidemiological aspects indicate the difficulty of finding a universal viral indicator. The sensitivity of all indicators tested was better for samples from the contaminated zone (site 2) than from site 1. As this area is highly affected by urban sewage and is contaminated year-round by fecal coliforms and F-t- RNA. these potential indicators could be present systematically when enteric viruses in effluent sewage are discharged into the marine environment during human viral gas- troenteritis epidemics. Thus, F-l- RNA sensitivity measurements could reach 100%. However, as this area is strictly forbidden for shellfish commercialization on the basis of £. coli criteria, a viral indicator would not be inore satisfactory than fecal coliforms. These findings are consistent with those already published on ma- rine water contamination in highly contaminated areas (Chung et al. 1998. Legnani et al. 1998). Concerning areas in compliance with microbiological sanitary criteria, our results differ from those in previous reports recom- mending F+ RNA as a viral indicator (Chung et al. 1998, Dore et al, 2000). probably because of variations in F-i- RNA contamina- tion levels. Moreover, one of these studies (Dore et al. 2000) concerned the use of F+ RNA to evaluate the sanitary quality of market-ready oysters after purification (generally 48 h according to regulations) and not. as in our study, to monitor the water quality in shellfish harvesting areas. During the purification pro- cess, the E. coli concentration in shellfish is drastically reduced within a 48-h period as compared to phage and virus levels. This short purification period may not be long enough to eliminate viral contamination (Schwab et al. 1998). In such conditions, F-i- RNA phages should be reliable indicators of enteric \'iruses in oysters (Dore et al. 2000). In a long-term environmental survey, the occurrence and con- centration of F-f RNA, as compared to enteric viruses, would prob- ably differ throughout the year. This would be particularly true for slightly contaminated areas in which sewage discharge is sporadic. The F-l- RNA results observed in the present study are directly comparable with those of enteric virus detection because both relate to the analysis of digestive gland in which enteric viruses and F-t- RNA are mainly concentrated (Atmar et al. 199.'). Dore & Lees 199,'i). Similarly. Croci et al. (2000) found no conelation between enteroviruses, hepatitis A virus, bacteriophages, and E. coli in mussels collected in waters in accordance with microbio- logical sanitary criteria. These authors, as well as Toze (1999). recommended that viruses be detected directly in shellfish to pre- vent health risks related to their consumption. Our results indicate that the traditional and proposed indicators tested did not fully prevent the viral risk related to shellfish con- sumption, especially in slightly contaminated shellfish-growing areas. Direct detection of the real pathogen by molecular tools could provide another means of assessing viral risk for shellfish consumers. This recommendation needs to be considered during high-risk seasons when enteric viruses are likely lo be present in the population and enter into wastewater and sewage subsequently discharged into the marine environment (Burkhardt & Caici 2000. Miossec et al. 2000). Further studies are required to confirm these findings. ACKNOWLEDGMENTS The authors are grateful to J. C. Sauvagnargues and collabora- tors (IFREMER) for sampling and bacteriological analyses and to P. Gros (IFREMER) for helpful advice and comments on the manuscript. Financial support was provided by the French Minis- try of Environment and the Languedoc-Roussillon Region. Atmar, R. L., F. H. Neill. J. L. Romalde. F. Le Guyader. C. M. Woodley, T. G. Metcalf & M. K. Estes. 1995. Detection of Norwalk virus and hepatitis A virus in shellfish tissues using the polymerase chain reac- tion./Ipp/. Environ. Microbiol. 61:3014—3018. Burkhardt. W., Ill & K. R. Calci. 2000. Selective accumulation may ac- count for shellfish-associated viral illness. .A/jpl. Environ. Microbiol. 66:1375-1378. Chung. H., L. A. Jaykus, G. Lovelace & M. Sobsey. 1998. Bacteriophages and bacteria as indicators of enteric viruses in oysters and their harvest waters. Wat. Sci. Tech. 12:37^4. Croci, L., D. De Medici, C. Scalfaro, A. Fiore. M. Divizia. D. Donia. A. M. Cosentino. P. Moretti & G. Costantini. 2000. Determination of en- teroviruses, hepatitis A virus, bacteriophages and Escherichia coli in Adriatic Sea mussels. J. Appl. Microbiol. 88:293-298. LITERATURE CITED Desenclos. J. C. F. Dubois, N. Mariotte & A. Goudeau. 1997. Analyse des strategies de depistage oriente de I'infection par le virus de I'hepatite C. Gastroenterol. Clin. Biol. 20:S25-S32. Desselberger, U. 1998. Viral gastroenteritis. Cnrr. Op. Inf. Dis. 11:565- 575. Dore. W. J., K. Henshilwood & D. N. Lees. 2000. Evaluation of F-specific RNA bacteriophage as a candidate human enteric virus indicator for bivalve niolluscan shellfish. Appl. Environ. Microbiol. 66:1280-1285. Dore, W. J. & D. N. Lees. 1995. B'ehavior of Escherichia coli and male- specific bacteriophage in environmentally contaminated bivalve mol- luscs before and after depuration. Appl. Environ. Microbiol. 61:2830- 28.^4. Dowell. S. P., C. Groves. K. B. Kirkland, H. G. Cicirello. T. Ando. Q. Jin. J. R. Gentsch. S. S. Monroe, C. D. Humphrey, C. Slemp, D. M. Dwyer. Indicators of Viral Shellush Contamination 1227 R. A. Meriwelher & R. I. Glass. 1995. A multistate outbreak of oysler- associated gaslroenleritis: Implications for interstate tracing of con- taminated shellfish. ./. /;;/: Di.y 171 : 1497- ISO.V Dupont. J.. D. Menard. C, Herve, F, Chevalier. B. Beliaff & B. Minier. 1 996. Rapid estnnalion of Escherichia coli in live murine bivalve shell- fish using automated conductance measurement. J. Appi Microlvul. 80:81-90. Efron, B. 1982. The jackknile, the bootstrap and other resamplmg plans. Soc. Ind. Applied Math. Reviews 38:92. Evans, H. S.. P. Madden, C. Douglas. G. K. Adak, S. J. Brien, T. Djuretic, P, G. Wall & R. Stanwell-Smith. 1998. General outbreaks of infectious mtestinal disease in England and Wales: 1995 and 1996. Coinmuu. Dis. Piihl. Hcallh 1:165-171. Ganlzer. C A. Maul. J. M. Audic & L. Schwartzbrod. 1998. Detection of infectious enteiovnus. enterovirus genomes, somatic coliphages and Bacleioiiles fiiif>ili.s phages in treated wastewater. Appl. Environ. Mi- crobial. 64:4307-4312. Griffin, D. W.. C. J. Gibson III, E. K. Lipp. K. Riley. J. H. Paul 111 & J. B. Rose. 1999. Detection of viral pathogens by reverse transcriptase PCR and of microbial indicators by standard methods in the canals of the Florida Keys. Appl. Environ. Microbiol. 65:41 18—1125. Havelaar, A. H. 1993. Bacteriophages as models of human enteric viruses in the environment. .A.S.M. News. JR. 59:614-619. Havelaar, A. H. & W. M. Pot-Hogebooni. 1988. F-specific RNA- bacteriophages as model viruses in water hygiene: Ecological aspects. Wat. Sci. Tech. 20:399-407. Havelaar. A. H.. M. van Olphen & Y. C. Drost. 1993. F-specific RNA bacteriophages are adequate model organisms for enteric viruses in fresh water. Appl. Em: Microbiol. 59:2956-2962. Hollinger. P. B. & J. R. Ticehurst. 1996. Hepatitis A virus. In: B. N. Fields. D. M. Knipe. P. M. Howley et al.. editors. Fields virology. 3rd ed. Philadelphia: Lippincott-Raven Publishers, pp. 78,V810. Kapikian. A. Z., M. K. Estes & R. M. Chanock. 1996. Norwalk group ol viruses. In: B. N. Fields, D. M. Knipe. P. M. Howley et af. editors. Fields virology, 3rd ed. Philadelphia: Lippincott-Raven Publishers, pp. 783-810. Leclerc. H., S. Edberg, V, Pierzo & J. M. Delattre. 2000. Bacteriophages as indicators of enteric viruses and public health risk in groundwaters. ./. Appl. Microbiol. 88:5-21. Lee, J. v.. S. R. Dawson, S. Ward, S. B. Surman & K. R. Neal. 1997. Bacteriophages are a better indicator of illness rates than bacteria amongst users of a white water course fed by lowland river. Wat. Sci. Tech. 35:165-170. Lees. D. N.. K. Henshilwood. C. 1. Gallimore & D. W. G. Brown. 1995. Detection of small round structured viruses in shellfish by reverse transcription-PCR. App/. Environ. Microbiol. 61:4418—1424. Legnani. P., E. Leoni, D. Lev. R. Rossi. G. C. Villa & P. Bisbini. 1998. Distribution of indicator bacteria and bacteriophages in shellfish and shellfish-growing waters. J. Appl. Microbiol. 85:790-798. Le Guyader. P.. E. Dubois. D. Menard & M. Pommepuy. 1994. Detection of hepatitis A virus, rotavirus, and enterovirus in naturally contami- nated shellfish and sediment by reverse transcription-scminested PCR. Ap/il. Environ. Microbiol. 60:3665-3671. Lc Guyader. F.. L. Haugarreau. L. Miossec. E. Dubois & M. Pommepuy. 2000. Three year study to assess human enteric viruses in shellfish. Appl. Environ. Microbiol. 66:3241-3248. Le Guyader. F.. L. Miossec. L. Haugarreau. E. Dubois. H. Kopecka & M. Pommepuy. 1998. RT-PCR evaluation of viral contamination in five shellfish beds over a 2 1 -month period. Wut. Sci. Tech. 38:45-50. Maillard, J. Y. 1996. Bacteriophages: A model system for human virus. / Appl. Microbiol. 23:273-274. Melnick. J. L. 1996. Enteroviruses: Polioviruses. coxsackieviruses, echo- viruses, and newer enteroviruses. In: B. N. Fields. D. M. Knipe. P. M. Howley et af, editors. Fields virology, 3rd ed. Philadelphia: Lippincott- Raven Publishers, pp. 655-712. Metcalt, T. G., J. L. Melnick & M. K. Estes. 1995. Environmental virol- ogy: From detection of virus in sewage and water by isolation to identification by molecular biology-a trip of over 50 years. .Annn. Rev. Microbiol. 49:461-487. Miossec, L., F. Le Guyader, L. Haugaueau & M. Pommepuy 2000. Im- portance de la pluviometrie sur la contamination virale du milieu lit- toral lors de phenomenes epidemiques dans la population. Rev. Epiileni. et Same Puhl. 48:2S62-2S71. Otsu, R. 1999. Outbreaks of gastroenteritis caused by SRSVs from 1987 to 1992 in Kyushu, Japan: Four outbreaks as.sociated with oyster con- sumption. Europ. J. Epidemiol. 15:175-180. Ricca. D. M. & J. J. Cooney. 1999. Coliphages and indicator bacteria in Boston Harbor, Massachusetts. Environ. To.xicol. 14:404^08. Schwab, K. J.. F. H. Neill. M. K. E.stes. T. G. Metcalf & R. L. Atmar. 1998. Distribution of Norwalk virus within shellfish following bioaccumula- tion and subsequent depuration by detection using RT-PCR. J. Food Prot. 61:1674-1680. Sobsey. M. 1997. Detection and occurrence of enteric virus pathogens and indicators in water. In: M. T. Martins. M. 1. Z. Sato. J. M. Tiedje. L. C. N. Hagler, J. Dohereirier & P. S. Sanchez, editors. Progress in microbial ecology. Sao Paulo. Brazil: SBM/Brazilian Society for Microbiology Press, pp. 469-474. Toze. S. 1999. PCR and the detection of microbial pathogens in water and wastewater. Wat. Res. 33:3545-3556. Tree. J. A.. M. R. Adams & D. N. Lees. 1997. Virus inactivation during disinfection of wastewater by chlorination and UV irradiation and the efficacy of F-H bacteriophage as a viral indicator. Wai. Sci. Tech. 35: 227-232. Vinje. J.. S. A. Altena & P. G. Koopmans. 1997. The incidence and genetic variability of small round-structured viruses in outbreaks of gastroen- teritis in The Netherlands. ./. //;/: Dis. 176:1374-1378. Jinirmil o) SluUfish Rcscimh. Vol. 20, No, ,^, l22y-l2,V\ 2001, NATURAL OCCURRENCP: OF VIBRIO SPP. AND LISTERIA MONOCYTOGENES IN MOLLUSCAN SHELLFISH IN PORTIKIAL F. S. PEREIRA.* M. M. GUERRA, AND F. A. BERNARDO CIISA/Lalwniuhio dc Inspec^-ao Saniuirin. /■'cuuklade de Mcdiciiia Vcleiinaiia-ilnivcrsidadc Teciuca de l.ishdd. Rim Prof. Cid dos Scintds. Paid Univcysihlrio da Ajiidii. 1300-477 Lisbon. Porliii^al ABSTRACT The naiiiral (Kx'Lirrence of Vibrm spp. and Ijstcnu iiuniocxtdticncs in shelllish u as analyzed o\er a period ol I S months, Slxtv-one samples of 14 different species of mollusean shellfish were collected ni local Lisbon markets (Portugal). Vibrio species were detected in 28 shellfish samples (45,y'/( ), Different vibrio species were identified m varying percentages of the shellfish samples tested: I', lyaraluiemolylicu.s (11.49r), V. algiiiolyriciis (6,59f). V. vttlnifuii.s (22,99f ). V. ilainsela (9.8%), V. metschnikovii (1,6%). and Vibrio spp. (21,3%), Listeria monocytoi^enes was isolated from eight samples (13,1%), Oxhev Li.sieria species, L. iiinoaia and L. grayi. were lound in 3,3% and 1,6% of shellfish samples, respectively. L. monocytogenes was found simultaneously with other different Listeria species in two samples. All L monocytogenes isolates were anti-sera poly O 1^+, Five ,samples (8,2% of total) were simultaneously contaminated with Vibrio spp, and Lisleriu spp. These results show that Vibrio spp. and L. monocytogenes are frequently found in shellfish products approved as safe, and the public health significance of these results should not be overlooked, KEY WORDS: Vibrio spp,. Listeria spp,, Portugal, occurrence INTRODUCTION Seafood has traditionally been a popular pari ol the diet in many regions of the world and in some countries, such as Japan, constitutes the main supply of animal protein. Today even more people are turning to fish as a healthy ahernative to red meat (Koenig et al, 1991 ), although there is still sotne resistance to this habitude. For example, between 1973 and 1987 in the United States, the meat consumption was appro.ximately 10 times that of fish (Huss 1994), This tendency of considering fish as healthy is related to the low fal content of many fish species and the effects on coronary heart disease of the W-3 polyunsaturated fatty acids found in fatty fish species, which are extremely important where cardiovascular disease mortality is high. However, consutiiption of fish and shellfish has been implicated in many cases of food-borne diseases outbreaks (Huss 1994), The various disease agents that have been associated with the consumption of seafood may conveniently be divided into two groups: indigenous bacteria and nonindigenous bacteria (Kueh & Chan 1985), Vibrio spp, and Listeria spp, are included in the mentioned group first (Huss 1994), Pathogenic Vibrio spp, are a human health hazard that occur naturally in estuarine waters. They are frequently found in filter- feeding niolluscan shellfish, which are generally the source of infection. Disease resulting from raw oyster ingestion results in fatality rates exceeding 50% for V. vtiliiiftcii.s infections (Wittman & Flick 1995, Hiady 1997), The incidence of pathogenic Vilvio spp, has been correlated with elevated water temperatures (Kaneko & Colwell 1973. HIady 1997), Typically, pathogenic Vilvio spp, multiply rapidly at temperatures between 20 C and 40°C, This is reflected in the large numbers of the organisms isolated from mol- lusean shellfish when water temperature rises to 30°C, and their virtual absence from molluscs taken from cold waters (ICMSF 1 998 1, For example, V. paruliaeinolyticus is widespread when wa- ter temperatures exceed 15''C (Kaneko & Colwell 1973), Never- theless they also have been i,solated from crabs taken from cold waters (ICMSF 1998), Vilvio spp, levels in the estuarine environ- ment are also dependent on the time of the day. depth, and tidal *Corre.sponding author. levels, and each of those factors must be taken into account to get an accurate estimate of vibrios in the environment (ICMSF 1998), V, paruliaeinolyticus. Vibrio cholerae. and Vibrio vulnificus are routinely part of the microflora in crustaceans captured from es- tuarine waters. Vibrio species cause a variety of human infections, w hich can usually be classified as gastrointestinal or extraintestinal infections (Kudoh 1988), Species most comtnonly associated with dianhea or gastroenteritis include V. cholerae (Ol and non-Oi) and V. parahaemolyticus. However, V, fluvicdis. V. furnissii. V. mimicus. and V. hollisae have also been shown to cause or be significantly associated with gastroenteritis (Kudoh 1988, Witt- man & Flick 1995), Interestingly, some of these Vibrio species may also cause extraintestinal infections such as wound infections and secondary septicemia (Kudoh 1988). V. vulnificus is the well- known cause of primary and/or secondary septicemias and wound infections with high mortality. The hazards associated with inges- tion of seafood contaminated with this potential pathogen have resulted in a precautionary statement issued by the U,S, Food and Drug Administration against consumption of raw or undercooked seafood by people with hepatic disorders (Anonyinous 1992), V. alginolyticus and V. damscla also cause several types of soft-tissue infections and wound infections, V. inetsclmikovii has been impli- cated as a cause of opportunistic infections, but its pathogenic significance in the illness needs further investigation. According to Wittman and Flick (1995), between 1984 and 1993. the greatest percentage of known death (95%) due to shellfish-borne disease was caused by noncholera vibrios; noncholera vibrios also ac- counts for the second highest causative agents of disease. Fish and seafood are considered potential sources of Listeria spp, because this microorganism is ubiquitous in nature (Farber & Peterkin 1991) and salt-tolerant (Dillon & Patel 1992), Fisheries and aquaculture products are directly exposed to water contatiii- nation. In fact. Listeria spp,, and particularly L. monocytogenes. have been isolated from a variety of fish and shellfish products (Ben Embarek 1994, Dillon et al, 1992, Dillon et al, 1994, Jor- gensen & Huss 1998, Weagant et al, 1988), Although not very involved in listeriosis, the presence of L, monocytofieiies in sea- food and particularly in shellfish has been implicated in some cases, A described epidemic of perinatal listeriosis was related to the consumption of raw fish and shellfish (Lennon et al, 1984), and 1229 1230 Pereira et al. some sporadic cases have been identified as associated with shell- fish, precisely smoked mussels and raw oysters (Plusquellec et al. 1998). Also a small outbreak of listeriosis associated with smoked mussels has been reported (Brett et al. 1998). To minimize bacte- rial growth and exposure to this organism, it is important to have adequate control of the environment for food production as well as to take precautions for adequate cooking and handling food (Rob- erts 1994). In Portugal, like in man\ other countries, food-borne diseases are not appropriately reported, thus the epidemiological impor- tance of seafood-borne diseases is lacking. Although some refer- ences have been made on the isolation of L monocytogenes in molluscan shellfish in Portugal (Pedro 1996), there is not much information on this subject and the occurrence of Vibrio spp. is totally unknown. The consumption of molluscan shellfish is very popular in Portugal, and has increased in the last ten years by 4Vf (SNI 1997). Additionally, raw or undercooked molluscan shellfish are the favored ways of eating it. The aim of this work was to perform an evaluation of the occurrence of Vibrio spp. and Listeria spp., particularly L. monocytogenes, in molluscan shellfish com- mercialized in Portugal to obtain some information about the sig- nificance of these hazards. MATERIAL AND METHODS Samples were collected between March 1998 and July 1999. Fresh molluscan shellfish were obtained from local Portuguese markets and were investigated for the presence of Vibrio spp. and Listeria spp. A total of 61 samples included three oysters {Cras- sostrea cmgulata). one hen clam (Mcictra coralina). five furrow shell (Scrobicidaria plana), seven mussels {Mytihts ediilis). seven wedge shell (Donax tnmcidus). six carpet shell {Tapes deciissa- tus). seven purr (Veneropsis cornigata). two purr (Veneropsis ri] 2 0 0 L. litorea 4 u 0 0 0 Buccinum spp. 4 1 1 0 0 5. inarginatiis 4 1 1 0 0 S. officinalis -) 0 0 0 0 P. cornucopia -1 1 1 0 0 Total fil 91 U.S'fl Sll3.1^r) 2(3.3<*i 1 ( 1 .6^<- ) enough to kill Vibrio spp. Other outbreaks have been attributed to contamination of cooked seafood by raw seafood followed by inadequate refrigeration (Blake et al. 1980. Hobbs & Roberts 1987). The levels from the inner tissue of the animal are lower than the ones reported previously from molluscan shellfish collected in Hong Kong markets that ranged between 10"* and 10'' cfu/g (Chan et al. 1989). This might be due to the lower temperatures of the seawater in Europe (between 13°C in winter and 22 "C in summer). On the other hand, it is known that enrichments plated on CPC agar are more appropriate for V. viiliiificiis. as well as including the addition of polimixin to the enrichment medium (Dalsgaard & Hoi 1997). This might explain the lower levels obtained once the initial research procedures were performed with nonspecific culture me- dia. However, media designed for one vibrio could delay or inhibit the growth of other Vibrio species. The incidence of V. viiliuficiis. V. algiiiolyricKs. and V''. damsela was not characterized by a sea- sonal distribution, probably due to the low temperature amplitudes in Portuguese seawater. On the other hand. V. parahaemolyticus was isolated only during the vsarmer months. L/.sr('nfl-contaminated samples were collected on March and April 1998 and May and July 1999 (Fig. 1). which reveals the tendency of this microorganism to develop in warmer temperatures (Dillon and Patel 1992). In a previous study conducted in Portugal. Pedro (1996) isolated Listeria spp. and L. monocytogenes in li'Jc and 20%. respectively, of samples from 3.'i shellfish. Isolations were only performed on clams and fuiTow shells. These results and this study's results obtained in Portugal reveal some differences when compared with those cited in literature where the presence of this microorganism is mostly found in oysters and mussels, de Simon et al. ( 1992) found Listeria spp. in 22.59f of the 40 samples of mussels analyzed and detected L. monocytogenes in three of them (7.5'>^). These authors also isolated L. innocua and L. seeligeri in 12.,'i';f and 2S'7r of the mentioned samples, respec- tively. In oysters, Weagant et al. (1988). Buchanan et al. (1989), and Colburn et al. (1990) did not find Listeria in any seafood sample. Motes (1991) did not find Listeria spp. from 75 oysters samples, all collected from estuarine environments along the U.S. Gulf Coast. Wilson (1995) isolated Listeria spp. in 11.0% of the molluscan/crustacean samples analyzed (8/74) in Ireland. In Fin- land. Hartemink and Georgsson (1991) found Listeria spp. in samples of raw and smoked fish, but no isolation was made in 1 1 shellfish samples. Jeyasekaran et al. ( 1996) isolated Listeria spp. in 44.4% ( 16/36) of shellfish samples in India, and L. monocytogenes was found in 12.1% of the contaminated samples. These authors also isolated L. innocua and L. welshimeri in 36.1% and 5.6% of the tropical samples, respectively. In our study. L. monocytogenes was the predominant species (13.1%) and L. innocua the second most common (3.3%). In the literature L. innocua is seldom referred to as being the most com- monly isolated species (Weagant et al. 1988. Ryu et al. 1992, Wilson 1995. Jeyasekaran et al. 1996). This fact is related to the shorter generation time of L. innocua in enrichment broth, which may lead to false-negative L. monocytogenes results in cultures v\here the two species are both present (Petran & Swanson 1993, Curiale & Lewus 1994). On the other hand, the coexistence of these two species in some samples (Table 3) indicates that these microorganisms may share the same ecological niche. Further- more, the presence of nonpathogenic Listeria spp. is considered to be an "indicator" of the presence of the pathogenic species (King et al. 1989). The nonselective enrichment step followed by a se- lective enrichment in two different media might have contributed to the slight enhancement of L. monocytogenes, because it consti- tutes a resuscitation step that may favor its recovery. This proce- dure has been recommended especially when examining products where Listeria spp. cells are expected to be injured (McCarthy et al. 1990). Vibrio spp. and Listeria spp. were isolated simultaneously in t"ive samples. These results have never been reported in Portugal. The analysis of the epidemiology of some food-borne diseases reveals a significantly higher rate of coinfection in patients due to naturally occurring pathogens (Hlady 1997). Hence, these results suggest the possibility of a higher risk with the consumption of these particularly molluscan shellfish. Considering the direct utilization of molluscan shellfish in hu- man diet and the deficient cooking habits applied to shellfish. L. monocytogenes and Vibrio spp. occurrence must be regarded as a relevant hazard to public health. Therefore, due to the importance of seafood on the Portuguese alimentary habits, it would be im- piirtant to establish the real hazards to humans. ACKNOWLEDGMENTS Financial support for authors F. Pereira and M. M. Guerra was provided by PhD fellowships within the PRAXIS XXI Program. Medida 4. administered by the National Board for Scientific and Technological Investigation (FCT. Portugal). Partial funding of this project was provided by the Interdisciplinary Investigation Animal Sanitation Centre (CIISA) of the F.M.V. of Lisbon. LITERATURE CITED Anonymous. \992. Bad Bug Book. U.S. Food and Drug Administralion Centre for Food Safety and Applied Nutrition |oii linel- Available through internet via wvvvv.cfsan.fda.gov/-mow/intro.html. File down- loaded on ()2.0.V 199(1. Vibrio spp. and L. monocytuuenes in Shellfish 1233 Ben Embarek, P. K. 1994. Presence, detection and growth of Listeria nioncicMiii^enes in seafoods: A review. Iiir. J. Food Microbiol. 23:17- M Blake. P. A.. R, E. Weaver & D. G. Mollis. 19S0. Disease of humans i other than cholera) caused by vibrios. Ann. Rev. Microbiol. 34:341-367. Brett. M. S. Y.. P. Short & J. McLauchlin. 1998. A small outbreak of listeriosis associated with smoked mussels. Iiii. .1. Food Microbiol. 43:223-229. Buchanan. R. L.. H. G. Stahl. M, M. Bencivengo & F. D. Corral. I9S9. Comparison of lithium chloride-phenylethanol-moxalactam and modi- tied Vogel Johnson agars for detection of Listeria spp. in retail-level meats, poultry and seafood. Appl. Environ. Microbiol. 52:59-63. Chan, K.. M. L. Woo. L. Y. Lam & G. L. French. 1989. Vibrio pani- haemolyticiis and other halophylic vibrios associated with seafood m Hong Kong. J. .Appl. Bactcriol. 66:57-64. Colburn. K. G.. C. A. Kaysner. C. Abeyta & M. M. Wekell. 1990. Listeria species in a California coast estuarine environment. .\ppl. Environni. Microbiol. 56:2007-2011. Curiale. M. S. & C. Lewus. 1994. Detection oi Listeria monocytoi>enes in samples containing Listeria iniiocua. J. Food Prot. 57:1048-1051. Dalsgaard. A. & L. Hoi. 1997. Prevalence and characterisation of Vibrio vitlnificiis isolated from shrimp products imported into Denmark. J. Food Prot. 60:1132-1135. de Simon. M.. C. Tarrago & M. D. Ferrer. 1992. Incidence of Listeria monocytogenes in fresh foods in Barcelona (Spam). Int. J. Food Mi- crobiol. 16:153-156. Dillon, R.. T. Patel & S. Ratnam. 1992. Prevalence of Listeria in smoked fish. J. Food Prot. 55:866-870. Dillon. R., T. Patel & S. Ratnam. 1994. Occurrence of Listeria in hot and cold smoked seafood products. Int. J. Food Microbiol. 22:73-77. Dillon, R. M. & T. Patel. 1992. Listeria in seafoods: A review. J. Food Prot. 55:1009-1015. Farber, J. M & P. I. Peterkin. 1991. Listeria iiionocyloi;enes a foodborne pathogen. Microbiol. Rev. 55:476-511. Hartemink. R. & F. Georgsson. 1991. Incidence of Listeria species in seafood and seafood salads. Int. J. Food Microbiol. 12:189-196. HIady. W. G. 1997. Vibrio infections associated with raw oyster consump- tion in Flonda. 1981-1994. J. Food Prot. 60:353-357. Hobbs. B. C. & D. Roberts. 1987. In: E. Arnold, editor. Vibrosis other than cholera. Food poisoning and food hygiene, 5th ed. London: Edward .'\mold. Huss. H. H. 1994. Assurance of Seafood Quality, FAO, Fisheries Technical Paper No. 334. Food and .Agriculture Organisation. Rome: Viale delle Termi di Caracalla. 1-00100. ICMSF (International Commission on Microbiological Specifications for Foods). 1998. Microorganisms in foods. Microbial ecology of food commodities. London: Blackie Academic & Professional, pp. 134- 160. Jeyasekaran. G.. I. Karunasagar & I. Karunasagar. 1996. Incidence of Listeria spp. in tropical fish. Int. .1. Food Microbiol. 31:333-340. Jorgensen. L. V. H. H. & Huss. 1998. Prevalence and growth of Listeria monocytogenes in naturally contaminated seafood. Int. J. Food Micro- biol. 42:127-131. Kaneko. T. & R. R. Colwell. 1973. Ecology of Vibrio parahueiiiolxtHiis m Chesapeake Bay. / Bacterial. 1 13:24-32. King. W., S. Raposa, J. Warshaw. A. Johnson. D. Halhert & J. D. Kliger. 1989. A neu colorimelric nucleic acid hybridization assay for Listeria in foods. /;;;. ./. Food Microbiol. 8:225-232. Koenig. K. L.. J. Muller& T. Rose. 1991. Vibrio viilnifuKs: Hazard on the half shell. Western J. Med. 155:400-404. Kudoh. Y. 1988. Laboratory diagnosis of infectious diseases, principles and practice. New York: Springer-Verlag. p. 533-539. Kueh. C. S. W. & K. Chan. 1985. Bacteria in bivalve shellfish with special reference to the oyster. J. Appl. Bacieriol. 59:41-47. Lennon. D.. B. Lewis. C. Mantell. D. Becroft. B. Dove. K. Farmer. S. Tonkin, N. Yeates, R. Stamp & K. Mickleson. 1984. Epidemic peri- natal listeriosis. Pediatr. Infect. Dis. 3:30-34. McCarthy, S. A.. M. L. Motes & R. M. McPhearson. 1990. Recovery of heat-stressed Listeria monocytogenes from experimentally and natu- rally contaminated shrimp. / Food Prot. 53:22-25. Motes. M. L., Jr. 1991. Incidence of Listeria spp. in shrimp, oysters and estuarine waters. J. Food Prot. 54:170-173. NACMCF. 1992. Microbiological criteria for raw molluscan shellfish. Recommendations by the National Advisory Committee on Microbio- logical Criteria for Foods. J. Food Prot. 55:463—180. Pedro. S. C. N. S. C. 1996. Ocorrencia de Listeria em pescado e sua ecorresistencia em meio aquatico. Dissertaij-ao para a obtenij-ao do graus de Mestre em Ciencia e Tecnologia de alimentos pela Universidade Tecnica de Lisboa. Petran. R. L. & K. M. J. Swanson. 1993. Simultaneous growth of Listeria monocytogenes and L. innoctia. J. Food Prot. 56:616-618. Plusquellec. A., P. Monfort, G. Piclet & L. Rio. 1998. Listeria innociia and Salmonella panama in mussels: A comparative study. J. Food Prot. 61:1137-1142. Roberts, D. 1994. Listeria monocytogenes and food: The U.K. approach. Daily Food Environ. Sanil. 14:202-204. Ryu, C.-H., S. Igimi, S. Inoue & S. Kumagai. 1992. The incidence of Listeria species in retail foods in Japan. //;/. / Food Microbiol. 13:97- 104. S.N.I. (Statistic National Institut). 1997. Portuguese Consumption Balance 1990-1997. Weagant. S. D., P. N. Sado, K. G. Colburn, J. D. Torkelson, F. A. Stanley, M. H. Krane, S. C. Shields & C. F. Thayer. 1988. The incidence of Listeria species in frozen seafood products. J. Food Prot. 5 1 :655-657, Wilson, I. G. 1995. Occurrence of Listeria species in ready to eat foods. Epidemiol. Infect. 115:519-526. Wittman. R. J. cS: G. J. Flick. 1995. Microbial contamination of shellfish: Prevalence, risk to human health, and control strategies. Ann. Vcr. Pub. Health 16:123-140. Jcnirnul of Shclllhli Research. Vol. 211, Nd. 3. 1235-1240. 2ll()K THE PREVALENCE OF NONCLLTIVABLE BACTERIA IN OYSTERS (TIOSTREA CHILENSIS, PHILIPPE 1845) JAIME ROMERO AND ROMILIO ESFEJO* Lahorcitorio dc Bioingenicrui. Insiitiito dc Nntricidn y Tecuoldi^ia de li>\ Aliiuciitos. Universidud de ChUc. Maciil 5540. Santlai;i>. Chile .ABSTR.ACT Gencnypic analysis based on molecular methods lias become a iiseliil tool to determine the composition of bacterial populations and communities. We used this tool to study the microflora of oysters. Analysis of the electrophoretic mobility of the I6S-23S rDNA spacers of the bacteria, isolated from oysters and cultured in marine agar, revealed the presence of three main groups of bacteria. Within each group the bacterial strains had identical I6S rDNAs. as judged by the electrophoretic mobility of the heteroduple.xes formed between amplified 16S rDNA. The extrapolation of these findings to the total bacterial population, however, seems unwarranted because the number of colonies grown in agar was less than 0.00 1 9r of the bacteria present in the oysters. Consistently, the electrophoretic mobility of the 16S-23S rDNA spacers of the total bacterial DNA extracted from oysters had a very different pattern of that extracted from the bacterial cultures. Analysis of bacterial DNA obtained from individual oysters showed the presence of some spacers in most of the oysters examined. These spacers may belong to bacteria of a common microflora that are noncultivable in marine agar. We conclude that most bacterial strains present in oysters are not cutivable in marine agar. KEY WORDS: microtlora. noncultivable bacteria, oyster, intergenic spacer INTRODUCTION Animals harbor large, active, and complex cotimiunities of mi- crobes. Among them is the normal microflora, a population of microbes that despite its conliiuial contact with different tissties causes no harm to the host. On the contrary, the normal microflora may have beneficial effects. It may participate in the host nutrition by complementing the process of food digestion or by contributing to the metabolism of endogenous and exogenous compounds. Nor- mal microflora may also play a role in pathogenesis by preventing the host colonization by pathogens, controlling their growth, and modulating the immune response of the host. The composition of each particular microflora is probably determined by host proper- ties, such as its contact history with colonizing bacteria and the dietary and environmental conditions that presail in its habitat (Kirjavainen & Gibson 1999, Savage 1986). At harvest oysters normally contain 10' to 10"" CFU/g (De Paola et al. 1990. Wright et al. 1996). which may consist of either nomial microflora or bacteria present in recently ingested food (Cook 1991, Vanderzant & Thompson 197.^) or both. Bacterial flora in oysters has been described by phenotypic characterization of the strains isolated by culture, usually in marine agar (Colwell & Liston 1960. Murchelano & Brown 196S). Vihiio and Pseudomonas spp. are the bacterial species most frequently iso- lated from oysters. They account for 20-30% of the total CFU found in these bivalves (Colwell & Liston 1960. Murchelano & Brown 1968. Kueh & Chan 1985). The load of vibrios in the oyster's normal microflora varies according to several factors in- cluding the environmental temperature (Kaspar & Tamplin 1993) and the oyster health status (e.g., the presence of parasites) (Tail et al. 1999). Among the vibrios, important human pathogens such as v. vulnificus and V. panihaemolytkus are occasionally found in oysters (Motes et al. 1998. Kaspar & Tamplin 1993. De Paola et al. 1990. Cook 1994, Wright et al. 1996). However, since the only bacterial microflora thus far characterized in oysters is that able to grow in consentional culture media, it is likely that important strains unable lo grow in such media have been missed by this kind 'Corresponding author. E-mail: respejo@uec.inta.uchile.cl of analysis. This situation has been described in many different bacterial habitats (Amman et al. 1995. Ueda et al. 1999. Suau et al. 1999). An example of a noncultivable bacteria found in oysters is Crisli.spirii peclincus. a large spirochete that is commonly identi- fied by its shape and large size in the crystalline style (Paster et al. 1996. Margulis et al. 1991). The availability of methods for the detection and genetic char- acterization of bacteria irrespective of the microorganism's ability to grow in culture inedia offers the opportunity to better charac- terize the bacterial microflora of different animal species including the oyster (Amman et al. 1995). Two of the most commonly used methods for genetic characterization of bacterial communities are the analysis of the nucleotide sequence of 16S rDNAs (Schmidt et al. 1991) and the study of the size of the spacers found between the 16S and 23S rRNA genes (Gurtler & Stanisich 1996. Jensen c 550 bp JB P — Figure 3. 16S-23S rDN.\ spacer pattern obser\ed after PCR amplifi- cation of DN.A extracted from the colonies grown in marine agar plates (CFU) and directb from Ihe o\ster pool (I)NAl. Arrows indicate the position and size of the main spacers ol)served in each sample. 1238 Romero and Espejo bacteria of individual oysters by amplification of DNA extracted directly from the oysters without prior cultivation. The premise behind the idea of analyzing individual oysters was that those strains belonging to the normal microflora should be present in most of the oysters. DNA was extracted from six indisidual oysters and the patterns of the I6S-23S spacers, obtained after amplifica- tion, were compared. Despite the fact that different individuals had different spacer patterns, some spacers with similar migrations were present in most of the oysters. Three specific spacers were readily observable in four of the six oysters, though in different propoilions (Fig. 4A). Their approximate sizes were 1000, 55(1. and 450 bp. In addition, the 550- and 450-bp spacers were present in six and five of the oysters, respectively. Furthermore, the 1.000- and 550-bp spacers were also detected in pooled samples of DNA obtained from the homogenales prepared from oysters harvested both in October and August (Fig. 4B). DISCUSSION Increased knowledge on the bacterial ecology of the oyster microflora would greatly contribute to the understanding of the role of bacterial microflora on the host digestion, metabolism, and protection from pathogens. This knowledge may be applied to hatchery management and also to postharvest processes to improve oyster quality and safety. A first requisite to ecological studies is a comprehensive knowledge of the community composition. Our results show that, as it has been observed in many other marine habitats (Amman et al. 1995). the bacteria obtained after cultiva- (A) 43«kv (B) Aug Oct Ladder ^ . , lOObp M**"" »ll#ilKlM» i 1000 bp 550 bp 450 bp Figure 4. 16,S-2.^S rDNA spacer observed after uinpliflcation of DN.\ extracted from individual oysters (A) and from pools of six oysters harvested in August and Octoher 1999 (B). The resuhs of only four oysters are shown (lanes oysters 1—4). Arrows and asterisks indicate the position and size of three spacers observed in most of the individual examined. tion represent only a minor fraction (<0.001%) of the total bacte- rial population present in the oysters. This recovery is one of the lowest reported in marine habitats (Amman et al. 1995. Schmidt et al. 1991 ). While the reason for this low recovery is not known, it is possible that a large part of the total bacterial count included nonviable microorganisms such as ingested allochthonous bacteria inactivated by the oyster digestion process. On the other hand, the absence of cultivable putative pathogenic vibrios, those able to grow at 37°C. may be due to the low coastal water temperature in this region ( 12-14"C year-round) as it is known that the presence of cultivable vibrio is greatly diminished at low temperature (Kas- par & Tamplin 1993. Motes et al. 1998). We cannot disregard, however, their presence in a viable noncultivable state as it has been described in some pathogenic Vibrio spp. like Vibrio vulnifi- cus (Oliver 1995) and Vibrio paraliueinolxticus (Jiang & Chai 1996). Cultivable bacterial strains may not only represent a minute fraction of the total nuinber but also of the diversity of the micro- organisms present in these bivalves. Detailed phenotvpic or geno- typic analysis based exclusively on the study of the bacterial com- munity obtained after culturing seems to be of little value for a general understanding of the bacterial ecology in oysters. This does not imply that the results obtained from studies using culti- vated bacteria are worthless. This type of study does give infor- mation about bacteria actually present in the hosts of study. Its limitation, however, is one of lack of generalizability. A more complete tinderstanding can certainly be obtained from studies that combine both the analysis of cultivated bacteria as well as that of bacteria directly obtained from the host without prior cultivation. Bearing in mind the limitation of culture studies, our results with cultured bacteria provide interesting information regarding the genetic features of these strains. The fact that most of the cultured bacteria could be grouped in three clusters according to the overall band pattern and the presence of common bands indi- cates that these three groups may prevail among the cultured strains. The high sequence similarity of the 16S rDNAs within the same spacer group and the dissimilarity observed between groups suggests that spacer patterns may be a useful tool to distinguish species in the oyster ecosystems. Interestingly, strains in group C showed the largest intragroup heterogeneity in spacer patterns and rendered more than one band after amplification of their 16S rDNA. As previously shown, these bands correspond to heterodu- plexes formed during the last cycle of amplification by annealing between single strands of different sequences (Espejo et al. 1998). These heteroduplexes are probably due to the presence of I6S rDNAs differing in sequence. Our results add to the increasing number of reports of intracellular heterogeneity in 16S rDNA (Ueda et al. 1999). According to the results obtained with cidtiired strains, the observation of particular spacers may be used as indication of the possible presence of specific bacterial strains. Considering that cultured bacteria correspond to such a low fraction of the total bacteria present in oysters, the large differences in the spacer pat- terns observed between bacterial DNA extracted from oysters and DNA obtained from cultured bacteria is not surprising. This large difference suggests that the groups of microorganisms observed after cultivation in marine agar do not represent the prevailing bacteria of oysters. The spacer patterns obtained from indi\idual oysters indicate the presence of a complex and different community of microor- ganisms in each individual. However, the spacers cominon to most Common Bacirria in Individual Oysters 1239 (if the oysters analyzed (coiresponding to approximate si,^es of l.OUO. 5f)(), and 430 bp) may correspond to bacterial strains that constitute part of the normal oyster microflora and as such are present in most individuals. The apparent presence of these same spacers in pools of oysters har\ested at different times of the year also supports this hypothesis. These spacers would correspond to bacteria unable to grow in marine agar because they are not ob- served after amplification of the DNA obtained from the cultiued bacteria. These results are similar to those obtained on the human mtes- tinal microflora, one of the most extensively studied microbial habitats. Intestines contain 10'" to lO" CFLl/g with an overall composition that differs among individuals but that may be stable for months in the same person (Zoetendal et al. 1998). Some bacterial species appear to be present in most individuals and are considered part of the normal microflora (Zoetendal et al. 1998, Moore & Moore 1995, Matsuki et al. 1999). Many of these species can only be detected through direct DNA analysis because it has not been possible to culture them (Suau et al. 1999). In the above- mentioned studies, however, the fraction of cultured bacteria is much larger than in our study. The variations in the composition of normal microflora be- tween individuals due to environmental and dietary differences are likely to be less pronounced in oysters than in human beings. Oysters grow in a much more homogeneous environment, which makes it more likely that they share the same diet and conse- quently similar microflora. Our results suggest that in oysters there are some bacterial strains that may form part of the normal microflora. These strains, however, are not observed when the bacteria are cultured in marine agar, as is commonly the case. For an appropriate and comprehen- sive description of the microflora, needed for understanding the bacterial ecology in oysters, it is necessary to distinguish between bacteria of the norinal microflora and allocthonous bacteria, such as opportunistic bacteria or those ingested with food. We plan to accomplish this by examination of the prevailing spacers after depuration of the oysters to remove opportunists and ingested bac- teria. Once the putative normal microflora has been defined, cul- tivation of the bacteria containing the main spacers will be at- tempted using specially designed media and different culture con- ditions. Successful cultivation will be necessary to accomplish the phenotypic identification of the bacteria constituting the normal microflora of oysters. ACKNOWLEDGMENTS This work was supported by grants 1990765 and 2000064 from FONDECYT. Amman, R.. W. Liidwmg & K. Scheiler. ]W5. Phylogenetii. idcntilicatioii and in situ detection of individual micmhial cells wilhcuil cullualion. Microbiol. Rev. 59:143-169. Collins, C. P. Lyne & J. Grange. 1991. Microbiological methods, hili ed. Oxford: Butterworth-Heineniann. pp. 250-258. Colwell, R. & J. Listen. 1960. Bacteriological study of the natural tlora ol the Pacific oysters (Crassostrea gigas). Appl. Microbiol. 8:104-109. Cook, D. W. 1991. Microbiology of bivalve molluscan shellfish. In: R. Ward & C. Hackney, editors. Microbiology of foods products. New York: Van Nostrand Reinhold. pp. 19-39. Cook, D. W. 1994. Effect of time and temperature on multiplication of Vibrio viiliiificKs in postharvest Gulf Coast shellstock oysters. Appl. Environ. Microbiol. 60:3483-3484. De Paola, A., L. Hopkins, J. Peeler, B. Wentz & R, McPhearson. I WO. Incidence of Vibrio parahaemolyticiis in U.S. coastal waters and oys- ters. Appl. Environ. Microbiol. 56:2299-2302. Espejo, R., C. Feijoo, J. Romero & M. Vasquez. 1998. Page analysis of the heteroduplexes formed between PCR amplified 16S rihosomal RNA genes: Estimation of sequence similarity and rDNA complexity. Mi- crobiology 144:1611-1617. Gurtler, V. & V. Stanisich. 1996. New approaches to typing and idcnliti- cation of bacteria using the 16S-23S rDNA spacer region. Microbiol- ogy 142:3-16. Jensen, M. A. & N. Straus. 1993. Effect of PCR conditions on the forma- tion of heteroduplex and single-stranded DNA products in the ampli- fication of bacterial ribosomal DNA spacer regions. PCR Methods Appl. 3:186-194. Jiang, X. & T. Chai. 1996. Survival of Vibrio pariihaeiiiohiiciis al low temperatures under starvation conditions and subsequent resuscilalion of viable, nonculturahle cells. Appl. Environ. Microbiol. 62:1300- 1.^05. Kaspar. C. W. & M. Tamplin. 1993. Effects of temperature and salinity on the survival of Vibrio vulnificus in seawater and shellfish, ■^ppl. Envi- ron. Microbiol. 59:2425-2429. Kirjavainen, P. V. & G. Gibson. 1999. Healthy gut nncronora and allergy: Factors intluencing development of the microblota. Ann. Meil. 3 1 :28S- 292. LITERATURE CITED Kueh. C. & K. Chan, l4X,s. Bacleria in bivalve shellfish with special reference to the oysters. / Appl. Bacleriol. 59:41-47. Kuwae. T. & Y. Hosokawa 1999. Determination of abundance and bio- volume of bacteria in sediments by dual staining with 4'.6-diamidino- 2-phenylindole and acridine orange: Relationship to dispersion treat- ment and sediment characteristics. Appl. Environ. Microbiol. 65:3407- .^412 Maigulis, L.. L. Naull & J. Sieburth. 1491. Criiispira from oyster styles: Complex morphology of large symbiotic spirochetes. Sviiibio.^i.^ 11:1- 19. Matsuki, T, K. Watanabe, R. Tanaka. M. Fukuda & H. Oyaizu. 1999. Distribution of bifidobacterial species in human intestinal microflora examined with 163 rRNA-gene- targeted species-specific primers. Appl. Environ. Microbiol. 65:4506—1512. Moore, W. E. & L. Moore. 1995. Intestinal Horas of populations that have a high risk of colon cancer. Appl. Environ. Microbiol. 61:3202-3207. Motes, M. L., A. De Paola, D. W. Cook, J. E. Veazey, J. C. Hunsucker, W. E. Garthright, R. J. Blodgett & S. Chirtel. 1998. Influence of water temperature and salinity on Vibrio viilnificu.s in Northern Gulf and Atlantic Coast oysters {Cra.tsoslrea virginicii). Appl. Environ. Micro- biol. 64:1459-1465. Murchelano, R. & C. Brown. 1968. Bacteriological study ot the natural flora of the Eastern oysters. Crussoslrea virginica. J. Invert. Pathol. 11:519-520. Oliver. J. D. 1995. The viable hut not cultuiable slate in the human patho- gen. Vibrio vulnificus. FEMS Microbiol. Lett. 133:203-208. Paster, B., D. Pelletier, F. Dewhirst, Vv". Weisburg, V. Fussing, L. Poulsen, S. Dannenberg & I. Schroeder. 1996. Phylogenetic position of the spirochetal genus Cristispira. Appl. Environ. Microbiol. 62:942-946. Pizarro, J., E. Jedlicki, O. Orellana. J. Romero & R. Espejo. 1996. Bacterial population in samples of bioleachcd copper ore as revealed by analysis of DNA obtained before and after cultivation. Appl. Environ. Micro- biol. 62:132.VI328. Savage, D. C. 1986. Gastrointestinal microlloia in mammalian nutrition, Aimu. Rev. Nutr. 6:155-178. 1240 Romero and Espejo Schmidt. T., E. De Long & N. Pace. 1991. Analysis of a marine pico- plankton community by 16S rRNA gene cloning and sequence. J. Bac- h'liol. 173:4371-4378. Suau. A., R. Bonnet, M. Sutren, J. Godon. G. R. Gibson. M. D. Collins & J. Dore. 1999. Direct analy.sis of genes encoding 16S rRNA from complex communities reveals many novel molecular species within the human gut. .4/)/)/. Environ. Microbiol 65:4799—1807 Tall. B. D.. F. La Peyre. J. Bier. M. Miliotis. D. Hanes. M. Kothary. D. Shah & M. Faisal. 1999. Perkinsus marinus extracellular protease modulates survival of Vibrio vulnificus in Eastern oyster (Crassostreo virginica] hemocytes. Api^l. Environ. Microbiol. 65:4261^263. Ueda, K.. T. Seki. T. Kudo, T. Yoshida & M. Kataoka. 1999. Two distinct mechanisms cause heterogeneity of 16S rRNA. J. Bacicruil. 181:78- 82. Vanderzant. C. & C. Thompson. 1973. Microbial flora and level of Vibrio puraluiemolyticus of oysters [Cra.s.soslreii virginico). water and sedi- ment from Galveston Bay. / Milk FooJ Tcchuol. 36:443-452. Wright. A. C, R. T. Hill. J. A. Johnson, M. C. Roghman. R. R. Colwell & J. Morris. Jr. 1996. Distribution of Vibrio vulnificus in the Chesapeake Bay. Appl. Environ. Microbiol. 62:717-724. Zoetendal. E. G.. A. D. L. Akkermans & W. De Vos. 1998. Temperature gradient gel electrophoresis analysis of 16S rRNA from human fecal samples reveals stable and host-specific communities of active bacteria. Appl. Environ. Microbiol. 64:3854—3859. .I.ninuil ol Shellfish Research. Vol. 20, No. 3. 1241-124.1. 21101. ENTERIC VIRUS CONTAMINATION OF SHELLFISH: INTERVENTION STRATEGIES GARY P. RICHARDS United Slates Department of Agrieiilture. Ai^nciiltural Research Service. Delaware State University. W.W. Baker Center. Dover. Delaware 19001 ABSTRACT Enteric viruses, such as hepaUlis A. Norv\alk-like caliciviruses. rotaviruses, and astroviruses. are responsible for outbreaks of food-borne illnes.s. There are an estimated 9.2 million cases of food-borne Norwalk-like illness in the United States each year. The portion of those cases associated with shellfish is uncertain; however, shellfish are a major vector of human caliciviruses. In addition to the classical viral illnesses transmitted by shellfish, hepatitis E may become a potential threat to the shellfish consumer, particularly in Asian countries. Intervention strategies to enhance product safety include increased industry and consumer education: changes in harvesting and water monitoring practices, product management, and processing technologies; immunizations; and the dc\elopnicnt of improved detection methods. A'£>' WORDS: enteric virus, hepatitis. Norvvalk. rotavirus, astrovirus. illness, shellfish, intervention INTRODUCTION Enteric viruses encompass a broad spectrum of pathogens transmitted to food and water by the fecal-to-oral route. They include hepatitis A and E viruses. Norwalk and related calicivi- ruses, rotavirus, astrovirus. enteric adenovirus, and coronavirus. Molluscan shellfish are filter feeders capable of bioconcentrating viruses that are present in seawater within their edible tissues. Data on the incidence of enteric virus illness in the United States were compiled by the Centers for Disease Control and Prevention and indicate that 23 million people in the United States contract Nor- walk-like viral illnesses each year, 40% of which are food-borne (Mead et al. 1999). In addition, 3.9 million people develop rotavi- rus and astrovirus infections, and 83,000 people develop hepatitis A (Mead et al. 1999). Only 1% of the astrovirus and rotavirus infections and 39f of the hepatitis A cases are estimated to be from foods (Mead et al. 1999). The number of illnesses caused by the ingestion of molluscan shellfish is uncertain, but shellfish are likely to be a major contributing factor for Norwalk virus illness and perhaps for hepatitis A as well. Outbreaks of hepatitis A and Norwalk-like illnesses have been linked to shellfish consumption (Richards 1988, Halliday et al. 1991). A major outbreak that was associated with the consumption of clams obtained from a mud flat outside of Shanghai, China, occurred in 1988, causing over 292,000 cases of hepatitis A over a three-month period (Halliday et al. 1991). Norwalk virus caused close to K.SOO oyster- and clam-associated illnesses in New York and New Jersey in 1982 (Richards 1985). To date, only sporadic cases of hepatitis E have been identified; however, hepatitis E is a major cause of morbidity and mortality in Asia and may eventually spread throughout the world. In the gen- eral population, hepatitis E has a mortality rate of 0.5-3% of the individuals infected; however, the death rate is substantially higher (15-25%) in pregnant women (Mast & Krawczynski 1996). CHARACTERISTICS OF ENTERIC VIRUSES Huinan enteric viruses are only known to replicate in human and some primate cells; there is no replication in shellfish. There- fore, there is no virus amplification within foods and temperature abuse does not increase virus numbers. Within the environment. viruses may persist for weeks or months and once in shellfish may last for several weeks (Gerba & Goyal 1978). They may be pre- served by cold temperatures and may be infectious in very low numbers. As few as 10-100 virus particles may be sufficient to elicit an infection (Cliver 1997). SOURCES OF CONTAMINATION There are numerous routes by which shellfish may become contaminated. In many cases, shellfish are impacted before harvest by contamination of the shellfish growing waters by runoff, boat waste discharge, or ineffective sewage treatment plants or septic tanks. At-harvest contamination has also been observed. Two out- breaks of Norwalk viius illness were associated with shellfish contaminated by ill harvesters who voinited overboard during har- vesting operations (Kohn et al. 1995. McDonnell et al. 1997). The lack of common sense and training in good sanitation and hygienic practices are major stumbling blocks to enhancing shellfish safety. Postharvest contamination of shellfish and other foods may occur from the unsanitized hands of infected food handlers, contami- nated processing surfaces, cross-contamination of the cooked w ith raw product, or by postharvest adulteration. Shellfish may be adul- terated postharvest through the use of unsanitary water, employed either as a food additive or for rinsing product or processing equip- ment; ice or food additives containing viral contamination; pack- aging materials; or unsanitary transport or storage conditions. Groups at risk of criminal or civil penalties in the event of a food-borne outbreak include shellfish growers, harvesters, and processors; transporters; retail and wholesale outlets; restaurants; and government regulatory agencies. One prominent outbreak of hepatitis A occurred from the consumption of contaminated oys- ters from Wallis Lake in New South Wales. Australia (Conaty et al. 2000) and led to a class action suit. The courts found the industry and regulatory agencies liable for millions of dollars in damages because they did not exercise adequate measures to pre- vent outbreaks of hepatitis A. STRATEGIES TO REDUCE ENTERIC VIRUSES IN SHELLFISH Strategies to reduce the incidence of shellfish-related illness must incorporate preharvest. at-harvest, and postharvest interven- tions (reviewed by Richards, 2001). Some piactical approaches directed toward the reduction of preharvest contamination are: ( I ) improved operation and tnonitoring of sewage treatment plants that impact coastal areas, (2) elimination of improperly functioning septic tanks from coastal areas, (3) better enforcement of shellfish 1241 1242 Richards harvesting and tagging regulations, (4) enhanced enforcement of ocean-dumping restrictions, and (5) improved shellfish monitor- ing. Shellfish tagging should be monitored to ensure that shellfish are only obtained from areas approved for harvesting. At-harvest contamination may be reduced by preventing sick shellfishermen from harvesting or handling shellstock. In addition, the discharge of boat wastes, particularly feces, vomitus, and urine, should be strictly forbidden. Postharvest strategies to reduce illness should involve various types of processing to reduce or eliminate potential viral contaminants. Processing may include: ( 1 ) thorough washing of shellstock; (2) heating/cooking of shellfish, either in-shell or shucked; (3 ) irradiation or dehydration of product; and/or (4) depu- ration and relaying. Thorough cooking is effective in eliminating enteric viruses, but the shellfish may lose many of their desirable characteristics if they are fully cooked. They may become tough, dry, and lose their flavor. In addition, high protein and lipid levels, as are found in shellfish tissues, tend to protect the viruses from thermal inactivation (Filippi & Banwart 1974. Bidawid el al. 2000). Irradiation can effectively eliminate enteric viruses from shell- fish, loiii/ing radiation, microwave energy, and ultraviolet irradia- tion are effective in reducing virus levels. Ionizing radiation was effective in substantially reducing enteric viruses in shellfish, but required a dose that caused rapid shellfish mortalities (Mallett et al. 1991 ). The use of ionizing radiation may be limited to the disin- fection of products destined for shucking. Microwave energy is effective in eliminating shellfish pathogens by virtue of cooking. Uneven heating and the problems associated with marketing a fully cooked shellfish product reduce the feasibility of microwave processing for shellfish. Ultraviolet light can inactivate only sur- face contaminants, but may be applied in a depuration setting to inactivate pathogens purged from the shellfish during the depura- tion process. Depuration is the commercial process where shellfish are placed in tanks of clean seawater and allowed to purge con- taminants for a short period of time. Ultraviolet light is often used to disinfect the depuration water as it is recirculated through the tank. Over a two- to three-day processing cycle, molluscan shell- fish will expel many of the bacteria and viruses, and much of the sediment from their digestive diverticula. Unfortunately, shellfish are unable to purge all the enteric viruses. Some viruses migrate from the digestive diverticula into the epithelial cells surrounding the lumen of the gut and become sequestered within the shellfish tissues (Hay & Scotti 1986. Richards 1990). Since only a few viruses may be sufficient to cause infection, these entrapped vi- ruses limit the effectiveness of depuration in producing a safe product. Relaying is where shellfish are removed from contaminated shellfish growing areas and placed in clean shellfish growing wa- ters for an extended period (weeks to months) (Richards 19S8). Through natural virus inactivation processes, relaying offers some advantages in eliminating viruses that might be sequestered within the shellfish tissues. However, relaying is only effective if the waters remain clean during the entire relay process. This is often difficult to establish since heavy rains, malfunctioning treatment plants, or illegal boat waste discharge could occur at any time during the extended relay period. In addition, shellfish harvesters, pressured to meet customer demands, may be tempted to prema- turely harvest shellfish. Intervention strategies may alst) include immunizations of food processors and handlers, the general population, children, or just high-risk individuals. There is no consensus on who should be immunized, and the costs and logistics of immunization programs can be daunting. Today, there are vaccines for hepatitis A virus, but the vaccine for rotavirus illness was taken off the market to await further testing. There is no vaccine currently for Norwalk- like viruses; however, vaccines are emerging for hepatitis E virus, which may become a serious public health concern in many coun- tries over the next decade. Immunization programs for food han- dlers would be extremely difficult to monitor since much of the workforce consists of teenagers who quite frequently move from job to job. The shellfish industry is increasingly subscribing to the provi- sions of the hazard analysis critical control point (HACCP) pro- gram. The key to HACCP is the delineation of the critical control points in the passage of product from the ocean to the consumer. Under HACCP, the critical control points are monitored to ensure the points are maintained within acceptable limits. Monitoring is practical for some aspects of shellfish sanitation and processing; however, methods to monitor viral contaminants are limited. Con- sequently, HACCP falls short of protecting the consumer from enteric viruses. Virus extraction and analytical methods have been developed for a variety of enteric viruses in shellfish. Over the past decade, most of the procedures developed have been based on molecular biological methods, principally reverse transcription-polymerase chain reaction (RT-PCR). Unfortunately, molecular biologically based tests do not differentiate between infectious and inactivated viruses and are only marginally effective in the evaluation of shell- fish (Richards 1999). More conventional cell culture-based assays are capable of detecting infectious viruses, but only if the viruses are capable of being propagated in cell or tissue culture. To date, no one has successfully propagated Norwalk-like viruses in vitro. and most of the wild-type hepatitis A viruses are incapable of being propagated in vitro. Consequently, further research is needed to identify better methods to monitor for the presence of infectious viruses in shellfish. Overall, there are several measures that can be employed to increase the virological safety of shellfish. The regulation and monitoring of harvesting areas by State and Federal agencies are crucial in providing reasonably clean shellfish in the marketplace. Shellfish must be obtained only from shellfish beds approved for harvesting based on sound sanitary criteria. Proper tagging of raw products will facilitate tracing back to the original source in the event of an outbreak, thus allowing contaminated products to be taken off the market. Efforts to educate shellfish harvesters, pro- cessors, and handlers should be enhanced to stress the importance of following the regulations. Thorough cooking is the most effec- tive and simplest method to inactivate enteric viruses in shellfish. If a raw product is desirable, extended relay of moderately or lightly contaminated shellfish into clean water is the most practical procedure to decontaminate large numbers of shellfish. Remedia- tion of shellfish growing waters may be accomplished by continu- ously monitoring and improving sewage treatment plants and sep- tic systems and by strict enforcement of laws against boat waste discharge. New molecular biological- and cell culture- based as- says for the detection of enteric viruses are under development. Future efforts to avert outbreaks of enteric virus illness will likely rely on a combination of enhanced enforcement of existing laws, improved processing and handling techniques, and direct virus testing. Enteric Virus Interventions 1243 Bidawid. S.. J. M. Farber. S. A. Sattar & S. Hayward. 20()U. Heat inacti- valion of hepatitis A virus in dairy foods. J. Food Prot. 63:522-528. Cliver. D. O. 1997. Virus transmission via food. FnocI Techiiol. 5I:7I-7S. Conaty. S.. P. Bird, G. Bell. E. Kraa, G. Grohmann & J. M. McAnulty. 2000. Hepatitis A in New South Wales. Australia from consumption of oysters: The first reported outbreak. Epideiniol. Infect. 124:121-130. Filippi. J. A. & G. J. Banwart. 1974. Effect of the fat content of ground beef on the heat inactivation of poliovirus. J. Food Sci. 39:865-868. Gerba. C. P. & S. M. Goyal. 1978. Detection and occurrence of enteric viruses in shellfish: A review. J. Food Prot. 41:743-754. Halliday, M. L.. L. Y. Kang. T. K. Zhou. M. D. Hu, Q. C. Pan. T. \. Fii. Y. S. Huang & S. L. Hu. 1991. An epidemic of hepatitis A attributable to the ingestion of raw clams in Shanghai. Cliina. ./, Infect. Dis. 164: 852-859. Hay, B. & P. Scotti. 1986. Evidence for intracellular adsorption of virus by the Pacific oyster, Cyussostiea i^igti.s. N.Z. J. Mtiiaic Frc^hnnier Res. 20:655-659. Kohn M. A., T. A. Farley. T. Ando, M. Curtis, S. A. Wilson. Q. Jin, S. S. Monroe, C. Baron, L. M. McFarland & R. I. Glass. 1995. An outbreak of Norwalk virus gastroenteritis associated with eating raw oysters. Implications for maintaining safe oyster beds, JAMA 273:466-471. Mallctt. J. C. L. E. Beghian, T. G. Metcalf & J. D. Kaylor. 1991, Potential of irradiation technology f"i improving shellfish sanitation. / Food Scifet}' 11:231-245. LITERATURE CITED Mast. E. E. & K. Krawc/ynskl. 1996. Hepatitis E: An overview. Ai»!. Rev. Med. 47:257-266. McDonnell, S., K. B. Kirkland. W. G. Hlady. C. Aristeguieta. R. S. Hop- kins, S. S. Monroe & R. 1. Glass. 1997. Failure of cooking to prevent shellfi.sh-associated viral gastroenteritis. Ardi. lutein. Med. 157:111- 116. Mead P. S., L. Slutsker. V. Diet/. L. F. McCaig. J. S. Bresee, C. Shapiro, P. M. Griffin & R. V. Tauxe. 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5:607-625. Richards. G. P. 1985. Outbreaks of shellfish-associated enteric virus illness in the United States: Requisite for development of viral guidelines. J, Food Prot. 48:815-823. Richards, G. P. 1988. Microbial purification of shellfish: A review of depuration and relaying. / Food Prot. 51:21 8-25 1 . Richards. G. P. 1990. Shellfish depuration. In: D. R. Ward & C. R. Hack- ney, editors. Microbiology of marine food products. New York: Van Nostrand Reinhold. pp. 395^28. Richards. G. P. 1999. Limitations of molecular biological techniques for assessing the virological safety of foods. J. Food Prot. 62:691-697. Richards. G. P. 2001. Enteric virus contamination of foods through indus- trial practices: A primer on intervention strategies. J. Indnst. Microbiol. Bioleclmol. 27:117-125. Joiiniul I'fSliclllisli Rcscairli. Vol. 20, No. }. 1245-1249, 2(101, DETECTION AND CHARACTERIZATION OF TOXIGENIC BACTERIA ASSOCIATED WITH ALEXANDRWM CATENELLA AND AVLACOMYA ATER CONTAMINATED WITH PSP MONICA VASQUEZ' * CAROL GRUTTNER,' SUSAN GALLAGHER,- AND EDWARD R. B. MOORE' ' Lcihoratorio dc Blolnin'iiieiia. INTA-Universidad de Chile, Av. Maciil 5540. Santiago. Chile; -Marine Laboratory. P.O. Box 101. Victoria Road. Aberdeen ABU 9DB. Scotland. United Kingdom: ^Division of Microbiology. German Research Centre for Biotechnology, Mascheroder Weg 1. 38124 Braunschweig. Germany ABSTRACT The panilytic shellfish loxins (PST) are potent neurotoxins. Among them, saxitoxin (STX) is one of the most dangerous for humans and animals, functioning hy hlocking sodium channels in mammalian nerve cells. Marine algae, such as the dinotlagellate AlexandrUim catenella. are generally believed to be responsible for the production of STX. and if PSTs are produced by the plankton, they can be concentrated hy filter-feeding organisms, effectively mcreasing the risk of poisoning for consumers of shellfish. However, it is not clear whether A. catenella itself, or the bacteria associated with it. produce the toxins and, furthermore, whether toxigenic bacteria are associated, as well, with shellfish contaminated with PST. Bacteria able to produce sodium channel blocking (SCB) toxins were isolated from a toxin-producing A. catenella culture and from shellfish (.Aiilacomya ater) contaminated with PST from the Magallanes region of southern Chile. Quantitative levels of PST and bacterial SCB toxins were ascertained by mouse assay and mouse neuroblastoma assay, respectively. More than 60% of the bacteria enriched and analyzed from Aulacomya ater and A. catenella were observed to produce detectable amounts of SCB. The SCB toxin-producing bacterial isolates were characterized by amplified I6S ribosomal DNA restriction analysis ( ARDRA), and selected isolates, representing the different ARDRA groups, were identified by I6S rDNA sequence comparisons and estimations of their phylogenetic relationships with validly published species. Most of the SCB toxin-producing isolates from A. ater and from A. catenella represent new species of genera within the Proleabaeteria and, in some cases, demonstrate close phylogenetic relationships with previously reported toxigenic bacteria isolated from other dinoflagellate cultures. However, the isolates from A. ater represent different taxonomic lineages from those associated with the dinoflagellate A. catenella. These observations suggest that SCB toxin contamination of shellfish may occur independently of algal contamination. KEY WORDS: toxigenic bacteria, Alexamhium catenella. dinoflagellate, shellfish, PSP, saxitoxin INTRODUCTION Paralytic shellfish toxins (PST). comprising at least 20 chemi- cally similar derivatives (Oshima et al. 1993) including saxitoxin (STX), are potent neurotoxins that cause paralytic shellfish poi- soning (PSP) in humans, PSTs act by blocking sodium channels in mammalian cells, thus preventing conductance of signals along neurons and effectively paralyzing muscle activity (Baden & Trainer 1993). Much evidence has suggested thai the PSTs origi- nate from marine dinoflagellates (e.g., of the genera Alexandriiim and Gymnodiniiin) (Anderson et al. 1990. Cembella 1998). If tox- ins are present in the plankton, they are concentrated by filter- feeding shellfish and subsequently conveyed to humans via their consumption. Along the southern Chilean coast, between 1972 and 1998, 348 cases of PSP, including 23 deaths, were reported (Suarez-Isia & Guzman 1998). The incidences of PSP. in all cases, were associ- ated with the consumption of PST-contaminated mussels. Moni- toring also detected the dinoflagellate Ale.xaiidriiim catenella in the PST-contaminated regions, and this species is believed to be re- sponsible for the toxicity in the shellfish (personal communication. Dr. L. Guzman. Instituto de Fomento Pescjuero, IFOP Magallanes. Chile). Sodium channel-blocking (SCB) toxins (i.e., STX) has been detected also in cultures of several species of cyanobacteria: .Apliaiiizonienon flos-aquae (Mahmood & Carmichael 1986), Ana- haena circinaUs (Humpage et al. 1994), Lynghya wollei (Car- michael et al. 1997). and CyUiulrnspenniipsis racHwr.skii (Lagos et al. 1999). Further, Negri and Jones (1995) reported high levels of PST in a freshwater mussel when fed with the neurotoxic cyano- *Corresponding author. E-mail: mvasquez@uec.inta.uchile.cl bacterium Anabaena circitialis. Recently, reports have described heterotrophic bacteria as well, isolated from cultures of toxic di- noflagellates, that are capable of producing SCB toxins (Kodama et al. 1988, Doucetle & Trick 1995. Franca et al. 1996. Gallacher et al. 1997). These bacteria have been associated with different dinoflagellate species from different geographic regions and rep- resent different taxonomic groups. However, few toxigenic bacte- ria isolated from shellfish have been described to date (Freitas et al. 1992). To define further the diversity and potential of toxigenic bac- teria associated with PST contamination of shellfish along the southern Chilean coast, isolated bacteria were enriched from mus- sels iAiilacomya ater). collected along the Magallanes coastline of southern Chile, and from a culture of a toxic dinoflagellate {A. catenella). The cultivable bacteria from the shellfish and from the dinotlagellate culture were analyzed for SCB toxin production, and the bacterial isolates determined to be toxigenic were characterized systematically. METHODS Isolaliiin and Cultivation oj Bacteria from PST-Conlaniinaled Mussels (Aulacomya ater) and from a Toxic Dinoflagellale Culture (Alexandrium catenella) Mussels (Aulacuniya ater) were collected from the Magallanes region of the southern Chilean coast, near Seno Europa (50°03'S, 74°2rW). After collection, the mussels were maintained at 4"C until analysis. The toxicity of the shellfish, as determined by the AOAC mouse bioassay (Williams 1984), was 560 ± 200 (xg PST/ 100 g of mussel flesh (personal communication. Dr. L. Guzman, Instituto de Fomento Pesquero, IFOP Magallanes, Chile). Mussel 1245 1246 Vasquez et al. tissue was homogenized in NaCi ( 1%, w/v). using a Tissue Tenor. Dilutions of the homogenate were then inoculated into marine broth (Difco Laboratories. Detroit, MI, USA), and 0.1 inL of each dilution was plated onto marine agar medium, prepared with 28'7f sea water. These cultures were incubated at 17"C for at least one week. One hundred colonies with different morphologies were picked and plated individually on marine agar plates, followed by incubation at 17°C for 72 h. A SCB toxin-producing culture of Alexandrium catenella. de- rived from clone ACC07 from the Aysen region (45°32'S, 73°34'W) in southern Chile, was obtained from Dr. Myriam Seg- uel (Instituto de Fomento Pesquero. Puerto Montt. Chile). Dilu- tions of the A. catenella culture were prepared, and 0. 1 niL were then plated on marine agar medium. These cultures were incubated at I7''C for at least one week. Twenty colonies with observable morphological differences were picked, plated, and incubated as previously described. Analysis of Toxin Production by Bacterial Isolates Each bacterial colony was inoculated into .30 niL of marine broth. After incubation at 23-25°C for 36 h on a rotary incubator (100 oscillations per minute), the culture was centrifuged (10.000 X g. 20 min), and the supernatants were collected and stored at -20°C until analyzed. SCB activity in bacterial culture supernatants was measured using a cell-culture assay with mouse neuroblastoma (MNB) (Gallacher & Birkbeck 1992). An STX dose-response curve, using a saxitoxin standard (NRC, Halifax, Canada), was prepared for the range of 0 to 500 nM. All isolate supernatants were tested in triplicate, and controls were done in quadruplicate repeats, as de- scribed by Gallacher et al. ( 1997). The supernatant samples used in the assays were prepared at 1/10 dilutions for all screenings. SCB toxin levels were compared to STX dose-responses to estimate SCB activities with respect to nM STX-equivalents. Phcnotypic and (knotypic Analyses of Toxigenic Bacterial Isolates Bacterial isolates were characterized by their colony morphol- ogy, cultivated on marine agar medium. Staining and microscopic characterization were carried out using standard microbiological methods (Murray & Robinow 1981). API 20 NE strips (BioMerieux Viteck, Inc., Hazelvvood. MO. USA) were used as described in API 20 NE protocols for determination of a standard set of phcnotypic characteristics. The genotypic diversity of SCB toxin-producing bacteria isolated from the mussel (.4. ater) tissue samples and from the dinoflagellate (A. catenella) culture was investigated using amplified I6S ribosomal DNA restriction analy- sis (ARDRA) (Vaneechoutter et al. 1992). Bacterial DNA was purified from the isolates as described previously (Espejo et al. 1998). I6S rRNA genes were targeted and amplified by polymer- ase chain reaction (PCR) using primers and the reaction conditions described by Moore et al. ( 1996). Subsequently, the amplified I6S rDNA was restricted overnight with 2.5 U of HaeWl (Gibco BRL, Eggenstein. Germany) or Msp{ (Gibco BRL) at 37°C or with 5 U 7"(/l(>geiiftic/ta\oni>niic relaliunships amung toxic marine shellfish-associated bacterial isolates. Bacterial Isolates Closest Phylogenetic Relative* 16S rRNA Gene Sequence Similarity ( % )** Phylojjenetic Taxon CH03 CH04 CH08 CH22 CH25 CH32 CH34 CH47 CH61 CH64 CH7I CH72 CH88 Psychrobacter Psychrobacler Psychrobacter Psychrobacler Psychrobacter Psychrobacter Psychrobacter Psychrobacter Psychrobacter Psychrobacter Psychrobacter Psychrobacter Psychrobacter i^lanciiicolti ghTiuiiicola glancincola pacificensis pacificensis glancincola glancincola glancincola glancincola glancincola glancincola glancincola glancincola 97.5 95.7 97.9 96.2 93.5 98.1 98.1 95.7 96.7 98.1 96.4 98.1 99.6 ProteobjLtena y Proteohacleria y Proteobacteria y Proteobacteria 7 Proteobacteria 7 Proteobacteria y Proteobacteria y Proteobacteria y Proteobacteria 7 Proteobacteria 7 Proteobacteria 7 Proteobacteria 7 Proteobacteria 7 * Validly published species. ** FASTA analysis (Pearson & Lipnian 1988). 26 gL^' and water temperatures of 6-7°C (depths 0-10 rii). It is probably not surprising, at these latitudes, to find psychrophilic or psychrotolerant bacteria associated with the shellfish. Further- more, the low water temperature may be one reason to have found a lower diversity of cultivable bacteria associated with mussels than what one might expect in more temperate or tropical en\'i- ronments. This is the first case of SCB toxin production observed in species of Psxchrohactcr and Siilfitolxicter genera, w hich suggests that toxin-production in bacteria are not limited to species oi Mo- raxella. Alteroininuts. and Psciicl/nnoims. as has been suggested from the results of earlier studies (Doucette & Trick 1995. Kodama et al. 1988). The high percentage of SCB toxin-pioducing bacteria isolated from the PST-contaminated mussels and from the di- noflagellate culture in this study, as well as results in other previ- ous reports (Gallacher et al. 1997). suggest that this property is not uncommon among marine bacteria. However, SCB toxin is not necessarily related only with saxitoxin. and further analyses (e.g., by other methods such as HPLC and mass spectrometry ) need to be included to reliably identify the toxin patterns. Interestingly, in this study, the range of bacterial diversity ob- tained from the toxin-contaminated mussels and from the toxic dinoflagellate culture were relatively limited, although past studies using enrichment cultivation methods (Kodama et al. 1988. Dou- cette & Trick 1995) also recovered only a limited range of bacte- rial species associated w ith marine algae. Questions about whether specific species or communities of bacteria are associated with certain algal species or with algal species in defined environmental situations have never been addressed to any degree. Furthermore, the observed high percentage of toxigenic isolates among the cul- tivable bacteria from the mussels and the dinoflagellate culture raises interesting questions with respect to the influence the asso- ciated bacteria may have on production of toxin by the algae and accumulation of toxin by the shellfish. However, the toxin- producing bacteria obtained from the PST-contaminated mussels comprised species from different phylogenetic lineages than those isolated from the toxin-producing dinotlagellate. These observa- tions suggest that the accumulation of PSTs in shellfish may not be dependent only upon the occurrence of toxin-producing marine algae. While it is clear that the shellfish are filtering phytoplank- ton. including SCB toxin-producing dinotlagellates such as A. catenella. the shellfish may have also been obtaining bacteria, the majority of which are found on particulate matter in the marine environment. In the case of our study, the shellfish were collected at a time when monitoring was not detecting significant numbers of A. catenella in the water column. However, the shellfish con- tained high levels of PST. Thus, although the Magallanes region was infected with .4. catenella blooms at earlier monitoring times, there was no direct influence of marine algae on the toxin le\ els of the shellfish at the time of our study. It should also be noted that there is usually a lag between the marine algal bloom (in water column) and the detection of PSTs in the shellfish. Results also suggest that toxin-producing bacteria may play a much more im- portant role in PST-contamination of shellfish than has been ac- cepted in the past. Obviously, much more extensive sampling and analysis must be carried out for such a working hypothesis to be tested with any reliability. ACKNOWLEDGMENTS We acknowledge Dr. Leonardo Guzman, of the Institute de Fomento Pesquero (IFOP). Region Magallanes. Chile, for provid- ing the means for sampling and for the analytical information on the toxin levels in mussels: Paulina Uribe, of the Institute de Ciencias Biomedicas (ICBM), Universidad de Chile, Santiago, Chile, for providing the enrichment from the culture of Alexan- driiiin catenella from which the bacterial strains were isolated; Fiona MacKintosh. of the Marine Laboratory, Aberdeen, United Kingdom, for the technical expertise in the analysis of toxin levels in bacterial isolates: and Annette Kriiger and Blanca Moller. of the Gesellschaft fiir Biotechnologische Forschung (GBF). Braunsch- weig. Germany, for the technical assistance in the nucleic acid fingerprinting and sequencing analyses. The project was supported in part by a grant ( 1990763) from the Chilean Fondo Nacional de Ciencia y Tecnologia (FONDECYT). an exchange grant (CHL 98/009) from the German Internationales BUro des Bundesminis- terium fiir Biotchnologie und Forschung (BMBF) and from the Chilean Coniision Nacional de Investigacion en Ciencia y Tecno- logia (CONICYT). and a grant (A/3057-1 ) from the International Foundation for Science (IFS). Toxigenic Bacteria in Dinoklagellate and Shellhish 1244 LITERATURE CITED Anderson. D. M.. D. M. Kullis. J. J. Sullivan. S. Hall .*^ C, Lee. 1 99(1. Dynamics and physiology of saxitoxin production b\ Ihe dmoflagel- \axei Alexandriiiin spp. Mar. Biol. 104:511-524. Baden. D. G. & V. L. Trainer. 1993. Mode of action of toxins of seafood poisoning. In: I. R. Falconer, editor. Algal toxins in seafood and drink- ing water. London: Academic Press, pp. 49-74. Baker, W.. A. van den Broek. E. Canion. P. Hingamp. P. Sterk. G, Stoesser & M. A. Tuli. 2000. The EMBL Nucleotide Sequence Database. Niui. Acids Re.i. 28:19-23. Bowman. J. P.. J. Cavanagh. J. J. Austin & K. Sanderson. 1996. Novel Psychrobacter species from Antarctic ormthogenic soils. //;/. J. Syst. Bacterioi 46:841-848. Bowman. J. P.. S. A. McCamnion. M. V. Brown. D. S. Nichols & T. A. McMeekin. 1997. Diversity and association of psychrophilic bacteria in Antarctic sea ice. Appl. Environ. Microbiol. 63:3068-3078. Carmichael. W. W.. W. R. Evans. Q. Q. Yin. P. Bell &. E. Moc/.ydlowski. 1997. Evidence for paralytic shellfish poisons in the freshwater cyano- bacteriuni Lxngybya wollei (Farlow ex Gomont) combnos. Appl. En- viron. Microbiol. 63:3104-3110. Cembella. A. D. 1998. Ecophysiology and metabolism of paralytic shell- fish toxins in marine microalgae. In: D. M. Anderson. A. D. Cembella. & G. M. Hallegraeff. editors. Physiological ecology of harmful algal blooms. NATO ASI Series Vol. G41. Berlin. Heidelberg: Springer- Verlag. pp. 381^03. Doucette. G. & C. G. Trick. 1995. Charactensation of bactena associated with different isolates of Allexcmclriiim tamarense. In: P. Lassus. G. Arzul. E. Erad-le Denn. P. Gentien. & C. Marcaillou-le-Baut, editors. Harmful marine algal blooms. Paris: Lavoisier, pp 33-38. Espejo. R. T., C. G. Feijoo. J. Romero & M. Vasquez. 1998. PAGE analysis of the heteroduplexes formed between PCR-ampIified 16S rRNA genes: Estimation of sequence similarity and rDNA complexity. Microbiology 144: 161 1-1617. Franca, S., L. Pinto. P. Alvito. I. Sousa. V. Vasconcelos & G. J. Doucette. 1996. Studies on prokaryotes associated with PSP producing di- notlagellates. In: T. Yasumoto. Y. Oshima & Y. Fukuyo. editors. Harmful and toxic algal blooms. Intergovernmental Oceanographic Commission of UNESCO. Paris. Intergovernmental Oceanographic Commission of UNESCO, pp. 347-350. Freitas, J. C. M. Ogata. M. Kodama. S. C. G. Martinez, M. F. Lima & C. K. Monteiro. 1992. Possible microbial source of guanidine neurotoxins found in the mussel Pema perna (Mollusca. Bivalvia. Mytilidia). In: P. Gopalakrishnakone, editor. Recent advances in toxicology research. Proceedings of 10th World Conference. Singapore, pp. 589-596. Gallacher, S. & T. H. Birkbeck. 1992. A tissue culture assay for direct detection of sodium channel blocking toxins in bacterial culture super- nates. FEMS Microbiol. Utr. 92:101-108. Gallacher. S.. K. J. Flynn. J. M. Franco, E. E. Brueggemann & H, B. Hines. 1997. Evidence for production of paralytic shellfish toxins by bacteria associated with Alexundnum \pp. iDinophyta) in culture. .\ppl. Envi- ron. Microbiol. 63:239-245. Gonzalez, J. M.. R. P. Kiene & M. A. Moran. 1999. Transformation of sulfur compounds by an abundant lineage of marine bacteria in the a-subclass of the class Proleobacteria. Appl. Environ. Microbiol. 65: 3810-3819. Gonzalez, J. M. & M. A. Moran. 1997. Numerical dominance of a group of marine bacteria in the a-subclass of the class Proleobaaenu. .Appl. Environ. Microbiol. 63:4237—1242. Humpage. A. R.. J. Rositano. A. H. Bretag. R. Brown. P. D. Baker. B. C. Nicholson & D. A. Steffensen. 1994. Paralytic shellfish poisons from Australian blue-green algal (cyanobacterial) blooms. A».sf. / Mar. Freshwater Res. 45:761-771. Kodama. M., T. Ogata & S. Sakamoto. 1988. Bacterial production of saxitoxin. Agric. Biol. Chem. 52:1073-1077. Lafay. B.. R. Ruimy, C. R. de Traunbernberg. V. Breittmayer, M. J. Gauthier & R. Christen. 1995. Roseohucler ulgicola sp. nov., a new marine bacterium isolated from the phycosphere of the toxin-producing dinofiagellate Prorocenlrum lima. Int. J. Syst. Bacterioi. 45:290-296. Lagos N.. H. Onodera. P. A. Zagatto. D. Andrinolo, S. M. Azevedo & Y. Oshima. 1999. The first evidence of paralytic shellfish toxins in the fresh water, cyanobacterium Cylindrospermopsis raciborskii. isolated from Brazil. Toxicon. 37:1359-73. Lane. D. J. 1991. 16S/23S sequencing. In: E. Stackebrandt & M. Good- fellow, editors. Nucleic acid techniques in bacterial systematics. Chich- ester: John Wiley, pp. 115-175. Mahmood. N. A. & W. W. Carmichael. 1986. Paralytic shellfish poisons produced by the freshwater cyanobacterium .Aphanizomenon flos- aquae NH-5. To.xicon. 24:175-186. Moore, E. R. B.. M. Mau, A. Arnscheidt, E. C. Bottger. R. A. Hutson. M. D. Collins. Y. van der Peer, R. De Wachter & K. N. Timmis. 1996. The determination and comparison of the 16S rRNA gene sequences of species of the genus Pseiidomonas (sensu stricto) and estimation of the natural intrageneric relationships. Syst. Appl. Microbiol. 19:478—492. Murray, R. G. E. & C. F. Robinow. 1981. Specimen preparation for light microscopy. In: P. Gerhardt, R. G. E. Murray, R. N. Costilow. E. W. Nester, W. A. Wood, N. R. Krieg & G. B. Phillips, editors. Manual of methods for general bacteriology. Washington, D.C.: A.S.M. Press, pp. 1 7-20. Negri. A. P. & G. J. Jones. 1995. Bioaccumulation of paralytic shellfish poisoning (PSP) toxins from the cyanobacterium .Anabaena circinalis by the freshwater mussel .Mathyria condola. Toxicon. 33:5667-5678. Oshima. Y., S. I. Blackburn & G. M. Hallegraeff 1993. Comparative study on paralytic shellfish toxin profiles of the dinofiagellate Gymnodiniiim catenatum from three different countries. Mar. Biol. 116:471—176. Pearson. W. R. & D. J. Lipman. 1988. Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85:82444-82448. Prokic. 1.. F. Brummer. T. Brigge. H. D. Gortz, G. Gerdts. C. Schtilt, M. Elbrachter & W. E. G. Miiller. 1998. Bacteria of the genus Roseobacter associated with toxic dinotlagellate Prorocentrum lima. Protist. 149: .^47-357. Suaiez-lsla. B. A. & L. Guzman. 1998, Floraciones de Algas Nocivas. Santiago. Chile: Editorial Universitaria. Vaneechoutter, M., R. Rossau, P. De Vos. M. Gillis. D. Janssens. N. Paepe. A. De Rouck. T. Fiers. G. Claeys & K. Kersters. 1992. Rapid identi- fication of bacteria of the Comainonadaccae with amplified ribosomal DNA-restriction analysis (ARDRA). FEMS Microbiol. Lett. 72:227- 233. Williams. S. 1984. Paralytic shellfish poison. Official methods of analysis. 14th ed. Ariington. VA: AOAC. .Iiiiirihil ,'/ .S7«///n7j Rcscanh. Vdl, 20. No. .?. 125I-I2.'S.S, 20(11. PHYTOPLANKTON COMPOSITION AND PYRODINWM BAHAMENSE TOXIC BLOOMS IN MANILA BAY, PHILIPPINES RHOUORA V. AZANZA* AND LILIBETH N. MIRANDA Murine Science Institute. College of Science. University of ilie Philippines. Dilinhin, Quezon City 1 101. Philippines ABSTRACT PyroJiniwn haliami'n.si' var. chiui'ivsmiiii has caused appro.ximately 1,992 paralytic shellfish poisoning (PSP) cases and 1 16 deaths in the Philippines and has remained the mam cause of PSP in the tropical world. The first toxic bloom of the organism in Manila Bay. Phdippines was in 1988. Since then the recurrence of the event (usually beginning in April/May and ending in September/October) has been experienced every year except in 1997 and 1999. The latest incident of the phenomenon was in 1998 followed by the absence of the species in 1999 during the months it was present in previous years. Eight stations in Bataan and Cavite, two red tide prone areas in Manila Bay. were sampled during three seasons from 1997 to 1999 for phytoplankton abundance and species composition. Bataan and Cavite, though situated within the same bay, have distinct differences during the three seasons with respect to dinotlagellate population. There is a pronounced increase in dinoflagellate population in Bataan from northeast monsoon to tradewinds to southwest monsoon. In Cavite, dinotlagellates constitute a minor portion of the phytoplankton community during the whole year. In 1998, Pyrodinium blooms in both areas started in April/May and terminated in October. A bloom of the heterotrophic dinoflagellate Noctihica scintillaiu in Bataan and the dominance of the diatom Chaewceros spp. in Cavite succeeded the PyrocHnium bloom. N. scintillans was observed to be present in the two sites during the entire duration of the study. A'£l' WORDS: Pynulinium Inihaimnsc. paralytic shellfish poisoning, toxic blooms. Manila Bay INTRODUCTION Toxic red tides are recurrent events in the Philippines and have been caused so far only by one dinoflagellate species. Pyrodinium huhiimense var. compres.sum. Manila Bay is one of the 16 bays affected by this Paralytic Shellfish Poisoning causative organism. Analysis of the phytoplankton population is an important tool in the proper management of toxic red tides in the country. This helps in the identification of Pyrodinium and other potentially harmful algal species that dominate the phytoplankton population during different seasons and the conditions favoring their bloom. Bajarias (1999) has provided an inventory of the dinoflagellates present in 1 1 marine coastal areas in the Philippines including Manila Bay. There has been no published work on phytoplankton distribu- tion in Manila Bay. This paper for the first time provides the relative abundance of dinoflagellates and diatom species in two toxic red tide affected areas in the bay for the three different sea.sons for the period 1997 to 1999. MATERIALS AND METHODS Manila Bay is a semi-enclosed bay located at the western coast of Luzon Island, Philippines (lat. 14°53'N: long. 120°76'E) and bounded by the provinces of Cavite in the south, Metro Manila and Ri/al in the east, Bulacan and Pampanga in the north, and Bataan in the west and northwest. The bay was first affected in 19S8 and annual occunences of toxic red tides have been recorded (Azanza 1997). Bataan and Cavite are two areas within Manila Buy that are commonly utilized for shellfish farming and harvest from the wild. Both these areas have a history of red tide occurrences. Phy- toplankton monitoring in eight stations in Bataan and Cavite (Fig. 1) was conducted three times for three years covering different seasons (northeast monsoon, tradewinds. and southwest monsoon). Climatological data was obtained from the Philippine Atmo- spheric, Geophysical and Astronomical Services Administration (PAGASA). Data for Bataan was obtained from Port Area, which is the weather station nearest to the study site. Samples for phytoplankton analysis were collected using a plankton net with mesh size of 20 |j.m towed vertically in the water '> MAP OF THE ^C^j PHILIPPINES _ii ImANILA BAYf <^)°'^'-^ pm^ * CD " '0 J)l)l BULACAN -^^^ 1 1 BATAAN \2 Limoy \ ^ MANILA BAY I \ METRO MANILA Lamao\ 3 2 1 CAVITE CITY^ 3 /pa.anaque . Mariveles / K^' Las Pinos ^^ _^ / J Bacoor ^ ^orregldorf ^ ^^^,j^ *Corresponding author. Figure I. Map of Manila Bay showing the sampling stations in Bataan and Cavite. 1251 1252 AZANZA AND MiRANDA column. The cells were preserved by the addition of \Vr Lugol's solution. Quantification of plankton was carried out using a Sedgewick Rafter chamber following the technique of Azanza (1997) and Corrales et al. (1995). Identification of phytoplankton was done using the taxonomic keys of Taylor (1976) and Tomas ( 1997). RESULTS Climactic Condition from 1997 to 1999 The climatological conditions during the three years of study are summarized in Figure 2. Cavite had a relatively higher monthly mean air temperature than Bataan. The highest mean air tempera- ture recorded in both places was in the range 34-35°C, occumng during tradewinds in the months of April to May (Fig. 2a). The general rainfall conditions in 1997 and 1998 were relatively simi- lar for the two sites. The amount of rainfall was higher during May to August of 1997. while from 1998 to 1999 there was little varia- tion in the amount of rainfall in both sites (Fig. 2b). The wind force in Cavite ranges from 2-6 nips, while in Bataan it ranges from 2—1 inps (Fig. 2c). Phytoplankton Composition The phytoplankton community was composed mainly of dia- toms and dinoflagellates. The diatoms were more dominant a. Temperature b. Rainfall S S wzS52"'<"Zg!55^"Z than dinoflagellates in all samples for the three seasons in the three years of study. There were 10 families of dinotlagellates. nine families of diatoms, and two marine flagellates (raphido- phytes and chrysophytes) found in Manila Bay (Table 1). Under dinoflagellates the families Gonyaiilacaceae. Ceratiaceae. Diiio- physicueae. Gyiiiiuuliiiiiuciw, Pvorocentraccae. Ncculucaceae, TABLE I. List of phytoplankton species identifled in Manila Bay, Philippines. Family Dinoflagellates Gonvaiilacaceae Ceratiaceae Dinophysiaceae Gymnodiniaceae Prorocentraceae Noctilucaceae Perijiniaceae c. Wind force Figure 2. Climatological conditions from 1997 to 1999. Pyrocystaceae Pyrophacaceae Diatoms Chaetocerotaceae Coscinodiscaceae Rliizosoleniaceae Melosiraceae Naviculaceae Bacillahaceae Thalassiosiraceae Thalassionemalaceae Oscillatoriaceae Flagellates Chatonellaceae (Raphidophytes) DiclyoLliaceac (Chrysiiphytes) Species Gonyaiihix spin if era Alexandriinn sp. Pyrodiniiim baluiiiwnse var. compressum Goniodoma sp. Ceratiwn fiirca Ceratium macroceros Ceniliiim tripos Ceratium trichoceros Ceratiwn fusiis Ceratium breve Ceratium carriense Ceratium vultur Dinopliysis caudata Dinophysis rotundata Dinopliysis miles Gyrodinium spirale Gvrodinium sp. Gymnodinium catenalwn Gymnodiniwn sanginneum Cochlodinium sp, Polykrikos kofoidii Prorocentrum micans Prorocentrum lima Noctihica scintillans Protoperidinium depressum Protoperidinium leonis Protoperidinium divert;ens Protoperidinium conicum Protoperidinium pelliu idiiiii Protoperidinium sp. Diplopsalis sp. Pyrocystis sp. Pyrophacus stenii Bacteriastruin spp. Chaetoceros spp. Coscinodiscus spp. Rhizosolenia spp. Guinurdia spp. Melosira spp. Pleurosigma spp. Navicula spp. Pseiidonilzscliia spp. Skeletonenui spp. Thalassiosiru spp. Thalassionema spp. Thalassiothrix spp. Trichodesmium spp. CImtonella Dictycicha spp. Pyrodinivm Toxic Blooms and Phytoplankton Composition 1253 TABLE 2. Corelation anal>sis between Pyrndiiiiiim cell coiiiil and eliniatold^lcal eondllions. Balaan Cavite Temperature Rainfall Wind force 0.2418 0.0320 0.3793 0.0203 -0.0998 Pcrkliiiiiuciic. P\i<>c\s!iiccac. and Pyinpliacavcac were repre- sented. The diatoms belonged to the families Chaetocerotaceae. Coscinodiscaceae, RhizDsoleniaceae. Melosiraceae. Navicii- hiceae. Bacillariaceae. Thalassiosiraceae. Thalassumcnuuaceae. and Osdllatonaceae. ChatoneUa. a raphidophyte. and Dktyocha. a chrysophyte. were the flagellates present in samples from 1999. In the majority of the samples, diatoms such as Skclennicmn spp.. Chctewceros spp., Coscinodiscus sp.. and Thalassiosira sp. occurred in high densities. Dinoflagellates such as Noctiliica scintillans. Protoperidinium spp., and Ceratium fiirca dominated the dinotlagellate population. Noctihua scintillans was con- sistently present during the three years of study (Fig. 5) with a distinct increase in population during northeast monsoon and tradewinds before the onset of rainy days. C. furca formed blooms during the southwest monsoon in 1999 in Bataan waters. Co- iiidildiiiii sp. was present in Bataan waters during April and July 1999. Noctiluca scintillans bloomed in Bataan and the diatom Chactocerns spp. in Cavite. both of which succeeded the Pyind- iniiiin bloom. Relative Frequency within Three Seasons Northeast Monsoon Diatoms dominated the phytoplankton community in Cavite and Bataan from 1997 to 1999. The dinotlagellates were repre- sented in higher percentage in Bataan (179f ) compared to Cavite {19c) (Fig. 3). Tradewinds Diatoms dominated this season in Cavite and Bataan with an increase in percentage of dinoflagellates (33%) in Bataan without an increase in Cavite (7%) (Fig. 3). Southwest Monsoon Diatoms dominated this season in Cavite, while in Bataan. diatoms (54%) were only slightly more dominant than dinoflagel- lates (46%) (Fig. 3). Pyrodinium bahamense var. compressum Bloom Pyrodininni bloomed in 1998 from April to October (Fig. 4). The percentage composition of phytoplankton (Tables 3 and 4 1 showed that in 1998 Pyrodinium dominated the plankton commu- nity in Bataan, while in Cavite, the organism also bloomed but not at high enough concentrations to dominate the other phytoplankton species. Other Potentially Harmful Algal Species Phytoplankton composition showed the presence of other po- tentially harmful algal species aside from Pyrodiniiiin. These a. Bataan 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% \ 0% J— D Dinotlagellate Q Diatom Northeast Tradewind Southwest b. Cavite 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% ^1 Tradewind Southwest Northeast Figure 3. Percentage composition of diatoms and dinoflagellates In 3 different seasons from 1997 to 1999 (al Bataan and (hi Cavite. harmful organisms include Ale.xandriiiiii sp.. Cimyaulux sp.. Cyin- nodinimn catenatnin. Dinophysis spp., Prorocentniin spp., and Noctiluca scintillans. A N. scintillans bloom is not toxic but may cause fish kills due to oxygen depletion. DISCUSSION Time series data showed that diatoms dominated the phy- toplankton population in Manila Bay throughout the three seasons of study (Fig. 3). Bataan and Cavite. though situated within the same bay. have distinct differences during the three seasons with respect to dinotlagellate population. There is a pronounced in- crease in dinotlagellate population in Bataan from northeast mon- 10000 : 1000 i 100 , -Bataan -Cavite a> oj a> Figure 4. Pyrodinium bahamense \:\r. compressum cell count from 1997 to 1999. 1254 AZANZA AND MiRANDA TABLE 3. Yearly Composition of Phytoplanliton in Bataan TABLE 4. Yearly Composition of Phytoplankton in Cavite Percent (%) Percent { '^^r 1 1 C omposition per Year Dinoflagellates Composition per Year Dinoflagellates 1997 1998 1999 1997 1998 1999 Pyrodinium bahameiise var. Pxnidiiuiini haluimensf \ar. compresswn 89.58 compressum 6.94 Alexandi'ium spp. 0.03 Alexandiiiim spp. 0.01 Cochlodinium spp. i:.9i 0.02 0.08 Chatonella spp. 0.01 Dinophysis spp. 2.00 0.27 1..54 Cochlodinium spp. 1.36 0.01 Goniodoma spp. 1 .50 Dinophysis spp. 1.67 1.09 0.19 Gonyaitla.x spp. 0.17 Goniodoma spp. 0.40 Gxinnodinium spp. 0.10 Gonxoitlax spp. 0.0! Gyrodinium spp. 0.02 Gxmnodinium spp. 0.10 3.65 0.25 Noclihtca scintillans 36.25 5.15 13.37 Gyrodinium spp. 2.86 0.01 Prurocentrum spp. 0.81 Noctduca scintillans 6.75 9.30 1.02 Protoperidiniwn spp. 3.14 0.44 2.41 Prorocentrum spp. 0.35 0.50 Pyrocystis noclihtca 0.02 Protoperidiniwn spp. 1.37 2.28 1.62 Pyrophacus spp. 0.26 Pyrocystis noctiluca 0.01 Scripsiella spp. 0.25 Pyrophacus spp. 0.07 Diatoms Scripsiella spp. 0.19 Bacleriastrum spp. 0.69 1.38 Dicryocha spp. 0.01 Ceraliiim spp. 4.28 0.51 33.42 Diatoms Chaeloceros spp. 17.27 1.07 14.52 Bacleriastrum spp. 0.60 0.82 1.87 Coscinodiscus spp. 4.73 0.75 3.07 Ceratium spp. 2.83 1.97 2.92 Dictyocha spp. 0.03 0.09 Chaeloceros spp. 37.42 35.20 9.24 Guinardia spp. 0.14 0.59 Coscinodiscus spp. 16.16 16.05 7.63 Navicida spp. 0.03 0.02 Guinardia spp. 0.99 0.31 Pleiirosigma spp. 5.21 1.19 0.34 Melosira spp. 0.06 Pseiidimitzchia spp. 0.41 Navicula spp. 0.10 Rhizosolemii spp. 3.95 0.38 2.05 Pleurosigma spp. 5.69 3.44 0.63 Skelelonemu spp. "'2 9"* Pseudonitzchia spp. 4.68 0.10 0.95 ThaUissionema spp. 2.44 0.19 0.07 Rhizosolenia spp. 1 .63 8.52 1 .35 Thalassiosini spp. 4.48 0.22 1.00 Skeletonema spp. 64.42 Thallasiotlirix spp. 2.24 Thalassionema spp. 6.35 7.74 4.95 TOTAL 10()fr 1(11)'; lOO'/r Thalassiosira spp. Thallasiothrix spp. 13.39 1.36 0.17 1.18 soon to trudewinds to soiilhwes t monsoon, while in Cavite. di- TOTAL lOO'r 100'* 100M2025->wzg>SS-'« 10000 tSS"»wz»ss~>tnzg55-»y3 Figure 5. Climatological conditions in relation with Pyrodinium cell density. LITERATURE CITED Azanza. R.V. 1997. Contribution^ to the understanding ol the bloom dy- namics of Pyrodinium baltamense var. compressum: A toxic red tide causative organism. Science Diliman 9:1-2. Azanza, R. V.. R. O. Roman & L. N. Miranda. 1998. Shellfish toxicity and Pyrodinium cell density in Bataan. Philippines (1994-1997). i. 5/if//- fisii Res. 17:1619-1622. Azanza-CoiToles. R. & S. Hall. 1993. Isolation and culture of pyrodinium balmmense var. compression from the Philippines. In: T. J. Smayda & Y. Shimizu. editors. Toxic phyloplankton blooms in the sea. Elsevier Science Pub. pp. 725-730. Bajarias. F. F. A. 1999. Distribution of marine dinollagellates in the Phil- ippines. In: G. Vigers, K. S. Ong, C. McPherson, N. Millson. 1. Watson, & A. Tang, editors. Proceedings, ASEAN-Canada Technical Confer- ence on Marine Science. EVS Environment Consultants Ltd & Dept. of Fisheries, Malaysia, pp. 458^67. Bajarias, F. F. A. & J. R. Relox. 1996. Hydrological and climatological parameters associated with the Pyrodinium blooms in Manila Bay, Philippines. In: T. Yasumoto, Y. Oshima & Y. Fukuyo, editors. Harm- ful and toxic algal blooms. Paris: IOC-UNESCO, pp. 49-,52. Corrales, R. A. & R. Crisostomo. 1996. Variation in Pyrodinium cyst density. In: T. Yasumoto, Y. Oshima & Y. Fukuyo, editors. Harmful and toxic algal blooms. Paris: IOC-UNESCO, pp. 181-184. Corrales, R. A., M. Reyes & M. Manin. 1995. Notes on the encystment and excystment of Pyrodinium habumense var. compressum in viiro. In: P. Lassus. G. Arzul. E. Erard-Le Denn, P. Gentien & C. Marcaillou-Le Baut, editors. Harmful marine algal blooms. Paris: Lavoisier Science Publ. pp. 573-578. Taylor. F. J. R. 1976. Dinoflagellates from the Indian Ocean Expedition. Bibl. Botamcu 132:1-234. Tomas, C. R. 1997. Identifying manne phytopkinkton, London: Academic Press Limited. 835 pp. Velasquez, I. B., G. S. Jacinto, C. I. Narcise & N. C. T. Cuaresma, Jr. 1997. The role of dissolved nutrients and other abiotic factors in red tide episodes in Manila Bay. In: G. Vigers. K. S. Ong, C. McPherson. N. Millson, 1. Watson & A. Tang, editors. Proceedings. ASEAN-Canada Technical Conference on Marine Science. EVS Environment Consult- ants Ltd & Dept. of Fisheries. Malaysia, pp. 111-67-78. Villanoy. C. L.. R. A. Corrales, G. S. Jacinto, N. T. Cuaresma, Jr. & R. P. Crisostomo. 1996. Toward the development of a cyst-based model for Pyrodinium red tides in Manila Bay, Philippines. In: T. Yasumoto, Y. Oshima & Y. Fukuyo, editors. Harmful and toxic algal blooms. Paris: IOC-UNESCO, pp. 189-192. Yniquez, A. T., R. V. Azanza. B. Dale & F. Siringan. 2000. Dinotlagellale cyst record on sediment cores from two sites in Manila Bay, Philip- pines, with different degrees of toxic red tide intluence. Poster paper presented during the 9th International Conference on Harmful Algal Blooms. Tasmania, Australia. 7-1 1 February 2000. Juiirmil oj Slu'lljish Kcsainh. Vol. 20, No. 3, 1237-1261, 2001. DISTRIBUTION OF GYMNODINIUM CATENATUM GRAHAM AND SHELLFISH TOXICITY ON THE COAST OF SUCRE STATE, VENEZUELA, FROM 1989 TO 1998 AMELIA LA BARBARA-SANCHEZ* AND JESUS F. GAMBOA-MARUEZ INIA-CIAE Siicn'/Niu'va Esparla. P. O. Box 236. Ciiimmd. estado Sucre. 6101. Venezuela ABSTRACT Gyinnodinium catemiliim blooms were first detected on the Venezuelan coasts in 1989. Since then, monitoring along the northeastern coast of Sucre state and the Gulf of Cariaco was carried out to detect the presence of this organism and to measure en\'ironmentaI parameters (temperature and salinity) that may be associated with its appearance. G. catenahim has been obser\'ed every year along the coast, with maximum abundance between June and October. Blooms of C. catenatiim (> 1.000 cells ■ ml"') have been recorded only in the northeastern coast during the rainy season (July - Nov.) when water temperature was above 25"C and salinity was .^6.5 to 38%f. The lowest abundance (< 100 cells ■ ml"') occurred from Dec. - June in a wide range of temperature (22-28'C) and salinity ( 33.2-39 %c). The species has been less abundant in the Gulf of Cariaco and has not formed blooms yet. The dynamics of the marine environment at the north coast seem to favor the growth of C. caiemilum populations. Shellfish toxicity showed different PSP levels and did not always coincide with the presence of G. calcmiuim KEY WORDS: G\iiiiiiuliiiiuiii calciuiiiini. shellfish toxicitv. Sucre state. Venezuela INTRODUCTION Algal blooms in the coastui waters of northeastern Venezuela occur with yearly frequency. Many blooms are associated with toxigenic species that induce toxicity in bivalve mollusks, which can have severe impacts on public health, on the regional economy, and upon the marine environment. Feiraz-Reyes (1992) reported that red tide events off the northeastern coast of Venezu- ela occur mainly during the dry season and occasionally during the rainy season. One of the species responsible for blooms along the coast of Sucre state has been Gyiniiodiiiiiiiu catenatiim (Graham 1943). Its appearance has been associated with PSP toxicity of bivalve mol- lusks of commercial value (La Barbera-Sancliez et al. 1993). The species was first detected in Venezuelan waters in 1989, Since then it has reappeared several times, causing PSP on the northeastern coast of Sucre state. This athecate dinoflagellate has a widespread distribution. It has been reported from the Mexican Pacific Coast (Morey-Gaines 1982, Mee et al, 1986), the coasts of Argentina (Balech 1964), the Atlantic Coast of Spain (Estrada et al. 1984, Fraga et al. 1988), the Mediterranean Coast (Carrada et al. 1991 ), and Australia (Oshima et al. 1987, Hallegraeff et al. 1989). The presence of G. cateiialiim in Venezuelan coastal waters is associated with certain environmental conditions, particularly with intrusion of warm seawater and periods of upwelling relaxation. Fraga et al, ( 1988, 1990) and Fraga ( 1996) associated the presence of G. catenatiim with the introduction of surface warm waters from offshore toward coastal inlets during the period of upwelling re- laxation after summer time. In this study, the presence and abun- dance of C. catenatiim is compared with observed trends of salin- ity and temperature of coastal waters in Sucre state, and toxicity events during the monitoring program performed from 1989 to 1998. MATERIALS AND METHODS A iiwnuoring program to detect the presence of toxic algae in the plankton and toxicity episodes in bivalve mollusks was iniple- *Corresponding author. E-mail: alabarbe@sucre.udo.edu.v mented from 1989 to 1998. Water samples were taken at sea surface and 3 m deep every 13 days, in different localities along the coast of Sucre state where bivalve mollusks beds are found. Three sampling sectors were established: the northeastern sector with stations at Bahia Patilla, La Iglesia. and San Juan de Las Galdonas; the northwestern sector with two stations at Chacopata; and the southwestern sector with stations at La Chiea and Penas Blancas within the Gulf of Cariaco (Fig. 1 ). Water samples were obtained using a Van Dorn bottle, with a portion fixed with Lugol solution for quantitative cell counts by the Utermohl ( 1958) method. The other portion was used to mea- sure water temperature with a thermometer (A = l°C) and salinity with an ATAGO refractometer (A = 0.1%f ). Transparency at each sampling station was measured using a Secchi disc. Phytoplankton samples were also obtained using a net with 23 (xm mesh for species identification. The concentrations of PSP toxins in ark shells (Area zebra) and mussels iPenia perna) were measured by mouse bioassay (AOAC 1984). RESULTS AND DISCUSSION Gyinnodinium catenatiim has a wide distribution on the coast of Sucre state. The species was first found in the coastal zone of Sucre state within waters of the Gulf of Cariaco during 1986. in very low density. In April-May 1988, G, catenatiim appeared in coexistence with Alexandrimn tamarense, Ganyaiila.x cf. soiissa, and Gonyaulax polygramma. in a bloom caused hy the latter spe- cies, which caused toxicity in the mussel. Perna perna. cultivated in the Gulf of Cariaco (La Barbera-Sanchez et al. 1993). In 1989, blooms of G. catenation were recorded in the north- eastern and northwestern sectors of Sucre state, with abundance larger than \Q x 10" cells • ml"' in February and April (Fig, 2), G. catenatum coexisted in low abundance with A. tamarense and A. iniinikitiim. The presence of G. catenatiim in these sectors was associated with toxicity events of PSP, which reached levels higher than 600 p.g STX/IOO g meat and affected several species of bivalve mollusks of commercial value (La Barbera-Sanchez op 1237 1258 La Barbara-Sanchez and Gamboa-Maruez ii° e'' LMofgenta Figure 1. Location of sampling stations in Sucre state, southwestern sector: Penas Blancas (1), La Chica (2); northwestern sector: Chacopata (3); northeastern sector: La Iglesia (4). Bahia Patilla (5), San Juan de Las Galdonas (6). 1000 1989 100 • 1994 -.,00 » V*' \ .♦. ? 1 1 5,0' 1 ° ■■■■■■• o o >' » 1-. ■10 % i o 1 1 1 ■ D / O) / 1 m 10 1990 ,0 □ 1995 / E E 1 1 / iC • %' ■ A A o • ° i \ c.\ \ 1 o o A A 1 . i V ■ ■ 10 1991 100 g 1996 \ • •. E 0 E ^10 • - ^ '. • o .5 \\, -'a'-. •* -■ 1 ^ o--a » "♦' o ■ ■ 100 1 1992 .■■♦. 1000 ■ 1997 'e ' * ■' OOO [ Q 5 1 ■■♦■■■ '■■♦■ D .*' ^ * O10 1 \ = 0 \ \ D 1000 • .♦. 1993 1000 1998 ■: 100 E ' ;' ° '■ \Afr-^ ^100 E • k A ; fi'\ M ^ u O 10 ' :\ V 0 10 ■ •♦ ■ 8 V • V A ^^ ^ * .i s J ■ A o ^ /^ 1 • 1 1 1 -' 1 s E i & 1 I 1 1 s 1 ? 1 t o 1 3 — ■ — Chacopata - -c- -Iglesia ■■■♦•- Patilla — ■ — Chacopata — D- — iglesia ■■■* • Patiila O Cariaco Guif A Sa 1 Juan G. o CariaiM Gulf A San Juan G. Figure 2. Abundance variation of Gymniidiniiim calenuliim on the coast of Sucre state, Venezuela: 1989-1998. Distribution of GYMNoniNii'M catenatum 1239 p O O X h 0 J 3 2,5 2^ 1,5 rooT-Mco'^fincDSoo C0O)O)G)0^O)05O)(J)(3) 0)0)050)0)0)0)0)050) Figure 3. Variation annual mean of the tovin concentration on tlie coast of Sucre state. cit.). In the northeastern sector, toxicity levels in the mussel. P. penui. and the clam. TivcUi mculrniiles, reached values higher than 1.000 ^Lg STX/100 g meat. The death of inussels in natural banks was also observed. In the northwestern sector, the levels of PSP toxin in ark shells. Area zebra, were beyond those permitted for human consumption. These events caused intoxication by con- sumption of bivalve mollusks in 15 persons, and great economic impact in the region. During the period 1990-92, G. cateiniliiiii was observed in low abundance (Fig. 2). but the number of observed organisms (from < 1 to 10" ) in the northeastern sector did not correspond with tox- icity levels recorded in the mussel. P. penia. during 1991 and 1992 (Fig. 3). In 1993. C. catenaluiii v\as observed during the entire year in the northeastern sector with variable densities (from 1 to 1.342 cells • ml"' ). The highest abundance was recorded in May. August, September, and October, while the lowest abundance was ob- served during the first months of the year (Fig. 2). The period of highest densities coincided with episodes of toxicity in November and December (1.233 p,g STX/lOOg meat) that extended until October 1994. at toxicity levels of 100 (jtg STX/lOOg meat. In 1994. G. catenatum registered low abundance and there was toxicity in mussels almost all year long, with maximum levels between January and April. However, there was no correspon- dence between the toxicity levels and the scarcity of vegetative cells of the toxic species. January and April are months of strong upwelling along the coasts off Sucre state; therefore, it is possible that the toxic event may have been originated by cysts re- suspended by upwelling in the water column, as has also been pointed out by La Barbera-Sanchez and Estrella (1996). The years 1995. 1996. 1997, and 1998 (Fig. 2) were notorious for the low number of cells of G. catenatum. except in the north- western sector, where there were increases in April 1997 (984 cells • ml"') at surface level. Otherwise, the levels of toxicity for PSP stayed low. with a total absence in 1995. For the years 1989 and 1993 the toxicity levels in bivalve mollusks of Sucre coasts were higher than the allowable maxi- mums (> 600 jjig STX/lOO g meat) (Fig. 3). The toxic levels for 1 990. 1 99(1. and 1 998 were lower than 80 |jLg STX/ 1 00 g meat. The levels for the remaining years in the study yielded values between 90 and 190 pg STX/lOO g meat. G. caleiuuiuti was always present in the area, at varying tem- peratures and salinities. It appears in higher density from July to February in coincidence with the rainy season (from June to No- vember) and with the onset of trade winds and upwelling (Novem- ber. December. January, and February) (Figs. 4 a, b. c). C. catenatum recorded low densities within a wide range of temperatures (22-28°C) and salinity gradients (33.2 -39%r). Blooms of the organism (> 1.000 cells ■ ml"') were observed in 1989 (February), 1993 (May, August, September, and October). 1997 (January. April), and 1998 (September), in waters with tem- peratures above 24"C and salinity gradients between 35.5 and 38%f during the rainy season, contrasting with the southwestern sector (Gulf of Cariaco), where the species has not bloomed yet (Fig. 4 a). The species was more abundant and frequent in the northwestern and northeastern sectors (Figs. 4 b and c). However, this behavior is more evident in the northeastern sector (Fig. 4 c), where the species has a spatial and seasonal distribution. The northeastern sector is characterized by strong swelling and bv the intluence of the low salinity and influx of nutrients from the Gulf of Paria and adjacent regions (Okuda et al. 1974). It has, in addition, increased orthophosphate content during the rainy sea- son, which favors blooms of diatoms and dinoflagellates (La Bar- bera-Sanchez et al. 1988). Figueiras and Pazos (1991) reported dense overgrowths of G. catenatum in water bodies of intense agitation, similar to what this study found for the northeastern region. Thus, the increase in the number of cells may be influenced by soil and wastes washed off by rain. This phenomenon affects the saline concentrations and incorporates organic and inorganic elements that serve as nutrient sources for the species. Peperzak et al. ( 1996) reported that an increase of freshwater in coastal areas may result in a transient stratification of salinity and reduced tur- bulence. This allows for greater stability of the water column and favors the growth of dinoflagellates. However, it has been found that it is thermal stratification, rather than salinity stratification, that leads to dinotlagellate accumulation and dominance. Like- wise. La Barbera-Sanchez and Estrella (1996) found a significant correlation between proliferation of G. cateiuitnm during the rainy season with temperature and transparency in the northeastern sec- tor of the coast of Sucre. Results have revealed that the distribution of G. catenatum along the coast of Sucre state does not seem to be related to the temperature and salinity. However, the abundance and the periods of toxicity do point to a transient relationship with temperature increases and rainy seasons. The dynamics of the marine environ- ment along the northwestern and northeastern sectors in the coast of Sucre state, which need to be fully studied, seem to favor the growth of G. catenatum populations, perhaps due to a greater availability of nutrients or to the presence of some oligoelement in the water during the rainy season. ACKNOWLEDGMENTS We acknowledge the cooperation of technicians Oswaldo Gal- lardo and Simon Silva from the Toxicology Laboratory at INIA Station in Cumana-Sucre. We thank Jose Alio (INIA. Sucre) for assistance in the production of this manuscript. This project was financed by the Technological Development Program, which is subscribed by the Venezuelan government and the Interamerican Development Bank. 1260 La Barbara-Sanchez and Gamboa-Maruez 4U 2. 38 36 * * i U) 34 32 30 28 o 26 24 ■ • ■ H 22 7(1 V-, 1000 -g Temperature ▲ Salinity ■ -Abundance Figure 4. Relationship between Gymiuidiniiim caienatum abundance with salinity and temperature in the sectors: A) southwestern, B) north- western, and C) northeastern of Sucre state. LITERATURE CITED AOAC. 1984. Official methods of analysis. Paralytic shellfish poison. Bio- logical method: 18.086-18.092. Baiech. E. 1964. El plancton del Mar del Plata durante el periodo 1961-62. Bol. Insl. Biol. Mar. Univ. Nac. Buenos Aires 4:1-49. Carrada. G. C. R. Casotti. M. Modigh & V. Saggiomo. 1991. Presence of Gymnoclinium catenalum (Dinophyceae) in a coastal Mediterranean lagoon. / Plankton Res. 13:229-238. Estrada. M.. F. J. Sanchez & S. Fraga. 1984. Gyinnotliniuni catenalum (Graham) en las Rfas gallegas (NO de Espafia). Inv. Pesq. 48:31^0. Ferraz-Reyes. E. 1992. Fitoplancton de la Ensenada de Canguas, Peninsula de Paria. Estado Sucre. Venezuela. Bol. Inst. Oceano^r. Vene-nela. Univ. Oiiente 31:17-26. Figueiras. F. G. & Y. Pazos. 1991. Hydrography and phytoplankton of the Ria de Vigo before and during a red tide of Gvinnodinium catenatiini Graham. / Plankton Res. 13:589-608. Fraga, S. 1996. Wintering of Gymnoclinium catenatum Graham (Dino- phyceae) in Iberian waters, pp. 21 1-214. In: T.Yasumoto, Y. Oshima & Y. Fukuyo, (eds.). Harmful and Toxic Algal Blooms. IOC- UNESCO, Sendai. Japan. Fraga, S.. B. Reguera & I. Bravo. 1990. Gymnoclinium catenatum bloom formation in the Spanish rias. In: E. Graneli. B. Sundstron. L. Edler & D. M. Anderson, editors. Toxic marine phytoplankton. New York: Elsevier, pp. 149-154. Fraga. S.. D. M. Anderson, I. Bravo, B. Reguera, K. A. Steidinger& C. M. Distribution of Gymnodinium catenatum 1261 Yentsch. 1988. Influence of upwelling relaxation on dinoflagellates and shellfish toxicity in Ria de Vigo. Spain. Esnitir. Cdii.si. Shelf. Sci. 27:349-361. Hallegraeff. G. M.. S. O. Stanley. C. J. Bolch & S. I. Blackburn. 1989. Gynmodiiuum calenatiim blooms and shellfish toxicity in southern Tas- mania. Australia. In: T. Okaichi, D. M. ."inderson. & T. Nemoto. edi- tors. Red tides: Biology, environmental science and toxicology. New York: Elsevier, pp. 411-414. La Barbera-Siinchez. A. & G. Estrella. 1996. Occuirence of toxic di- notlagellates and PSP on the northeast coast of Sucre State. Venezuela: Relationship with environmental parameters. In: T. Yasumoto, Y. Oshiiria. & Y. Fukuyo, editors. Harmful and toxic algal hloiims. Sendai. Japan: IOC-UNESCO, pp. 117-120. La Barbera-Sanchez. A.. S. Silva & O. Gallardo. 1988. Variacion estacio- nal del fitoplancton del Golfo de Cariaco y costa norte del Estado Sucre durante 1987. Acta Ciem. Venezolana 39 (Supl. 1 ):62. La Barbera-Sanchez, A.. S. Hall & E. Ferraz-Reyes. 1993. Alcxaudrium sp., Gymnodinium catenatum and PSP in Venezuela. In: T. J. Smayda & Y. Shimizu, editors. Toxic phytoplankton blooms in the sea. De- velopments in Marine Biology, 3. New York: Elsevier, pp: 281-285. Mee, L. D.. M. Espinosa & G. Diaz. 1986. Paralytic shellfish poison with a Gymnodiniimi catenatum red tide on the Pacific coast of Mexico. Mar. Environ. Res 19:77-92. Morey-Gaines G. 1982. Gymnodinium catenatum Graham (Dinophyceae): morphology and affinities with armoured form. Pliycologia 2/:l54- 163. Okuda. T.. J. Beni'tez. J. Sellier de Civrieux. J. Fukuoka & B. Ganiboa. 1974. Revision de los datos oceanograficos en el Mar Caribe Sur- oriental, especialmente el margen continental de Venezuela. Cuadeniox A-ules No. 15 Oceanogr. III. Conf N. U. Der. Mar. Caracas. 1-177 p. Oshima, Y.. M. Hasegawa. T. Yasumoto, G. Hallegraeff & S. Blackburn. 1987. Dinoflagellate Gymnodinium catenatiun as the source of para- lytic shellfish toxin in Tasmanian shellfish. Toxicon 25:1 105-1 111. Peperzak, L., G. J. Snoijer, R. Dijkema. W. W. C. Gieskes, J. Joordens, J. C. H. Peelers, C. Schol, E. G. Vrieling & W. Zeverboom. 1996. De- velopment of a Dinophy.sis acuminata blooom in the River Rhine plume (North Sea). In: Y. Yasumoto, Y. Oshima, & Y. Fukuyo, editors. Harmful and toxic algal blooms. Sendai, Japan: IOC-UNESCO, pp. 273-276. Utermohl, H. 1958. Zur Vervollkoiiiniung der quantitatixen Phytoplank- ton-Methodik. Mitt. Int. Vcr. Tlieor. Anf;c\v. Lunnol. 9:1-38. Journal oj Slwlljhh Kc.scunh. Vol^ 20, No. 3. 1263- 1 :6X. 20(11. PIGMENT PROFILE AND VIOLAXANTHIN CYCLE OF HETEROSIGMA AKASHIWO ( RAPHIDOPHYCEAE) YUTAKA OKUMURA,' * MAKOTO VAMASAKI,- TOSHIYUKI SUZUKI,* KAZUHIKO ICHIMI,' AND OSAMU OKU' 'Tohokii National Fisheries Research Institute. 3-27-5 Shinhama. Shioganui. Miyuiii 9H5-000I. Japan: -National Research Institute of Aquaciiltiire. 422-1 Nakatsuhamcmra, Nunsei. Waiarai. Mie 516-0193, Japan: ^Japau Science and Technology Corporation 4-1-8 Honchou, Kawaguchi. Sailama 332-0012. Japan ABSTRACT Pigmenl profiles of the RLiphidophycean Hcfcrasitiiiui iikit.'ihi\io were analy/ed by high-performance liquid chroma- tography (HPl.C) with a diode array detector. Chlorophyll f, + c,, fucoxanthin. violaxanthin. antheraxanthin, zeaxanthin. chlorophyll a. and (i-carotene in three strains of//, ukasliimi (NIES6, 10. 293) were identified. The pigment profiles of H. akushiwo in a light:dark cycle were also analyzed by HPLC. The peaks of antheraxanthin and zeaxanthin were higher in the light, while the reverse was recorded for violaxanthin. This is the first evidence of the existence of a violaxanthin cycle in Raphidophyceae. KEY WORDS: Raphidophyceae, Ht'lfiosigimi aka.shiwa. pigments INTRODUCTION Since the analysis of microalgal pigments by high-performance hquid chioniatography iHPLC) was developed (Mantoiira and Llewellyn 1983. Wright and Shearer 1984. Wright et al. 1991. Suzuki et al. 1993. Latasa et al. 1996. Furuya et al. 1998). the pigment profiles of many classes of microalgue have been ana- lyzed by this method (Kohata et al, 1991, Sn/Aiki and Ishimaru 1992. Garrido and Zupata 1998). In particular. HPLC analysis of pigments has been conducted to investigate marine microalgal spe- cies that contribute to primary production (Gieskes and Kraay 1983, 1986. Goericke and Repeta 1993. Letelier et al. 1993. Andersen et al. 1996. Bidigare and Ondrusek 1996. Roy et al. 1996. Suzuki et al. 1997). Analysis by HPLC is also useful to study the synthesis of microalgal pigments, such as in the diadinoxanthin cycle and violaxanthin cycle (Goericke and Welschmeyer 1992. Porra et al. 1997. Lohr and Wilhelm 1998. Oku and Kamatani 1999), In recent years, several reports iFiksdahl et al. 19S4a, Fiks- dahl et al. 1984b. Yamaoka and Takimura 198.^. Kohata and Wa- tanabe 1988. Kohata et al. 1991 ) and a review (Jeffrey and Vesk 1997) of pigment profiles of harmful algae of the family Raphi- dophyceae have been published. It is interesting that these profiles, in particular that of Helerosigma aka.sluwo. differ among the re- ports (Fiksdahl et al. 1984b, Yamaoka and Takimura. 198.^, Ko- hata et al. 1991). suggesting variation in the pigment profiles of such species. In the present study, detailed pigment profiles of three strains of H. akashiwo were obtained and the violaxanthin cycle of this species was investigated. MATERIALS AND METHODS Pigment Analysis by HPLC The following marine microalgae with known major pigments were cultured; Rhodophyceae. Porphyridium crucniKiu (CSIR025); Bacillariophyceae. Phaeodaclyliim triconuiuim (CSIR029); Chlorophyceae. Dunaliella tertiolecla (CSIR017-^); and Dinophyceae. Amphidiniitm canererae (CSIR0212). all ob- tained from the CSIRO Microalgae Research Centre. CSIRO Ma- rine Research Laboratories. Tasmania. Culture conditions were: temperature 21 ± L'C. illumination 8()(>(.) ± 1000 lux and a light: dark cycle of 14:10 hours. Test tubes (25 x 200 mm, 64 ml) containing 30 ml off/2 medium (Guillard and Ryther 1962) were used as the culture vessels, Thiily milliliters of the growth phase medium were filtered on glass-fiber filters (Whatman GF/F), The filters were then extracted with 10 ml N. N -dimethylformamide (DMF) (Suzuki and Ishi- maru, 1990). Chlorophyll (/ standard (Chi o; Wako Pure Chemi- cals, Ltd.) was also dissolved in DMF. Each algal sample extract and the Chi a standard were chroniatographed on an STR ODSII column (Shimadzu Chemical Industries, Ltd.) with a Shimadzu Raphidophyceae HPLC system (system controller. SCL-IOAVP; auto injector, SIL-IOADVP; liquid chromatograph, LC-IOATVP; column oven. CTO-IOAVP; and diode array detector. SPD- MIOAVP). The temperature of the column was kept at 40°C. The HPLC solvent system followed the modified method of Suzuki et al. ( 1997) as shown in Table 1. After pigment separation by HPLC. individual pigments were identified from published data of their diode anay spectra (Jeffrey et al. 1997). Analysis of Pigment Distribiitinn in Three Strains of H. akashiwo Three strains of H. akashina (N1ES6, NILS 10, NIES293) ob- tained froin the Microbial Culture Collection, National Institute for TABLE L HPLC solvent system program Time Flow rate (min) (ml/min) ^Al VcBl Conditions 0 1.0 100 0 Injection 1 1.0 0 75 Linear gradient 8.5 1.0 0 100 Linear gradient 28.5 1,0 0 100 Equilibration 28.51 1,0 100 0 38.50 1.0 100 0 Equilibration *Corresponding author. 1. = 80:20 methanol: 0.5M ammonium acetate (pH 7.2, v/v) 2. = 70:30 methanol: acetone 1263 1264 Okumura et al. Environmental Studies (NIES Collection). Environment Agency. Japan, were cultured and pigments analyzed using procedures as mentioned above. Identification of the pigments was made by comparison of their retention time with standard and diode array spectra of pigments of the other cultured microalgae. Response of Pigments of H. akashiwo to a Light :Dark Cycle (Light Intensity) H. iikiisliiwo NIES6 was pre-cultured in three test tubes with ."^0 ml of f/2 medium. 30 ml of medium from the exponential growth phase was added to three 300 ml glass Erienmeyer tlasks contain- ing 200 ml of f/2 medium. The microalgae were cultured under illumination of 8000±I000 lux with a light:dark cycle of 14:10 hours for four days. The microalgae were harvested in their growth phase after 1 1 and 13 hours of light period (Lll h, L13 h), 10 hours of dark period (DIO h), 4 and 1 1 hours in the next light period (NL 4 h and NL 11 h) and 8 hours in the next dark period (ND 8 h). The samples were frozen at -SO^^C immediately after filtering the cultures using glass-fiber filters. After the samples were har- vested, they were extracted in DMF for 24 hours and analyzed by HPLC. RESULTS AND DISCUSSION Pigment Analysis by HPLC The absorbance iiiaxinia (nm) of the pigments (Table 2) were close to those reported by Jeffrey et al. (1997). This also indicates that there was no significant influence of variations in the extrac- tion solvent and HPLC mobile phase on the spectra of the pig- ments. Analysis of Pigment Profiles in the Three Strains of H. akashiwo The major pigments of the three strains of W. akushin-o (NIES6. 1(1. 293) were eluted from the column in the order: chl c, + C2. fucoxanthin. violaxanthin. antheraxanthin. zeaxanthin. chl a. and P-carotene (Fig.l). The major pigments of Raphidophyceae, in particularly H. akashiwo, have varied among the published reports. Jeffrey and Vesk (1997) reported that the major pigments in Raphidophyceae are fucoxanthin. violaxanthin. and p.p-carotene, while Fiksdahl et al. ( 1984a) mentioned that the major pigments of TABLE 2. Spectrum and retention time of several carotenoids extracted in DMF. Pigment Retention time (min) ,\bsorhance maxima (nm) Ob.served Literature (Jeffrey et al. 1997)* Pendinin {Aiiiphiiliiiitini carterae] Fucoxanthin {Pliiu'chhicrylKiii tricdniulum) Neoxanthin {Duuuluila tertiolecta) Violaxanthin (Dunaliella leniolechi) Diadinoxanthin ( Phaeodacrvliim triconiiiuiin ) Antheraxanthin {Dunaliella leriiolecta) DiatoxantliMi {PluiciHiactxIum rritnniiitiiiii) Zeaxanlhin {Pnrphyridiwn Liiicntuni) lOJIO 472 09.98 465 10.00 471 10.67 450 10.67 448 — 10.68 449 — 10.87 413 437 465 10.88 413 437 465 10.89 413 437 465 11.21 415 439 470 ii.:i 417 440 469 11.22 416 440 470 11.68 446 476 11.68 — 446 476 11.69 — 446 476 11.98 446 473 11.99 — 446 474 12.00 — 446 474 12.51 451 479 12.52 — 453 480 12.53 — 452 481 12.83 451 479 12,82 — 452 479 12. SI — 452 479 474 449 413 417 427 421 427 428 468 439 441 448 444 454 455 468 471 477 471 482 483 * Extraction in 90% acetone Pigments of Heterosigma aic^shiwo 126? so- 3 " < E (D » O c nj 1- (A 4.. - •; ■• ■■ ■"■'•'" • : : I : : ; ■| 1 ) l'- \ \ \ \ \ \ \ \ \ N1ES6 ; 2"\ \ i \ j ['"""; : ] ] 1 -J 1 \ L 6 ] \ } 1 { \ j < u £ ni o m <,« Retention time (min) NIES10 ;2 I ...;.. • .• M. i 4l|lM5 i 6 i _Lu_--ik|.l Uk. Retention time (min) so- ~ : ■■ f ; 1 i : „ 3 i i ; is NIES293 3" ■f t 1 ;"■ ■: i \ t — < E o H 1 1 h c 1 ■ i i i ro -e» j - -^ -i ■: '.-y Hl-i i ; r- o to ' \ \ ^nM : 1 \ ^ - Retention time (min) Figure 1. Typical thromatograms (435 nm) of three strained //. akashiwo (MESA, 10,293). Individual peaks arc (1 1 Chlf|+c,. |2| fucoxanthin. (3) viulaxanthin, (4) antheraxanthin, (5) zeaxanthin, (6) Chi a, and (7) (i-carotene. Raphidophyceae are chl a. c , + c-,. fucoxanthin, violaxanthin. /.e- axanthin. and 3-carotene. Fiksdahl et al. (1984b) also reported that the major caroteiioid in H. akasliiwo is peridinin; fucoxanthin was not observed in their study. Yamaoka and Takimura (1985) ob- served that in H. akasliiwo. fucoxanthin and ^-carotene are present as major pigments while zeaxanthin was absent. Carotenoids of C. aiuiqiia were fucoxanthin, violaxanthin. and p-carotene in Kohata and Watanabe's work (1988). Kohata et al. (1991) reported that fucoxanthin. violaxanthin. antheraxanthin. zeaxanthin. and (J-caro- tene were the major carotenoids of H. akashiwo. The pigment 1266 Okumura et al. profile of H. iikashiwo in the present experiment was similar to that reported by Kohata et al. ( 1991 ). Response of Pigments of H. akasliiwo to a Liglit.Dark Cycle (Light Intensity) The peaks of antheraxanthin and zeaxanthin (LI 1 h, L13 h. NL4 h and NLl 1 hi in the liaht were higher than those obtained under conditions of darkness (DIO h and ND8 h) (Fig. 2). The reverse, however, was the case for violaxanthin. It is known that pigment profiles in Bacillariophyceae and Dinophyceae vary with light intensity. There is the so-called "diadinoxanthin cycle" in which diatoxanthin is epoxidized to diadinoxanthin in darkness and de-epoxidized from diadinoxanthin to diatoxanthin in the light (Goericke and Welschmeyer 1992. Lohr and Wilhelm 1998). The "violaxanthin cycle" is another reported cycle relating to photo- < J u -e o < >■- : : ; : / : : : : i I \ \ I \ i 1 i 1 Lllh i \ I \ \ ( > i i 1 i i i ; 1 ■ 1 ■ — b r— t { ^ r; 3i i ! i Jylll/uii 1 ;l.|A|/'j Il|i\ 1 ■ 1 ' ■ '■ : /V 1 ' : 3 < o c O M < V3T 1: L,/>: 'V^l.l'l I I I .\. JJL DIOh Retention time (min) Figure 2. Typical chromatograms (435 nm)of //. akashiwo (N1ES6I under a Iight:dark cycle. Individual peaks are ( 1 ) Chlf , +c,, (2)fucoxanthin, (3) violaxanthin, (4) antheraxanthin, (5) zeaxanthin, (6) Chi a, and (7) (i-carotene. Pigments of Heterosigma akashiwo 1267 < 0) o c (0 1. o (0 < -•1 >; 1 1 : 2l i r 61 1 1 j 1 1 • ■ I i M i 3' 5 i ^ 1 '\-'H 1 i-^-^ \ i M i 1 1 i , NL4h 10 12 Retention time (min) ID* < o c -s = o w < < '^' o c TO Si . ^ - o 05 < i\L 1 ;rif-1-:5 _k fclilui „.i/JaJ iiHi—i NLIIh Retention time (min) : T ■ i 2 ; I i ; ■ ; ; M 1 ; lj X ' f'"i 3"! 1 1 1 1 1 ; ; ; \\ ': : : i„.. H 1 — -1 — -t^--^ H — ^^ \ \ \ \ . ND8h i 1 "1 \ ^ :' \ i \ : 1 i i i i 1 i i ..i ; 1 i i : i i i 1 i 1 i 1 7 i i Retention time (min) Figure 2. Continued. chemical epoxidation : zeaxanthin and antheraxanthin are epoxi- dized to antheraxanthin and violaxanthin, respectively, in the dark. and these epoxidized materials are de-epoxidized to the original materials in the light (Porra et al. 1997. Lohr and Wilhelm 1999). Although diadinoxanthin and violaxanthin cycles have been ob- served in Bacillariophyceae. Xanthophyceae. Haptophyceae. and Dinophyceae (e.g.. Willenioes and Monas 1991. Arsalane et al. 1994. Oku and Kamatani 1999. Lohr and Wilhelm 1999). the violaxanthin cycle in H. akashiwo has not previously been re- ported. Our observation is the first evidence of the existence of the violaxanthin cycle in Raphidophyceae. ACKNOWLEDGMENTS The technical assistance of Hiromi Suzuki is gratefully ac- knowledged. We thank Tsurumura Setsu (Tokyo University of Fisheries) for technical advice throuszhout this work. 1268 Okumura et al. LITERATURE CITED Andersen. R. A.. R. R. Bidigare, M. D. Keller & M. Latasa. 19%. A comparison of HPLC pigment signatures and electron microscopic ob- servations for oligotrophic waters of the North Atlantic and Pacific Oceans. Deep Sea Res.II 43:517-537. Arsalane. W.. B. Rousseau & J. C. Duval. 1994. Influence of the pool size of the xanthophyll cycle on the effects of light stress in a diatom; competition between photoprotection and photoinhibition. Photochem. Photohuil. 60:237-243. Bidigare. R. R. & M. E. Ondrusek. 1996. Spatial and temporal variability of phytoplankton pigment distributions in the central equatorial Pacific Ocean. Deep Sea Res.II 43:809-833. Fiksdahl, A., N. Withers. R.L. Guillard & S. Liaaen-Jensen. 1984a. Carot- enoids of the raphidophyceae-A chemosystematic contribution. Coinp. Biochem. Physiol. 788:265-271. Fiksdahl, A., N. Withers, & S. Liaaen-Jensen. 1 984b. Carotenoids of Het- erosigma akashiwo: A chemosystematic contribution. Binchem. Sys- tem. Ecol. 12:355-356. Furuya. K.. M. Hayashi. & Y. Yahushita, 1998. HPLC determination of phytoplankton pigments using W.W. -dimethylformaniide. J. Oceaiiogr. 54:199-203. Garrido, J. L., & M. Zapata. 1998. Detection of new pigments from Einil- iania hu.xleyi (Prymnesiophyceae) by high-performance liquid chroma- tography, liquid chromatography-ma.ss spectrometry, visible spectros- copy, and fast atom bombardment mass spectrometry. J. Phycol. 34: 70-78. Gieskes. W. W. C. & G. W. Kraay. 1983. Dominance of Cryptophyceae during the phytoplankton spring bloom in the central North Sea de- tected by HPLC analysis of pigments. Mar. Biol. 75:179-185. Gieskes, W. W. & G. W. Kraay. 1986. Analysis of phytoplankton pigments by HPLC before, during and after mass occurrence of the microtlagel- late Corymbellus aureus during the spring bloom in the open northern North Sea in 1983. Mar. Biol. 92:45-52. Goericke, R. & N. A. Welschmeyer. 1992. Pigment turnover in the marine diatom Thalassiosira weissflogii. II. The '"'C02-laheling kinetics of carotenoids. J. Phycol. 28:507-517. Goericke. R. & D. J. Repeta. 1993. Chlorophylls u and h and divinyl chlorophylls a and /' in the open subtropical North Atlantic Ocean. Mar. Ecol. Prog. Ser. 101:307-313. Guillard, R. R. L. & J. H. Ryther. 1962. Studies of raanne planktonic diatoms. 1. Cyclorella nana Hustedt. and Detonula confervacea (Cleve) Gran. Can. J. Microbiol. 62:229-239. Jeffrey, S. W. & M. Vesk. 1997. Introduction to marine phytoplankton and their pigment signatures. In: S. W. Jeffrey, editor. Phytoplankton pig- ments in oceanography. France: UNESCO, pp. 37-84. Jeffrey, S. W., R. F. C. Mantoura &. T. Bjornland. 1997. Data for the identification of 47 key phytoplankton pigments. In: S. W. Jeffrey, editor. Phytoplankton Pigments in Oceanography. France: UNESCO, pp. 449-559. Kohata, K. & M. Watanabe. 1988. Diel changes in the composition of photosysnthetic pigments and cellular carbon and nitrogen in Chat- limella anliqua (Raphidophyceae). J. Phycol. 24:58-66. Kohata, K., M. Watanabe & K. Yamanaka. 1991. Highly sensitive deter- mination of photosynthetic pigments in marine in situ samples by high- performance liquid chromatography. J. Chromatogr. 558:131-140. Latasa, M., R. R. Bidigare, M. E. Ondrusek & M. C. Kennicutt II. 1996. HPLC analysis of algal pigments: a comparison exercise among labo- ratories and recommendations for improved analytical performance. Mar. Chem. 51:315-324. Letelier, R. M., R. R. Bidigare. D. V. Hebel, M. Ondrusek. C. D. Winn & D. M. Karl. 1993. Temporal variability of phytoplankton community structure based on pigment analysis. Limnol. Oceanogr. 38:1420-1437. Lohr. M. & C. Wilhelm. 1998. Pigment synthesis and xanthophyll cycle in diatoms under high light stress and during low light recovery. In: G. Garab. editor. Photosynthesis: Mechanisms and Effects. 3. Amsterdam. The Netherlands: Kluwer Academic Publishers, pp. 2313-2316. Lohr, M. & C. Wilhelm. 1999. Algae displaying the diadinoxanthin cycle also possess the violaxanthin cycle. Proc. Natl. Acad. Sci. USA 96: 8784-8789. Mantoura, R. F. C. & C. A. Llewellyn. 1983. The rapid determination of algal chlorophyll and carotenoid pigments and their breakdown prod- ucts in natural waters by reverse-phase high-performance liquid chro- matography. Anal. Chim. Ada 151:297-314. Oku, O. & A. Kamatani. 1999. Resting spore formation and biochemical composition of the marine planktonic diatom Chaetoceros pseuJocur- visetus in culture: ecological significance of decreased nucleotide con- tent and activation of the xanthophyll cycle by resting spore formation. Mar. Biol. 135:425^36. Porta, R. J.. E. E. Pfundel & N. Engel. 1997. Metabolism and function of photosynthetic pigments, pp. 85-126. In: S.W. Jeffrey (ed.). Phy- toplankton Pigments in Oceanography. UNESCO, France. Roy, S., J. -P. Chanut, M. Gosselin, & T. Sime-Ngando. 1996. Character- ization of phytoplankton communities in the lower St. Lawrence Es- tuary using HPLC-detected pigments and cell microscopy. Mar. Ecol. Prog. Ser. 142:55-73. Suzuki. K., N. Handa, H. Kiyosawa & J. Ishizaka. 1997. Temporal and Spatial distribution of phytoplankton pigments in the Central Pacific Ocean along 175°E during the boreal summers of 1992 and 1993. J. Oceanogr. 53:383-396. Suzuki, R. & T. Ishimaru. 1990. An improved method for the determina- tion of phytoplankton chlorophyll using N. A'-dimethylformamide. J. Oceanogr. 46:190-194. Suzuki, R. & T. Ishimaru. 1992. Characteristics of photosynthetic pigment composition of GymnoJinniin mikimotoi Miyake et Kominami ex Oda. / Oceanogr. 48:367-375. Suzuki, R., M. Takahashi. K. Furuya & T. Ishimaru. 1993. Simplified technique for the rapid determination of phytoplankton pigments by reverse-phase high-performance liquid chromatography. J. Oceanogr. 49:571-580. Willemoes, M. & E. Monas. 1991. Relationship between growth irradiance and the xanthophyll cycle pool in the diatom Nilzschia palea. Physiol. Planta 83:449^56. Wright, S. W. & J. D. Shearer. 1984. Rapid extraction and high- performance liquid chromatography of chlorophylls and carotenoids from marine phytoplankton. J. Chromatogr. 294:281-295. Wright, S. W.. S. W. Jeffrey, R. F. C. Mantoura, C. A. Llewellyn. T. Bjomland, D. Repeta & N. Welschmeyer. 1991. Improved HPLC method for the analysis of chlorophylls and carotenoids from marine phytoplankton. Mar. Ecol. Prog. Ser. 77:183-196. Yamaoka. Y. & O. Takimura. 1985. Chemical composition of red tide phytoplankton (in Japanese). Chem. Soc. Japan 1985:1488-1494. Journal of Shellfish Riscanh. Vol. 20. No. 3, l264-i:72, 2(){)l. IMPACTS OF THE HARMFUL DINOFLAGELLATE, HETEROCAPSA CIRCULARISQUAMA, ON SHELLFISH AQUACULTURE IN JAPAN Y. MATSUYAMA,' * T. UCHIDA.' T. HONJO," AND S. E. SHUMWAY' ' Hanufitl Algal Bloom Division. National Research Institute of Fisheries and Environment oj Inland Sea. Maruishi. Ohno. Saeki, Hiroshima 739-0452. Japan: 'Faculty of Agriculture. Kyuslui University, Hakozaki. Fukuoka 812-85HI. Japan: Department of Marine Sciences. University of Connecticut. Groton. CT06J40 ABSTRACT In the last decade, the hioom-forniing dinoflagellate Heterocapsa circularisquama Horiguchi has caused red tides along the Japanese coast. Because this dinoflagellate shows a detrimental effect on moiluscan shellfishes such as bivalves and gastropods, almost all red tides have associated catastrophic death of farmed animals. Recent proliferations of H. circiilarisi/uiinia throughout the western Japanese coast have devastated the shellfish aquaculture industries and are a cause for concern due to the consequent economic losses. KEY WORDS: Heterocapsa circitlarisqnaina. red tide, shellfish, death, toxicity INTRODUCTION Current proliferation of harmful algal blooms causes serious problems for public health and fisheries industries (Okaichi 1989. Smayda 1990. Shumway 1990. Hallegraeff 1993. Anderson 1994. Honjo 1994). On the Japanese coast, the novel marine dinoflagel- late Heterocapsa circularisquama Horiguchi (Horiguchi 1993). appeared in 1988 and then rapidly expanded over the western area (Matsuyama et al. 1996). Red tides due to H. circularisquama have damaged shellfish aquaculture in most of the region (Yamamoto & Tanaka 1990. Yoshida & Miyamoto 1995. Matsuyama et al. 1996, Etou et al. 1998). Although H. circularisquama blooms mainly *Corresponding author. affect shellfisheries aquaculture. no harmful effects on wild and cultured finfish. other marine vertebrates, and public health haz- ards were recorded. Incidence of blooms of this species has in- creased recently, and the economic losses in aquaculture have been a cause for concern for the shellfisheries industry and society (Matsuyama et al. 1996). In the present article, we review the damage caused to aquaculture and the toxicity of the organism. DAMAGE TO SHELLFISHERIES AND TOXICITY Red Tide and Shellfish Damage in Aquaculture Industries The first incidence of red tide due to H. circularisquama and subsequent death of shellfish occurred in Uranouchi Bay, on the southern part of Shikoku Island (Fig. I ) in 1988. Red tide due to is\^ s^\roa C^^^ QV^ Z' n Suo-nada (0 97, •98) Fukuoka Bay (• 98) J Hiroshima Bay (• 95, • 97, • 98) Iman Bay (0 95,0 96) IT^-^ 1 Uchinoumi Bay (O 98) y^y\ Lake Hamana (• 93) \ Kusu-uraBay(»94) _^' guzeti Sea (• 97, O 98) Ago Bay (• 92, O 93, • 94, O 95, •Se, 0 97) V ^^ Gokasho Bay (• 94, O 95, • 97) / Kuroshio Current O No damage to shellfisheries # Damage to shellfisheries Figure 1. Records of red tide Old'' cells L"') due to Heterocapsa circularisquama and two major currents in Japan. 1269 1270 Matsuyama et al. H. ciixularisquama has also been recorded at Fukuoka Bay in 1989, and at Ago Bay in 1992, resulting in mass mortality of shellfish (Fig. 2). Until 1998, 28 cases of W. circiiUiristjiiaiiui red tide (maximmii cell density SIO*' cells L"'), including 14 incidences leading to fisheries damage, had been recorded in 14 locations of western Japan (Fig. 1 ). The red tide due to H. circularise/Kama has been associated with massive killing of commercially important bivalve species: manila clam Riulitapes phllippiiianim. Pacific oyster Crassostrea gigas. pearl oyster Pincladafucata. blue mussel Myti- Ins galhiprovinciaHs. etc. The current proliferation of H. circitlar- isquaiud throughout western Japan has essentially destroyed mol- luscan shellfish aquaculture. Economic losses of shellfish aqua- culture have been estimated to be at least 93 million US dollars in the last decade (Table 1). On the other hand, there have been no records of death of finfish and crustacean species or public health hazard due to the consumption of shellfish and other seafood products in association with the red tide of H. ciirularisquama. This type of biohazard in marine animals is markedly different from previous reports of damage caused by harmful algae responsible for PSP, D.SP. ASP. NSP, ciguatera poisoning, and ichthyotoxicity. Behavior of Shellfish Affected by H. circularisquania The effects of H. Lirciilarisquaina on bivalve molluscs have been described in previous studies. Matsuyama et al. (1996) have observed that exposure of pearl oysters to 5-10 x 10"^ H. ciixu- larisquama cells L~' resulted in death within several days, al- though the level of dissolved oxygen was not critical. The dead individuals have been characterized by a marked shrinkage of the mantle, decrease of glycogen lobe attached to the mantle, and gut discoloration. The symptoms clearly reflect the direct cytotoxic effect of H. circiilarisqiiama on pearl oyster physiology. Similar harmful effects on the oyster C. gigas and the mussel M. gallo- provinciatis have been shown during the red tides which have occurred elsewhere (Etou et al. 1998, Matsuyama et al. 1998). Laboratory Exposure Experiment i'siiig Cultured Material Nagai et al. ( 1996) have shown that the mortality of pearl oyster spat caused by H. circiilarisqiiama depends on the cell density of this alga. Pearl oysters exposed to H. circiitarisqucmia cells at a density above 10'' cells L~' have showed vigorous contraction of the mantle and gills, clapping, sustained valve closure, paralysis, and heartbeat stoppage within 24 hours. Furthermore, the mussel Figure 2. Photographs of Heterocapsa circiilarisquama and affected shellfish. (1) and (2): Light micrographs of//, circiilarisquaina; (3) Dead shellfish: mussel Mytilus gallnprovincialis , razor clam Solen strictus, manila dam Ruditapes philippinarum due to red tide of//, circularisquama (Fukuoka Bay, 1989, photograph is provided by V. Tanaka): (4| Dead pearl oyster I'inctada fiicata (Ago Bay, 1992); (5) Dead oyster Crassostrea gigas (Hiroshima Bay, 1995): (6) Farming-ground of manila clam Ruditapes philippinarum affected by the red tide of //. circularisquama (Hiroshima Bay, 1995, Photograph is provided by T. Hamasakil. Impacts of Heterocapsa circular/squama 1271 TABLE I. Damage to fisherits industries caused l)> harmrul alual hlooms in Japan. Causative Agents Periods Amount (million $US) Cluilliinella spp. 1969-1999 185 Helemciipsa ciniiliin.si/iuiii}ii 1988-1999 92 G\mnodinium mikiinotoi 1972-1999 83 Heterosigma akashiwu 1972-1999 15 M. i10* death Gastropods iC-io-' Unusual locomotion >10'' death Acidians >10" Feeding inhibition Jellyfish >10'' Tentacle shrinkage Protozoa ' 10'- 10'' Feeding inhibition >10'' Death Dinotlagellales' >10'' Death bj cell contact Diatoms >10^ - Copepods >10^ - Finfish >10<' - Crab >10^ - Lobster, shrimp >10' - Lobster, shrimp >10^ - Gammarid >10' - Star fish >10' - Sea urchin >10^ _ Mouse" - - -: not affected ' .species-specific " intraperitoneal injection ( 10" cells/mouse) poisoning has never been observed in samples collected from red tide areas. Direct HPCL analysis has tailed to detect PSP toxins or DSP toxins in the cells of H. circularisquama. No death or symp- toms have been observed in 5 mice to which a cultured cell pellet of H. circularisquama had been injected in intraperitoneally at a rate of 10" cells/mouse (Table 2). Characterization of Heterocapsa circularisquama Toxicity The toxicity of H. circularisquama to bivalves has been shown to be mediated by a chemical agent. The toxic effect of H. circu- larisquama on bivalves is not due to extracellular metabolites, cell exudates, and "naked cells" prepared by sonication and centrifu- gation (Fig. 3A). Furthermore, trypsin and SDS (sodium dodecyl sulfate) treatments have been found to decrease drastically the toxicity of H. circularisquama cells (Fig. .^B). Therefore, a labile protein-like complex localized on the cell surface presumably ex- erts a detrimental effect on bivalves. However, purification and characterization of toxic fractions have not been successful be- cause this agent is highly labile under neutral conditions. Recently, exposure of larval manila clam to H. circularisquama has demon- strated acute toxicity on their survival and development, which is associated with a significant increase of intracellular calcium con- centration. SUMMARY Blooms of the dinofiagellate H. circuUirisi/muua have been associated with mass mortalities of various bivalves in western Japan since 1988. Toxicity of this species is acute and specific to 0 Clearance rate (%) 50 100 Control l^^^^^^^^^^^^^^^^^^^^^^^^^y?^^^^^^^ Intact Filtrate ^^^^^^^^^^^^^^^^^^^y?^^^^^^^ Sonication ^^^^^^ Centnfugation ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Clearance rate (%) 50 100 Control ^^^^^^^^^^^^^^^^^;?^;S^^^^^^^^^^^^^^^^ Intact ^S-^ Trypsin ^^^^^^^^^^ SDC EDTA ^^^ SDS ^^^^^^JS^^:?^^^^^^^^^^^^^^^^^^^^^^^^ ^^ B Figure 3. Relative clearance rates of .Mytilus galloprovincialis feeding to physically (Al and chemically (B) treated Heterocapsa circularis- quama cells (Control: Isoclirysis galhana, 8 x 10' cells L"'). Initial cell density of//, circularisquama ranged from 2.5-2.8 x 10"' cells L"'. SDS: sodium dodecylsulfate. SDC: sodium deoxycholate. DTT: dithiothrei- tol. Each chemical was used at a concentration lower than that which vtould Inhibit the swimming of //. circularisquama. Error bars shov* ±S. D. 1272 Matsuyama et al. bivalves and gastropods. Unfortunately, no successful damage pre- vention strategy has been developed until now. Therefore, reloca- tion and/or early harvest of organism are the only methods to reduce the shelirisheries damage. In some locations, early warning systems by local governments that are based on frequent field monitoring of H. ciniilurisquama cells have successfully reduced the shellfisheries damage. H. circiilarisquama blooms do not cause finfish killing and mammals illness. However, a rapid decline of demand for products due to misinformation, rather than actual shellfish poisoning, could occur. On the other hand, toxicity of H. circidarisquaina can easily be reduced when the cells are disturbed with physical and chemical treatments (unpublished). This fact will provide crucial information for improvement of damage pre- vention in the future. LITERATURE CITED Anderson. D. M. 1994. Red tide. Sci. Am. 271:52-58. Etou. T.. K. Kuwamura & H. Satou. 1998. The occurrence of a Heiero- capsa cirvuhiriscjumna red tide and subsequent damages to shellfish in the Buzen Sea in Autumn 1997. Bull. Fukiioka Fisli. Mar. Tecnol. Re.\. Cent 8:91-96 (In Japanese with EngUsh abstract). Hallegraeff. G. M. 1993. A review of harmful algal blooms and their apparent global increase. Phycologia 32:79-99. Honjo, T 1994. The biology and prediction of representative red tides associated with fish kills in Japan. Rev. Fish. Sci. 2:225-253. Horiguchi. T. 1995. Helerocapsa ciiridurisqiiama sp. nov. (Peridiniales. Dinophyceae); a new marine dinoflagellate causing mass mortality of bivalves in Japan. Phycol. Res. 43:129-136. Matsuyama, Y., T. Uchida, K. Nagai, M. Ishimura, A. Nishimura, M. Yamaguchi & T. Honjo. 1996. Biological and environmental aspects of noxious dinoflagellate red tides by Helerocapsa circularisquama in the west Japan. In: T. Yasumoto, Y. Oshima & Y. Fukuyo, editors. Harm- ful and toxic algal blooms. Paris: Intergovernmental Oceanographic Commission of UNESCO, pp. 247-250. Matsuyama. Y., T Uchida & T. Honjo. 1997. Toxic effects of the di- nofiagellate Helerocapsa circularisquama on clearance rate of the blue mussel Mylilus galloprorincialis. Mar. Ecol. Proi>. Ser. 146:73-80. Matsuyama, Y., T. Uchida & T. Honjo. 1998. Effect of harmful dinoflagel- lates, Gymnodinium mikimotoi and Helerocapsa circularisquama. red tide on filterins rate of bivalve mollusks. Fish. Sci. 65:248-253. Nagai. K.. Y. Matsuyama. T. Uchida, M. Yamaguchi. M. Ishimura. A. Nishimura. S. Akamatsu & T. Honjo. 1996. Toxicity and LDs,, levels of the red tide dinotlagellale Helerocapsa circularisquama on juvenile pearl oysters. Aquacutlure 1-14:149-154. Okaichi, T. 1989. Red tide problems in the Seto Inland Sea, Japan. In: T. Okaichi, D. M. Anderson & T. Nemoto. editors. Red tides, biology, environmental science, and toxicology. New York: Elsevier, pp. 137- 142. Shumway. S. E. 1990. A review of the effects of algal blooms on shellfish and aquaculture. / World .Aquacullure Soc. 21:65-104. Smayda, T. J. 1990. Novel and nuisance phytoplankton blooms in the sea: evidence for a global epidemic. In: E. Graneli. B. Sundstrom. L. Elder & M. Anderson, editors. Toxic marine phytoplankton. New York: Elsevier, pp. 29^0. Tangen. K. 1977. Bloom oi Gyrodinium aureolum (Dinophyceae) in north European waters, accompanied by mortality in marine organism. Sarsia 62:123-133. Yamamoto, C. & Y. Tanaka. 1990. Two species of harmful red tide plank- ton increased in Fukuoka Bay. Bull. Fukuoka Pre). Fish. Exp. Sin 16:43^(4 (In Japanese). Yoshida, Y. & M. Miyamoto. 1995. Growth of Helerocapsa circularis- quama population in Kusu-ura Bay and variation of diurnal motion of red tide in 1994. Rep. Kumamolo Pre)'. Fish. Res. Cent 3:31-35 (In Japanese). Miiiriial of Shellfish Research. Vol. 20. No. 3, I27.V127S. 2001. GILL STRUCTURE, ANATOMY AND HABITAT OF ANODONTIA EDENTULA: EVIDENCE OF ENDOSYMBIOSIS MA. JUNEMIE HAZEL L. LEBATA* AND JURGENNE H. PRIMAVERA Acjuavulliire Depaiiinent. Southeast Asian Fisheries Development Center, Tigbaitan, lloilo ?II2I. Pliilippines ABSTRACT Surveys and interviews were eonJucted to determine sources and habitat of Aniiil(miiii edcuuilii. Results showed that they inhabit muddy substrate of mangrove areas or the adjacent mudflats, burying at 20-60 cm deep in the mud. They are strategically situated in the sulfide-rich. low-o.\ygen layer of the substrate but have access to oxygen through their inhalant tube: both sulfide and o.xygen are essential for their survival. Study of the clam's gross anatomy revealed thick, fleshy, deep purple to blackish brown gills; reduced digestive structure; and a highly elastic foot capable of extending several times longer than its body length. These observations conform with the anatomy of fellow lucinid clams. Furthermore, scanning electron micrographs showed coccoid or spherical bacteria occupying bacteriocytes in the clam's gills. Intermediate cells separating bacteriocytes observed in other lucinids were also noted in the SEM. KEY WORDS: Anndontia eilcntida. anatomy, bacteria, gdl. endosymbiosis. habitat INTRODUCTION AiuhIoiuui cJciinda (Liiine. 1758) is one of the poptilar bi- valves iti the coastal areas of central and southern parts of the Philippines. It grows to a maximum size of 8-9 cm shell length (SL), 180-210 g total weight (TW) and is an important source of food and livelihood in coastal areas where it is abundant (Lebata 2000). It is one of the few expensive bivalves in the region as it is being sold per piece and not per kg or container like the common bivalves, i.e. mussels, oysters and scallops. It is a local favorite becau.se of its sweet taste, especially when eaten |-a\v. and men deeply believe in its aphrodisiac properties. Its popularity led to overexploitation as manifested by the decline in catch, in terms of quantity and size, through the years. A. edenlula inhabits sandy-muddy substrate near mangrove ar- eas or the adjacent mudflats (Sotto & von Cosel 1982). It belongs to order Veneroida. family Lucinidae (Poutiers 19981. to which eulatiiellibranchs containing sytnbiotic bacteria predominantly be- long (Schweimanns & Felbeck 1985). Animal-bacteria symbioses have been observed in at least five animal phyla but are most widespread among the bivalves (Distel 1998). Symbioses have been observed in marine tnollusks thriving in organically rich seditiients (Janssen 1992) and hydrogen sulfide- rich habitats such as anoxic basins (Felbeck et al. 1981); seagrass beds (Cavanaugh 1983); mangrove swamps (Vetter 1985. Frenkiel et al. 1996); and sewage outfalls (Felbeck et al. 1981 ). Numerous studies have been conducted on this aspect in other lucinid species (Berg & Alatalo 1984, Giere 1985, Schweitnanns & Felbeck 1985, Vetter 1985. Reid & Brand 1986, Distel & Felbeck 1987. Frenkiel et al. 1996, Gros et al. 1997, Gros et al. 1998). Studies on A. edentiila have been limited to species identification and descrip- tion (Sotto & von Cosel 1982, Poutiers 1998), light microscopic observation (Janssen 1992). paleontology (Cooper 1996), and just recently its elemental sulfur content (Lebata 2000). In the Philippines, the first article on clam symbiosis was pre- sented by Janssen (1992). Common bivalve species such as Codakia lifieriiui. A. edentula. and Fiinhria Jhiihriata were ob- tained from the Mactan Island fish market and observed under a *Corresponding author. E-mail: jlebataCS'aqd.seafdec.org.ph light microscope. Samples were fixed and brought to Germany for light and electron microscopy. However, only F. fiinbriata was studied in detail using electron tnicroscopy. Symbiosis was dis- cussed based on literature review but no actual field observation or laboratory experiments were conducted to prove this in A. eden- tula. A new study conducted by Lebata (2000) showed that the clams have elemental sulfur in their gills. Elemental sulfur, ac- cording to Vetter ( 1985) is stored in the periplasmic spaces of the bacteria and represents a novel inorganic energy reserve for ani- mal-bacteria symbiosis in the absence of sulfide. This may there- fore support the claim of Janssen (1992) on the existence of en- dosymbiosis in this lucinid species. Considering its potential for aciuaculture and the declining state of its population, it is necessary to have a clear understanding of this clam-bacteria relationship before undertaking culture studies in this species. The objectives of this study were to describe the habitat of the clam; to study its gross anatomy and relate it to its habitat; and to check for the presence of bacteria in the gills using scanning electron microscopy. MATERIAL.S AND METHODS Habitat and Collection Surveys and interviews were conducted from 1995 to 1 997 in Panay, Guimaras, and Negros islands in central Philippines (Fig. 1 ) to determine sources and habitat of A. edentula. The most acces- sible area to the laboratory, Brgy. San Roque in Estancia, lloilo, central Philippines ( 1 f I4'N, 123°8'E), was chosen as the field study site and collection site of clams for laboratory studies. The area has been described in Lebata (2000). A. edentula was collected during the ebb of a spring tide during the day with the help of a clam collector who can recognize the opening of the clams" anterior inhalant tubes (Fig. I ) on the surface of the substrate. A specialized gear made of flattened iron attached to a wooden handle was used in digging the substrate. The tube that led to the clam was then carefully traced. During collection, depth occupied by the clam was measured from the surface of the substrate, and distance frotn the mangrove forest was noted. Physico-chemical parameters were monitored /;; situ at 30 cm deep. Temperature was measured using a field mercury thermom- eter; salinity with an Atago refractometer; pH with WTW pH 1 92 1273 1274 Lebata and Primavera PHILIPPINES I Figure 1. Map of the Philippines show ing Panay, Guiniaras and Negros islands surveyed during site selection; collection site chosen was Brgy. San Roque in Estancia, Iloilo. central Philippines. Inset shows the opening of Anodoiitia edentiila's inhalant siphon as seen on the surface of the suhstrate. meter; D.O. with a YSl Model 51 B dissolved oxygen meter; and sulfide with a Lovibond photometer PC 22. For laboratory experiments, collected clams were transported live to SEAFDEC Aqnaciillure Department in Tigbauan, Iloilo. approximately 160 km from the collection site. During transport, they were placed in plastic baskets cushioned with papers soaked in seawater. Upon arrival in the laboratory, they were placed in I -ton fiberglass tanks provided with 30-50 cm deep newly col- lected mangrove mud and seawater (32-35 ppt). Clams were har- vested from the tanks as needed. Analoniy Several clams were dissected and the gross anatomy studied following Allen (1958). They were then examined for the presence of gut using an Olympus SZ-ST dissecting microscope. This was done because absence or reduction of the digestive system is a common condition in most clams containing endosymbionts in the gills (Giere 1985). Histological sections of the gonad were also made and observed under an Olympus BH2 compound micro- scope. Live clams were also placed in a glass aquarium provided Evidence of Endosymbiosis in A. edentula 1275 with seawater hut no siihstrate to ohsene the extension and re- traction of the foot. Scanning I'kclron microscopy Newly eoilected li\e clams ranging from 50.5-62.6 mm .shell length were transported to the National Institutes of Biotechnology and .Applied Microbiology (BIOTECH) at the University of the Philippines. Los Banos. College. Laguna in northern Philippines (Fig. I ) on 13 May 1998 for scanning electron microscopy (SEM). Upon arrival, the biggest clam (62.6 mm SL) was dissected and the gills processed for SEM. The specimens were fixed in 3% gluter- aldehyde with 0.1 M cacodylate buffer at ph 6.8 for 3 h. washed with 0.1 M cacodylate buffer 3 times at 15 min each, post-fixed in 1% osmium tetroxide for 1 h. rewashed with 0.1 M cacodylate buffer, dehydrated in graded ethanol series, and substituted with isoamyl acetate. Then they were dried in the Hitachi HCP-2 critical point dryer, mounted on supports, coated with gold and palladium using the Hitachi ElOl ion sputter, and viewed under the Hitachi S-500 scanning electron microscope. Photomicrographs of the specimens were taken and compared with scanning electron mi- crographs of bacteria from other endosymbiont-containing inver- tebrates (Karl et al. 1980. Fiala-Medioni & Metivier 1986. Fiala- Medioni et al. 1986) and lucinid clams (Berg and Alatalo 1984). RESULTS Hahilal A. cilciuiila thrive in the muddy substrate of mangrove areas or the adjacent mudflats. The closer the area to the mangroves the more clams can be harvested (within 20 m from the edge of the forest) and the number of clams collected decreased as collection moved seaward. Pre\ ious and recent surveys showed that they are marine and can be found nowhere in rivers or estuaries. Clams were deeply buried in the mud ranging from 20-60 cm below the surface, but mostly in the deeper portions. Other non- lucinid bivalves inhabiting tlie area did not buo' deeper than 15-20 cm. The temperature ranged from 27-3 1°C (mean = 28.7 ± 0.28). the salinity from 32-36 ppt (mean = 34.5 ± 0.31), pH from 5.15-6.55 (mean = 5.95 ± 0.10). dissolved oxygen from 0.2-1.0 ppm (mean = 0.3 + 0.05). and sulfide 1.5-30.0 |jlM (mean = 25 ± 1.78). Figure 2. (a) Anatomy (pI a nc'Hl> dissiiiid .\ii(i